NPFMC SSC October 6, 2026 - Day 2

Okay, good morning everybody. We're going to open our SSC session this morning and we're going to catch up on a few things we didn't finish from yesterday. So I think first We're going to get to meet Jamal Moss. I don't know if Council staff wants to introduce you, or you're welcome to come on up. Welcome, Dr. Moss.

Thank you, and good morning. It's great to be here. I just wanted to take a moment to introduce myself. My name is Jamal Moss. I work for the Alaska Fisheries Science Center.

So I'm acting director right now. I know most of you, which is great, but there's a few that I don't, and so looking forward to getting to know you and listen in, as well as provide any information you need from the science directorate. So really looking forward to working closely with you all. I wanted to just give a little bit about my background for those that might not know me. I, um, Let's see, I started working at the Auke Bay Lab in 2003 as a research fisheries biologist right out of grad school and have been there for about 20 years, mostly as a research fisheries biologist, but then had a stint as the deputy lab director there.

And then from there, I moved to the regional office where I served as the deputy regional administrator. And so I was there for about 3.5 years and got plucked and asked to move back to the Science Center, which has been a great treat. It's been a lot of fun to think about science and enable science and work with the amazing scientists there. So that's really all I wanted to say, just to kind of put a face with the name and encourage you all to just reach out to me anytime you want to, whether it's at the Council or offline. So thank you.

Thanks, Dr. Moss. Welcome, and we look forward to working with you.

Okay, we're now going to restart where we left off on our agenda from yesterday, which is still on our C2 crab specifications, and we will pick up with St. Matthew's blue king crab. So welcome back, Katie, and thanks for staying another day.

Alright, so this is the final assessment for St. Matthew's blue king crab. This is a stock that we do an assess— full assessment for on a biannual basis. And let's see if it likes the arrows. Nope.

It is currently under a rebuilding plan and has been since June of 2020. It is one of our assessed stocks that's done in a Tier 4 framework, framework, but is still in GMAX. The directed fishery has been closed since the 2016-2017 crab season. So since this stock is a biannual stock, in May the, the models that we saw for proposed models included the year of fishery and survey data, and it did actually include an ADFNG survey also. There's a pot survey there for St. Matt's every 3 years.

And so the May models included those new datas, and then the September models will include— did include the new year of fishery data and then the most recent EBS survey data point. So the models explorations that we saw include the base model, which is the one— the last accepted model from 2024. There was some GMAX version bridging. This stock was brought up to the most recent GMAX versions the same version that Bristol Bay red king crab is currently in. There were no differences there, so that was good.

There weren't any expected to be. And then, then the author brought forward a model-based index to account for the loss of the corner stations in the EBS survey. And so you can see here in the models considered underneath that, the base model with updated GMAX and data, the model that has the model-based index for the EBS trawl survey, and then a research model was brought forward by the author just to try and get some feedback from the CPT, wasn't brought forward for specifications on including model-based size comps. So I think we provided her some good feedback on that and potentially a path forward for, you know, bringing it forward the next round.

So the corner stations— I didn't put in here what year they were removed, but they haven't been done the last 2, 3 years, I think. So it makes the EBS trial survey footprint inconsistent over time.

Especially for St. Matt's. So the author's been working on a model-based index. This has been in development for a while now, and it's currently, the CPT believes, and the author is ready to be adopted in the base model. The, the figure that you see on the right compares the index taken from the design-based, as you can see in the black dots, to model-based indices, index with either year or year and depth as Effects there, fixed effects. So the CPT, the model that we saw at CPT just has year as a fixed effect, not depth, but the author did explore a wide range of different options pretty thoroughly.

The, they are pretty well detailed in the SAFE report if you have more interest in those, but we did thoroughly go over those same methods in May and The main difference between these models is the scale of MMB, but the biggest difference there, if you look at the right— sorry, right top right figure, it was actually with the— yeah, model doing— yeah, both model-based indices scale the stock slightly higher than the accepted model. However, due to this being a Tier 4 stock and how we calculate status and BMSY, the overall status of the stock is pretty similar, so it doesn't really affect overall interpretation of the status of the stock.

Model-based—. Model—. The model that included— I keep saying model too many times— 26.2 includes the model-based index, and the CPT felt that this included consistent survey input for every year. So the CPT did endorse that model, and that was the author-recommended model moving forward. The author did provide some projections— that's what you see on the bottom right here— to look at rebuilding progress, and they are largely dependent on our expectations on recruitment.

So they— we did the recruitment using the entire time series of possible recruitments, random draws from what has been observed in the entire time series versus same time series that's used for BMSY, and then random draws from, from the last like 10 years or so. And you can see that your expectations for the stock largely depend on if there's large recruitment events, which is not anything new for the stock. That's primarily the premise of the rebuilding plan that was done in 2020 was that the stock is highly dependent upon expected recruitment. The last completed crab year, the stock status is actually 0.518. I guess we could round that to 0.52, but anyways, so the stock is actually not overfished.

So that is a good status for the stock. Glue Bay has made some slow progress in that rebuilding plan, and this is the first year where our status is above 0.5.

Oh, I went too far, sorry.

Okay, so the table here has all the, the both models in it, but the model that we're looking at is 26.2. So the resulting OFL that comes out of that model is, is slightly higher. The, like I said, the directed fishery has been closed and there's pretty minimal bycatch for this, the stock in groundfish fisheries. The CPT and author both recommended to model 26.2. And we recommended an ABC buffer of 25%, which is consistent to what we've used in previous years.

I will say that, you know, and I'm co-author on this, so I'll take the heat for this one, but we did not do a risk table for this stock. There was still a lot of things that we have to work out with the risk tables as far as formatting and things like that, so it's on our plan to do it next time it comes up, but it just felt like it was a little premature for the stock to, to do it right now. So while we're still figuring out risk table usage and formatting, um, in a kind of iterative process, the, the buffer considerations are mainly that, you know, this is a low— the stock has been— is in a rebuilding plan. We have some ecosystem concerns as far as it's a cold water stock and potential eventual warming of temperatures potentially being concerning for, for rebuilding potential, um, and just overall not a lot of, a lot of information on this stock too. It is, it is our Tier 4 stock because of the less information that we have that are specific to St. Matt's and a lot of life history information that we're borrowing from other stocks.

So I think that's all I had on St. Matt's, but I can entertain any questions you have.

Okay, looks like we have no questions. Um, we have this broken up so that we have the option for public testimony now and then discussion, and then we'll go ahead and take Western Aleutian Islands, REDD, and the rest to balance the report afterwards.

So, not seeing any signups for public testimony, we'll go right to our discussion. And start with our first lead, Fabio.

Good morning, everyone. Thank you, Mr. Chairman.

So the SSC reviewed the presentation on the 2026 St. Matthew Island blue king crab stock assessment from CPT Chair Katie Peloff. The SSC thanks the authors for their substantial effort in developing the spatial-temporal model-based EBS trawl survey biomass index and the model-based estuarian seed trawl survey size composition estimates. The SSC also appreciates the response to previous CPT and SSC recommendations. There was no public testimony. Uh, the author presented 3 models.

The assessment was updated to DMAX version 2.20.44. And the bridging analysis between Models 26.1 and 26.1a resulted in similar model fits with the overall negative likelihood unchanged. The models presented were Model 26.1b, that's the new base model, which is the Model 26.1a updated with 26 Estimberensee trawl survey biomass and size composition data, and updated groundfish by catch data. Model 26.2, that's the model 26.1B with a model-based EBS trawl survey biomass index. And model 26.2A, and that's the model 26.2 with model-based EBS trawl survey size composition.

Model 26.2a is a research model and was not recommended as a candidate for harvest specification. The model-based heavy esterol survey biomass index accounts for differences in sample stations across years and provides a constant time series that includes both years with and without the higher density core stations. The SSC considers this is an important improvement for maintaining the continuity of the biomass index under the revised survey design. The SSE supports using year as fixed effect, given a limited tap contrast available to support that as an explanatory variable, while a continued evaluation of model fit, residual variability, and potential spatial covariates is warranted.

Models 26.1B and 26.2 produced similar temporal patterns in recruitment and mature male biomass, but Model 26.2 resulted in a substantially higher biomass scale.

Out of—. Uh, relative depletion trajectories were similar. The difference in biomass scale is important because the survey index directly influences estimated recruitment, the mature male biomass, and management quantities.

The San Matthew Island Blue King Crab is a Tier 4 stock under rebuilding plan following its 2018 overfishing designation, and the direct fishery has been closed since 2016-2017 crab season. The stock's not overfished and overfishing is not occurring.

The stock was above the MSST in 2025-2026, but both harvest specification models projected decline in the mature male biomass in 2026 and 2027. And model 26.2 estimates the B over BMSY below 0.5. Given the projected decline in the mature male biomass, the stocks continue low biomass relative to the BMSY. Rebuilding will depend strongly on successful recruitment. Projections indicated that under recent recruitment conditions, the stock is unlikely to reach the MSST by 2031, even without direct fishing mortality.

The SSC supports Elders and CPT's recommendation to use Model 26.2 for status determination and harvest specification, resulting in an OFL of 182 metric tons and an ABC of 136 metric tons with a 25% ABC buffer, unchanged from previous— from the previous year. Although the retrospective pattern has improved and the model-based—. Sorry—. Index has been developed, substantial uncertainty remains regarding life history parameters, surveys catchability and selectivity, recruitment and rebuilding. The SSC supports retaining the 25% buffer given this remaining uncertainty and the stocks continue low biomass relative to BMSY.

The SSC supports retaining the biannual assessment cycle through 2028 assessment, allowing time to further develop the model-based size compositions and approach for integrating the EBS Trial Survey and ADFNG PODS Survey Index before the next full assessment. In addition, the SSC recommends the following prioritized research and model developments: continue the development of integrated model-based index incorporating the EBS Trial Survey and ADFNG PODS Survey, and evaluate alternative approach for integrating information from the two surveys. Continued evaluation of the model-based index EBS trawl survey include the model fit and residual variability, survey catchability, and selectivity parameterization.

Continued development of the model-based EBS trawl survey composition, size compositions, using abundance rather than biomass, and evaluate correlation among, among size-specific spatial fields before incorporating this estimates into the assessment, develop a risk table framework to provide a more systematic and transparent basis for evaluating the ABC buffer, evaluate whether expanding the model size ranges to include a smaller crab, potentially down to 50 millimeters, carapace length, that would provide useful information on the pre-recruit portion of the population and the transition into the size class 1.

And finally, continue research on the life history and population process important to rebuilding, including growth and molting, natural mortality, spatial distribution, and recruitment. That's what I have. Thank you.

Thanks, Fabio. Next up is Kerry. Yeah, thank you, Fabio. Um, I concur with, uh, supporting both the author and plan team recommendation for model 26.2 that incorporates the new, uh, model-based, uh, index, uh, for the NPS survey. Um, I have to say I appreciated the description of the cross-validation, uh, for exploring out-of-sample predictive performance for the different spatial-temporal models.

I think the authors made a clear case for why the inclusion of, of factor depth rather than smooth depth made more sense in this case, and including depth was important. And all of the, the sort of residual and diagnostic plots that I would expect from any spatial-temporal model fit were included. So appreciate that very much.

In terms of our recommendation, I believe I heard that we're— we'll support the 25% buffer as recommended by the author and the plan team, which continues what we've seen in the past. And as Katie articulated, is primarily based on 2 factors, sort of continued low stock size. I think we could add to that the absence of larger recruitment events in the recent past and projected warmer conditions for this cold temperature preferring species.

I appreciate the list of research objectives that my colleague articulated and support all of those. In addition to the plan team recommendation that in thinking about, so it was a research model in this case, Um, but, uh, 26.2A was exploring, uh, model-based, uh, size composition expansion, um, and there was a recommendation to use abundance and not biomass, which was captured in the CPT minutes.

Um, I guess one question I had, I just wanted to make sure, so there was, I think, a recommendation from the SSC that this be moved to a 3-year cycle. The author's recommending that it remain on a 2-year cycle just to facilitate further model development.

I guess I am generally supportive of an author who wants to get more done, but I wanted to make sure I understood what your recommendation was to continue or 2-year cycle or 3-year cycle moving forward. I support the author's recommendation to keep with the 2-year cycle.

Excellent. Yeah, I would support that as well.

Thanks, Kerry. Other comments from the SSC?

Okay. I'll just briefly summarize. Thanks for a very clear summary from the two of you. We're going to thank the authors for their work to develop the geostatistical model and the response to previous SSC comments. We're going to note the three models that we saw, including 26.1B, 26.2, and 26.2A, the research model.

We will support Model 26.2, including the new model-based survey index accounting for different survey designs over time, noting our appreciation for the cross-validation and model diagnostics that were provided. We'll note that the stock is classified as Tier 4 under rebuilding, and the directed fishery has been closed since 2016-2017. We will support the OFL ABC and the ABC with a 25% buffer from the maximum permissible, in agreement with both the author authors and the plan team. We'll note that this buffer is based on continued low stock size and projected warm future conditions. We will note that overfishing did not occur.

The stock is not currently overfished, but is still in rebuilding, and we will highlight that rebuilding will depend strongly on future recruitment. We will support the biennial assessment moving forward to be seen next in 2028. And we have a number of research recommendations, including continued work on the integrated modeling of the trawl survey index, development of model-based size— model-based survey size compositions using abundance, not biomass, development of a risk table, evaluation of whether to include additional small crab size bins in the size compositions, and additional biological research for this stock.

Okay, that concludes our discussion of St. Matthew's blue king crab. We'll now go back to Katie for Western Aleutian Islands red king crab and the balance of the crab planting report.

All right, um, so Western Aleutian Island red king crab, uh, we had made a decision to roll over this, uh, specifications for another year and to move the, the final say for the stock to next year, 2027. So we just have a slide here that the specifications that we're rolling over are an OFL of 56 tons, an ABC of 14 tons. These are the same specifications that we've had for the past 3, I think, years. And these rollover specifications, I will note, that include a pretty large buffer. There's been a large buffer on this stock for a while now.

It is a Tier 5 stock that will be assessed next spring. And, and it— this last season, the total catch mortality was actually zero. That's, that's not a typo. There was actually no bycatch mortality of Western Long Island red king crab, and there is no direct fishery mortality. So there's actually no mortality, so therefore overfishing did not occur.

And we don't assess status because it's a Tier 5 stock, so even in the years when we have an assessment, we don't have a status determination. So that's what we're recommending as a rollover specification for this upcoming '26-'27 crab year. Any questions?

I have one, just to clarify. So when we talk about a rollover, because we have a lot of different things going on with different assessments this year with regard to groundfish. Did I hear you say clearly that this is a rollover, but our specs don't change on an annual basis anyway? Yes, so our specs don't change. They actually haven't changed in a while for the stock just because we don't have a lot of information.

So yeah, we don't— just, we're just taking the same specifications that we had put in place for a 3-year cycle and just extending it one more year. Thanks for clarifying. And I can use different terminology if it's better in the future. It works. Normally it would be the only rollover stock we're talking about.

Other SSC questions?

You have more material. Yeah, go ahead. Always shouldn't be— yeah, so the stocks that we do not do assessments for on an annual basis or those that we did assessments for before the crab year was completed, we do assess over fishing status. And so this is just to show that for both Pribilof Island king crab stocks, blue and reds, the directed fishery is closed for both the stocks. Bycatch mortality was pretty minimal and similar to what it's been in previous years.

Overfishing did not occur for either of those stocks. The total catch mortality was less than the OFL. And then Pribilof Island golden king crab, we already— that's on a— that one was already completed, but we just threw it in here because it's what we get in the table. Their total catch mortality was less than the OFL. We We can't tell you the exact number because it's confidential.

Sorry. And then Aleutian Island golden king crab, that total crab year wasn't completed when we did that assessment in May, so we just have a more— a final number for total catch mortality there that was also less than the OFL. So no, overfishing did not occur. So I believe— look at— I believe—. Yep, there you go.

So that's all I had for the balance of the report. I'm happy to take any other questions that you might have on anything I didn't cover in my oral report that's in our written report. That you potentially have questions on.

Looks like you satisfied all of our questions yesterday. Okay, thanks again for an excellent set of presentations. So we will now check for public testimony.

Noting that Western Aleutian Islands is influential scientific information. I see no signups, so we have no testimony. We will go then into our discussion, and our first lead on this is Ali. Thanks, Ian.

Um, and I will start out my comments by thanking, um, Katie and Anita, um, and also Emily who was online yesterday for their presentation, and thanks to, thanks to everyone for their work on C2 Bering Sea crab specifications.

I'm going to go out of order in my notes. So for our overfishing updates, we received status, overfishing status updates for 4 stocks. To cut to the chase, total catch mortality for each of the stocks was below the OFL, so overfishing did not occur in the last year's crab year. For Western Aleutian Islands red king crab, this is a Tier 5 stock assessed on a triennial cycle. The previous assessment was completed in 2023, but the plan team recommended rolling over specifications due to capacity constraints.

We actually agreed with this approach in June and we now look forward to seeing a full SAFE report that is currently planned for April 2027. The directed fishery was closed and there was no bycatch, so overfishing did not occur for that stock either.

I do have comments on the survey presentation, which we actually received at the beginning of the crab plan team report yesterday. So we received that presentation from Emily Risner. Thank you to her for being available for that. I think our comments can start off by commending the survey team and the crab assessment authors for, as usual, providing the survey data and updating model runs and assessment documents on the extremely short time frame that's necessitated by the timing of the survey and our management process.

Um, survey results were generally positive, um, for most crab stocks in 2026, with the exception being EBS snow crab, of course. Um, I think while there were positive signs for most stock, I think we can continue to express some general concern over the low stock status of many of our Bering Sea crab stocks.

I will provide some summary information in our minutes, but I think in the interest of time, I don't need to go over that. One thing I did want to note with regards to Bristol Bay red king crab, the plan team noted that the Bristol Bay red king crab assessment authors are planning a retrospective analysis to better evaluate the effects of resampling on size composition and abundance estimates at the January meeting, and I think we should support that effort and look forward to those results.

The only other recommendation I have for the survey section is that we should support the crab plan team recommendation that survey data with the updated net spread values be incorporated into all the crab stock assessments as soon as possible. It sounds like only snow crab was able to include this, so I think we should support that update, those updates.

Yesterday, to move into another section, Yesterday we had, we discussed hybrids a lot yesterday, and there's a lot of information on hybrids in the survey section as well. And so what I am going to propose here is that we have a new section dedicated specifically to hybrid crab under our general crab minutes in that report, and I have some information on hybrids and a couple of recommendations here, but I would propose that we combine that with our very wide-ranging discussion that we had yesterday. Um, and so some of this information will be in addition to that discussion that we had yesterday. Um, so that is my proposal currently. Um, so I want to note that last year you know, we had this unprecedented spike in hybrid crab in the trawl survey.

In 2026, the abundance declined across most size-sex categories, but with the notable exception being the industry preferred males, which continued to increase beyond the record highs that we noted last year.

And the plan team noted that there were still a lot of hybrid crab present in the system despite these declines, and the relative abundance of snow and hybrid crab greater than 101 millimeters was very similar to last year, but for the first time hybrids actually exceeded the number of snow crab out there. ADF&G opened up a million-pound hybrid quota in last year with an estimated catch of 0.68 million pounds, and the crab plane team noted that there were challenges associated with estimating those landings of hybrid crabs specifically. In the past, both the plan team and the SSC have noted the need to understand more about hybrid biology, genetics, and population dynamics. This year, the survey team was able to collect information on, or data on, hybrid weight, which should improve hybrid-specific catch estimation in the future.

And so the events of last year's survey really spurred a lot of discussion over the last year In October last year, we agreed that hybrid— agreed with the plan team that hybrid crab didn't need to be included in the assessments, and we also noted that we have a mechanism for tracking them in our survey— through our survey reports. We had a large number of hybrid identification and catch accounting questions, and many of those were the subject of a May CPT meeting earlier this year. And as we noted in June, um, earlier this year, that if hybrid abundance continues to be elevated, that our approach to monitoring and assessment may need to be reevaluated.

Uh, I have just a couple recommendations, um, in agreement with the plan team. I think we should strongly support further efforts to collect data on hybrids while they continue to be present. In our survey samples as capacity allows. Survey folks are very busy out there on the water, but I think we would suggest that this is a relatively high priority.

The other, the other recommendation I have is actually a request for clarification on the presence of hybrids in the total catch that is included in the SSC harvest specifications tables, I think we are all still a little confused about where various catch streams go and whether or not they include hybrids. So it would be helpful to circle back with the CPT on this.

And the plan team noted in— not under the CPT report, but under the snow crab comments that it would be useful to have more information on the policy and regulatory pathways to potentially allow for future harvest and management of hybrids. Public testimony strongly supported a process with the end goal of harvesting crab, hybrid crab when available, but in a manner that doesn't risk overharvesting snow crab, and a workshop that included managers, industry, Committee and Scientists was proposed in public testimony to initiate this process. And I think our minutes will reflect the wide-ranging discussion that we had under the snow crab agenda item yesterday in response to this.

That's all I have.

Oh, no, wait, sorry. I have a couple other general comments. Didn't scroll down far enough. Um, as a general comment, um, and as the SSC has noted previously, we would support the expansion of technical expertise on the crab plan team, and in particular the need for a second co-chair to help address overall capacity constraints on the crab plan team so that Katie no longer has to be the lone co-chair. Um, with regards to SAFE guidelines, um, I just wanted to, uh, note our appreciation for the efforts efforts to streamline and standardize the crab safe reports and the other associated documentation that they provide to us, to the extent practical.

And that's all I have for the moment. Thank you. Great, thanks, Ali. Next up is Andrew.

Great, thank you. Thanks, Ali, for that. Thorough summary. I actually don't really have much to say. I think you covered everything that I would have—.

I was scrolling through my notes and you hit everything that I had noted too. So I think I agree with your recommendation for the separate section in the general CRAB section on hybrids. I think that's definitely an important topic and we talked about it a lot. Yesterday. So, yeah, nothing else to add.

Thanks.

Thanks, Andrew. Comments from other SSC members? Sherri?

I just wanted to follow up on the comment on risk tables that I started during Bristol Bay yesterday and then postponed to this section.

My suggestion after seeing the rest of the presentations is just to note that the SSC appreciates the draft risk tables that were provided by CRAB authors, and we look forward to the work next April when the CPT reviews these risk tables for consistency. And the SSC will— I think the SSC should recommend that the CPT include a discussion of how to name levels in this two-tier risk table and ensure that the levels included in the table correspond or represent the overall buffer. So, for example, in the Bristol Bay SAFE, the authors included level 1 for the current year, which doesn't necessarily explain the historical buffer of 20% overall. So the SSC— I think we can note that the method used by the Tanner authors is promising, where there were levels for the historical level to represent the historical buffer, and then on the current year, there was a change or no change notation for the current year considerations to clarify that there was no change to the historical buffer in the current year. And this approach has the benefit that if the uncertain— uncertainty decreases in the current year, so in other words, if a retrospective pattern is resolved in the current year, there can actually be a decrease in the overall buffer.

So the current year can result in increases or decreases relative to the historic, whereas putting Level 1 doesn't necessarily indicate a possible decrease. So that's my suggestion. I don't— I suggest that we not be explicit but let them work through this as long as somehow you can make a connection between the levels and the overall buffer at the end.

Thanks, Sherri. Other comments from the SSC?

Curry?

Yeah, I think we might highlight our continued support for research to better understand the processes regulating recruitment in crab. I see this as a real high priority, understanding both the density-dependent processes and the environmental processes. This is particularly evident when we look at the rebuilding analysis for St. Matthew blue king crab. Whether you resample from the entire time series of recruitments or just from the last 10 years, determines whether or not that stock is rebuilt with no fishing by 2031. And so even for stocks for which— that aren't under a rebuilding plan, the assumption of the time series for recruitment is important for our reference points.

And so I guess to the extent that we can articulate some continued support for understanding recruitment processes, either through analytical work or to the extent that experimental work is practicable, I think we should.

Thanks, Kerry. Other comments from the SSC? Any follow-ups?

Okay, seeing none, we'll summarize this section.

So we will thank the presenters for all the various sections that we saw in the crab plan team report, including council staff, Katie, and the survey section. We'll note that we did not receive additional public testimony on Western Aleutian Islands red king crab or the overfishing updates. For the overfishing updates, we saw that overfishing did not occur for Pribilof Island red king crab, Pribilof Island blue king crab, Pribilof Island golden king crab, although that one was confidential, and Aleutian Island golden king crab. For Western Aleutian Islands red king crab, we'll note that this is a Tier 5 on a triennial cycle and that what we saw this year were rollover specifications for the OFL and ABC with no changes from previous values. Those included a 75% buffer from the maximum permissible, and we look forward to seeing a full SAFE document in 2027.

We'll note that for Western Aleutian Islands red, the fishery is closed. There was no bycatch estimated to have occurred, therefore overfishing did not occur. And the status is not estimated, not assessed for this stock. On the survey presentation, we thank the authors for the extensive information provided to the crab authors, the crab plan team, and the SSC on a short timeline. We'll note that there were mostly positive signs across many stocks.

However, we remain generally concerned about all the crab stocks, and particularly snow crab, with the decline seen in the survey information there.

We will request and support further retrospective analysis for Bristol Bay red king crab on resampling of size and abundance estimates, as has been already proposed. We will support the crab plan team recommendation that survey data with updated net spread be incorporated as soon as possible for all the assessments for which it has not already. Which I think are just say a snow crab.

We will capture, well, with regard to our discussion on hybrids, we had an extensive discussion on hybrids that ranged across our Tanner and snow crab items as well as our general crab and scallop plant team report. We will likely move the bulk of that discussion into our general crab minutes, although we'll probably reference it in our Tanner and snow crab sections. Generally, we will note that there's still a large abundance of hybrids, hybrid crab in the Bering Sea. We will note the efforts by Alaska Department of Fish and Game last year to add additional hybrid quota to, to the snow crab TAC in the form of an extra million pounds last year, of which I think around two-thirds were landed. We will note that there was some additional biological sampling of hybrids this year, including weight information, and we will support additional collection of biological data for hybrids as is possible in the future.

And we will highlight that although we have provided or supported the current path forward, which is to not include hybrids in either the Tanner or the snow crab assessments, we'll note that continued elevated hybrid abundance may mean that the current approach to modeling will need to be updated. We will specifically request clarification on the inclusion of hybrids in the catch that's reported in our specification tables, as well as the catch streams in the stock assessments. And we will note that it would be, based on public testimony and everything we've, we've heard and discussed, that more information on the regulatory pathways for harvest of hybrids in a manner that does not increase the risk on snow crab would be quite helpful. We will support the expansion of— the continued expansion of technical expertise on the crab and scallop plan team, and especially highlight the need for additional support in the form of a co-chair for Katie.

With regard to risk tables, we support the work that's been done so far and note the continued development in risk tables this year. We will recommend further discussion of how to assign levels to the risk table elements such that they represent the overall buffer to include both the historical and persistent factors, as well as the more recent information. We will note that buffers can go up or down depending on the change in both the short and the long-term factors, and we would like to see the levels reflect that on a year-to-year basis.

Finally, we will support continued research on processes regulating recruitment across all crab stocks. We'll note that this is critically important for both rebuilding and for reference point calculation, and that potential approaches could be both analytic or experimental.

I think that rounds out our crab items. I missed anything important? Speak up.

Great. So that will close off C2 for us, and I think we will take a short break before we start on groundfish. However, just for the for the public, the way we're going to take up groundfish, we're going to be in 3 sections. We're going to have the Joint Groundfish Plan Team report, and that's going to cover the report itself, an update on harvest control rule development, as well as sablefish proposed models. After we've done those 3 sections, we will take public testimony on all those sections and then have our discussion on all 3 sections separately.

So if you're here waiting on sablefish, that So that public comment period will be rolled into the full Joint Groundfish Plan Team. We will then have our discussion and finish off the Joint Groundfish Plan Team items and then move on to the GOA and the Bering Sea after that. So let's take an early break this morning and come back at the top of the hour, which would be 9 o'clock Alaska time.

Okay, if everybody could come back to the table, about ready to start again.

Okay, welcome back. We are now on to C3 Groundfish Joint Plan Team Report, and I'll turn it over to staff to lead us through. Welcome. Good morning, Mr. Chairman, members of the SSC. I'm Sarah Cleaver.

With me is Diana Stram. We're the plan team coordinators on council staff for Gulf of Alaska and Bering Sea Aleutian Islands Groundfish Plan Team. We just have a few slides at the beginning of this presentation before we get into all of the other presenters who will introduce momentarily, uh, just to orient you on what the SSC actions at this meeting are. Some of these are a little bit different than typical, so I'll just walk through. The first one is to provide recommendations for proposed rulemaking for all of the groundfish stocks.

And this is your typical rollover for proposed rulemaking from the final rule that was published in March, which is based off of the December SSC recommendations, as well as based off of your January recommendation for Gulf of Alaska Pacific cod. And that proposed rulemaking is essentially to notify the public that they can expect upcoming final rulemaking later on after you provide your final recommendations for all of the rest of the groundfish stocks in December. So that's mostly procedural, and, and you should be familiar with that from, from past years. The, the second action here is to provide final harvest specification recommendations for select stocks. And typically, I guess maybe in the past 1 to 2 years, you have received harvest projections at this October meeting since we started doing those a little bit earlier for, for review timing purposes.

And this year, the, the difference is that you will also be receiving some final assessments at this meeting. This includes not only harvest projections but also some full and update assessments, and you will be providing final recommendations for those stocks, which we will kind of just keep in our back pocket in a separate table to be included with your final recommendations at— in December. And I'll highlight on the next slide which, which stocks those are for. Uh, number 3 here is to provide your recommendations on proposed models that the authors will bring back in their assessments, um, at the November plan team and the December SSC meeting for review. And then Dr. Stram will provide an update on compiled SSC requests on the HCR work, and she has a few slides as well.

So you'll receive, receive that update and the joint plan team recommendations during the joint plan team report.

This is included in the action memo. We just put this together so that you are aware of which stocks we are, um, reviewing at, at which meetings and what stage they're at. So the stocks where you're reviewing the proposed models are on the left, and then the stocks for which the SSC is reviewing the final SAFE chapters and providing your final recommendations at this meeting are on the right in orange here, and you will not be reviewing those stocks again in December. And mostly for the public, but the AP and Council won't be providing TAC recommendations for those stocks at this meeting, so it'll just be the SSC looking at those stocks on the right, and then the AP and Council will be using that information that you provide at this meeting to do the, their TAC setting process. In December along with the specifications for all the other groundfish stocks that you'll be look— looking at in December.

I just wanted to include the Council's motion to make sure we kind of close the loop with the SSC on updates from the Council regarding spatial apportionment, apportionment for groundfish. This is the Council's October motion. From last year, and it's included on your e-agenda, and it describes the, the council's, um, it provides the council's written, uh, guidance for spatial apportionment of ABCs, and we've renamed those, um, BRDs, or BIRDs, biologically informed recommended distributions, instead of the subarea ABCs. So you will see in your, um, tables and the plan team's recommended tables, the updates to the formatting there to account for the name change in order to communicate that these are not true ABCs and they're spatial apportionments of that. The piece that is probably most pertinent to the SSC here is the bottom bullet, which if the council anticipates that there are any stocks or stock complexes where it it may want to deviate their TAC recommendations from the BRDs that come out of, um, the assessments and your, uh, BRD recommendations.

The council is intending to try to use our kind of two-part harvest specifications process to receive either quantitative or qualitative input from both the plan teams and the SSC BRD, and the Council has noted that useful information could include whether the previous year's BRDs or small changes from the SSC's current recommended BRDs, whether those would be likely to result in biological impacts to the stock or stock complex. So providing some information on that topic, and the only stock for which you would be handling this at this meeting is Gulf of Alaska big and longnose skates. Those are the only stocks that we have at this meeting to go through this process, and so the plan team has gone through this process with big and longnose skates, and we've included in our plan team report what we're hoping is helpful for the SSC in their discussion, and hopefully helpful for the council when it comes to tax setting. For those stocks, which will happen in December. I imagine that there, there could be some SSC questions on this topic, and I'm happy to take any now, or we can kind of address this specifically when we go through the exercise for Gulf of Alaska skates.

Question from Sherri.

Thank you for that. I actually hadn't thought of this before the meeting, but as you went through that right now, um, so the bottom part said— so there were a couple of things. The bottom part said that, um, useful information could include whether a previous year's BRD or small changes, um, would result in biological impacts. So my understanding from that is that we don't have to do both of those. We don't have to define what small is at this meeting so that we're consistent.

We can do sort of as much as we're able to in this cycle, and I'm, I'm wondering if that's also your understanding.

Through the chair, Dr. Dressel. Yes, and thank Thanks for bringing attention to that. Because we are— and the SSC specifically is really time limited this cycle because of the shutdown and all of the assessments that are all— that we're doing all assessments for all stocks this year. I think any streamlined process for looking at BRD flexibility is absolutely welcome this year, and you don't— certainly don't have to address both of those, and we can— we are hoping to have some time on the SSC's agenda in the future to try to delve into a process if the SSC chooses to kind of address this process in a little bit more detail.

Thank you. I've got actually two others. So I I believe that there was another council motion under GOA Specs from last December that named 4 stocks, 4 or 5 stocks that the council was interested in whether there was flexibility. And I guess my question is if you, if you remember which ones those were, and then also whether we should go ahead with that recommendation from last October unless we hear, unless there's new guidance from, you know, the council at this meeting and look at flexibility for only those in December. Through the Chair, Dr. Dressel.

So the motion is, that other motion I believe is also on your E-agenda, but it was speaking to Gulf of Alaska thornyheads, ruffeye blackspotted rockfish, and shortraker in addition to big and longnose skate. States. And so in December and November for the plan teams, those stocks— Shortraker, ruffye blackspotted, and thornyheads— will come up, and that discussion can be held during the November plan team and the December plan team unless the council provides any different guidance at this meeting.

All right, and then my last one is the one that I hadn't thought about, and that is, so for those 3 species in December, does the SSC need name at this meeting what they want authors to bring back? So in other words, if they want authors to bring back last year's apportionments plus any new apportionments, for most I'm guessing that— well, no, there will be updated data. So in other words, do we need to specify as part of our minutes at this meeting things to bring back for those 3 species in December? Through the chair, Dr. Dressel, That's a good question. I, I guess my first expectation was that based on what has been described in the past as far as helpful information on stock structure for those stocks, um, that council staff would work with the authors to make sure that they've included, um, available information or information that has already been included in past assessments.

Um, when they bring forward those documents in November. Um, I think the— because those stocks aren't on the plant team and SSC agendas at this meeting, um, there wasn't an expectation that there'd be an opportunity to assign any additional work to bring forward for November. I think if the SSC has specific recommendations, we can work with the center to figure out about what level of information to bring forward, if there is any beyond our, our general guidance that we already have, which is any existing information on, on stock structure or genetics for those, for those stocks. But we did not— we haven't discussed whether or not authors would be able to bring forward anything additional that they aren't currently expecting to do between now and November.

Last one, um, does that mean that authors will automatically bring forward— since it's mentioned in the motion, they would automatically bring forward— could include previous years' BRDs. Will that automatically come to us? Through the chair, Dr. Dressel. So we, we have that information and can bring that forward. Um, that's, that's easy enough for us to do.

Thanks much.

Uh, I think we will go ahead and just introduce everyone that is involved in the Joint Groundfish Plan Team report. Um, online we have all of our co-chairs, Dr. Steve Barbeau, Jimmy and Ellie, Chris Lunsford, Clay Shotwell, and Cindy Tribuzio. And then I believe we also have Dr. Gaethel and Dr. Chang online when we get to the sablefish assessment. So I will pass this off to Dr. Shotwell and just thank everyone for, for joining us here this morning.

Thank you, Sarah. Can you hear me okay? We can, thanks. Okay, great. You ready for us to start the joint team report?

Yes, please proceed. Excellent. All right, good morning, everyone. Clay Shotwell, co-chair of the Bering Sea and Aleutian Islands Plan Team. Yeah, definitely want to thank all the other co-chairs that are shown here today that Sarah introduced.

Thanks for their help and support in this meeting and also for putting together this presentation. Next slide. The Groundfish Plan Teams met in person on September 8th through 11th, and here's the list of the plan team members on the slide by team. I want to thank them all for their time and contributions during the week. To help conduct a successful meeting.

Uh, I also want to mention that this was Sophia Wasserman's and Steve Whitney's last plan team meeting. Um, I wish to extend our heartfelt appreciation for their time serving on their respective teams and also for providing their expertise in their respective disciplines. They'll both be very missed. Uh, next slide.

Joint team met on Tuesday, September 8th through 9th. The link to the agenda is on the slide if you need it. Links to the topic presentations are all in the headers of the plan team minutes. They're also in the slide headings for the rest of the presentations and following individual team presentations. At the start of the joint meeting, we discussed timing of future meetings.

The plan team meeting will be held next month on November 16th. The 20th. We also discussed potential meeting dates for the September 27 meeting, noting that there's multiple options and that the November meeting will likely be the 15th to the 19th. Both of these November meetings deal with the issues surrounding our holidays and proximity to the council meeting.

Next slide.

This is just the summary table for the topics we're going to present for this presentation. On the joint team. Um, the general organization is a brief summary of the topic followed by the plan team discussion and any recommendations we may have. Uh, the exception to this is the ESR update, um, which we're not going to review because you already received the overview by Dr. Siddin, uh, yesterday. The table also includes the presenter or lead author, the type of presentation, whether or not there are any recommendations and the total models recommended for the November where that's applicable.

Next slide.

So I'll just jump into it. Um, in the beginning, we— Dr. Melissa Howtuk provided an information-only presentation on general updates from the AFSC. They proposed a new process for adding September model updates. First showing maybe stepwise changes of any model structure adjustments that the authors are going to make, followed by adding all the available data that's available, and then finally making any model changes resulting from them adding that new data. The process— they think the process would be most useful for some of the less frequent assessments, such as our 4-year cycle stocks.

It should also reduce workload in October and potentially mitigate any model issues. With fitting new data.

Another update they provided was to simplify the standard set of projection scenarios. We have 7 right now and they're suggesting to winnow them down to 4 and to also standardize the projection model output for next year. They also suggest that authors will provide a standardized set of summary slides as part of the plan team presentation. And that should help for this presentation. For the co-chairs as they give this to the SSC in October and December.

And then they finally— they provided a table on notable reduction in age data samples for 2026. Next slide, please.

So following that presentation, the team discussed the need for data update process and protocols. And they agreed with that, and the new process they suggested. They also stated that it should be up to the author discretion on data, on when they're going to pull data and the timing of that. We all supported the standard projection scenarios and the summary slide template for the authors. We also have concerns regarding the reduction in aging capacity, and it's been reduced by 40%.

Recommendations from the team include that AFSC leadership should work with the authors to provide general protocols in the guidelines for September data retrieval and that, you know, very much support increasing AFSC aging capacity and maintaining age sample reading as possible.

Through the terror, do you want me to stop after each of these or just keep rolling?

We'll keep an eye out for questions, but if we don't stop you, go ahead and keep rolling. Thanks. Okay, thank you. Next slide.

Okay, next up, I provided an update on the ecosystem and socioeconomic profiles, or the ESPs, and our formal integration of structural causal inference and the use of dynamic structural equation models, or DSIM, in our ESP framework. We provided a brief overview of progress on the national front for ESPs and also our responses to several plan team and SSC recommendations regarding our indicator categorization, and you've heard a lot about that through the crab plan team and through some of the ESPs that you saw yesterday from Katie. We then walked through a 5-step workflow, which you see on the top right figure, on where to include structural causal theory into established structures of an ESP, um, namely the mechanism table and any associated expert-driven hypotheses. The next steps walk through building the directed acyclic graphs, or our DAGs— those are the box and arrows diagrams— testing trophic pathways in step 3, and then model validation and finally model selection in steps 4 and 5. We also put forward a suite of diagnostics for each step.

We didn't really have time to go through them specifically, but you'll see— start seeing these diagnostics in several of the ESPs that have DSym in them. I think 2 of them are in the Gulf presentation today. And then finally, we show 4 levels of application for this new workflow, from level 1, which is just building the ESP that has a mechanism table, through level 4, which is a DSym integrated within the operational stock assessment model. And you'll see that with Gulf Pollack later on. These levels map to the data availability for the stock.

And so you'll— I think, yeah, you'll see an example for a Level 3 through Gulf arrowtooth and the Level 4 in the Pollack presentation. Next slide, please.

The team discussed the timeline for integrating ESP products into the stock assessment And they included evaluating stability measure when adding new data and appropriateness of the DAG structures. The team also discussed development of DSAM prior to finalizing an ESP, but we suggested that the collaborative approach is key to developing vetted mechanism tables and associated DAGs. We also discussed limitations of staffing capacity to complete ESPs and to review the products from ESPs.

The recommendations from the team include adding a step to the ESP DSM workflow that I just showed you to document the available evidence for linkages among variables. We talked about growth and natural mortality. And for both teams to review the workflow jointly because at this time we only have these examples in the Gulf team.

But we will have examples in the Bering team soon. They also recommended expanding or creating new ESPs, that they should be supported by dedicated staff time rather than relying on the stock assessment authors.

Kalei, we have a question here. Kaylan, go ahead.

I think this is a good time to fit this in. I was a little confused. I was reading the GOA Groundfish Planting Report and in a couple sections I think twice, there's language around incorporating more socioeconomic information into the DSHEP framework. And I was wondering if you could talk about that, because I think in the past when that was presented, and actually in your— in the prior figures here, there— I would be hard-pressed to call any of it socioeconomic. It seemed like maybe connections on more of the environmental to ecological side.

Through the chair, thank you. Oh, go ahead. Oh, no, go ahead. I just say through the chair, thank you for your question.

Um, at this time, we don't have any examples of using the socioeconomic indicators in a DSAM, but we do know that the the structural causal framework sort of started with economic information. Dr. Thorsen has provided us some examples, some like preliminary examples of how to do that. We haven't pursued that yet, but I'm hopeful that, and that we will be able to in the future. And this being able to use DSSEM allows us to do that. And we also have some folks from the, our economic and social sciences team that are interested in pursuing, creating some DSM models with the indicators that we have identified so far.

I think this also allows us to include more indicators than we've thought before because we didn't really know how to use them in the past. So it kind of opens up a, a brave new world for how to use this information and then report it through the ESP. Does it help?

Maybe I can just ask a follow-up question. So it, it sounds like I'm right that the SSC hasn't seen that, and maybe the question would be, would the SSC see that at some point, or am I missing something in these, um, these reports where it's been described in more detail than I just sort of read off, or Where would I find more information on that? No, through the chair, you have not seen it yet. We're still, we're still developing those. And I didn't go through this, but we also, in the presentation to the plan team on this topic, we are starting trainings on how to use DSAM.

And in those trainings, we're including several social scientists and economists so that we can bring them into this, into what we're doing in the methodology and try to get some examples on you know, through collective brainpower on how to do this. But no, you haven't seen it yet, and I'm hopeful that within the next few years you'll start seeing examples of how to use these indicators through this framework.

We have another question from Sherry. Hi, Clay. Were you done with the ecosystem profiles slide, or are you still going? Um, I am finished if you don't have any other questions on it. Okay, I actually wanted to go back.

I'm a little slow, um, on the AFSC updates. Um, so it's like 2 slides back. Um, uh, the minutes— I know that in the joint planning, the topic came up of whether or not in the assessments, the projection scenarios, whether those were all and the minutes say that they recommended a standardized set, which I couldn't tell whether that meant keeping it as it is now or changing it. And this one says standard, which I would assume is sort of what it is now. Do you remember, did the joint plan team— was there a recommendation to change the projection scenarios that are currently in the stock assessments?

Through the chair, thank you, Dr. Russell. I don't think there was a recommendation to change the scenarios, just to reduce them. So there's 7 right now, and I believe they were suggesting to reduce them. We only really use 2, 3, 6, and 7, and so that was what we would reduce them to rather than— we're not changing them.

Okay. I can see if my co-chairs want to jump in on this. So there is a plan to drop 3 scenarios then? I believe so, if that seems reasonable. Okay, thank you very much.

And I think we also talked with GC on this to make sure we weren't dropping something that was required. Thank you.

Okay, no more questions right now. Please proceed.

Okay, thank you. The next topic we have is on the longline survey. Pat Mallica provided an update on survey logistics, also went through the methods of the survey and reviewed the standard data collections for catch, length, otoliths, temperature, and whale depredation, and then he also reviewed some special projects for sablefish. As you know, the survey design has changed to sampling the Gulf in odd years and the Bering Sea, Aleutian Islands in even years.

This resulted in this year sampling 16 stations for the Bering Sea, 14 for the Aleutian Islands, a lot of hooks fished, 32,000 links, 1,100 sablefish otoliths collected. They tagged 1,300 sablefish, 36 shortspine thornyheads, and 8 Greenland turbot. They also produced 56 temperature depth profiles, and they produce sort of a subsurface temperature indicator at about 250 meters, and that has been near average in both the Bering Sea and Aleutian Islands. Overall, I showed some of the trajectories on the right there for some of the stocks that we use in several of our assessments. Pacific cod was down in the Aleutian Islands, oceans.

It was up in the Bering Sea. Greenland turbot was down in both areas. Shortrigger was up in both areas. And ruff-eye blackspotted was up in the oceans and down in the Bering Sea. And then the teams discussed whether increasing sablefish abundance could increase hook competition and affect the RPNs for other species.

But the authors stated that the uncertainty estimates are sufficiently large that correction for hook competition may be difficult. That's all I have for the longline survey. Next slide. Oh, and then I'll pass it over to Dr. Schramm to review harvest control rules unless there's any other questions on what I've shown.

Okay, seeing no questions, go on to our HCRs. Okay, thank you. Good morning, members of the SSC. Before I get started, so this is a slightly more expanded version of, um, what we would normally report back during the Joint Plan Team report. And I'm just going to go through, um, what you saw in May, uh, the results of the council action from that.

And then I'm going to walk through, um, briefly a, a high-level summary of, of the compilation of all of your SSC comments to date that we are in the process of, um, addressing during the course of this work. And then we'll give you a, a brief snapshot of the results presented to the Groundfish Plan team. But before I get started, I do have a list of team members that are on the line. So if you have questions during the course of this or on the results, I just want to walk through. We didn't put them all on the title slide, but on the, on the line we have Dr. Kirsten Holzman, Dr. Megan Bryan, Dr. Paul Spencer, Dr. Angela Al-Bahassani, Dr. Andy Whitehouse, Dr. Melissa Haltuch, Dr. Carrie McGilliard, Dr. Jim Ianelli, and all of them are working on— and I'm sorry, and Dr. Bia Diaz as well.

So all of them are on the HCR team. They're all available for questions. I'm giving a really high-level summary, but if you want to ask more questions, we provided a much more deeper dive into sneak peek results at the groundfish plan team, but that was not the purpose of bringing this to the, to the SSC at this time. I'm mostly going to focus on on the SSC comments to date and what we're intending to bring back to you in May. We do understand that you have time on your May agenda for us to come back to you, so I'm looking for advice on what the most useful thing to bring back to you at that point.

I would just point you— the— on the slide here is the link to the Harvest Control Rule paper, Risk and Uncertainty, from the May— from May 2026 that was presented to the SSC. And one of the results of presenting this to the Council was the request by the council for a joint plan team presentation on the draft results, as well as an overview of the sablefish MSE. So that is what we took up at the joint plan team meeting in September, both of those requests. Um, so additional details, um, with respect to, uh, the council's motion then, um, continued support for this effort and for the work that's leading to, um, what we will be bringing back in May, as well as the joint plan team presentation on those draft results. The council supported the overarching objectives that staff put forward in the paper, with the understanding that as we move forward, we'll be refining those to be more specific and probably more specific by stock.

We have a priority that has been expressed by the SSC as well as endorsed by the council on the evaluations for pollock and Pacific cod in the Bering Sea and the Gulf of Alaska, as well as Gulf of— as well as Alaska sablefish. Specifically, the council endorsed a communication and engagement strategy. We had a little bit of a discussion with the Ecosystem Committee yesterday, and I'll touch on some of the engagement strategy that the Science Center will be pursuing as well as we move forward towards finalizing results by next fall. And then again, addressing the SSC recommendations throughout 2025 and 2026, specifics. We are working on having an explicit kind of like a response to the SSC like you have in a stock assessment.

That's what we'll be compiling for you for May. We don't have that at this point, which is why I'm just going to touch on the overarching SSC request, but please understand that we are in the process. We have a very long table of all of your recommendations. We've also matched them with the joint plan team recommendations, most of which are complementary. There's some few additional things from the joint plan teams the SSC, but you'll be getting a listing of all of those when we come back to you in May.

So just to touch on that, I grouped your high-level recommendations into 5 different groupings— 4 or 5— and the first one I'm just going to touch on is the grouping of harvest control rule structure and design. And just to give a sort of an overview of some of the big higher-level things that you've asked of us: testing a small type targeted subgroup of HCRs rather than all of the ones that have been brought forward. That has been the focus in, in the recent evaluations that we've looked at, specifically HCR 1, 5, 10, and 7. You asked us to compare the cap-like HCRs such as 5 and 10 against more of a flat fishing mortality HCR such as status quo, as well as the environmentally varying HCR, HCR 7, We had a discussion specific to HCR-7 at the joint plan team meeting with parameterizing that and more of a discussion of how best to do that. You did suggest that we combine HCR-5 and 10 into a single HCR since they are redundant in shape and intent.

I think we'll come back to you in May with our thoughts on that, that there are reasons to be able to tweak them separately. But again, We'll be responding to that much more explicitly when we come back to you in May. You suggested not limiting the harvest control rules to climate indicators alone, as with HCR 7, and that the framework for HCR should be flexible to tack trends, biomass trajectories, recruitment singles, risk table concerns, and socioeconomic factors, since the management risk isn't only ecosystem-driven. We're working on that. Some of those those are going to be addressed more in performance metrics than they would be in the actual HCR itself, specifically when we're talking about socioeconomic factors.

And that you favor harvest controls that are robust against a range of conditions rather than over a whole range of conditions, as opposed to— if you remember when we started talking about should we have a trigger that would move you to a different HCR, and so we've kind of stepped back from that with looking at— we are— the more, um, the more data we have on the draft results, it definitely indicates that these are species-specific. And so we may not be looking at a wholesale different suite of HCRs for all of our stocks, but just for a few targeted ones for which it makes the most sense. And therefore, it wouldn't be so much a trigger, but something that we would just be looking at a different, more robust HCR over time. And then the other 3 categories that I have these grouped into, climate and environmental scenarios. You asked that we specifically clearly define and communicate the scientific basis for warming scenarios, the most likely scenarios plus mild to extreme bracketing scenarios.

You'll see in the, the draft results how that's been addressed. Looking at timeframes that span both near-term and longer-term forecasts. And grouping specific reference points reflecting life history and heterogeneity across species. Um, with respect to modeling, you specifically requested that you review the experimental design before modeling begins, and this is going to be something that we'll touch on when we're talking about what we would bring back to you in May, such that we could maybe bring back to you some of these modeling designs before we move forward with the full analysis. Um, you asked to clearly identify what platforms are being used and where approaches diverge substantially.

We have had discussion at the joint plan team as well as with the SSC before about looking at specifically Gulf Pacific cod and trends under high warming or other scenarios. And that as we move forward in our timeframe and our timeline, that it allows both the SSC and the plan team to review both initial and final analyses, and the team came back requesting that as well. And then finally, with performance metrics and communication, We had this discussion with you, um, just this past June about, um, a long list of performance metrics that we put forward and the suggestion that we keep the public-facing ones of those few and targeted to high-level key trade-offs. I think we will continue to explore a long range of performance metrics, but we understand that when we're coming forward in a public-facing, especially in communication, that we should be targeting just a few of those. Those.

We are still working on adding economic metrics, and we do have Angela Al-Bahassani on, as well as Sarah Wise, looking at some of those metrics that are being developed, specifically with catch stability, total value over time, and total catch. We're also— and they're working on exploring social indicators and how to communicate them for decision-making. And again, as with the Council, you also have suggested that we develop a communication plan, and we are working on our communication and engagement plan for this analysis.

Just as a quick platform, this then is— whoops, I think there's a—. Hang on. Okay. So the 3 main modeling platforms, and again, this is something that we would be suggesting to come back to you in May with more details. The multi-species assessment model used to be called Seattle, now it's called Synergy.

That's being used, as well as food web modeling, and then specific to the HCR-7, an environmentally enhanced pollock model. And I, I can pause at any point if you have more questions for all the analysts. Again, we have a large group of people that are online, so I'll try to slow down a little bit.

And so I just have a high-level summary. This is the only slide I have on the, the draft —oops, I'm sorry. We do have one question. Go ahead, Franz.

Uh, yeah, thanks. Uh, I'm not sure if you were going to get into any more of those, uh, detailed comparisons here. I meant to scroll forward, but I'll ask now. Um, on the, um, the multi-species assessment model, we've seen repeatedly the single species kind of enhanced Pollock model. We've seen sort of versions of that.

The, the use of a food web model is pretty new in, you know, this kind of context and how we might deal with that in, in the harvest control rule setting. And I guess I have some questions specific to that model. Is now a good time to ask those, or should I wait until you show some of the results? Thank you. Through the chair, Dr.

Mutter, you can ask them now. Um, it may be that that's something that we're bringing back to you in May when when we get to the slide after this, it's with a suggested list of things that we would like to try to bring back to you in May. And so that might be a better time to delve into depth, but you do have the, the right people online. So if you have specific questions about the food web modeling or the synergy or the other, you certainly have the right people online to ask them of now. I'll wait until we kind of get through that part.

Sherry?

Sure, I've got two questions which I don't think you're going to hit on again, but if you do, just let me know and I'll wait. And you actually, you touched on both. One that I was going to ask, somewhere I saw that results in May were going to be final, but I've also seen them not final, and when we had asked whether it was possible to be part of the experimental design, I was going to ask if that was going to be possible given the timeline if the results in May need to be final. So just clarification. Sure, thank you, and apologies for any confusion about that.

We've had a lot of discussion about what constitutes draft results versus what constitutes final results, and I'll get to this in the next slide after this. So this is the sneak peek of the actual— of the results to date. The next step moving forward, and again, we have lots of people on the line that can, can better explain this, is to try to be engaging with the stakeholders on what the draft results are indicating in order to then finalize for— so right now the, the timeline is, is suggesting that we would be bringing final results back in the fall of 2027. And what we're working towards right now is what can we bring you in May of 2027 to help move that forward as we engage in what the Science Center is starting to look at as informal engagement with stakeholders. So rather than the council having any kind of workshops, we're trying to figure out what are the best time frames and, and already existing stakeholder meetings— Comfish and Kodiak, some cooperative research meetings that are intended to occur in the spring— so that we don't burn out stakeholders with asking too much of them as we're trying to take advantage of those existing meetings to try to begin to discuss more informally with stakeholders.

This, this planning meeting is really the first time that we were able to engage stakeholders and to try to get their thought process on what's a good time to start to bring this information to you so that we can help fold that in and have the right targeted analyses. I'll pause to see if any of my colleagues online want to augment to that discussion.

Yeah, this is Kirsten Holzman here. Can you hear me okay?

Yes, we can. Thanks. Yeah, thanks. I think Diana did a great job of summarizing that, but we are definitely in the input stage for designing some of those informal discussions. And so this discussion that we're having now is very helpful towards that, and we can revisit and adjust course as needed.

Sherry. So since we aren't seeing detailed results, that wasn't your plan at this meeting, so we'll still have time to be part of the initial review or the, um, experimental design of that if we review in May, and then stakeholders, hopefully you'll be meeting with them before May if possible in some way. Through the chair, yes. That's what our intention is. We're still working on, which is what my next slide is, but we're still working on how best to go about that coming off of what we've provided to the joint plan teams and then trying to— we understand that from a communications standpoint, this is a difficult thing to bring forward.

And so we're really trying to work on making sure that we can best communicate draft results, get feedback from the public in the most meaningful way possible. People, which is more likely to be more informal than it is in a workshop-type setting.

Go ahead, Sheri.

Thank you. You also talked about the economic metrics and the indices that you're working on, so that sounds, that sounds great. I did see in the slides to the joint plan team that it said that socioeconomic objectives were included in the FMP, but op— I don't know if I can say— operationalization in metrics is less clear. And I was wondering if— does that indicate that you're not going to include them, or that you need— there are questions that we need to answer at this meeting, I guess? Thank you.

Through the chair, no, they just sit there in development. That's all that's really meant to indicate. Both community-level metrics as well as socioeconomic metrics. And so we have a whole subset of the HCR team, Angela, Dr. Al-Bahassani, and Dr. Wise, that are both working with Kirsten Holzman and others to develop those at this point.

And so we're— what we'll get to on the next slide is what we're hoping to have, a more fleshed-out proposal on how we're engaging and moving forward to present to the groundfish plan teams at the beginning of the November meeting, which would then obviously also come to you through the plan team report. And so that's what we're aiming for, so that we have the opportunity to have stakeholders there and we can more advertise what our intentions are at that point.

Jason?

Yeah, I'm just wondering about the timeline for fall of 2027 and sort of setting expectations for stakeholders so they know where to engage at. I'm, I'm curious, you had a communications plan element that we had recommended last time. Um, I wonder if you could sort of speak to that, how the next 8 to 10 months sort of might look to help the stakeholder set expectations on engagement. Sure, it's through the chair. If, if you don't mind, if I advance one slide, it might make it a little easier.

We can go back to draft results. Because this is where we're trying to figure out how to move all these, these pieces together. So with the understanding that the idea being that the final results would be available in the fall, we're trying to move forward to what's the best use of everyone's time between now and May. So we do know that our discussion paper to bring to you in May will include our written responses, as you would have in a stock assessment, to the compilation of all of the SSCs recommendations as well as the plan team recommendations over 2025 and 2026. I would suggest, and we can back up to the results, but some of the conversation that we had at the plan team as well and with stakeholders was that right now, and I would look to those online to correct me if I'm wrong, I believe that right now the draft results do not include forward projections of MOM-6.

They include kind of a bookended short-term warming, longer-term or extreme warming, but I believe that MOM6 forecasts will be available possibly as soon as this fall. And so one of the things that has come up through the public is in order for them to really buy into looking at those kind of forecasts and projecting forward under harvest control rules, they'd like to hear more about what comprises the MOM6 and the, the the validity of those forecasts going forward. So we were— we wanted to put forward to you at the SSC, would May be an appropriate time for you to be reviewing in more depth the MOM6 forecast as well as the other modeling platforms? And this is getting back to Dr. Muter's question about the food web modeling platform. Is— would May be an appropriate time to have a more in-depth SSC review of the different modeling platforms, including MOMSIX.

So that's just a suggestion to you for your discussion, in addition to us coming back to you with the written responses to SSC requests. And then separately, we also took up the Sablefish MSE. The Sablefish MSE, and I know that Dr. Gaethel and Dr. Chung are also online, and they'll be providing you the different Sablefish reports. Report. But with respect to the sablefish MSE, they still have one more stakeholder meeting to hold, and that is— hopefully we talked a little bit last week— ideally maybe January, February.

Not sure, Dr. Cunningham can speak to this better than I. And so we might— we were trying to take these up all together, but the more that we hear from draft results, both from from Sablefish and from Klim, it's that the HCR considerations are likely to be species-specific. We aren't talking about a wholesale revamping of our tier system. We're not necessarily talking about HCRs for every single one of our stocks, and so it might just be that we do take this up iteratively. And so Sablefish might lend itself more for something that we might be able to bring back results of their stakeholder meetings and a further discussion of how, what, whether the SSC, the public, the council desires to take the research results from Sablefish and turn that into some form of management action that what you would like to see in the stock assessment for this 2027.

So right now, as far as our engagement efforts that we're doing on the council staff side, is we're trying to focus our engagement and our communication on what is robust about status info so that everybody understands what our current process is, where it's robust to environmental changes, everything that we've been doing all along with risk tables, everything that's been very progressive about it, and why it is robust and we're really just looking at some specific changes to certain stocks. And that's what we're trying to make sure that people understand as we begin to engage. Hopefully that answers some of your questions about what, what what's coming and where we're focusing. And if you want, I can back up one if you want to talk about the draft results so far, but I can pause.

Franz, go ahead. I didn't realize that was still on.

Well, are you going to go back to the previous slide and go over that? No, I'll hold off until we get through that. Okay, so sorry to do this out of order. I didn't mean to. Um, again, the just, just like a high-level, um, draft results of what we've been talking about so far, um, focusing on, on 1, 5, and 10.

And then again, we had a lot of discussion of parameterization and objectives for HCR7. Um, like I mentioned before, the draft simulations are using assumptions of status quo warming, an optimistic slight warming, um, scenario, a pessimistic increased warming scenario, and then no change. So like mean of historical conditions. Again, the MOM6 model climate predictions will be used in the future. And again, as we just were talking about, that did come up with stakeholders at the Joint Groundfish Plan Team meeting that they would like to have a better understanding of the robustness of that.

And before we are actually formally using forecasts from the, those modeling platforms. Model platforms. And that does align with the SSC recommendations on reviewing model platforms and the forcing scenarios. Again, generally, all draft results are indicating that in general, our harvest control is robust, but there might be benefits from the use of, so far, what we see in draft results from HCR5 and HCR10. And we are looking at different, so HCRs with the B20 cutoff for, uh, steller sea lion purposes, um, tend to result in higher spawning stock biomass but also higher volatility in catch with that cutoff.

Um, the cap-like harvest control rules, HCR 5 and 10, tend to stabilize catch and spawning biomass and reducing the volatility. They also result in reduced mean catch. Um, draft results for HCRs 5 and 10 seem to increase age-class diversity That is also something that you saw in the results from the sablefish MSE as well. And the HCR7 is the one that's more explicitly linked to climate and increases the volatility, but also increases catch and is more nuanced parameterization. That's what we mostly spoke about at the groundfish planning team meeting is looking at how, how best to address HCR7.

And we have some team discussion of some of that that's included in our minutes. So that's really what I had for draft results, but I'm happy to pause. And again, everyone's on the line for questions.

Franz? Uh, yeah, thanks for that. Um, I'll just ask a couple, uh, questions. Um, and, and I, I guess I kind of look forward to maybe seeing more of that in, in May. But, um, as we, uh, I mean, I guess I'll start with the MOM6, and I realize that's kind of longer term, but some of my questions relate to the NPZ output, like the predictions of phytoplankton and zooplankton from the ROMS NPZ model.

And so my first question is just, as we transition to using MOM6, Is that going to have an NPZ model component added onto it that is going to be available for both forecasts and forecasts down the line? Again, realizing this is more longer term.

Thank you. Through the chair, I'm going to pause and see if Kirsten wants to address that, if you can, Kirsten. Yeah, thank you so much, Franz, for the question. Yes, it will have an NPZ. It is different than the— there were actually a couple of different flavors of NPZ that we had previously had with the ROMS modeling platform, and the team is evaluating those alternatives for NPZ.

So there's various stages of skill testing, and I think it would be great to— I think the timing is good to bring that forward, especially in terms of some of the physical variables. There's a lot better performance on things like mixed layer depth, for example, from MOM6. But the continued evaluation around the NPZ is a good topic for discussion. And Kelly Kearney and Wei Cheng and, and Darian Pilcher and others that have been involved in the ROMS modeling are involved in that MOM6 evaluation as well. So yes, the, the short answer is there are MPZ extensions, the forecasts and, and how those work on the scale is something that is hot off the press and, and being evaluated right now.

Okay, thanks for that. Um, so my other question then is really about, since most of the model runs, or the model runs that will be seeing and that are in development, are going to be based on the ROMS NPZ And as I understand it, like right now, the, the two things that are used from that is the phytoplankton, um, phytoplankton index of some index of overall phytoplankton productivity. I'm not sure exactly what's used in the, in the RPATH model. And then the, the copepod, some form of the copepod index, a ratio of large to small copepods. Again, I, I forget the exact use in the multi-species model.

But I think we had asked about that previously, and I think there was a paper in 2025 looking at kind of validation of some of those indices. And I was just wondering, I mean, I think at the time there was, you know, kind of an acknowledgement that that the models differed pretty widely in their kind of the patterns of simulated primary production and that that would, you know, affect anything kind of upstream.

And so my first question related to that is, has there been— since then, has there been sort of additional validation of the model estimated trends in the particular index that the RPAS model is using? Its reliability in terms of capturing interannual variability, and maybe more importantly, capturing trend as we, you know, run the, run the projections. If— so the question, I guess, is, was, has there been any kind of more vetting of that since the CUNY '25 paper, I think?

Yeah, through the chair, this is Kirsten again. Thanks for the question. And actually, Diana, it might help to go back to the three model slide. I think it might be the previous one. There we go.

Yeah, if we undo the sneak peek, there we go. Now we can see. So yes, Franz, you're correct in remembering that there are different couplings for deployment connections, and that's part of why we have the suite of models that we're looking at. And in fact, there are— there is a larger suite. For example, goACclaim also has the Atlantis model.

There's different experiments in place to test the sensitivity of the model outcomes to those assumptions of coupling, and that's specifically something we want to capture and characterize as part of this effort. On the end— on the, the endpoint of this is to understand how an HCR performs given different assumptions or information about mechanistic connections between the climate and environment and stock productivity. So we're designing it to be able to capture that. That said, there have been new versions, as you mentioned, the 2025 paper of, of RPATH does couple differently than the previous ones. So productivity drives bottom-up dynamics.

Recruitment is not specifically parameterized, it's emergent out of those relationships. And there's been a considerable amount of effort to validate the higher trophic level patterns so that the model is tuned to match those higher trophic level patterns. And the validation in, in one respect is through that iterative fitting. And that's just a different modeling approach than than a statistical model like the other two on this slide, which fits statistical relationships for things like recruitment. So it gives us a good contrast to look at.

If a particular HCR performs well in all three of these, that tells us that it's robust to some of the assumptions about climate mechanisms propagating through relationships in the food web.

Yeah, thanks. That helps a lot. And that, you know, that context for just, you know, sort of having these contrasting models. And one of the reasons I ask that is because from what I saw, some of the sort of preliminary results are maybe quite different, you know, results in terms of the projections of future biomass and catches and so forth from the different modeling approaches. And so one question that's specific to the Food Web model is, like, as I understand it at this time, it, you know, it basically includes the phytoplankton time series as sort of a bottom-up driver of the system.

Are there other— in that food web model, are there other, um, does it include any other temperature or climate information besides the phytoplankton output from the NPZ model as sort of a driver of— as a bottom-up driver?

Yeah, thank you, Franz. Um, through the chair and anybody else online who is able to unmute and wants to step in, I'll My understanding is that temperature also drives the predator-prey energetics in that model. So in a way that is based on laboratory experiments and understanding of temperature effects on demand of the predator, and that, that, so that also comes in directly in those food web interactions.

Yeah, thanks. So, um, yeah, that's good to— good to know that there's both aspects then perhaps captured in, in that model as well, sort of the temperature, um, driven interactions as well as the bottom-up temperatures or the bottom-up phytoplankton, um, productivity. Um, I guess, I mean, one of the reasons I was kind asking about this is to, you know, as you go forward and do these sort of separate analyses, how are you thinking then, and you spoke to that a little bit already, just wondering how you, you know, let's say by the target date of next fall, what sort of plans you have to kind of reconcile the different sets of analyses to, you know, understand why the differences may arise, you know, if there are going to be big differences, and how do you then think about integrating results for kind of selecting a best-performing HCR in some sense? Have you kind of thought that through, how that would work once you you know, once the individual models have been completed? Yeah, through the Chair, thank you, Franz, for the question.

That is one of the— it is something that we're actively working on. So there is a— and this is in collaboration with GOA-CLIMB as well. So there's a paper where we're outlining the decision points in comparing those. The most important thing is to be able to articulate the amount of confidence associated with, or uncertainty associated with, those couplings. So it's perhaps easier if all the models find the same HCR performs well, but it's more insightful and more knowledge generating if the models have some disagreement around that.

So an example might be if status quo works very well in a climate naive and single species model framework, but doesn't work as well in the food web model. There may be feedbacks within the food web that we can articulate that kind of compound effects of that HCR and a different model— sorry, a different HCR that can, for example, expand age class diversity or preserve predator-prey interactions might perform better. And so we can look at that particular emergent difference and identify how much of it is due to the structure of the assumptions behind how species and environmental drivers interact. So that's our goal right now, is to articulate that and summarize that for next fall. But we, we would love feedback on how we're framing that.

And so that's something we could also discuss in One more from France. Yeah, thanks. Again, that, yeah, that looks, that sounds really promising, and I kind of look forward to that and maybe having some more discussions on that. I don't have any probably good ideas for our discussions today to contribute that, but that's obviously going to be fairly important. My final question My question actually relates to the caps, and I was a little hesitant to bring that even up here, but I know, you know, you've done some analyses of the effect of the ecosystem cap, and I was just wondering how at present you're thinking about the role of the ecosystem cap in informing, you know, the development, ultimately the choice of HCRs, if at all.

Is, you know, should the choice of the ACR, HCR, or the best performing HCR be informed by the ecosystem cap? Is that kind of part of these discussions too?

Through the chair, thank you for that question. Actually, I think that would be a good one for additional discussion. The modeling approach right now is that we are running these HCRs with and without a model that was developed to mimic the cap in the Bering Sea. Goa Climb is similarly evaluating different CAP levels, and so we will have some information to kind of talk about that interaction between the two. But for example, HCR-10 draws from that emergent— the CAP effects on Pollock creates a certain shape, and that shape was used to inspire HCR's— HCR-10, for example.

And so We are trying to run them. It's a, it's a, it's a lot of simulations, as you might imagine. We're running it with and without the cap effects, and the cap does have— that's part of why that was on the previous slide from our previously published work. We know that the cap has a large effect on the outcome of these simulations for some species and not for others. And so it does also depend on which species and stocks we're talking that.

One little follow-up. So thank you. So is that in the projections then? Is there sort of a feedback already now that the simulations, the model runs are, you know, the annual tax that you specify in the simulations are constrained by that cap, Yes, we have a—. There's a variety of switches you can flip on and off, and one of the switches we are providing is flip on and off of the status quo cap effects as we modeled in.

I think we did present this previously many years ago to the Council on SSC, but we had a model that had been developed of the TAC the ABC to TAC and TAC to catch process that was developed by Amanda Feag and Dr. Haney, and that, that can be applied to all of these. Um, one note is that depending on if a species increases or decreases under future projections, that can, that can also interact with the effectiveness of that cap because that's a hard cap. It's a number that doesn't change. So there are some a whole series of another discussions around that we could have. But right now what we're doing is running these with and without those cap effects.

So we can look at that. Yeah, thank you. That's it for me.

Good. Dana next.

Yeah, thanks, Diana and team. On your previous slide, you were describing the climate scenarios. And I think we've had discussions at previous meetings about how RCP 8.5 585 has sort of become a non-credible scenario. I was wondering if you're planning on carrying that forward through May.

Yeah, through the chair there.

So there is a— we should have perhaps a summary of the dialogue around RCP 585. Sorry, SSP 585. There are differing It is— all of these are supposed to be plausible but bounding. And so if you're having a discussion about which future trajectory are we most likely on, that's where a lot of the dialogue around SSP585 has, has emerged. And thankfully the trend is not towards SSP585, but the— for the purpose of understanding the bounding conditions around what might, might occur, it helps us to have those outer, outer bounds.

And so you can run the ACRs under that outer bound that's significantly warm and the outer bound that's optimistically cooler, and you can look at the relative performance. Ideally, you also would have intermediates between those so you could map more closely to some of the global trends based on recent trajectories, and that's the suite of intermediate SSPs. For example, um, the, the kind of middle of the road one that Go A Climb also includes. And so we do actually have the bounds, but we also have the intermediates as well. It becomes a lot of information to try to communicate, and so I think a good conversation for May would be, do we want to to look at how sensitive those outcomes are to different warming levels, because that's a different, um, uh, sort of way to summarize the emergent information.

Next question from Jason.

Thanks, Kirsten, and, um, all the presenters. I'm, I'm curious about the Steller sea lion control rule you mentioned, the B-20. I'm wondering If maybe you could just briefly describe how you're treating the catch in that scenario, given that that rule acts on directed fishing. Yeah, through the chair, thank you so much. Yeah, we have also in the suite of HCRs also have a switch for turning on and off the Abeminate AB20 cutoff and looking at just what that does to performance metrics like total catch or mean average catch or volatility and catch.

As you might expect, if you have a stock that is declining in trends and hits that B20, it cuts off the fishing. It has a perhaps a conservation and sea lion benefit that does cut off the fishing, so it introduces volatility. And that's exactly the types of performance metric trade-offs we'd like to get some feedback on.

Okay, I don't see any more questions right now. Go ahead.

Okay, at this point I would turn it over to Dr. Shotwell to provide have the plan team report back on this.

Thanks, Dr. Schramm. Um, yeah, this is going to be fairly short. The teams recognize the incredible amount of work done by this team and greatly appreciated the presentation and opportunity for our discussion. Uh, we did discuss the HCR7 formulation and the uncertainty in projecting environmental covariates. Um, thought that less frequently assessed stocks are a bigger concern given the longer period necessary for forecast.

We also reviewed performance metrics and discussed those and discussed the initiation of a stakeholder engagement strategy and then ways to evaluate outcomes of having alternative HCRs in place during outlier events, such as marine heatwaves. And then regarding the sablefish MSE, the teams discussed the stability of the movement and economic data limitations in the management strategy evaluation. Recommendations from the team included considering 3 additional types of performance metrics, and we listed how often the stock collapses, how quickly the stock rebuilds if below a certain threshold, and how often CATCH or ABC falls below a defined threshold or if it's zero. And then we would like the team to present a progress report on the stakeholder involvement following any engagement strategy they start. I think that was maybe following the May meeting.

And then the team would like a simulation of conditions similar to this, to the something like the marine heatwave event to determine performance of the HDRs at the extremes, if that's possible.

And I think that's all we have.

Question from Sherry. Thanks, Clay. I think this question may be for Diana and Crew. I know that in previous SSC minutes we've mentioned the marine heatwave and the retrospective perspective, and we even mentioned it as sort of a simple way to look at things. And I know that it also came up at the joint plan team meeting, you know, this recommendation here to look at it if possible.

I did understand Kirsten's response that it's a little— it's tricky to look at something that happened in the past because if you're looking at a time series, you're also looking at the impact of what actually happened. And so I'm curious if you had But I'm curious to hear your thoughts on how, if and how, something like that might be possible now that you've been faced with it a couple of times and you might have had a little time to think about it since the joint plan team. Thank you, through the chair, and I'll turn to you in a second, Kirsten. We have been talking about it since the joint plan team meeting. You know, our initial response is that we really can't do that in any way, but understanding that that's really where the rubber meets the road.

That the public and the stakeholders are requesting, how we can formulate that. We've been trying to talk as a team as ways that we might go about that. And I don't know, Kerr, if you want to provide a little bit more on kind of our discussions that we've had since then. Yes, through the Chair, it was such an excellent articulation of the needed information. And so what we've been doing is setting it up so that we can emulate marine heatwave events and then and look at the performance through those marine heatwave events.

And we have, we have capacity to do that with the current modeling framework that will allow us to simulate a marine heatwave-like event. And then those are at different scales, so we can see how things would progress through it with the HCR. So we've been using it to help frame drilling down into some particular aspects of the the modeling sets where we have those emergent patterns.

And a follow-up question to you.

Yep, so that totally, that makes sense. I know that it sounded, you know, we've heard from the public that for them one way to sort of get a handle or a hold on what difference this would make would be to look at sort of how things might have turned out differently. If you're simulating an event like that, is something like that possible where it might be tractable for people, or will it need to be more like a forward simulation, which seems more possible from a programming perspective but might be a little bit harder to grasp?

Through the chair, yeah, thank you. I think that's actually a very good delineation. So there's the near term, what we can do relatively quickly with what what we've built so far. The true counterfactual is complex, and so it would involve— we talked in a couple other slides about the social and economic performance metrics. We would want to make sure we have that connection between catch that comes in, how it propagates through community stability, and some scenarios or some mechanistic linkage for how that feeds back into how catch is executed in following years.

The other piece of it is the spatial distribution that can happen under these different marine heatwave events also affects accessibility to the resource, access to the stock, the distance that fishers have to travel, and the cost of fuel. So all of those things need to be built in order to answer the full complex counterfactual. It's a, it's a very important question, and I think really understanding would something we put in place in the future help prevent what we experienced before is exactly the right question to ask. There's just a couple of faster and more in-depth ways to answer that question. So we're trying to build out the capacity to do both, but answer it perhaps with what we have in hand first and then build out from there the pieces that that might be needed for that full answer to the question.

I hope that makes sense. Yeah, appreciate that.

I do have another one. This may be more for Diana.

That was— let's see, just a second, let me scroll down.

So in June when we talked, we talked about the SSC supported having different timelines for crab analyses, recognizing that the harvest control rules that are in place and what's used for tax setting would sort of need to be worked on, and so there's extra steps there. I'm wondering if you have an idea or a timeline for these next steps. I recognize that under the IRA funding, things are in a rush, and so I'm kind of curious in how CRAB might fit in. We, just as a tiny bit of background, we heard from the CIU review in Tanner some topics about harvest control rules, and so it's come up in different contexts. So I'm curious.

Sure, thank you. Through the chair, I guess two, two different things. One The modeling platforms do have crab built into them, so we're not, we're not really losing anything yet by not focusing on it. And I think the way we've been trying to think about it is that as we go through this process for certain groundfish stocks, I think we'll learn a lot more about how best to structure these analyses that will, when we circle back to crab, help us, you know, if things look more promising or if we're still running these analyses and we can still include red king crab in them. We still have them built in for Tanner crab and for snow crab.

So they're not, they're not getting lost in the modeling aspect of it. It's more the ability of the plan team and the SSC from the crab perspective to have the time and energy to focus on that with all the host of other issues, including hybrids, which has been one of the things that has slowed down the desire to look at harvest control rules is that there's a lot of other issues, as I understand it, in front of the crab team and the crab authors, besides trying to say, how do we take what you're using as an ABC control rule for all intents and purposes as a percentage, and how do we model that against what are some other harvest control rules that might be more adequate for crab? So I guess my answer to that would be that we're certainly intending to circle back with CRAB at the appropriate time when it seems like the team wants us to participate more in trying to structure their timeline going forward. So it's, it's not lost, it just seems that there's other things that have risen to more importance in the CRAB world than tackling this particular aspect of it. But again, we're happy to circle back —any point.

Thank you. The last that I think that we— the last that I recall was I think we had the February meeting where we both had both groundfish and crab, and crab was very interested, and they actually wanted to stay on the same timeline. And so that's the last that we heard was that there was real interest in this. Fully recognize there's only one co-chair now, but I'm curious, did they— have you received the feedback from the crab plant the team that they are— that this is not a priority for them? Through the chair, and again, I don't want to look behind me in case someone's going to differ with me, my understanding was that the reason why we were not talking to the CRAB team at their May meeting before we were presenting this paper to the SSC and to the council was specifically that as a priority measure it was not necessarily— I don't want to, again, I don't want to speak for the crab team.

I think that we're realizing the extent of this analysis and that we have to prioritize some aspects of it forward. And so for us, where we're at right now, it is easier to prioritize groundfish, understanding that we still want to circle back to crab, but there different issues for crab that we still have to figure out before— right now we're looking at how do you compare a Tier 3 control rule for groundfish against HCRs 5, 10, 7. For crab, we have an OFL control rule and we have a buffer system for ABC. We don't have an ABC control rule for crab. For which we can compare as status quo.

So we have to make a lot more decisions for crab than we currently have for groundfish, and I think when the team has time for us to come back to them, I think we should begin the conversation of what do we assume is the status quo for crab for comparison as HCR-1 is for groundfish as a Tier 3 control. I don't know if that makes sense, but I, again, I don't want to speak for them that they they're deprioritizing us. I think that the crab team is very interested in, in being a consideration in this. I think we're just trying to take things in the same way that I personally was hoping we could do all of the stocks together at one time. I think we're realizing we're going to have to do them separately, and so in the same manner, we're going to have to tackle crab and the issues for crab separately from what we're looking forward to for groundfish.

Great, thank you. Um, I've kind of got a two-part question, and the first part is related to this, but I see as part of the council motion that they're asking for the scalability of work under different budget scenarios. Um, and so I guess I have two questions on that, and I can repeat them if you want, but One of them is, one of them is sort of how are you thinking of doing scalability? So in other words, do you have a vision? I recognize that they suggested this in June and may not be figured out, but it— how does that fit in?

So in other words, would we cut out certain harvest control options. Would we cut out different species? Would you drop crab, like, completely? I'm wondering if you have information on that, and I ask it because it was sort of linked to crab. But the other one that the council mentioned was given different data scenarios with surveys changing, and I know that the SSC raised, uh, quite a while ago suggesting that you do a whole lot of work on different data scenarios, and the response at that time was, we're going to use the current, um, because I realized that opened Pandora's box.

But I'm curious if you understood what the council meant about the surveys and your response, both for the surveys and then how to approach the budget.

Um, through the chair, thanks. I'm not sure that I'm going to have an answer for you. It has been an ongoing concern that we've discussed as a team. Um, as you're all aware, resources can become more limiting. We don't necessarily know, and I, I guess I'm going to stumble a little here, so Kirsten, I might look to you to maybe provide the more of the big picture viewpoint of where we see the overall team and the resources that we have available under the these certain dwindling resources?

If I'm not putting you too much on the spot, I apologize.

Through the Chair, with respect to crab, I assume is the context. Yeah, I think we, we had to necessarily prioritize the initial set of runs. We do, as Diana had said, we do still have crab. They're in the food web models, they're in the larger modeling suites, and so if it becomes something that needs to be prioritized, that's something we could discuss and circle back and, and provide a timeline for that.

Kirsten, I think it's a broader issue. I don't think it's specifically related to CRAB. I think it's more related to specifically in the council's motion in June. It is how is our work and our effort moving forward affected by the potential for dwindling resources resources and team members and contracts going for— I mean, if I'm assuming that that's more the context that you're talking about, as if financial resources become a problem for us going forward. Thanks, Diana.

And through the chair, yeah, absolutely. The, the— there's various stages of this that depend on survey information. For validation. It requires— if individuals are doing 3 or 4 jobs instead of 1, the progress continues. It just continues slower at a rate that's spread across multiple daily needs.

And so there we're constantly trying to continue to provide best available science, um, on reduced resources and Sometimes that turns into delays in delivery. And so I hope that that helps. We're trying to make sure we prioritize what's needed most and still provide that information. But it is a— it's an ongoing challenge.

Thank you, Kristen and Diana. I actually had not fully understood what the council was getting at in that motion. I thought they were asking you to look at scenarios if different surveys were missing because they mentioned surveys. So that's why I was wondering in a scenario-based how that would affect you. I realize now that it was how does budgets affect you, which is an incredibly broad question, and thank you for the response that you gave.

Through the Chair, just to follow up, I don't think that you're wrong and that I think they were also asking about what happens under different survey moving forward. And maybe I hadn't thought of this before, but maybe that is something that we can come back in May as part of that discussion paper to, as a team, give more thought to, okay, these are all the things that we're assuming going into these models, these data streams. Let's look at the— let's look at some worst-case scenarioing, because I think that is also something that the public is asking us for, the stakeholders are asking for. Is how are we supposed to have a lot of confidence in your forecasting forward if we know that certain pieces of data might not be available in the future? So I do think the council was getting at that too.

And as a team, we haven't really done that sort of scenarioing, but I, I think that's something that we could probably put some time and effort towards and, and maybe have a, a more succinct answer to you when we come back in May. Hopefully I'm not overstepping Kirsten and the rest of my colleagues that are online here. Great, thank you very much. Yeah, whatever you have at that time.

Franz? Uh, yeah, just a separate question, but, um, you, you mentioned here again, uh, sort of the discussions on transparency earlier. Uh, those are often kind of separate from evaluating the HCRs, and I was just wondering, I was trying to look up the full list of performance metrics. Do we actually have a performance metric that measures transparency or simplicity in some way? And have you had discussions around that, if that even makes sense?

Um, through the chair, Dr. Mader, we have an objective. Um, so one of the overarching objectives from the council is directly related to transparency. Um, we are in the process, and then part of the engagement that we are hoping to have with the public is to refine some of those objectives so that they, they can be more measurable. And so we haven't tackled the transparency one yet, but that is the intention, is looking at all of those objectives and sort of matching them to metrics that would allow you to, to, um, have some way of measuring that. And so transparency would by nature be part of that.

Okay, seeing no further questions, go ahead. I think that we are moving on to sablefish now then, I believe, and we would be switching to Dr. Gaethje's presentation online.

All right, can you hear me? We can. Welcome and thanks. Yep, no problem. And you're seeing my slides?

We can see the slides. Your audio is a little bit weak, so if you can be closer to your mic, that would Great, all right, is this a little bit better? Yes, it is, thank you. All right, perfect. All right, let's get this party started.

So I'm going to talk about the Sablefish model changes and the CIE review that happened this year, and I just want to, you know, point out Matt Chang, who is a key part of the Sablefish Assessment Team A lot of this work wouldn't be possible without him, especially all the work that he's put into developing the Spork stock assessment package.

So just some quick background. So the last actual operational assessment, like for most species, was in 2024. That model was a bespoke ADMB model. So that was model number 23.5. So in September in 2025, we came with a model development document and kind of introduced the SPORK platform, which is an RTMB-based stock assessment package.

And Matt did a really nice kind of this one-to-one model bridging between the ADMB model and the SPORK model, showing that the kind of the uncertainty and the parameter estimates, time series estimates, were all pretty much spot on between between the two models. So the plan team and SSC recommended moving forward to use this SPORC RTMD model. So for 2026, all the model changes just stayed within SPORC, and so we haven't gone back to the ADMD model. So I just wanted to point that out. And then kind of the important thing that I think has gotten overlooked is that in 2025, because we didn't adopt what the new model would have been, which which was what we were proposing was Model 2512.

Again, that was kind of recommended in September. Everyone looked at it, everyone reviewed it, um, but because we had the government shutdown, we didn't actually, um, get to use that model operationally. And so the key reason I point this out is that the quotas from that 2512 model were actually about a 17% decrease from the old 23.5 model, at least for 2025. '25. Pretty similar, I think, for 2026.

And so I just wanted to point that out because kind of some foreshadowing that people keep that in the back of their mind of what would have happened in 2025 if we had used this model. So for 2026, we did a lot of model development and kind of had a CIE review. So anyone that's looked at the documents, we have a whole slew of documents We essentially made an off-cycle draft SAFE report using all the data through 2025, presenting two new models, a single region model and the spatially implicit fleets as areas model, as well as Matt produced the outputs for the five region and three region spatial models. There's another appendix on some closed-loop simulations, and then there's the three Center for Independent expert reviewer reports that were provided. And then the final kind of summary September document that I produced kind of summarizes all of these changes, this model development, summarizes the CIE recommendations and reviewer reports, and also touches on a slight change in apportionment, which I'll get to at the end.

So just to Just to harp on this, because this is kind of the fundamental issue that we're dealing with, is that, like Clay mentioned, the longline survey design changed in 2025. Again, we didn't have a survey in 2024. And so we went from having the Gulf of Alaska surveyed annually to having it surveyed only in odd years starting in 2025. And then switching from the BSAI being sampled in odd and even years to the BSAI by being surveyed together only in even years. And really the issue here is that before we were sampling, you know, upwards of 75-80% of the sailfish stock across all of Alaska, but switching to this new survey design, we're really only getting about 50% of the stock being sampled in any given year.

And so the issue for the assessment is really that we have to make this sort of imputation to create synoptic longline survey index to put into the single region model to have a constant kind of scale of the population from the index. And, you know, in the past, I think historically people viewed it as, you know, the Gulf of Alaska is, you know, where most of the abundance is, where most of the biomass is. And so, you know, when you add in either Bering Sea or Aleutian Islands, you're getting a pretty good idea of what the size of the population is. So doing the imputation for either the Bering Sea or the Aleutians wasn't too big of a deal. But now, you know, it really causes some heartburn when you start imputing, you know, 50% of your index values.

And really, the other issue to think about there is— so on the lower left are the regional indices from the Longline Survey. And so, yeah, like, historically, the Gulf of Alaska was the bulk of the population. And so you're getting a good kind of idea of the size of the population from the index with those historical Gulf of Alaska surveys. But more recently, a lot of the abundance and biomass is becoming from the Bering Sea and Aleutians. So it causes this kind of new dynamic.

And then I think the other thing to keep in mind is that because we didn't have the survey in 2024, when we developed models using the 2025 data. So all of the models I'll present have data through the end of 2025. So these are the first kind of assessments that we're presenting that you're missing the longline survey age comps. And so obviously those are key indicators of recruitment. And so when we do the 2025 models and start looking at biomass and then the ABC recommendations, you know, that's a key part of the information that's missing is this information on recent recruitment.

And this is the first set of kind of model runs that don't have a year of longline survey age comps. Simultaneously, we've kind of been scaling down because of the age reader capacity issues. We've kind of reduced the number of fishing, longline fishery or fixed gear fishery age comps that are going into the assessment. So I'll touch on those later on in the presentation, but definitely things to keep in mind. We have a question here, Martin.

Go ahead. Yeah, yeah, thanks, um, um, Dana. Um, looking at the figure on the bottom left, um, at least it appears to me that, that, I mean, you've seen a really strong increase in all, all three areas, but it looks like there's been a sort of disproportionate increase in the Aleutian Islands and the Bering Sea relative to the Gulf of Alaska. And is that the case? And have you been able— have you looked at that in any sort of quantitative way?

I mean, it's definitely the case. In terms of the relative increase. I don't know exactly what you mean by in a quantitative way. I mean, the indices are the indices. They are quantitative.

I guess I just need more information on what you, what you mean by that. Oh, I, I, my first take would be something really ad hoc is that we'd scale all, all three of them to average to 1 over the same time period and just plot them over the top just to get a better gauge of how the relative trends have differed in the— yeah, the different areas. But that, that'd be a super simple thing to do, and maybe we can get enough from the particular figure here just to understand that case anyways. So yeah, these standardized indices are available in the, like, the SAFE document. There's a figure of the raw indices that are shown here, the RPNs, versus scaled to, um, the mean for each of the, um, for each of the regions, which gives you an idea of how they're increasing, yeah, relatively.

So hopefully that's what you might imply. I don't know what figure number it is, but it's in the— that SAFE document.

Um, so yeah, so essentially we We've spent a lot of time starting back in probably 2024 just trying to think of ways— we knew that the longline survey design was changing. We knew it had some big implications, and we were just trying to brainstorm ideas of how better to handle the data we have and put this into a model and not be relying on a lot of imputed data that kind of gave us that heartburn like I talked about. So kind of the best idea that we had and one that's been used in a lot of other fisheries or a lot of other species, I won't say a lot, a number of other species is what's called the spatially implicit fleets as areas model. So essentially you can disaggregate surveys or fleets by their kind of regional structure and then do a spatially implicit model and fit those disaggregated indices and age comps explicitly within a single population, single region model. The key benefit here is that, like I said, you don't need to impute any data.

You're fitting the regional indices and comps explicitly, so you don't have to combine them into a single value and, you know, account for that scaling issue that you'd have if you do a single index. The other benefit is that it really helps account for some of the regional heterogeneity availability issues, and also implicitly account for movement through the fleet structure through the selectivity estimates. And so it really adds like a kind of a limited amount of complex complexity because you don't have to go to a full spatially explicit model where you start needing to estimate movement. And if you're fitting tagging data, you know, you have issues with some of the parameters of the tagging data like tag reporting rates and things like that. The limitations are that you kind of need to adjust how you interpret the dynamics of some of the model outputs explicitly.

The selectivity estimates become a little bit more confusing because they're not just gear selectivity. They're now kind of accounting for availability across regions of different age— different cohorts and things like that. And then the other issue is obviously as you start partitioning data, you reduce the sample sizes. Sablefish is already kind of sample size limited because we're talking about the entire Alaska population. We have two fishing fleets, the fixed gear fleet, the trawl fleet, and so when you start doing regional kind of breaking up into the regional indices and the regional, especially the age comps, the sample sizes definitely start getting a little bit thin.

And on top of that, you have the sex-specific models, so you're already splitting out by sex. So yeah, but in the end, you know, a spatially implicit model for sailfish I think is helpful because it lets us fit the raw data explicitly, mainly the longline survey indices and age comps, without fitting the imputed index values. And so that's the key part there, that last bullet, which I probably harp on too much in this presentation, but that's really what we wanted to get to. We have a couple questions here in the room. Fabio, go ahead.

Yep. Okay, thank you, Mr. Chairman. Then I have a question regarding the, the regional disaggregation. So does the regional disaggregation of the age compositions in the fleet-as-era model suggest that the conflict between the composition data and the Long Line Survey indices is spatially structured, particularly given the stronger recent shift towards younger spatially explicit fish in the Bering Sea and Aleutian Islands? Is it— so the conflict between fitting the age comps and fitting the longline survey indices, is that a spatially explicit conflict?

I don't—. I mean, I think it's just fundamental. I mean, I think, like, you look at each region and there is a conflict between fitting the index and the age comp simultaneously. And so I guess in that regard, yes, I think it is spatially explicit, but it occurs whether or not you do a spatial— like a full spatial model, which Matt develops but I won't talk about today, whether we do the fleets as areas model or if we just aggregate everything into the single region model. That conflict is just throughout all of these models.

So I hope that answers the question.

And a question from Martin.

Yeah, thanks.

Specifically about the splitting out the survey into the 3 areas, and it sort of follows up on my previous question. The way I understand a fleet is area's model for the survey, that some— you're essentially would, for each of those areas, you'd be estimating a catchability and then a selectivity curve. But the assumption would be that for each of those area surveys, as interpreted by those, the selectivity and catchability, that the stock, you would be assuming that the survey is proportional to overall abundance abundance. Um, and, and that seems like, given what we've seen about redistribution, that may be a fairly strong assumption. So I'm just wondering, under the list of limitations, if you would consider that to be a limitation of the fleets' area approach for specific— specifically for surveys.

It's not really an issue for the fisheries, but when you're sort of driving a survey and relating it to overall abundance, it seems like that might be a fairly strong assumption.

Yeah. So when we started this, I did have that concern, and I was wondering how the model would kind of handle that, essentially scaling the Gulf of Alaska index or the Bering Sea index to the entire population. And essentially what happens is the Qs in the model are completely freely estimated. So like you said, each kind of— now each fleet, which is each regional index, survey index, has its own Q that's estimated. And essentially what you see is that the Q adjusts for the size of that index to kind of get the total population scale.

And so there are kind of differences in those Q estimates. That implicitly account for kind of what you're seeing in the index versus what the model wants the complete scale to be. So I mean, I think it's definitely a concern to think about, but I think it does a pretty good job of kind of interpreting those different regional indices and scaling to the entire population size. And I did do— I was kind of worried about that, so I did some sensitivity runs like like taking and only using the Gulf of Alaska survey data, put that into the Fleeces Areas Model, you know, have it estimate Q and H comps and see, like, what happens. And essentially, what it does is it shrinks the size of the population and basically predicts an ABC value that is on par with the Gulf of Alaska ABC from the full Fleeces Areas Model.

Model after you do the apportionment. So I think it's doing a pretty good job of internal scaling, but, you know, it's definitely one of the uncertainties for sure.

Yeah, just—. Thanks. Yeah, that's helpful. I guess the real sort of stickler for me is if we've seen something like an expansion of the stock into the Bering Sea, and the Aleutian Islands, whether that would sort of— there'd be sort of a time change in how that Q relates to the overall population. Because if it sort of built up, so it's like the catchability wouldn't necessarily be stable over the whole time period of the assessment model because of that sort of disproportionate increase in the Aleutian Islands and the Bering Sea.

And it seems like that's probably the fundamental concern about the fleets as areas sort of approach.

Um, well, so yeah, we've had a lot of conversations about that. And, you know, when you start getting time varying catchability, That's kind of a slippery slope. And if you think that's happening, I think in the single region model you'd also need a time-varying catchability. In my mind, we partially take care of that through the time-varying selectivity aspect. So we have time blocks in there to account for differences in availability of different age classes over time, which I think is more of what's happening.

It's just— it's hard to— say exactly because there's not really an indication necessarily that more fish are moving into the Bering Sea and Aleutian Islands. It's more of an indication that there's more recruitment in those regions and that when you look at the spatial model, those fish, once they, once fish get to 6, 7, 8 in the Bering Sea, Aleutian Islands, they tend actually to move out of those regions and back into the Gulf. And so There's some conflicting information, I think, in terms of that aspect. I also think if you started using time-varying catchability, it would decrease the size of the population quite a bit because you're saying you're catching more in the survey instead of that being fundamentally an increase in recruitment. So I don't know, there's, yeah, there's uncertainty.

Okay, that's it for our questions for the moment. Go ahead and proceed, please. Okay.

So, yeah, we did a pretty extensive model development, kind of trying to figure out what the best structure for the fleets as areas model would be, basically starting with the minimum complexity, which is how to— being able to handle the 3 longline survey regional fleets so that we didn't have to do any data imputation So we also, once we had that structure, we then looked at various aspects of increased complexity, mostly through spatially disaggregating the fishery fleets, allowing for alternative selectivity parameterizations like dome selectivity, and also a little bit into time blocks and whatnot. I won't get into that. We basically took, I won't get into it too much, but we basically had a 3 kind of stage approach approach. The first one was just your basic simulation cross-testing to look at which fleets' scenarios' parameterizations were robust. The underlying— I should say the underlying operating model for all of the simulations was a conditioned 5-region spatially explicit operating model that was based on the 5-region assessment model that Matt developed.

So then we apply through our basic simulation approach, just apply different fleets-as-areas structures just to see what models worked, which were robust.

We went from there to then applying some of the more robust fleets-as-areas models to the real-world data, ended up with, I don't know, 100, 150 different model parameterizations, saw what really the key there was, what was stable, what we thought could be used for operational management. Assessment, and then kind of developed some of the better models from that and then put them into the spatially explicit closed-loop simulation analysis to really look at the general performance and what the implications were for using different structures in the assessment. So the figure on the— in the middle on the right there just shows the actual realized spawning stock biomass from these closed-loop simulations when we applied each of the 4 different different model structures. So the black line is the single region model that uses the carryover index like we do in the one region model that I'll show in a second. The yellow line is the fleets as areas model.

And then the blue line is a spatially explicit three region model. And then the five region model is not really a model. It's just using the results from the OM to simulate forward. And then kind of what was a little bit eye-opening on the far right is then the realized the standardized error in the spawning stock biomass over time for different implementations of each of those assessment models in the feedback period of the closed-loop simulation. So as you add more data moving forward in time, you know, what does the estimate of spawning stock biomass, and here the relative error in that estimate, look like?

So the top panel is the single region model, and the middle panel is the Fleets of Xarias. The bottom panel is the three region, and you can really see the instability there. So we kind kind of moved away from that almost immediately. But the real take-home point for us was that one, it appeared that the single region model was positively biased, a lot more so than the fleets-as-areas model, and it was also relatively more unstable in terms of the time series estimates moving forward, which I kind of envision as a pseudo-retrospective analysis of the assessment model. Model, whereas the Fleet Cesárea's model seemed a lot more stable.

It was less biased. And so this was kind of enlightening and I think helpful for making some final model decisions. In the end, we kind of weighed all these stages, all these results, really looked at stability, parsimony, and performance when trying to choose a final model. And in the end, we wound up with kind of going back to the kind of the most basic spatially implicit model, which was just disaggregating the three longline survey fleets by regions as well as the age comps and using that one.

So what was presented to the CIE and was also presented in the September document are these two kind of updated models. So there's 26.1 one region, which is an updated version of the single region model model. So, taking Model 25.12 from last year and adding some changes to the data inputs, improving some of the parameterization, and really the big one is moving to this carryover longline survey index. Essentially, the old approach for the developing the kind of aggregate single region index involved taking the growth or the change in the Gulf of Alaska index and applying that percent change to either the Bering Sea or the Aleutian Islands in the off year for either of those and extrapolating that index and then summing across all the regions. Because we don't survey the Gulf of Alaska any longer every year, we can't use that approach.

So, the carryover just assumes that if a region wasn't surveyed in a given year, we use the value from the previous time it was surveyed and just hold it constant for the year it's not surveyed. We have a question here. Martin, go ahead. Yes, sorry, Dan. I actually had some questions about the closed-loop simulation, and I don't— I see you decided not to talk about it.

It was in the plan team figures, and obviously you provided a document, so I think it's okay for me to go into that a bit, but I don't know quite how to do it because it relates to the figure that you had that showed the relative bias for the different model configurations and then showed both your preferred model which involved both asymptotic fisheries and asymptotic surveys, as well as a bunch of other combinations. And I mean, would it be possible for you to bring that up so I can speak to it? Or I'm not quite sure what we can do here, but I think it's an important point that I'd like to make. So I think you're talking about the cross-testing, the simulation cross-testing with the box plot. Yeah, yeah.

Figure 3 of the closed-loop simulation document. Yeah, I don't know if I have it in this presentation, to be honest. Yeah, it's not, but I could— it's easier, I could just share my screen, but—.

Yeah, I mean, that might be the easiest at this point.

That may take a minute. Is it— would it be possible to just ask the question without the visual? No, it's important. Okay, sorry. No, it's just— I mean, it's, it's kind of a key, key thing, and it's not going to be obvious unless I can point to it.

Um, if it's in— is it in one of the plan team presentations? Yes. Then maybe staff could queue that up behind the scenes. We could continue on and then loop back to it once it's queued up.

Let's, let's do that. Dan, why don't you continue now, and when it's queued up, we'll loop back and put it on the screen and revisit the question.

Okay, I could ask an easier question now, if that's okay. Okay, go ahead. Um, no, it's, it's the sort of transition. It's on the next slide, I think, Dan.

Uh, yeah, the transition from the model that was accepted for use in October of 25 to the model that you're calling 26.1, and it's— it, um, you know, I have a list of probably, um, 15 or so model changes, um, from those— between those two models, and I wondered if you could sort of step through what those changes were, as well as, you know, what the impact was in that transition, because we do need to sort of have a set of steps that take us to these models that are being considered here. Well, I was told to make a 10-minute presentation, so I don't have all that stuff. I can— I have a bunch of extra slides, and I could go through all of that. I just don't know. I mean, it's going to take a while.

No, just, just sort of summarize what, what, what it was and what the overall impact was.

Um, let's see, where are they?

I don't even know if it happens here. Yep, so this is the full list of changes.

So yeah, I mean, there's quite a few. I'd say like the first one was updating some of the priors on selectivity and stuff to make it a little bit more stable.

Then we kind of used an input sample size that scaled the input sample sizes going into the age composition and length composition data to get some information on the intra— interannual variation in those. From then, the C was moving to a logistic function for the trawl. Fishery that was formerly a gamma dome that didn't really make sense given the data, the length composition, length composition data we had. From there, we updated biologies. So the model utilizes an externally estimated length and weight curve.

The last time it was updated was 2021, so we updated that. Then for E, we updated the estimation of natural mortality, basically just allowing it to estimate sex-specific natural mortality. Previously, it was a single-sex estimate with a prior, so now we just use the same prior but estimate sex-specific. In F, we removed the CPUE index, that's no longer being updated. In G, we removed the remaining length composition data for the NOAA longline survey in the fixed gear fishery.

2512 Removed length data for those when we had age comps, but it left in about 5 years of length composition data from roughly 1990 to 1996 because there's no age comps there. We found a lot of model tension on trying to fit the length comps and the age comps, and so those were removed. H switched the longline survey building approach to the carryover method that I just described. And then I— and I, we then removed the Western Aleutian Island extrapolated stations from the longline survey index.

And then— so this figure shows what the indices look like. Depending on how they're created. So, the green line is essentially switching from that Gulf of Alaska extrapolation to the carryover approach. Doesn't have that much of an impact, except in certain years. And then the blue line is then taking out the Western Aleutian Islands, which is a bigger impact in the recent years because the Aleutian Islands have a disproportionate increase, so it has a higher proportion of the total index.

So that has a bigger impact.

I thought I had the— I thought I had the SSB ones, but maybe not. Sorry, I didn't have these organized. So yeah, I don't think I have a figure in this presentation of the implications of each of those specifically, unfortunately.

So that's about all I have for that.

Okay, thanks. That's actually pretty helpful. So you'd characterize those as mostly— I mean, there's some important changes, but there's all— mostly it's sort of bookkeeping and sort of modernizing this, making it more consistent and streamlining the assessment. I mean, that's kind of the way to understand what you were trying to do here? Yeah, I think so.

I mean, for the most part, it was— a lot of it was recognizing that we had some model tension in a lot of the different data sources and kind of simplifying the approach a little bit. Like I said, the length comps were one that was really causing some instability, and so removing that that kind of helped increase stability.

The longline survey index changes were a necessity, essentially. But yeah, I think kind of bookkeeping and improving the assessment is how I look at it.

And I guess maybe to follow up a bit on the carryover approach, is that— did that have I guess historically it would have had less of an impact, but moving forward, potentially would have a larger impact. What sort of concerns are there generally about the carryover approach?

Well, I mean, larger impact versus what? The problem— I mean, the problem going forward is we can't do the Gulf of Alaska extrapolation approach. So there's no fundamental comparison. And I did an analysis in the 2025 document of what the potential impacts of using the carryover might be. And basically, I think the primary issue is that if we get to the top of the trajectory, you know, sablefish population starts to decrease and the longline survey index starts to decrease.

In the year or two after that increase, the carryover, which is using, you know, the last value for an index, is probably going to be too high. So we're going to have this lag in trying to see the decrease. We're going to have a lag in seeing when the population is decreasing. So we're going to— if you're using the single region index with the carryover and putting it into a single region model, you're probably going to be overestimating the ABC for a number of years before everything kind of catches up. And you're just going to keep overestimating because you're using the carryover, which is keeping the non-surveyed regions constant but the population's declining, right?

So I think fundamentally that's the most— from a kind of risk perspective, that's probably the biggest issue with that one, aside from just imputing data, which which obviously gives me a little bit of heartburn using imputed data in an assessment.

Okay, I think we should move on at the moment, and we'll get back to questions again. So we'll let you proceed, Dan. Okay. Yeah, so those— I didn't even talk about— so the Fleets as Areas model, I think at this point people get the picture. We split the longline survey into the 3 regions: Bering Sea, Aleutian Islands, Gulf of Alaska.

We estimate separate selectivity and catchabilities for each of those. We fit the index and the age comps independently for each of those regions, but it's still, like Martin pointed out, still a single population, so the data applies to the whole population, and that probably is the bigger uncertainty there.

But it doesn't rely on any imputed data. Data. So all of this was reviewed by the CIE. I mentioned those like 5 or 6 documents. Those were all presented to the CIE. We had a 3-day review in person, some good discussion.

In the end, they kind of recommended the Fleets of Zarias model as the best scientific information available. They, you know, they looked at the SPORC code base, and they seemed pretty satisfied with how it was designed. And in the end, they really, really harped on the issue that the sailfish assessment's kind of limited by the available data, not by any of the individual modeling structures or approaches. Basically, they pointed out that you can't model your way out of the data limitations. Like, more complex models, more complex statistical approaches aren't going to resolve the issue that we have limited age age composition data, and we have a survey that's been truncated in the recent years.

And so they really, really pointed out that we were kind of at the bare minimum for otolith reading for sablefish, given that one, it's an Alaska-wide population, two, we need to have a sex-specific model, and three, you know, we're losing data consistently over the last 5 or 6 years. And I'll show a slide later on that in terms of the fishery age comps, which have been reduced quite a bit recently. And then they had the same comments about the survey, really, is that we can't— if we truncate the survey any further, we're probably really not going to be able to run this type of Alaska-wide assessment anymore.

So yeah, just— Yep. Mark, go ahead. Sorry, and maybe it's a trivial point, but it is a question. It did concern me a bit, and it's really just that phrase, best scientific information available. And the way it's usually understood in the context of the council process is that that term is sort of reserved for the outcome you know, the full review process, which consists of both, you know, the CIE input, but as well the plan team, as well as the SSC input, followed by NIMP certifying that best available science, best scientific information available has been used for harvest specifications.

So, I, I, um, I assume that you're not saying that there is no role here now for the plan team or the SSC and, and sort of evaluating, uh, the work that's been done here. And so I would at least ask for that clarification. Um, you're going to have to take that up NIMS headquarters because that's the language they use in the CIE terms of reference, and they— it's boilerplate language, and that's what they ask the CIE to do. So I don't know, I don't really care either way. I mean, there's obviously room.

I mean, we're here doing more reviews, so it's not— this isn't set in stone, but that's the language that they use. That's all I know.

I guess the other somewhat of a concern was having the CIE pick a preferred model, which again is, is, is a little unusual, I think, in terms of how we've run CIE reviews in the, in the past.

Was there sort of a— yeah, okay. I mean, I think maybe that's for comment rather than for you to discuss.

Okay, back to you, Dan.

Okay, so yeah, I just have a few slides on just sort of the time series outputs of the two models. In terms of the recruitment trends, nothing too different. So the yellow line here is the Fleets of Zaria's model, the blue line is the single region model. Because we're now disaggregating the age comps and not, you know, making one single set of age comps for the Longline Survey, you do start to see some slight differences in the recruitment trends, especially recently. As I've noted before, a lot of the big recruitment events that we're seeing, we usually kind of see them in the Bering Sea and Aleutians first, and then, then they kind of move into the Gulf a little bit as had a little bit later ages.

So essentially, what happens, you get some switching in these age comps because the single region model essentially is just dominated by the Gulf of Alaska age comps, and so the signals from the other regions get a little bit swamped. So you just kind of see this transition. The 2014 cohort gets a little bit bigger in the fleets-as-areas model. There's some slight switching between the 2016 and 2017. And then kind of there's this higher recent recruitment in the single region model.

But the thing I like to— I need to point out there is just that, again, because we didn't have a 2024 survey, and so we have no age comps from the longline survey in these models since the 2023 longline survey, essentially a lot, you know, from 2020 to 2025, you're kind of seeing things kind of revert a little bit towards the mean in both of the models. So that one is— it's kind of a strange point in the time series because of that missed survey.

Yeah. And then this figure is just the total biomass. And so an interesting thing here is that the Fleets of Zaire's model actually increased the biomass quicker than the single region model in the recent time period because some of those recruitment events were actually bigger. It levels off because those last, you know, 4 or 5 years, you know, the recruitment was quite a bit lower. But again, a lot of that's impacted directly by the loss of the age comps from the longline survey.

And the reason I'm showing biomass, because this is a real driver of the ABC ABC scale. You know, the SSB is used in the harvest control rule to determine whether or not you're on the threshold there. But if you're not, they're both applying F40, and then the F40 is applied to the total biomass through the selectivity filter. So this is really what drives the ABC.

And then in terms of the SSB, again, the Fleets of Xerious model is actually increasing a little bit more rapidly than the single region model, but it does get to kind of this lower point historically than the single region model.

So, I mean, there's a lot going on, but in the end, one of the big things that I think people are definitely latching onto, which is to be expected, is that for the 20— and again, I need to point out, this is 2026 catch advice, so these models basically have data through 2025, And then when you're looking at catch advice, it's predicted for 2026. Essentially, the fleets as areas model is quite a bit lower than the single region model by about 9,000 tons. I know I usually bring up catch advice in September and things like that, but I think it's important to point this out and talk about.

So, yeah, and then the other thing about this is just looking at that historical management trajectory and kind of the SSB compared to some of our reference points. And I think generally the trajectory is the same, but the point of the fleets as areas model is that we probably were specifying some catches that are a little too high in that 2000, kind of late '90s, early 2000 period. And it looked like the SSB went lower than the fleet's— area's model suggests. So I think there's some interesting and important things in that to think about.

So that's kind of all I've got for the model building. Again, I was focusing on trying to get this as a quick presentation. There's a lot more in the plan team presentation, and obviously all the documents go through all the diagnostics and things like that. I think for me the main take-home point is that the fleets-as-areas model doesn't rely on any imputed data. When we did the fairly— probably one of the more robust kind of model development, model building, model comparison approaches that have been done in a while is that the fleets-as-areas model, one, was parsimonious.

We weren't moving to the full spatial model. It had kind of the limited complexity but enough complexity to handle kind of the changes in the data that we were trying to address. It was stable and generally it had better performance diagnostics than the single region model. And then as I pointed out in Martin's question about the carryover index, I think we really need to think and be careful about this lag and the inertia in the single region model from that carryover approach as you get to the top or the bottom of a population trajectory, because you are carrying over values that are either going to over- or underestimate the size of the population and the trend in the index.

So— Hey, Dan. Yep. Can I just add a point that In addition to the carryover for the single region model, there's also an issue where we're only having age comps from the Gulf in one year versus having age comps from the Bering Sea and Aleutian Islands in other years, who are effectively treating those as representative of the population. And so that's also an issue that, um, that we should probably be aware of as all. Um, yeah, and I mean, that applies to both models too, because moving forward, the age comps that go into either one are— yeah, they're going to be very specific to those regions.

So it's definitely a lot of changes issues in the kind of underlying data going into the assessment that are tough to deal with, I guess.

And so, yeah, I mean, obviously there's a lot of concerns about the changes in model structure, what the potential implications might be. I've been responding to emails from stakeholders for the last 2 or 3 weeks. There's a lot of good public testimony online that we've read, some good digs in there that give me a good laugh. A biologist playing with models, I get it. But I also had some good questions.

Obviously, you know, one of the hardest and probably best questions I had is just, you know, are you worried that the Fleeces Areas Model is overly conservative. And, you know, I'll just— the only way to answer that is that we know none of these models are correct. Like, we're never going to know the truth. And there is a possibility that the fleets as areas is a little bit conservative, and there's an equal possibility that the single region model is overly optimistic. And so, you know, based— you know, that's why I think the closed-loop simulations were pretty important and eye-opening to us where the fleets as areas model tended to produce estimates that were much more in line with the truth than the single region model.

And so I think that was kind of helpful. But yeah, I mean, there's a lot going on and it's hard to kind of keep track of it all, to be honest. And so that's what this figure on the data changes is kind of trying to point out is that, you know, we've lost the CPUE data back. Starting in 2023. The same year, there was a reduction in the fishery age comp reading.

Then in 2024, we didn't have a survey, so we didn't have an index or age comps. 2025, We got the survey back, but it was only in the Gulf of Alaska. But then we had a reduction in the fishery age comps down to 600. So yeah, there's a lot going on, and And yeah, and I think ultimately what I, what I wanted to point out is that ultimately these are changes that are structural to the assessment. We think they're more defensible.

We think the model is producing a better estimate of the trends in the population, but it's never going to be perfect for sure. And we're definitely aware of the fact that, you know, moving from the 2026 harvest specifications, which was about 47,000 tons for 2026, um, this would be about a 32% decline in what the ABC would have been for 2026. I'd also point out, like I did at the beginning, that if we had gone with Model 25.12, we would have been at 39.5, which would have been kind of a stepping stone. Um, and so yeah, there's just a lot going on. The shutdown didn't help things out for sure.

And then, yeah, just keep all that in mind.

I would say the last bullet there, that like I said, we have new survey data finally coming in for the 2026 assessments that we'll see in November, which finally gives us some more age comps from the Gulf of Alaska, which we don't have in these models because they're on a 1-year lag. And there's some pretty strong indications that we had some some higher above-average recruitment events in some of those years. And I would say in combination with the fleets-as-areas model predicting a faster, larger increase in SSB and biomass, and the fact that we've never really caught the full ABC, I think the— I think there's going to be maybe not so drastic of a change in the ABC as we've seen in this 2025 model, which is a little more data-limited.

So are there any questions on the models?

I have one for you, Dan. In many other assessments where we have biennial survey information, we see a RIMA model applied to create a time series, and I'm wondering, did you consider that that along with a carryover approach, and how did that turn out?

Yeah, we had a lot of talk about the best way to do it. I, I always get confused exactly what the REMA model would do because most of those assessments aren't annual, and so you're kind of smoothing between two data points, whereas here we have an annual assessment, we have an area that's not surveyed, and so you're not kind of imputing between two data points, you're not smoothing between two data points, you're just filling a data point. So my hunch is that goes back to kind of the average. I mean, is that not right?

If that's a question for me, I'm not sure. It was a question for you.

Um, we, uh, I mean, we talked about it and this as a group, and this was what we came up with as the best potential outcome. Someone else might have some more background with REMA, can probably chime in, but we can also follow up. Hey Dan, I mean, like, at the end of the day, REMA is interpolating or imputing those values with some sort correlation structure. And so I think it'll end up being the average of some sort of AR1 process interpolating between points, but I think the overall index that gets constructed is going to be very similar to the carryover. That being said, we didn't really consider the RIMA approach in this point, just given that we didn't think that it would be that different, but also that there's still the issue of dealing with the age comps and not necessarily being able to interpolate the age comps in a similar way.

Thanks for that response. Um, one, one other question, and it's maybe a broader question. I'm not even sure if something for you guys to answer, but the You, you mentioned early on in one of your slides that essentially this whole effort was precipitated by the change in the survey design, and I'm wondering, what are the prospects that the survey design is permanent and that we might not see a different survey design a couple years down the road? And/or was there discussion when the survey design changed about doing a biennial coast-wide survey versus a half the survey each year to type approach? And I'm sure there's some documentation on this somewhere, but I've just probably missed it in the background.

Yeah, that's probably above my pay grade. What I will say is we did discuss potential options. We do not really have control— we don't have much control over those options because it's a contracting issue of getting a boat to do the survey, I think it would have been better to be able to do a coast-wide survey, but fundamentally, being able to do that is very difficult. And I think it was a contract thing that probably prevented it. I know Chris might be online and he might be able to talk more to that.

Yeah, through the chair, Ian, this is Chris Lunsford. Can I respond to that? That? Yes, you can. Thanks, Chris.

Thanks, Ian. Uh, Chris Luntzer with AFSC. To respond to the longline survey, it was a combination of cost recovery aspects and the government having to fund the survey, along with the available vessels that were capable of doing the survey, which are our freezer longline fleet, mostly doing cod fishing. So it was, it I don't see our survey design changing because it's not just cost recovery. The government is now funding it, so spreading it out over the course of 2 years is much more palatable to the government.

It would be a pretty huge sum if we did one coast-wide survey every other year. And then secondly, just getting vessel time was one of our biggest hangups. So we came to this conclusion. We also investigated looking at alternating between all areas, but we just couldn't cover the ground, things like that. So this is the design we ended up at, and it's not ideal, but I think it's the one we're going to be in for a while unless, as Dan said, the powers above us can offer more funding and/or we can find some other aspect to do the survey with different vessel types.

Great. Thanks for that additional background.

Martin.

Are we going to be able to get back to that figure that I asked for?

Yes, I believe they've got it queued up. Hurry. Um, my question for you, Dan, we're pushing up against lunchtime, and I know you still have a few slides left, Is it going to be more efficient to finish your presentation on apportionment and then jump out and find that slide, or to do that now and then jump back in and try to finish up apportionment slides before lunch?

Uh, I have no idea, but Martin really wants to ask the question, so I should probably let him. Okay, let's go after this slide right now then. I believe you're going to have to stop sharing, Dan, to give staff control over the projector.

So this, I believe this is the full image that was a sub— a smaller version was in the planting presentation. Is that what you're looking for, Martin? Yeah, yeah, that's, that's it exactly. And so, so, um, do you have a cursor or something? Could you put it up first off, put it on the preferred model configuration, which is that— you went past it.

No, it's down. It's there, right? And then, is that correct, Dan?

I don't see a cursor.

Okay, it's there. Okay, I see it now. Okay, and so it's that one, the, the one that you're identified was it's— and I'm sorry if we're way into the weeds here, but I mean, I think it's an important issue. To the left, one to the left. Yeah, one to the left.

So it's the second facet, the second— so second column, second row, and it's the first box plot.

There, there you go. And so presumably that shows— it's the second box down there, um, which shows what you're saying, that it's a really tight, sort of almost unbiased configuration, um, which is kind of what— what— why I think you there are other reasons, but this is a setup where all the columns are the fishery selectivity options, and this option is the one that you identified, which assumes that all the fisheries are asymptotic. And then the row going across across is where all surveys are asymptotic. And so if you look at the column, there's different sort of configurations of assumptions for the survey, but across all of the survey assumptions, you have a pretty— that column is positive except for the one that the model configuration that you identified. So it's saying that generally asymptotic surveys are— tend to produce a positive bias except for the case that you identified.

And then as you look across the row, those are all options with respect to the I'm sorry, I can't quite see it, so I'm going to have to pull it up. It is posted to the SSC agenda, if that helps now, below the Sablefish presentation.

So that is looking at the options for different configurations of the survey, and you can see here that all of those particular options for that row, which is where the surveys are all assumed to be asymptotic, are negatively biased except for that particular one that you identified as the preferred model. And so I guess what I'm seeing here is that there's two things that are biased. Asymptotic survey selectivity is biased, is biased and asymptotic Fisheri selectivity is biased, but the combination of the two cancel out two respective biases. And the question, I guess, is, is that something that you've noticed? And it seems to me to be implying that the assumptions for that model are not sort of intrinsically appropriate, but they're the good properties are essentially because of canceling biases.

And I'm just wondering if that interpretation makes sense to you.

I mean, I guess partially. I mean, I would look at all of these and say that all the models are biased in some way. And this one's— this was the starting point for all of our investigations, to be honest. This was the first stage I think that for us the closed-loop simulation was a little bit more enlightening, but yeah, I mean, I guess the survey with survey selectivity logistic has a slight negative bias across the row.

And yeah, I mean, I guess I'm not exactly seeing what your point is.

And I think maybe the another helpful point is to look at the scale of the bias there. For the most part, they're all within less than 10%, whereas except for the upper top left panel where everything is dome-shaped. So I think, I mean, I think they're, yeah, they're all biased, but I think most of these models are relatively unbiased for the most part as well. And I think there's also issues with simulated data and real-world data when we go into that realm that we really needed to consider when we ended up picking a final model. It just ended up being that one of these models that we identified in simulation appeared to be a relatively good or robust option for what we were applying for in the real world.

Okay, follow-up. I'm sorry, we're—. Yeah, go ahead and do one more follow-up and then we'll get back into the presentation. Okay, um, the, um, so in, in the paper that, um, that we received, the option that was identified as the least biased option was the one with dome-shaped selectivity for both the Bering Sea fishery and the, um, um, survey where you had 3 fisheries and 1 trawl fishery, which is the bottom right corner, uh, bottom right corner. So I don't know if you can get your cursor over there now.

That box, but the first box in that square, the first one there, that was the best performing one which had, you know, was that you identified in the paper. And so I guess I'm wondering if you could— and I guess maybe the larger context is that is that a lot of the sort of research that's gone on with fleets as areas models in a simulation context finds that there are often dome-shaped selectivity patterns when you set up this fleets as areas configuration because of different local differences in fishing mortality, recruitment, and migration.

And it seems to me that this might be missing something here if you assume everything asymptotic, particularly for the Bering Sea where there is known to be movement into the Gulf. And I'm wondering why this particular model wasn't considered further, given that it was the best performing model.

Yeah, I mean, there's a lot to that. First of all, I'll go back to its best performing model, barely better than some of the others. Yeah, if you're looking just at this simulation context, um, technically that one's still positively, positively biased as opposed to maybe a slight negative bias. Anyways, I mean, I think you're splitting hairs if you're saying that any of these models that are in here are significantly better than others. But to your point, we then took these results and developed 100+ models with the real-world data.

We were originally hoping that, you know, we would find— I mean, that's— we took your line of reasoning when we were doing this thinking like, yes, there should probably be some dome-shaped dome-shaped selectivity to account for the movement dynamics, things like that. But including dome-shaped selectivity in a lot of these fleets and then also splitting the longline fishery into 3 separate fleets creates a ton of instability, but also reduces the sample sizes in the longline fishery age composition significantly. The Aleutian Islands especially has very limited age sampling, so estimating a selectivity curve for that region was almost impossible. Then adding things like dome-shaped selectivity, which are inherently more difficult to estimate, added more instability. All those model runs, we've— we included, I think they're in Appendix C. I think we discussed some of those.

But essentially, the instability aspect was really the defining part of why we didn't move forward further with those. We do show results with the 3 fishery fleets, but logistics selectivity, because that was the one that we could actually get to stabilize and get realistic solutions. So fundamentally, that was basically the argument, is that adding fishery fleets and doming was unstable. Matt did a lot of sensitivity runs that are also presented in 3C and elsewhere where we allow for time-varying survey selectivity. And Matt, I don't know if you wanted to talk about that quickly.

Yeah, kind of getting at the changes in availability for the survey, we did a run with basically semi-parametric survey selectivity, and it's able to fit the composition data better, but at a cost of not really fitting the indices. And I think importantly, when we started doing retrospective peels, the retrospective inconsistencies got really large. And so there, yeah, there were sort of other model variants we explored for sensitivity runs, But at the end of the day, it just became sort of this trade-off between model stability and parsimony.

Yeah, and that's why, I mean, like we said, we use this figure right here as kind of the starting point when we went down this road and we're like, okay, what are some of the options that are, one, likely to work well, and two, likely to be feasible? But again, like Matt started off with, I wouldn't read into a single simulation study as the end-all be-all, and that's why we had multiple stages. And so hopefully we've answered your question. It's definitely something we looked at in depth for sure.

Okay, I think we need to step back into the presentation. We may have some more follow-up. Thanks for jumping out and getting this extra slide organized, speaking to it. We are approaching lunch, so anything you can do to get us through this last little bit bit would be great. I would love to.

All right, so quickly, like I said, I got to touch on apportionment. We have to, you know, why do we do apportionment? We need to set the BRDs like Sarah talked about at the beginning. The main point of apportionment for sablefish is one, to avoid localized depletion in any given region, but also just to avoid fishing on only one, you know, segment of the population in terms of only mature fish or only small, recruits. Like, we want to have balanced harvest across the age structure and across the regions.

And I will point out that when we set this apportionment in the 2020, based on a bunch of simulation work by Kerry Fenske, our advice was that the catch distribution across these regions should not deviate too far from the survey estimated biomass proportions. And I think it's pretty obvious at this point, but the main issue now is that the old approach to developing that single region set of— essentially, we don't have a synoptic survey, right? So we have that problem of there's always a region that's not surveyed. How do we determine apportionment across all of Alaska when we're missing a survey region? And so historically, we've just used the old approach, which was to use the change in the Gulf of Alaska to then extrapolate the BSRAI.

Survey, and then put all those regional indices together to do apportionment. Basically, what we're proposing now, because of the change in the survey design, is that we just switch to the carryover approach for apportionment so that we're holding a region that's not surveyed. It just uses its previous value when we calculate the apportionment across all the regions. And then I will point out that one of the biggest impacts is just removing those western Aleutian Islands extrapolated because that's really what changes the apportionment or the regional indices the most for the AI. And so the figure on the right just compares what the apportionment would have been under the old approach versus what it is under the new approach.

So the blue lines uses the Gulf of Alaska extrapolation, the red lines uses the carryover, the boxes are the regional apportionment, and then it goes through each of the years, so 2016, '17, '18, blah, blah, blah. It goes up to 2023 because we can't use the Gulf of Alaska approach in 2025 because we don't have— yeah. Anyways, the point being essentially that the Aleutian Islands apportionment decreases. So you go look at the blue, which is the old approach, versus the red, the new approach, So Aleutian Islands decreases by, I think it's about 20%, and then the other regions essentially increase proportional to the decrease in the Aleutians, with the net impact being that the Gulf of Alaska regions increase by about 10 to 15%. The Bering Sea is kind of a net neutral.

And so I'm hoping that this change in the apportionment itself itself is relatively straightforward and there's not too much heartburn from doing this. So this is the only change and that's kind of it. So if there are any questions on that— Seeing none, go ahead. Okay, so I just wanted to point out note that there is an impact on the ABC if you're looking at what was enacted for 2026, so the harvest specifications versus what the single region or the FAA model, um, is saying the ABC should have been for 2026, and then applying the new apportionment to that value. Um, and so the last box there shows you what the change in the BRDs would have been if we had used used the 26.2 FAA and multiplied that by the, the new apportionment method.

So the ABC decreased by 32% compared to the 2026 harvest specifications. The regional impacts vary because of the new apportionment. So the Gulf of Alaska gets a little bit less of an impact versus, say, the Aleutian Islands, which has a very strong decrease.

By about 50%. So that's why I wanted to point this out. Obviously, I think everyone's aware there's a lot— like I said, a lot of stakeholder concern about this, and I totally understand that. But I just— this again is for 2026. This isn't for 2027.

So I just want to keep that clear. And so So I don't know how much to get into this. I don't— I'm not trying to suggest that we should do anything different. I'm just pointing out that given the amount of concern, I expect that in December when the final recommendations are out, there's going to be a lot of talk about what alternatives could be undertaken by the council and things like that to address this large impact on the ABC and also, you know, obviously on the fishery. And so I just wanted to highlight that it would be helpful to maybe have some alternatives that people might want to see so that they can be done beforehand and not be rushed at the last minute, because that's always where mistakes happen.

So I just wanted to point that out and have hopefully there be some conversations sooner rather than later if there's alternatives that want to be explored that require further analysis or, you know, just me coming up with the numbers. I don't want to be doing that on like December 9th or something like that and potential for, for making a mistake there.

So there's a lot going on here. This is more informational, but the two blue boxes are kind of two alternatives that, you know, I'm not recommending them, but they are something that could be used, which is just maintaining the, you know, the harvest specs for, say, the Gulf of Alaska from 2026, reducing the quota appropriately, and then adjusting the BSAI to account for that. Another option in the second blue box would be BRDs from the Gulf of Alaska from the one region model, which is a slight reduction from 2026, and again, reducing the overall quota to what it should be from the FAA model, and then adjusting the BSAI. I will say that, you know, we've done some preliminary runs and the red box is probably close to what things may or may not look like for 2027. So it's just something to keep in mind.

And I won't get into this too much, but Matt did some updated closed-loop analysis this to kind of mimic what Kerry did back in 2022, but through a closed simulation loop, and kind of looked at some alternative options for apportionment and what the impacts might be. Long story short, we didn't actually see too much of a negative impact by changing how things are apportioned. Mainly, if you increase the apportionment to the Gulf of Alaska, it didn't seem like it would have a too detrimental impact. I would say there's a lot of caveats to all of this simulation work, so I wouldn't read into it too much. But we did just want to kind of show this and do this quick analysis just so we kind of— before we made any suggest— not suggestions, but showed any alternatives, we wanted to make sure that they wouldn't have a huge negative impact and I don't think any short-term changes to how things are apportioned would have, you know, a huge impact on the biology.

And I have, I think, 2 more slides, and so I just want to— I show this slide every year and it's in the assessment every year and I really harp on it, and I just want to make the point. So the orange line here is the SSB, the yellow line is the catch, and the bars are are the recruitment estimates. This was from the 2024 SAFE. Sorry, I didn't have time to update them. But the point that I always make basically every year is that there's a, if you just look at the data and the trends in the population, back in time, anytime the catch goes above about 30,000 tons, the population declines.

And, you know, I think just from a, just looking at these, you know, for me, I look at this and I'm like, that seems like a pretty strong indication that catches under 30,000, around 30,000, under 30,000 are probably fine. But once you start getting up in those higher, you know, mid-30s, 40s, 50,000 tons, bad things tend to happen. We were definitely saved by these huge recruitment events in the last couple years, and we're back up at a high high level, which is great, but I think now is the time to kind of think about what the long-term implications are if we keep ratcheting up the quota and what we could potentially get out of it if we stay at a relatively moderate level. And so that's really what the, you know, we have the management strategy evaluation for sablefish. We've been working on that project for about 3 years now, and And a lot of the results from that kind of indicate that one of the best potential options is probably this harvest gap.

That's a whole different issue, but I just wanted to point that out.

And then Matt also did some alternative simulations with basically a constant catch for the next 12-ish years, what the population might look like under different recruitment scenarios. If we had constant catch of 10,000, 20,000, 30,000, 40,000, or 50,000 tons. And not surprisingly, you know, at the higher kind of end of that catch level, if we go back to kind of our long-term average recruitment, we can expect to kind of end up back where we were. I'm not saying that this is ever what is what's going to happen or what the future holds, because we do an annual assessment, we update the projections, we're pretty reactive, responsive, and so So, you know, not saying that's what the future holds, but, you know, there is this kind of, I'd say, slight risk that if you keep the quota high, you're going to end up kind of in a potentially in a bad spot. Conversely, if the recruitment stays high, we can probably fish even harder, but, you know, the odds of maintaining recruitment as high as they were in 2014, 2016, 2017, probably pretty low.

So anyways, this was just something that we've been passing around to kind of show some of the implications of different catch levels, things like that.

And so, yeah, I think this is the last slide.

You know, we're recommending 26.2, the fleets as areas model. We think it does a better job of handling the data that we have. It obviously leads to some moderate rescaling, but it also suggests that the population is growing faster than some of the other models since about 2018. We're proposing this new apportionment approach. It increases slightly the apportionment to the Gulf but reduces the apportionment to the AI by a pretty strong amount.

Like I said, I think alternative short-term apportionment approaches are unlikely to have major biological impacts as long as they're short-term and as long as you kind of utilize quotas that are appropriate. But I do also want to point out that the 2026 assessment does have the BSAI survey, and so the Bering Sea went down 14%, AI went up 26%. 36%. We also have the 2025 Gulf of Alaska longline survey age compositions in there, so we're starting to get indications of those recent recruitment strengths that look pretty positive for the 2020 to 2022 cohort. So I think that is it.

Okay, thanks for that presentation. This will be the This is the last chance to ask questions of the analysts.

We do have— we are going to go back to the plan team after this, or perhaps after lunch.

Go ahead, Martin. Yeah, just a question about— it seems like using the gamma distribution to model dom-shedding double normal shape selectivity seems like maybe it's been around a long time in the sablefish model, but it's not really widely used elsewhere. And I know that, I mean, the Spork implementation for, I can't believe I'm going to have to say Spork now, but for Rexel used the double normal, or at least part of the double normal, I'm wondering how much you've explored other parameterizations of dome-shaped selectivity that aren't as limiting as the gamma dome-shaped pattern.

Yeah, I mean, that's a good question. I mean, the— fundamentally, the issue is the stability one. The gamma only has two parameters, and that's kind of why it's been used for sablefish. Given that it's a sex-specific model. That's primarily it.

Like Matt said, he's done the time-varying selectivity, and so we've explored that one. But I think, Matt, I'll let you take it away. You've done most of the analysis there. Yeah, we've looked at the double normal, and I couldn't quite get it working well. Well enough to be like stable and something to present.

And that's why we didn't consider it for this round, but it's definitely something we'd be interested in looking at in the future.

Okay, I'm seeing no further questions. I— we are definitely due for a break, so let's take our lunch break and then we'll come back and finish out the plan team items. For the public who's been waiting patiently, we will get to public testimony as soon as we're done with the rest of the plan team report. And if you're here waiting for Bering Sea and Aleutian Islands or GOA items specifically outside the joint plan team, be aware that we may have to rearrange the schedule. The order of the GOA and the Bering Sea are still up in the air at the moment.

So we'll— stay tuned, we'll make that announcement as soon as we decide. So we will come back at 1 o'clock. Alaska time. Thanks.

Okay, welcome back everybody. We are going to restart on the Joint Plan Team report, and I think now we're going to get the comments from the Joint Plan Team on Sablefish. For the public, if you're just joining us, we're going to be— we have a couple more items to get through in the Joint Plan Team report, and then we'll be taking public testimony on all of the Joint Plan Team report items, including HCRs and Sablefish, following this next short section here. So back to you, Steph.

Yeah, thank you. Through the chair also, I don't know if you wanted to alert people that when we finish the Joint Plan Team, we'll take up the Bering Sea. Specs before we do golf.

Sorry, Mr. Chair, we're getting the PowerPoint back up and, uh, Dr. Shotwell will give the Sablefish report.

Thanks, Diana. Can you hear me okay? We can, thank you. Okay, this is going to be short and sweet. The team greatly appreciated the substantial effort and documentation provided by the sablefish assessment team.

We did discuss the persistent difficulties in fitting the plus group in all of the models, but that a sensitivity showed that the model was fairly robust to that lack of fit. The team also noted that this was primarily associated with a limited historical period and was not as pronounced in recent years, that lack of fit. The team also discussed natural mortality, that natural mortality remains an important source of model tension, and the teams discussed potential for investigating age-varying natural mortality as a future research direction. We also discussed the carryover assumption required to construct the Alaska-wide survey index, and that that was a strong assumption. As Dr. Stewart mentioned, we discussed— similar to what Dr. Stewart mentioned, we discussed potential use of REMA for future single region models to smooth the regional survey estimates.

But as Dr. Chang mentioned, that only really addresses the index and does not resolve the issues with the regional age composition.

We also appreciate the update regarding the change in apportionment method, but due to survey— or sorry, due to the survey design change. And we noted that decisions on ABCs and BRDs are evaluated in November. So recommendations from the team included for November: the team's agreed with the author to bring forward the Fleet-as-Areas model, the updated Single Region model, the Carryover model, and to also bring forward the Continuity-based model. And the model numbers are on the slide. The team's also recommended avoiding further reductions in longline survey coverage and that the age data collection should be continued.

We emphasize that maintaining the current level of sampling should be considered a minimum. I think that's all we have for the review of sablefish from the plant team.

Okay, seeing no questions, go ahead and proceed. Okay, we have, I think, Just a couple more topics.

We received a presentation from Dr. Steve Barbeau and Pete Holsen provided on the Pacific cod research, recent research, and they mainly focused on connectivity and development of spatial assessment approaches. Their main focus was on the new SPORK model, but they also showed us a few SS3 models They also talked about tagging and the long-term goal of a management strategy evaluation.

There's—. They show quite a bit about the connectivity between the Eastern Bering Sea and the western Gulf of Alaska, and that it's supported by all— not only conventional tagging, but pop-up tagging and genetics. And then Dr. Barbeau showed exploratory results from several Tuari SS3 models, but there's substantial confounding among catchability, movement, recruitment allocation, growth, and selectivity. And in stock synthesis, there is no easy way to rigorously test among various hypotheses on how spatial processes impact stock dynamics. So he also presented the SPARC model as a more flexible platform for continued model development and included direct— and being able to include direct movement data.

Initial explorations for SPORK are in development by Dr. Chang. And then Dr. Holson provided details on the Western Gola spatial assessment research, also using SPORK. And the PCODD team plans to continue development and testing of these SPORK models through 2027 assessment cycle. And I think they plan to have a serviceable model available for the '28 assessment cycle. They also emphasize limitations of tagging data across space and time.

Although the majority of historical conventional tagging were from the Bering Sea, the authors suggested that more tagging would improve understanding about movement and spatial proportionality.

And then finally, the teams emphasized the importance of incorporating the tagging data directly and by using, and by using the SPORK model, and they noted that there was a substantial change to the the Pacific cod, if there's a substantial change in Pacific cod management, it would require careful development and clear communication to stakeholders and being included through the council process.

And that's all I have for that. Next slide.

The team's received reports from in-season management. Steve Whitney provided an overview of of the BSAI in-season management, noting broad decline in fishery utilization in recent years and also continued reductions in the active Amendment 80 vessels. Abijan also provided an overview of the GOA in-season management, noting shifts in utilization for pollock and Pacific cod, and reviewed some of the rockfish and skate catches. The team did not have any substantial discussion or recommendations.

Next slide.

And then finally, we had a presentation from Mike Fay, and the teams adopted the halibut discard mortality rates based on the recommendations from the halibut DMR working group.

And that's about it for that. And I think that closes up the joint plan team presentation.

Excellent. Thank you for walking us through that. Have any final questions for the Joint Plan Team?

Okay. Then that takes us to public testimony on the Joint Plan Team items.

We have 9 signups. We will go through in order.

And first up is Mr. Robert Alberson.

Hello?

You're on. Am I on? Okay, my name for the record is Bob Alverson. I'm with Fishing Vessel Owners Association out of Seattle. This is a trade association of halibut and black cod IFQ holders and vessel owners, and we're commenting on agenda item C3 having to do with the models for, for black cod.

As you know, the stock assessment teams report provides two new stock assessment models, Model 26.1 REG and Model 26.2 FAA. The two models recommend an ABC of 41,000 metric tons and 32,000 metric tons respectively for 2027. The coast-wide ABC for 2026 is 47,000 metric tons based on the model models, both used in '25 and '26.

The stock assessment team mitigates certain risks associated with the models. Those risks apparently come from structure issues within the models, and they, they mitigate that by restricting commercial harvest to the equivalence of a harvest policy of F65. Which is pretty conservative. The Council on the SSC harvest strategy goal has been F40 for sablefish for a long time. The associated cost to cover the new risk with the new models is significant.

We anticipate that this would be about a 20 million— 20 million pound hit overall, most of that coming out of this— out of the Gulf of Alaska. And at a 63% addressed weight, that's about 13 million pounds at $4 a pound. That's what we're getting right now. And so you're talking about a $50 million hit for two models that are kind of all over the board. The preferred model from the CIE says carryover is good enough for apportionment but isn't good enough for the model because it produces 54.

4,000 Tons. So I don't know quite what that means. You, you guys probably do. Um, the other model, um, produces a 41,000-ton, uh, amount, but it doesn't have a carryover option. We would like to see something in December other than just these two models, if it would be possible to 26.1 done with the carryover, or whatever model you choose, or the stock assessment team chooses, did it be phased in?

A phase-in over a couple years would give the modelers a new set of data from the Gulf of Alaska for the '28 season. If it was a 4-year phase-in, it would give them 2 2 surveys there and 2 in the, in the Bering Sea, that might calm out that particular model, that second model that seems to want a 32% reduction. With regards to some of the data that is missing, I would note that the first time, that the new models do not use logbook data reflecting commercial catch rates. The captains and crews do considerable work to complete very detailed logbook entries. The SSC needs to insist the logbook data be reincorporated in the future.

It is likely that the current omission of this data, such as weight per unit of effort, is resulting in some of the problems associated with the two models. Additionally, now that Sablefish Fleet is harvesting 91% of the harvest with pots rather than hook and line, there needs to be a new metric for weight per unit of effort from the logbooks, perhaps weight per day or pot weight per day. We request the SSC encourage the plan team or the stock assessment team to update the weight per unit of effort metric from the logbooks, incorporate that into future models. It was suggested that, uh, You would not be able to see a downturn in the resource if you have this every other year surveyed as is planned, odd years being in the Gulf of Alaska. If you have the logbook data, you'll be instantaneously seeing what's happening, at least in the industry, if you were able to incorporate that.

And you wouldn't have that lag that was suggested by the stock assessment presenter. So in, in conclusion, Mr. Chairman, we would like to see, uh, one or two additional options for the public to discuss, or yourselves to discuss in December. One would be this Model 26.1 with a carryover, and if the— whatever model is, is the preferred model, that it be— we'd like to see an option that is a phase-in. And I guess that would be a phase-in on the existing model that it produced the 2026 harvest limit.

Thank you for your testimony. Would you be willing to take a question? I'll take a question. All right, Martin. Yeah, thanks, but, um, I mean, we're throwing a lot of models out here, but my understanding was that 26.1 already had the carryover built into it, and it was the 25.12 that didn't.

And, and so I'm not sure exactly what you're proposing here.

Through the chair, my understanding, it— the carryover is not in the first model that produces 41,000 tons. Okay, thanks.

Okay, thanks again for your testimony. Next up we have, uh, Miss Linda Banken.

Thank you, Mr. Chairman, members of the SSC. Linda Benken testifying for Alaska Longline Fishermen's Association. I will echo a lot of what Bob said, mostly in our request for alternative methods, models to be brought forward to the SSC from the stock assessment team in December to be the basis for the 2026 assessment and 2027 specification process. I would just note that there are more sablefish in the water right now than there has been since the 1970s.

So it feels to us like there is time to take a look at some— there's a lot of changes, and to really evaluate those changes and whether they're the right changes to make. And at minimum, there's certainly time to phase in some of these changes over a time period, um, as you ground truth them, and to minimize some of the impacts on the fleet and the communities that depend on this. I— we understand that, that a lot of these changes come from having to fill in gaps because we no longer have the annual stock assessment in all areas. We don't have a comprehensive annual stock assessment and not getting CPUE data because the IPHC isn't key punching it, even though the commercial fleet is filling all that data out for you. There's just a lot of changes that it really bothers us that there is just an acceptance that you are gonna always now have this lack of data.

And to me, the SSC responding to that and making a statement that it's really important to reestablish publish comprehensive annual surveys. This fishery is the most valuable on a per-pound basis finfish fishery in Alaska. It's a really valuable fishery, $63 to $93 million per year. You should have a survey. The cost of not having a survey, of having these uncertainties and having this kind of cut, is on the order of $50 million.

It's a poor trade-off. So we will be pushing to restore these surveys. We really hope the SSC will emphasize that as well. We will continue to fill out our logbooks and to provide you with commercial catch data. We hope it will get used.

It's sort of frustrating to be on this side doing our part and seeing it not incorporated into the stock assessment and apportionment process. I, I also just want to note that if you follow the logic of all these changes, when you saw that backcast, it suggests that between 2005 and 2015, there were actually half as much sablefish in the water, or at least that's what this current assumptions would suggest, as we thought there were at that time. And I would say from someone who was fishing then and talking to a lot of fishermen, fishing was still pretty good. In those days. It did not seem like there were half as much fish as we thought there were.

And also then you did have comprehensive annual surveys, you were getting CPUE data, you were using the Gulf trawl. It was just much more data rich. And to think that we were that far off, to me, is a real stretch. Um, I guess, you know, I just again remind you that we have said for years, be cautious about going up. If you go up too far and have to come down, it comes at a huge cost to the industry and to the communities.

That's exactly what we just did. It's not building resilience, it's the antithesis of that. It just feels like some of the worst management choices that could be made at this time. So again, we ask that you task or request that the stock assessment authors bring forward some alternative methods and models models that could be the basis for your 2027 recommendations, and that we think about phasing in over at least 3 years some of the changes that are being suggested, if this is in fact the direction that we're going. Thanks so much.

I'm happy to answer any questions.

Thank you for your testimony.

Next up, we have Mr. Bernie Burkholder.

Good afternoon, Mr. Chairman, members of the SSC. My name is Bernie Burkholder, and I'm speaking here today on behalf of Endurance Fish Company company, and Muckinam Fisheries, a quota company, and the 3 vessels that we use to harvest black cod with.

I want to voice concern about the models that we have in front of us. I don't propose to say I know what the answer is, but I think we've got a little bit of work work to do ahead of us. I, I just think that these are not really giving us the options. We've waited so long to get to a situation where we've got this kind of resource, this kind of CPU, I mean record catches per pot, per longline, whatever, and we are overly concerned, I think, at this point. I'm not saying we shouldn't keep an eye on it, we shouldn't be looking at it on a year-to-year basis, and I don't know that the surveys are ever going to go back to where they were.

I wish they would, but I guess that's kind of doubtful. So let's learn to work with what we've got. Certainly looking at the logbook information, I think would give us all a lot of real-time information that we would like to have.

Sablefish stocks are currently healthy with catch per unit efforts reaching all-time highs for the vessels I represent and the biomass setting in the upper end of historic limits. Despite this strong stock health, proposed models 26.1 REG and 26.2 FAA would sharply drop the acceptable biological catch to 41,000 and 32,000 metric tons respectively, a drastic reduction from the 47,000 metric tons the base model used in '25 and '26. While these new models are considered scientifically credible, and I believe they are, they remain untested. And I would like to see a period, a prove-up period, where we could watch these and see what they actually do. Um, an ABC reduction of this magnitude will cause severe ripple effects throughout my industry.

Market harm. It will derail hard-won progress in growing markets. I mean, we have worked with processors, processors in this room, to just boost bootstrap ourselves up and get markets that can take this increased fish and get the increased price. And now all of a sudden, to take this kind of a loss, those markets aren't going to wait there on the shelf, they're going to be gone. Financial impacts, these will change the quota price, they will change CDQ groups' performance, economic communities in the Alaska community system.

This is going to have some far-ranging impacts. And if we need to take a cut, we need to take a cut. We've always— we've all been here, and if it's needed, we'll do it. But I'm not sure that this really is needed at this time. Uh, I too align myself with the previous testifiers Bob Alverson, Paul Clampett, Linda Bacon.

Recommendations that I would, would ask that you consider would be to keep a base model with the established 25-26 ABCs. Require a prove-up period for any new models that come in so that we've got time to really look at it and ferret out what might be the problems. Request a CIE analysis on anything new coming up, which I think you're good at doing and you've already done. And then I would like to see you align quotas more closely where— with the areas that these fish are actually caught in. The Gulf of Alaska is different than the Bering Sea and the Aleutian Islands.

Balance and take that into consideration. Uh, thank you very much for your time and your consideration, and if you've got any questions, I'll be happy to try to answer.

Thank you for your testimony. We do have a question. Curry, go ahead. Yeah, thank you, Mr. Burkholder. Um, I'm wondering if you could clarify for me your last point about aligning quotas Spatially, I think this has to do with apportionment and— yeah, thank you.

Yes, through the chair, thank you. I think that when you look at carryover, I mean, and I've heard this talked about, if you're looking at regional carryover, where you're catching that fish is important, and that's the point I guess I'm trying trying to make. I don't think you can do that on a statewide system, but I think you can do it regionally.

Thank you.

Okay, thank you for your testimony. Thank you very much. We will go next to Lauren Howard. Welcome, Lauren.

All right, good afternoon everyone. Lauren Howard for the record. It's nice to be able to do this. Usually I'm in the AP meeting and I don't get to attend SSC, so I appreciate the opportunity to speak today about this. It's a very personal issue for me.

My husband and I, um, longline, and we are what you would consider the younger generation of fishermen still trying to get ourselves established in the industry. We bought our current vessel in 2020 that allowed us to longline ourselves, and we all know how that year went and how it went in succeeding years. So we've been doing our best to make our loan payments on our quota that we own primarily from other fisheries, and we're finally to a point this year where things have reached some sense of stability. You've heard a lot about that already in terms of markets and prices and what that's looked like for us, and we're really grateful that things have kind of reached that level stability. And over the last few years, we've been told over and over from, you know, this body and others in this process that, you know, the stock is looking relatively stable, that there are signs of good recruitment, the fishery was looking stronger.

So it led a lot of people to consider the options for, okay, maybe it's a good time to buy some quota. None of us really had the money to do it because we haven't been making much money in the fishery, but luckily the State of Alaska Loan Program gave the whole industry a rate decrease for 3 years, kind of to help out the industry, to say, I know you guys have had a hard time. So myself, along with actually quite a few other people I know this year, decided we're going to buy some quota. We're going to take advantage of this opportunity. It makes sense.

You know, the market's a little better, the price is better. It's better than leasing quota. You're going to actually be able to make your payment. So my husband and I just made a modest purchase of 5Q. So now I'm looking at an absolute, you know, 20-25% potential decrease in that before I've even had a chance to fish it.

And when I did the math, that decrease basically amounts to what my payment is. So it's like taking what the payment is twice against what it looked like. And so a lot of us made these decisions based on the science that we were being told and thought that it was a good time to continue to invest in the fishery and create more stability for, you know, our businesses and grow them. You know, we're in that point in our lives where this is our chance to do it. We weren't allocated quota.

We're buying into it. We're doing the best we can. So this issue is very personal for a lot of people. You know, there's a lot of permit holders in this fishery. It's a large fishery.

A lot of us are residents of coastal communities. So the ripple effects on this are big, especially when the markets respond and we see another decrease in those prices. So that's the personal side. A few concerns I had in reading these documents, specifically about the lack of data. I think there's a lot of choice there in the data that's being used.

As we've talked about, there's a lot of fishery-dependent data. I know the IPHC wasn't sharing data, but 91% of the fisheries using pots, and they use a daily fishing logbook which is controlled by NOAA and NIMS. So you guys have access to 91% of the sablefish fishery data, at least from last year. You know, maybe it was 84% the year before, but the majority of the fishery is using a federal logbook. There's a lot industry data.

I'm personally working on an IFWIF project right now, and we just applied for a second round to try to corroborate EM-reviewed data with electronic logbook data and put those together to kind of show the correlation and ensure that it's accurate. And I know there's been concerns about, you know, the accuracy of fishermen-collected data, but the reality is, is that if we can't be there on a survey, we have people out there on the grounds every here. So if there's a way that we can continue to corroborate this data and make this a more usable data set, I'm willing to continue to do that work and make that a possibility. There's been a shift towards collaborative science because we know that we're lacking funding. Things have gotten more expensive, and that's not going to change.

We're going to continue to see these funding gaps. And I've listened to presentations, like specifically last year— I'm on the Pacific State Marine Fisheries Commission Advisory Board— I remember Henio addressing us and saying, you know, we have to learn to live in, in these cuts. You know, we're not going to have the same amount of data. It doesn't necessarily need to translate to conservation cuts. We want to still see fishermen fishing and prosecuting the fisheries.

So there is a level at which we can look for data sources that we have to put into a model in the absence of the same survey system that we've had. So I don't have faith that we're going to maybe go back to the same amount of data that we had before, just because I look at the cost of everything and the state of our funding, our federal funding, and I just assume this is probably going to be how it is. So how can we look at these other data sources? And the last concern I had that I'll voice is around the lack of trawl survey data going into this stock assessment, especially when looking at recruitment. So often we don't see small fish on the hooks, you know, till they're 1, 2 pounds or 5 years old.

Where we see the small fish in the recruitment show up is in the trawl survey. So if we're not looking at sablefish in the trawl survey, How are we going to see the signs of new recruitment? So that was a concern I had when I saw that they were going away from using that data, because to me that's not going to show any strong year classes coming through. So thank you for the opportunity to comment, and hopefully we can come up with something else to use for HarvestX this year that takes it a little less of a cut. And if we need to phase it in, that's understandable.

We're the first ones to sit there and argue for a conservation cut, but when the planting report says in bold writing this is not a biological concern, this is just a model change. We struggle with that. Thanks.

Thank you for your testimony. Uh, next up we have Dr. Alexander Stubbs.

Oh, not seeing any. Thank you. Next up we have Dr. Alexander Stubbs, and I believe this is going to be virtual. Yeah, can you hear me? We certainly can.

Go ahead. All right, good morning, uh, co-chairs and members of the SSC. I'm Dr. Alexander Stubbs, PhD in integrative biology from UC Berkeley, and I'm the inventor of SlinkyPods that now take most of Alaska's fixed gear sablefish catch. I'm a Bering Sea quota holder and an active sablefish fisherman in the Bering Sea and the Gulf of Alaska. For the choice of models, I support Model 26.2 FAA as the for ABC, but with a phase-in if it can be done without a conservation concern, and it sounds like it can.

Tier 3 gives no phase-in of the ABC itself, so the instrument is the model basis. A time-limited ensemble of 26.1 and FAA with the weight moving to FAA over 2 or 3 years and modest weight on 26.1 since FAA run on carryover indices inflated bonding biomass. Have the authors report the probability of exceeding OFL and in the spawning biomass path for each year. For depredation, in December 2024, you asked the authors for any issues with the performance of slinky pots that may affect CPE. I'm just here to testify that there are, and they said they weren't aware of any.

I'm just here to testify there is one, and it's not unique to slinky pots. Killer whales have learned to open traditional and slinky pots alike in the Bering Sea and increasingly Western and Central Gulf. Health. Most operations set back rather than haul through a pod's attack, so the tonnage of this lost fish is probably small, but it is a kind of unaccounted for mortality, and the assessment's depredation estimate is frozen at 2022 values on the premise that pots ended depredation. Fishermen already report whale presence and depredation set by set voluntarily through the logbook program the assessment used until 2023.

Nearly 8,000 pot sets in 2022 2. Make these fields standard for the daily fishing log and process them for a depredation estimate by area and gear with a flag on depredated sets in any index. This is just another reason why logbooks need to be integrated and why it isn't a choice. On fishery data, pots take 85 to 90% of the catch. The IPHC stopped sharing logbooks in 2023, but every IFQ pot vessel keeps a daily fishing log and sends it to NOAA quarterly.

The catch rates are in the agency's hands, and an assessment of a pot fishery with no fishery index does not meet the best available science standard when the data exists and are simply not processed. Direct that logs be processed and EM effort entered. And with each region now surveyed every other year, a station lost to depredation leaves a hole that stands for 2 years. If that's affecting the index in some areas, have the survey program evaluate whether whale-resistant pots calibrated against gates could replace some of those stations. For apportionment, I support the survey-based BRDs with carryover for 2027.

Because the ABC cannot be phased, softening the Gulf cut means moving BSAI quota east, which is what the presented additional alternatives do, holding the Gulf whole and cutting the Bering Sea 54 to 57% and the Aleutian 64 to 67% on utilization. That's a tack deviation that comes back to you at the SSC for biological impact advice, and the biology does not support reported. The 2026 survey found the Bering Sea well above its long-term mean and the illusions near record high. In 2022, before killer whales learned to open pots— and, um, before killer whales learned to open pots, the Bering Sea fleet landed 86% of its allocation. In 2025, 27%, with deliveries down from 142 to 58, as my figure shows.

What changed was whales, not the fish, and fishermen me included, are investing heavily right now to solve it. A quote does not reallocate unused quota. Every IFQ permit is a fixed share of TAC, so 57% cut takes 57% from the boat that caught all of its quota and 57% from the holder who fished none. The active boats lose more than half their pounds with no way to replace them in this small block area, and trips stop penciling out. If idle quota is the problem, The council or SSC should ask for a discussion paper on reallocating quota that hasn't been fished in years, not a cut that only hurts those who fish and harvest their allocation.

The Bering Sea TAC has been held at 67 to 70% of its ABCs since 2024, with support of many IFQ holders there, including myself, so that we'd have room if there was a drop necessary. Dropping the extrapolated Western Aleutian Island stations already moved about about 5 points of coast-wide ABC, roughly 1,600 tons from the Aleutians to the Gulf, which I support because it follows the survey. The authors' closed loop finds similar outcomes even at 80/20, but its assumptions—high movement, perfectly known regional biomass, no density dependence, no spawner-recruit link—leave the split little room to matter, and the authors themselves say localized depletion and age truncation remains a concern if the imbalance is sustained. A reallocation built on utilization sustains itself. The realized split is already 80/20.

The Gulf caught 97% of its FAA share by September 2022, and these alternatives take removals near the 18-15 with the Gulf above F40. This body adopted survey-based apportionment, so removals follow the fish. Bias in the east just sets a ratchet where western attainment falls because a trip is no longer worth making justifies the next shift because attainment dropped and then falls again. Thank you.

Thank you, Dr. Stubbs. Um, seeing no questions here in the room, so thanks again for your testimony.

We will go now to Nels Evans. Welcome, Nels.

Hello, members of the SSC. Uh, my name is Nelson Evans. I'm Executive Director of Petersburg Vessel Owners Association, PVOA for short. I'm going to be repeating a lot of what has been said earlier by previous testifiers. I'm going to try and truncate just just for brevity here.

PVOA is a commercial fishing trade association based out of Petersburg, Alaska, representing a diverse group of mixed-year fishermen from Dixon Entrance to the Bering Sea. PVOA members are heavily invested in the IFQ longline fisheries, and Petersburg as a community has one of the highest sablefish IFQ ownerships in the state. All this is to say that sablefish is incredibly important to PVOA and to Petersburg. First time I testified to the Council 3 years ago was to maintain the status quo of the sablefish ABC. I again am testifying to maintain status quo in the fishery.

3 Years ago, the trends were— the assessment trends were the opposite. We were looking to push the ABC higher. Today, we're looking to drop it. The only thing that's maintained consistency over these 3 years is a healthy sablefish biomass that continues to grow and contribute contribute to the SSB. The new proposed models, to PVOA's understanding, have come largely due to decreased annual survey coverage.

The survey structure is new, and PVOA would encourage allowing this survey to run at least one more full cycle to continue establishing a baseline and work out the kinks before generating models that would significantly reduce the ABC with no connectivity to the actual biomass health. PVOA membership has expressed that they would be supportive of seeing reductions in the if it was stock-driven. This, however, has been reiterated to not be the case with these new proposed models. BVOA also has strong concerns on the removal of significant historical data inputs in the new proposed models, including the historical longline data and GOA trawl data. In a data-limited environment, we don't believe it's responsive to remove more data inputs when available.

Again, on that through line, same as other testifiers, We would also like to highlight the importance of finding ways to fit CPE data back into the assessment. Fishermen fill out the logbooks, they take the time, lose sleep to fill these out. We would like our data to, you know, go into the science, help inform you guys as we lose information. Again, BVOA encourages the SSC to recommend additional sablefish stock assessment models be brought forward to the November plan team and December council permitting. If we are to move forward with one of the new stock assessment models that have been presented on today, we encourage a phased-in approach as supported by previous testifiers.

Thank you for hearing our concerns and thoughts.

Thank you for your testimony. Would you be willing to take a question? Curry, go ahead. Yeah, thank you for testifying. Um, I just—.

I want to get it straight in my mind. So the logbook data was formally entered by the IPHC and provided to NOAA. My understanding is that that sort of link in, in the chain in terms of data provision is what's gone away. Um, I guess, what are your recommendations for how we bring that back? And do you, do you know if it's a lack of funding that's in place or why that's transpired?

Uh, not 100% why it's transpired. I kind of entered the, entered the process around the same time we lost that data, but I would look for, I don't know, formal letters, anything that could try and draw connectivity between the two organizations. I mean, they're both dealing with data limitations in their survey modeling. I think that we're in a time where more collaborative survey approaches are necessary. So reconnecting those ties is important.

Thank you.

Okay, thanks again for your testimony. Next up we have Mr. Glenn Merrill.

Thank you, Mr. Chairman, members of the SSC. Before I begin, I just want to let you know I'll be testifying on 3 issues: alternative harvest control rules, sablefish, and then one issue under Bering Sea which is dealing with the surveys, uh, and then I am willing to take questions should you have any at the end of my testimony. Thanks for the time today. On the harvest control rule issue, I guess I would like to begin— I am clearly not a scientist, I am not a statistician or a stock assessment scientist, I consider myself, uh, science adjacent. And I guess so when I look at this harvest control rule approach, really what I'm left with are a series of important questions I think that you should be asking of this endeavor before we continue.

I think first is, you know, how reliable do we think these new harvest control rule approaches will be, incorporating a lot of additional data sources that may in some case be contradictory or lack the continuity or robustness— robustness of information that we use in our current assessment process? In other words, when you amalgamate all of this data together and you're trying to throw a dart to hit the bullseye to understand whether or not you have a good representation of what's happening, are you going to hit the bullseye or are you going to hit the wall instead? I think the second question that really comes to mind is, why do we think this approach will be superior to our current harvest control rule approach? That has already been a question that's been raised. I think that's an important one to think about answering.

Third is, why would we not be explicit that the statutory BSAI 2 million metric ton OY cap remain in place under this modeling effort. I included two figures in my testimony. Uh, the first one, Figure 1, just provides you an overview of the large gap that exists in the vast majority of years between the TAC and the ABC for Bering Sea Pollock. It was a stock that was discussed during the Joint Plan Team as maybe a particular stock of interest in in terms of this modeling effort. I think there's a lot of information to suggest our current management system and the buffering that we are required to do under the 2 million metric ton of YCAP system already incorporates a pretty robust response to a wide range of climate responses for that stock over an extended period of time.

Fourth question is, uh, why would we not be— sorry, fourth question is, do we want to narrowly focus on stocks where there it is clear that TAC is set equal to ABC and catch is close to the TAC limits. And again, Figure 2 in my testimony basically gives you a quick overview of the fact that for a large number of stocks, there is a vast gap between catch, ABC, and then the OFL. We have a lot of stocks where there's a lot of buffering that already is included. How applicable will this effort be across the broad range of stocks that we have? Do we want to focus on stocks where we know that there's been some clear response to climate.

In other words, we have a lot of stocks where it seems to suggest that either our buffering or the stocks themselves may not be responding to the current conditions of climate. Do we want to focus our efforts? We have a clear example in the Gulf of Alaska with Gulf of Alaska Pacific cod. I think you may hear from other testifiers that perhaps focusing on simulating the potential applicability of this type of approach to that type of a stock may be useful as a first step? And then finally, how does this fit within overall scientific priorities?

I guess to roll all this up, the questions are: why this approach? For what stocks? Why now? And what is the opportunity cost for this investment? Second issue I wanted to talk about was sablefish.

Briefly align myself with a lot of the concerns that have already existed here. I think one thing in particular caught my attention during during the discussion with the analysts. Current model may be overly optimistic, but the fleet Cesarean model may be pessimistic. And I guess one question I would have is, should the model choice be informed by the fairly well-accepted fact that we have decent survey conditions reflecting a large abundance of sablefish currently? In other words, is pessimism warranted under the current conditions?

I think the SSC has flirted with ensemble models in the past with Bering Sea Pacific cod, that was proven to be problematic because it was such a complex issue with, I think, 5 separate models. I think you already heard from one testifier the idea of maybe an ensemble light approach here might be something worth considering, and perhaps some guidance on that could be useful. And then my final point that I wanted to make was concerning deep station coverage in the survey. This has come up, I think, repeatedly, we have not had a slope survey for a decade. We participate in the turbot fishery, we also harvest sablefish as well.

We are also constrained by OROC, other rockfish, which is a limiting factor in many of our deepwater fisheries. I think it's well understood that those deep stations can provide a lot of information on those critically important stocks. Having some additional survey coverage and the SSC providing guidance that continuing to pursue the ability to have some level of deep station coverage would be, I think, particularly valuable for understanding those stocks and potential impacts. Thank you.

Okay, seeing no questions, thank you for your testimony. We will in fact note your testimony on the Bering Sea when we get to to that section. So thank you for combining for efficiency.

Next in our list of public testimony, we have Mr. Jim Armstrong.

Good afternoon, co-chairs and members of the SSC. Jim Armstrong, with the freezer longliners. I'm going to comment on the HCR issue and then a couple thoughts on sablefish and the Pacific cod model overhaul.

We commented on HCRs at the council meeting in June. We didn't— I think we did not comment to the SSC at that meeting. Our concerns now continue to be on the need for a clear demonstration of the benefits that would come from any of these alternative HCRs. HCR 7 is the option that explicitly factors in environmental information, and that's been suggested for cod. As I understand it, harvest rates would be responsive to anticipated changes in stock productivity associated with environmental covariates under that control rule.

Um, it continues to be unresolved, at least to me, whether an alternative HCR would have led to a different outcome during the Gulf Cod disaster. If we're trying to find clarity on this issue, um, we encourage the SSC to, um, ask the analysts to demonstrate how things would have been different in the Gulf if HCR-7 had been in place. What alternative ABCs would have been set, when would those ABCs have diverged, how much biomass would have been preserved, and how would the trajectory of biomass have been different, and very importantly, how it impacts the stakeholders and communities have been different.

Maybe another bookend, since sablefish seemed to get a real boost under the same conditions, ask the same questions for that stock.

Um, we shouldn't be charting a timeline for alternative harvest control rules before we've demonstrated that they would in fact achieve climate resilience. I note that in June, several SSC members recognized the need for greater clarity in the conversation on this issue, and nothing would achieve that more effectively than running through some specific examples.

Whether or not HCRs are intended to make us more resilient to shocks like Gulf Cod, then perhaps the analyst could demonstrate the improvements that would come from factoring in long-term trends and climate conditions. Under our current process, harvest limits are adjusted annually based on most recent data collections. It seems, uh, to me anyway, that the additive effects of multi-decadal, uh, decreases in carrying capacity or recruitment are going to, uh, either be undetectable or, or be swamped by other near-term population effects, essentially adding complexity to an already complex process. Um, if we were trying to predict what average biomass would be in 50 years, long long-term trends would be directly applicable, but when we're talking about next year, it's less convincing.

Um, the SSC is very conscientious in its approach to complex fishery management issues, and the climate resiliency, uh, and climate resiliency is an ambitious aspiration. Nevertheless, we encourage you to fully consider the resiliency that existing processes already provide and clearly demonstrate evaluate the improvements that a given alternative would provide.

On sablefish, FLC members who participate in the sablefish IFQ fishery join others in asking the SSC to provide advice to the Council that would allow for a gradual stepwise transition toward the revised ABCs that the, the FAA model is generating. Much of that change reflects correct, uh, a correction in model structure rather than a newly observed decline in the stock. So a cut, um, in a single year would be very disruptive, and so we encourage a gradual transition. Um, lastly, FLC is tracking the overhaul of Pacific cod, uh, assessments. Don't have any comments on spork yet, but I'll note that it draws heavily from genetic and tagging data Our vessels have served as platforms for putting out satellite tags and collecting genetic samples.

In the past, uh, the Gulf Cod disaster created funding for a lot of tagging in the Gulf, but involving our vessels would help tagging and genetic sampling efforts be more consistent with the spatial distribution of cod harvest, especially when they're over spawning grounds in the A season. That's it. Thank you.

Great. Thank you for your testimony. Uh, sorry, we— looks like we do have one question, if you're willing to take it. Martin, go ahead. I can let you off the hook, right?

No, I was just trying to understand what your last point was. Are you suggesting that your boats be used as platforms for tagging? Is Is that the general gist of it? Through the chair, thanks, Dr. Dorn.

We have put tags out, satellite tags from our boats on a couple of different trips, and we've actually been collecting genetic samples as well in anticipation of the, you know, the need to know more about the stock of origin for what you're catching, I think, under the new model. So, if you look at maps of where genetic samples are coming from and also where tags have been deployed, it's really, it's not, doesn't look like where cod get caught. There's a, the survey has provided a primary platform for putting out tags and getting samples in the summer. I think the big interesting question is where, you know, what is the contribution of the different spawning areas to the overall catch? And so you really need to put the tags out, you know, during the A season.

Yeah.

Rob?

Yeah, just a follow-up question to your comment about the locations of samples for genetic sampling. And I'm wondering, are you referring to the figure that was shown where it seemed like the, for example, that the genetic, the bubbles, the, you know, for the genetic samples in the southeastern Bering Sea, they were all from on the inner shelf. And I don't know how location-specific those are, but there were no bubbles out on the outer shelf, shelf break, where I know a lot of the fishery occurs, at least in the fall. Is that what you're referring to?

Through the chair, yes, that's exactly it. If you look at, I think we would be the blue, light blue dots, if you have the colors in your mind as well, and those are like the smallest number of dots, so it just seems like we really need better coverage of that.

Okay, thanks again for your testimony. Now we have Mr. Paul Clampett.

Um, I thank you, Chairman Stewart, and then count— and committee members. Um, My name is Paul Clampett. I'm the president of the Sablefish and Halibut Pot Association and vice president of the Fishing Vessel Owners Association. I own the fishing vessel Augustine, which fishes its full— which fishes a full sablefish cap each year, almost entirely in the Gulf of Alaska. Thank you for your— thank you to the assessment team for a great deal of work under this— under difficult conditions.

My message is simple. For 2027, the committee should base sablefish catch on the single region model 26.1. It's credible, it's proven, and it produces the smallest reduction. And the author's own summary states that the change in ABC across these models is not due to any biological change or concern. It's simply an update of model structure.

There are 4 reasons. First, the stock is healthy and growing under every model before you. Spawning biomass is 67% of unfished and rising, far above the 40% target. Target. No model shows overfishing.

Second model, 26.1, meets every scientific test. The Center for Independent Experts review— stubby— found it— I hope I'm pronouncing that right— found it scientifically credible and passed every diagnostic. In, in the author's own simulations, all the models produce similar biomass and cash trajectories, and it agrees closely with the continuity— with the continuity model, the successor to the model behind the current ABC ABC: 40,900 tons and 39, 39,600 tons. Two independent runs point to the same answer. Third, the Fleet Cesare model is not ready.

It would cut the ABC by a third to roughly 32,000 tons. That cut rests on key assumption— one key assumption. The authors ran the same model using the carryover survey index instead. It produced an ABC of 54,000 1 assumption about how to treat unsurveyed areas swings the answer by 22,000 tons. Notably, the authors propose the same carryover approach for apportionment.

If carrying forward the last observed survey value is sound enough to divide the quota among areas, it should not be dismissed as unreliable when it points to a larger stock. The model also relies on very small age samples from the Bering Sea Nualuchians, collected only every other year. And reviewer Hartman found that under the reduced survey design, it is likely to underestimate recruitment during periods of strong recruitment, which is exactly where we are now. Fourth, the risks are not equal. With the stock— with the stock this healthy, choosing the higher estimate poses almost no risk to the resource.

Choosing the lower one, if it proves wrong, costs the fleet millions of dollars in harvest that could have been taken sustainably. These costs fall fall on the people who actually fish. The quota that goes unfished is in the Bering Sea and the Aleutians, not the Gulf. In 2025, the Gulf landed 85% of its sablefish. The Aleutians landed 2%.

For my vessel, the fleet says area model means a loss of $577,000 a year, and that is gross, which cuts the crew. Model 26.1 cuts that loss by more than half. It would also be the biggest one-year cut in the fishery's modern history by a wide margin. Since ABCs were first set in 1988, the largest single-year reduction was about 23% in 1996. A 32— in 1996, a 32% cut would be roughly half again as deep as anything this fishery has seen.

Even the declines of the early 1990s was taken gradually over several years, and when the big year classes first appeared in 2019 and 2020 Originally, this committee held the ABCs at about half the model's maximum and stepped it up over several years because it, it was not yet confident in the model. The same caution should apply now. My requests are these: first, base the 2027 ABC on Model 26.1 and ask the authors to present its full diagnostics in November, not just for comparison. Second, pair Model 26.1 with carryover apportionment. The authors have proposed.

My calculation from the author's own table, that combination keeps the Gulf within 3% of its 2026 near 22,000 tons, with most of the reduction falling in the Bering Sea and the Aleutians, where most of the quota goes unfished. Because that quota largely gets uncaught, actual Alaska-wide catch would likely stay near today's level of 23,000 to 25,000 tons. Third, defer any decision to replace 20— Model 26.1 with the fleets as area models until the model has been tested against the 2027 Gulf Survey. And fourth, ask the authors to update the natural mortality sensitivity analysis for the 2026 models and to present the recruitment bias and analysis as independent reviewers recommended. A healthy growing stock does not call for a one-third cut on an untested model.

Thank you, and I'm happy to answer any questions.

Thank you very much. Thank you for your testimony.

That concludes the signups for our oral public testimony. We will note there were attachments associated with several of those testimonies, and we also have 15 separate written comments which we will consider in our discussions.

Okay, so that brings us back to— we have 3 items on the table right now: the Joint Plan Team discussion, harvest control rules, and sablefish. And I think we will start with the Joint Plan Team discussion, which will be all the other issues that we saw in the report. And we'll go first to Rob.

Okay, thank you. And yeah, just Overall, thanks to the plan team for all their efforts and the details in their reports and their presentations and the recordings online that we were able to view. I'll start off with the ESPs and DCIM comments. So the plan team report included 3 questions that were requesting and requested comments relative to those questions. They weren't presented presented to us this morning, but I'm just going to run through them with some suggestions that I would make and just see if there's other SSE input.

So the first question being, should the ESP and DSIM workflow become a standardized AFSC, standardized for AFSC ESPs, and what concerns should be addressed? So I, I would propose that the SSC does support a a standardized workflow. And as far as concerns, I think it's— we often hear of just whether an ESP has a DSIM, but there's lots of different reasons for the DSIM for a particular stock or use of a DSIM for that stock. So identifying what what time-varying parameters is important and to be— that the DSEM would be focused on. Also, and this is based on partly which of those parameters has the best potential of being informed, such as reduced uncertainty or variance by additional or ecosystem information.

So to help identify which of those parameters are our priority, and so that this can inform these kind of causal predictive relationships. Metrics that are not found to be predictive or quantitatively informing the assessment regarded as contextual, as we've heard from crab plant teams and the groundfish plant team in the past too, that they're still valuable for risk tables and ESRs. And then decisions to assess predictive linkages could be seems like they should be annual and evaluated on an annual basis for annual assessments, or when, you know, whenever the, the cycle of the assessment is. And as far as the time frame for where the DAGs are assessed, the linkages, that could potentially be on a longer time interval. And I think with those decisions made by the assessment author and and in particular, subject matter experts collectively.

The second question: Are the proposed diagnostics sufficient to evaluate ESP DSMs, or additional diagnostics needed? I mean, so far, yes, but we'll need to evaluate those as more progress and more applications occur in the future and more details are provided for further evaluation by the SSC. Given the proposed workflow, should the current list of stocks identified for ESPs be revised, and which stocks should be prioritized or added? I don't— I think that's a broader question to be addressed that we weren't— we didn't really discuss enough information to look at that, but I think for the most part, you know, the annual assessment assessments are, are a priority, and where there's— and any assessment where additional information might be informative for, for specific parameters in the model. And I would recommend that SSC supports the plan team's recommendation on developing the workflows to document evidence for these languages for the DAGs among variables and create a unified process among the teams.

Also in supporting dedicated staff rather than relying on the assessment authors for this effort. But I would also emphasize that I think it's important that we, that, that we emphasize that there's a need to have a lead for these efforts. And so, I mean, in most situations, or in many situations I can think of, it seems like the assessment author best person to be at the helm of that, but regardless, it's important that somebody is leading the effort for a particular stock and the assessment author is involved. We heard a report of a longline survey, which there was a lot of what was discussed during the sablefish stock assessment review, and so I'm not going to go through too many of those, just a couple couple summaries of those sablefish RPM increased by 26% in the Aleutian Islands and decreased by 14% in the Bering Sea. Abundance remained well above the long-term mean and was at or near historical highs for the Aleutian Islands.

The Pacific cod RPM, the relative population number, decreased by 22% in the Aleutian Islands and increased by 33% in the Bering Greenland turbot RPM decreased by 58% in the Aleutian Islands and 41% in the Bering Sea, continuing a recent declining survey trend. So, and that's been consistent for a while. Short rakers are up in both areas for Aleutian Islands, Bering Sea, and then they're rough-eyed blackspotted are up and the AI are down, or Sorry, up in the Aleutian Islands, but down in the Bering Sea. I think it's worth noting that the temperature index that they provide is really valuable. It's approximately 250-meter depth temperature, and it's at near average over the time series.

And that's to be expected, kind of these more deeper, more isolated areas. So despite all the warming we're seeing in lots of other areas, this deep water habitat for at least in terms of a temperature field, seems pretty stable for sablefish.

And then I'll just—. A few statements about the Pacific cod research. The genetic structure within— there's clear genetic structure within the Gulf of Alaska and also within the Bering Sea, and that the Aleutian Islands are separate from those two. But the lack of genetic structure that's been discussed between the central grower Western Goa and Eastern Bering Sea is definitely a conundrum at this point. As what was just mentioned, additional genetic information would be valuable and a more— and improved spatial representation of those data.

But in reality, unless there's new ability to distinguish that Central Western Goa, Gulf of Alaska, and Eastern Bering Sea then there's going to be the, um, tagging data is going to be the only option for distinguishing movement within those regions if they cannot be distinguished genetically. I know they can't currently, and I don't know what the future holds.

Um, for the, um, Okay, so, and then a couple as far as some explorations into spatial models with stock synthesis for the Eastern Bering Sea versus the Spork approach. I think the, um, the what was presented by the Bering Sea author in terms of assessing stock synthesis versus a spork models. Um, there, there's a lot of similar capabilities, um, in terms of models allowing variability and annual catchability fit survey indices better, and models allowing movement fit size and age compositions better. And stock synthesis and spork produced both similar results. However, the stock synthesis approach was difficult for trying to understand spatial dynamics, so the team recommended using Sporc as a better platform for addressing spatial dynamics first framework and then potentially incorporating movement data, random effects, and simulations.

And then the, the work in the, um, with the Gulf of Alaska assessment author and the application of work. Um, this also, um, with the spatial model evaluation will continue and be included in, um, a CIE review that's going to be in spring of 2028 and then presented to the plan teams, I presume, in the fall of 2028. Um, they evaluated 4 combinations of 2 versus 3 population structures, and and alternative regional definitions were evaluated. So the goal is to use an MSC to evaluate the consequences of status quo, ABC and TAC, or alternative apportionments across seasons and between Eastern Bering Sea and the Western GOA. But it's important to have this spatial model developed before that management strategy evaluation.

And then the, The team noted primary data limitations of no fisheries independent data during the winter season, the A season, and then the movement data remain limited for them.

Let's see, in terms of in-season management, there was the reference to declining Amendment 80 vessel participation from 20 down to 14, which is quite substantial. And then overall for the Bering Sea, as of September 1st, overall catch was similar to 2025, with a broader decline in fishery utilization and continued reduction in, in Amendment 80 vessels, as I mentioned.

And they noted the importance of voluntary industry cooperatives in managing choke species and avoiding exceeding catch limits.

Let's see, for the GOA, for pollock, in Area 620 was fully utilized during the A season and new entries of vessels, while Area 610 was not fully utilized because of limited vessel participation, shifts in processing, and other species.

And for just the highlight also in for pollock, Area or for PCOD, for Pacific cod central gulf, fixed gear expected to be fully utilized due to favorable high prices. Western Gulf of Alaska Pacific cod A season was not fully utilized because of low catch rates and a pause in fishing, but increased effort was expected in the B season, including vessels moving from the Bering Sea. And I think I'll leave it at that for the in-season report since those are those are the primary stocks we're going to be talking about here at this meeting. The halibut DMRs is mostly just an update of, you know, the past 10 years using observer information assessments to estimate mortality rate for halibut, and the primary method change in 2017-2018 was removal of Gulf of Alaska rockfish, Kesho vechil operational group as a separate DMR category due to lack lack of support for prior concerns. And these observations will be incorporated into a broader non-pelagic trawl catcher vessel operation group.

And otherwise, most of the DMRs have changed little.

And I'll leave it at that.

Thanks, Rob. Next, we'll go to Jen.

Yeah, I want to thank the analysts and plan team for all their for the charts and reports and the very clear presentations, and also Rob for a detailed and thorough review of all of the main aspects of the report.

We, as mentioned, we received numerous, a detailed presentation on ESP and DSAM and how it interacts with ESP, and the ESP authors presented processes and requested some feedback from us that I thought Rob outlined nicely. The questions about structured workflow and formalized steps, structured questions about suite of diagnostics, and kind of the discussion of levels of application from start through full implementation and integration of the ESP and DSUMs.

I agree with Rob that the SSC should support the team's recommendation that stock-specific workloads be developed and that Where my notes are, that informate workflows, specific workflows should include information on how the various variables are linked as those are developed.

We should, so we should support structured workflow, formalized steps, standard processes.

Question about which stocks should be prioritized and evaluated, really didn't get into that in this meeting. So will be something we should consider at a future meeting with more information.

I do note that a couple of times across today's presentations, we heard heard, uh, issues of the impact of reductions in personnel and, um, both for handling and processing data or developing models or, um, actually conducting surveys in the latest case. Um, and I think it's really important that we highlight when those reductions are impacting, um, what and how we can consider things and when they're impacting, for example, like we just heard, potential quota and catch limits. In the case of ESP and DSIM, it's really clear that the staffing capacity to develop these models will be important, and I echo the recommendation that this not be just another task that is assigned, is given on top of as an overload, but that some careful consideration be made on how we do that.

On the— Longline survey information. Rob gave the summary of the various trends of species and overall patterns.

It's clear that there's a lot of good information coming out of those surveys. It's also clear that the lack of aging data is, has the potential to impact models moving forward. We really should consider how to address those issues.

And, um, my notes here.

We heard relative to, um, sailfish management strategies, in particular, I was struck by how the lack of changes in survey design and structure are having a variety of very real-world consequences. And yet it doesn't seem to be something that we can address through changes in survey design. Fine, but it does seem to be something that we should consider recommending inclusion of the fishery-specific data. So I think Rob otherwise hit most of my scribbles, so I will leave it at that for right now.

Great, thanks, Jen. I'll look to comments from other SSC members.

Sorry, go ahead, Jason. Yeah, I guess I just wanted to highlight my notes. I think, I think Rob and Jen covered it, but just, just we heard public testimony from several people about just the concerns about the loss of out, not only the longline, but also the deepwater slope stations. So I just want to make sure that, you know, that's highlighted in our minutes. I think it was said, but I just wanted to reiterate it.

Great, thanks. Chris. Yeah, I think this falls along the same line. I wanted to sort of amplify Jen's point on the value of cert—. The discussion we had about the value of the surveys and the testimony received about the value of the surveys.

And in general, emphasize that we should have a strategic approach to understanding the value of information in terms of what we're giving up in catch because we don't have that information. And then in the spirit of adaptive management, we should reevaluate whether the initial cuts were in the right place Thanks, Ali.

Uh, yeah, could we also, um, I guess, um, I suppose that this makes sense going on the last couple of commenters as, as a general comment rather than something in the joint Plan Team report as a general groundfish comment. Does that make sense to folks? I was also thinking that we should broaden it a little bit more to not just talk about surveys but also aging capacity, which has been limiting in a number of stocks and for a number of years.

So I would, I would make those two suggestions.

Rob, go ahead. Yeah, I agree, and especially since what we just brought up as far as the slope survey isn't the longline survey. I mean, that's the trawl slope survey in the front of Bering Sea. So yeah, it's, it's probably a broader topic for than just this effort. But we're— it's super important, and to also address, as was mentioned too, rather the, the— it all costs money, but what's more— what would be more valuable and realistic to implement too, in terms of, for example, the logbook data?

Okay, I will summarize then. We're going to thank the plan team and the analysts for their efforts Going through the list of items that were presented to us, first ESPs and DSEM. We will support this— the move toward a more standardized workflow and request identification of which time-varying parameters the DSEM would be focused on with the goal of reducing uncertainty and variance due to the addition of information.

We will suggest assessing the linkages on a timeframe that's created by the authors and the experts. We will suggest that the diagnostics as are currently presented seem to be reasonable to evaluate, but we will need to likely continue to evaluate that as we see more of these examples and case studies. We don't have a specific recommendation on stocks for priority. Based on the information at hand, but perhaps annual assessments would be considered a higher priority in the future. And we will support the development of workflows for creating the DAGs and adding staff time beyond just assessment authors.

However, we will note that there does need to be a lead for these efforts, and that's probably likely going to be the assessment author.

We will include— I think it makes sense to include a broader discussion of reductions in staff and money and effort in the field across both the GOA and the Bering Sea. This is going to include reductions in personnel for data modeling and surveys. And we will highlight that it would be really helpful to have these changes linked to specifically which products they impact and also the strategic value of loss of information. What are we losing compared to what's been cut?

We will note the trends in the Longline Survey and also the, the reduction in effort there, as well as the general reduction in aging capacity and the age data that's being generated from these surveys. We will note the— that this is not just a reduction issue in the longline survey, it also has to do with the loss of slope stations in the trawl survey.

And so for the public, this, this will probably turn up either in our joint minutes or even higher up just under all groundfish topics. We will, specific to the Joint Plan Team, Um, we will have some comment on the Pacific cod research, um, generally supporting additional genetic information, but noting that improved spatial representation in that case could help, as well as additional tagging data, which could prove to be a more useful tool in distinguishing movement beyond just genetic differences.

For the exploration into spatial cod models, We will support the Joint Plan Team recommendation to further develop the SPORK implementation and look forward to the, both the CIE review and the potential operational model in 2028.

We will summarize the in-season management trends and also note that we will check the box on the halibut DMRs for spec setting this year.

I think that's everything we had in our summary.

Okay, we'll go now into our discussion of the Harvest Control Rule update presentation.

And first off, we have Cheri, but I think we'll go right to Frans.

Uh, yeah, thanks. I'll try to, um, kind of, uh, blend what Cheri had sent me, and I, I haven't integrated that all, so there may be some duplication or discontinuity. Um, but yeah, we received an update on the ongoing Harvest Control Rule discussions, including a nicely condensed list of past SST, SSC requests and timeline for further work, and then the JPT recommendations associated with that. Uh, we should thank the presenters and the whole team really, and many groups, the many groups without really naming any that who have provided input over the years on this issue. We'll note that we received public testimony from Glenn Merrill and Jim Armstrong on this issue.

And I think we should just note that the SEC appreciates substantial progress in moving the HCR discussions from what kind of started out as broad concerns, really, whether the existing system is robust to non-stationarity in a changing climate and other issues that have been identified over the years in a changing environment toward a more quantitative evaluation of alternative management strategies.

And in that context too, I would note that the SSC might consider the ongoing sablefish management strategy evaluation as well as ongoing kind of pollock and Pacific cod simulations as important case studies for developing a general framework for evaluating future ACR modifications as we may apply this to other stocks.

I also took a quick look at the kind of the history of this issue in the context of all the different discussions over the years for kind of a bit of perspective, and it's nice to see a fairly coherent, if somewhat sinuous, progression from these broad concerns to now a much more concrete program of the stock-specific simulation, simulations and simulation testing that essentially ask for what stocks and objectives may alternative harvest control rules actually outperform what we do currently.

So again, just kudos to the teams for all the work.

Um, I have a list of specific comments here from Sherry that I haven't fully integrated with mine, so I'll do my best to characterize those.

The presenters had asked if May would be a good time to look at some issues around MLOM6 validations and forecasts and review the modeling platforms or approaches like the food web model and other issues. The SSC supports this as long as there's time for analysis— for the analysts to respond to SSC recommendations in the timeline.

Uh, presenters also stated that the team is focusing on communicating, uh, what's robust now about the current harvest control rules and addressing individual stock changes. And I think public testimony also reiterated this point. Um, the SSC agrees that the context is really important and supported this approach, and we point to our June meetings, which said the current management system has on the whole been robust and highly successful.

We appreciate our continued work on economic and community metrics and indices.

We support including the Groundfish Plan team performance metrics as practicable on metrics of sort of frequency of stock collapses, collapses or dropping to zero, and rebuilding after a shock, and others as listed in the Plante Minutes.

Like the plant team, the SSC would be interested in hearing conclusions from stakeholder input following an engagement strategy at some point as practicable.

And then, uh, um, Chariot asked about, um, kind of the timeline for, for CRAP and, um, noted some delays in, in that because of the, you know, needs for the team to start with some examples now where progress has been made and then reduced time and lower prioritization by the CPT to take up this issue.

But, uh, analysts noted they would present to the crab plan team, uh, when the, the time allows. Uh, the SSC still strongly supports including crab stocks in the HCR work and suggests continued conversations with the crab plan team and including crab in the timelines going forward.

The council motion on June 26 included a statement saying outline the scalability of this work under varying budget scenarios, which includes survey data. Uh, so the SSC might request the HCR team report on their responses or planned responses to the council to how the project would, uh, might change in our different budgets. For instance, under low budgets, what some kind of state projects that might be slowed under the same goals and coverage, or dropping some HCR scenarios, or dropping specific species and so forth. Similarly, when trying to address how project— the— how the project would address having varying levels of survey data, the SSC suggest the team consider some of the previous SSC recommendations, and I think Sherri was going to include some links or pointers to where we had previously commented on that.

Um, and then we support the retrospective comparisons with assessing impacts of marine heat waves on COD as recommended by the JPT and previously requested by the SSC and also mentioned in public comment. And specifically, we support the simulation approach to look at performance of HCRs under marine heatwave conditions, including analysis of how they would have performed in the past.

I would encourage us also to— and maybe I'm kind of looking for some advice on that, but, um, I think as a, you know, collective council family here, and I encourage us and the council to start thinking about, uh, kind of revising the overall tier structure in parallel with, um, while these evaluations are, are ongoing. Um, you know, there's lots of different options that we could think about because some of these proposed harvest control rules might, you know, require some changes, but it's also an opportunity to kind of step back and see what has worked over time and what hasn't worked and where can we potentially, you know, simplify given that we also add complexity. So options might be to, you know, say collapse Tier 1 and 2 into a single tier for stocks that have a built-in stock-recruit relationship, but then revise how ABC is set below the MSY. That could include climate-informed rules such as the HCR7, or, you know, collapsing Tier 1, 2, 3 in a single tier for stocks that have an age-structured assessment and then add flexibility to accommodate stocks with or without a stock-recruit relationships and the different harvest control rules, such as the ones that are currently being evaluated, or alternatively restructuring tiers to accommodate kind of multiple tiers for stocks with different information availability that kind of reflect where we currently are at, and that would, you know, take take some, take some thinking about, um, how to maybe do that better than in our current tier system. Um, so, uh, I guess I'm bringing that up to maybe say if they wish the Council could perhaps initiate a discussion paper to think through the particular practicalities of implementation, uh, now that we have an outline of what potential modifications are under consideration and, and what additional flexibility may or may not be warranted.

Um, this would seem like an appropriate step following, uh, the Climate Resilience Workshop, which among other things, uh, you know, asked for a review of the existing tier system. But I'll note that depending on the results from the analyses, we may also find that we don't want to make any changes. So that may also be somewhat premature.

With respect to the different HCR options, I suggest kind of following some of the points that Kirsten Holtzman had made in kind of separating the HCRs intended to be robust to climate shocks and surprises like HCR 5 and 10 from more climate adaptive HCRs like HCR 7. Um, and this is in part in response to a question that the team had asked to the, to the plan team as well. Um, you know, for most stocks, certainly at this point, we don't have any and may never have informative links between productivity and climate And so in those cases, it seems that the most important thing would be to develop a general HCR that's robust to all kinds of variability and shocks and, and leave that in place. So basically tweaks to our current HCRs to make them a little more robust, but not making them adaptive. Informative.

And then in case where we do have informative climate links, use more of the HCR 7 type rules that adjust elements of the HCR, such as, you know, max FABC or B40 or others.

And that, that could be applied to either the, you know, current shape of the, our HCR or current HCR, or to the, you know, proposed HCR 5 and 10 in similar ways by linking the parameters of the HCR to temperature, other variables like they are proposing.

So essentially, just sort of focusing on those two aspects, you know, building more robust harvest control rules that are relatively static and, and of building more adaptive control rules, but more of a bifurcation.

I think we already commented on this, so I'm not going to comment again on that.

Yeah, and then, um, you know, there were some questions about the objectives, uh, and, you know, I, I think we might, um, suggest for kind of ease of, uh, for facilitating the analyst's job to compare the different harvest control rules that the council maybe develop a smaller set of ideally prioritized objectives that would help them evaluate trade-offs among the different HCRs.

Um, with regards to the performance metrics, um, the June paper organized the performance metrics kind of into stock status, yield, biological stability, and societal and ecosystem categories. And, uh, we might suggest to add a category to evaluate transparency and/or simplicity to, you know, address some of the, you know, address that specific objective that has been formulated around transparency.

And then with regards to kind of how to organize the perform and communicate the performance, you know, there's kind of— I think we had kind of commented on that before and may want to reiterate that for communication purposes, to have sort of a small set of easily communicated metrics. And there's maybe a couple of ways to go about that. To perhaps select sort of one key metric or a couple metrics for each of the categories for presenting and communicating results to make sure that, you know, all the different objectives and categories are represented, or to combine metrics in each category into a single either unweighted or weighted, you know, mean to present sort of aggregated metrics. Metrics for each of the categories relating to the objectives. It's kind of two alternative possibilities to think about that.

And I already mentioned that we support the planning recommendations for the additional types of performance metrics around frequency of collapse and time to rebuilding Yeah, and then maybe finally, I suggest we recommend, you know, using really the current HCR plus the risk table process as an explicit benchmark in all the evaluations. Variations. I mean, generally, you know, the central question is whether the alternative HCRs, as was also highlighted in public testimony, provide, you know, demonstrably better and more consistent performance than the flexibility that we already currently have through the assessment updates and the existing HCRs and the risk table adjustments.

Further, I think we might note that we also recognize the challenge of simulating the inherently qualitative approach to determining buffers that we currently are using. Evaluating performance against the status quo may itself involve some kind of qualitative considerations around the use of the risk table. That's all I had.

Thank you. So, please don't complain for me. Thanks, Franz. And I should note that we are dealing with a rolling challenge of availability for SSC members, staff, analysts, and other people. So it's to nobody's discredit that if we miss a few people, we're trying to rearrange the schedule as best we can.

Are there other SSC comments or follow-ups?

Chris? Yeah, I guess I would just, uh, I'd like to reemphasize, um, Franz's comment about the objective of transparency. I think that's a super good thing. I think that would be super helpful for everyone. Uh, and I don't remember, but as we were having this conversation and that came up, um, it might also be good to have an objective of time and/or effort that entails that new method versus the status quo of an assessment assessment and a tier and a risk table.

I'm just kind of thinking that the whole thing through of what that really entails as a full, full new assessment with the associated HCR. Thanks.

I had two comments personally. One was just to make sure that we retain under the discussion of scalability under future budget scenarios. I think that we— there was definitely a secondary or maybe even primary issue about, um, are we creating rules that are more data hungry? And if we have less data in the future, is that going to be the issue? And that's, to me, that's slightly separate than budget issues for the team that's actually doing the analysis now.

This is more in terms of long-term participation. So hopefully we can capture both of those things in our minutes. The only other thing that I'd like to hear, maybe other opinions or a follow-up on, was the, the idea that we might suggest a council decision paper or note that one could be done on overall evaluation of the tier system in conjunction. And I'm just posing the question, is that, is that likely to be most productive before we see results of this exercise, simultaneously, or after? And I don't have the answer, but we may want to be fairly specific as much as we can there.

Dana.

It sounded like your, your last comment was about resource constraints. But this would be at least doubling down on resources if we want to do these simultaneously.

Yeah, I was actually just trying to reiterate what Franz had proposed and just noting that there are some challenges there. I think I agree with you. Jason?

I guess I wasn't clear exactly what Franz's proposal was in terms of the timing on the discussion paper and how it fit into this overall process. So, and what we're actually sort of recommending to the council at this meeting. So maybe, I don't know if there's some way to kind of clarify that or have that discussion, that would be helpful. Yeah, I mean, I initially suggested that, you know, it might be time to kind of start that in, in maybe in parallel and rethinking the overall structure while we still have flexibility in sort of shaping the harvest control rules as well. But I also caveated that by saying that, you know, if, you know, the results suggest that, you know, the harvest control rules that are favored all well within our our tier system now, or that we want to stick with what we are, um, what we are doing, uh, you know, now.

If that proves to be as robust as alternative approaches, then, then that might be kind of wasted time. Part of what motivated this too was just, you know, more broadly actually You know, we've found now with the Pollock stock and, you know, considering perhaps in December changing Yellowfin Sole tier assignment as well, that, you know, the Tier 1 and Tier 2 approaches seemed like a great idea at the time and made sense. And but I think we find that stock recruitment relationships generally just aren't that informative enough to really implement those as designed. So that would be another reason to kind of reconsider perhaps, at least the tiers 1 through 3. The others would, you know, obviously still address very, you know, different data levels or data availability.

Jason. Yeah, I guess maybe a follow-up and clarification then. I mean, so when the HCRs that we discussed today, we, we selected specific species to focus on for this group, um, and I guess it sounds to me like, you know, part of what you're describing is our current baseline, but then something else that's maybe beyond those species, if it's looking at the overall tier structure 1 through 3, which seems a little bit different maybe than what this group is working on at the moment. I don't know if there's some way to think about it though in terms of kind of the baseline for what this group is working on, and maybe it sort of follows that project plan. I guess I I guess what I'm nervous about is recommending a discussion paper to the Council but not really having a full understanding of what that is that we're recommending scope-wise.

Yeah, I'm certainly— I mean, I was kind of looking for feedback on that, and I'm certainly not— I don't think we need to, you know, make a strong recommendation to kind of initiate an analysis now along those lines, but I think we should— or, you know, as together, I think we should think about, you know, does it make sense to, um, to— given that we are in this process of identifying, uh, or, or testing a number of different harvest control rules that may or may not fit within our current system system, um, you know, does it make sense to maybe, uh, resync the tier system overall? But I'm okay with just, um, you know, highlighting that, that, you know, review of the tier system is kind of on the docket, you know, that was part of the, part of the plan.

Chris. Yeah, just thinking this through, kind of the end result. Will this— if we adopt or kind of move forward on some HCR difference, that's going to require an FMP amendment. Is that true? So we're going to end up opening that up.

Is that— so I'm trying to get at what Franz is getting at. It's like, if we're opening that up, do we take the time and effort to do it, or just— and then do we have HCRs that are just for a couple of stocks in the FMP? It's going to be a fairly awkward document edit amendment. And so I think that's, that's where I was getting at, where your comment was coming from, is like, if we're going to do this, do we need to be thoughtful about what the end result is practically going to be? And that's going to be amendments to the current FMP.

For both groundfish and crab potentially. Thanks.

Kaylan.

I think we've made this comment in the past, but I think it would be incredibly helpful in tying with the communication piece if there is a really clear articulation of what the goal of this process is and And I've sat through this a few times. I'm listening to this, these comments and questions now, and I think a lot of these harvest control rules, if I understand them right, the council has full sort of purview to choose the tack that they want that would be consistent with some of these harvest control rules. And so I think I'm just always a little bit confused about what this harvest control rule specifically would gain in this process. And I feel like then some of these other questions, to me, it seems like it might be easier to kind of follow through on those questions as well.

And I know I had asked for this before because I sit here and I never quite get it. And then I am looking at you all over there who seem really in deep. And so that would just be such an incredibly useful thing for me. And I, I just would tie it back tack too. I'm pretty sure all but one of these harvest control rules, the council could choose the tack that would be consistent with them, if I'm understanding things right.

So what does the harvest control rule specifically gain?

Jason?

I don't have any clarity to your question because I think— no, um, I'm just thinking about process, and I guess I'm, I'm a little, I'm a little concerned that we're sort of reopening up, um, ground that we've discussed in the past. But I think that's okay. I just, um, I guess personally I, I don't I'm not comfortable making a suggestion, you know, for a full discussion paper at this point on what a new tier system would look like without really understanding the scope of that. So I, I feel like I'd be comfortable making a comment. This might be an opportunity, um, for us to, to think about that, um, as going along with this harvest control rule situation, you know, this evaluation.

But I don't, I don't know that I'm personally feeling like a discussion paper is the right step right now. And I mean, I, I feel also that we just had a discussion about FMP amendments, but that seems like that's down the road if that happens. And so I don't know that that's necessarily where we're at right now. I think that that's— we haven't—. We were still going through this process and this evaluation, so I don't really know that we should be commenting on what would be an FMP amendment or not.

It's just my view on that.

Franz. Oops. Yeah, thanks. Yeah, I wonder if we can actually maybe get some clarity from Diana perhaps on just— I mean, I, I had in part motivated this by, um, you know, the climate resilience work plan as reviewing the existing tier system, and I wonder if that was sort of met by the review that you did of what harvest control rules could, could fall where, if that kind of satisfied the the intent of the review of the existing tier system under the climate resilience work plan.

Yeah, so we don't usually go back to staff, but I think this might help us dig ourselves out of this discussion. So please help us out. Uh, I'll do my best, uh, through the chair, uh, Dr. Mueller. I, I think not necessarily. That was sort of introducing what would require an FMP amendment versus what could require an FMP amendment.

I think all of this was predicated on a variety of stock assessments that weren't fitting into— some of this is predicated on a variety of stock assessments that weren't fitting naturally in our tier system. That was where we originally started coming forward with all of it, and that's even prior to when you then made changes for the tier level for Paul, like potentially for yellowfin sole, leaving us only one tier stock across both areas. So I think it, it does make sense to begin to look at what structural changes might be recommended under our FMP to streamline how we're actually using the tier system, if that helps. But that would be something in addition to possibly selecting ABC harvest control rules for different stocks. So I think you'd be doing it all kind of at the same time if you chose to incorporate an ABC control rule through the, through the tier system as opposed to in an individual stock assessment as a below-max ABC control rule.

And again, that slightly gets at the transparency aspect, and we can try to maybe bring some of this argument back in May, if it helps to try to think through what are the— if we're trying to address transparency and setting an ABC, what are the means by which you would do that? One transparent method is to modify, amend the FMP specifically so that it's clear to everyone that what HCR is being used, as opposed to other ideas we've talked about, which is it could be incorporated into an individual assessment. And you might come back with this when you're talking about sablefish in terms of you, you can incorporate it in an assessment provided it's not going above the current ABC control rule, and you would just be using it as a more explicit means of going below max ABC than just the— or in coupled with your risk table. So I think we've talked around all of those, and if it would be helpful for us to sort of dust off those different, those different mechanisms for how you would go about this. We could certainly try to fold that into a, to the discussion paper that would be coming back to you in May in conjunction with the other requests that you have.

But I think it would be a good time, given what we know now, to go back and start to look at what are the ways in which at the end of this whole route you would be looking at the modifications if you're going to amend the FMP, maybe you want to try to modify the tier system to be more appropriate for how it's being used as opposed to how it was conceived of originally. I don't know if that gets it or digs a hole deeper. I'm not sure.

Yeah, thanks. That certainly helps. And maybe, um, maybe this is a bit premature, and maybe what we can just ask for is you know, some considerations of as these analyses go forward and discussion of by the group, by the teams really, of how these, you know, what the analyses and how the different harvest control rules under considerations as they're being developed for sablefish, for other stocks, how they would or would not fit into the current tier structure to just at least be clear on what can be done now versus through an— what would, you know, I mean, we already got some clarity on, you know, what would require an FMP amendment in the paper that you had brought brought forward. But as we kind of work on these and other rules now more concretely, you know, I think it would be helpful to sort of put that in the context of the current harvest control rules and where they would fit, um, and if they would fit under the Tier 3, basically. Through the chair, if it helps, I think that's kind of along the lines of what we were thinking about going forward with Sablefish and having this next set of stakeholder discussions is trying to sort of pair it with where would you go with this?

How would you move from research into potential use and management? So it might be that we are maybe bringing back some of that argument by virtue of using the sablefish example and carrying it through to what you would be talking about in May for recommending how that would go forward if it was recommended to go forward.

Okay, thanks. I think you did help us get out of that rut. Thank you.

I feel like we've kind of settled on a solution here on that. Are there additional comments?

Okay. I think I'm not going to try and summarize everything. Frans did a really nice organized job of covering all the bullet points. I'll just try to touch on the things that we had substantive discussion on. And of course, the basics.

We're going to thank the presenters and the entire team working on this project. And we really appreciated seeing the updated and condensed list of past SSC requests. That was really helpful, as well as the update on the timeline for future work. Noting that this— we appreciate seeing the substantial progress in reviewing the existing system and the new development. Developments and really turning this from a bunch of ideas over a number of years into this quantitative evaluation.

Like I said, I'm not going to try and get into most of the specific summary points. Franz did a great job on that already. I think the big one was this question of— we have had a pretty solid discussion now of how the harvest control rules and the tier tier system can and can potentially interact and could potentially be revised together. I think we will flag that the tier system and the HCRs are part of the same system, and we have several examples of where there's an interaction between the harvest control rule that's being applied and the tier that the stock is in. We will look forward to more discussion and/or summary of how we're currently using the tier system and the mechanisms for how potential new HCRs could be integrated in or along with the tier system.

And we just heard that this is perhaps part of the plan for a description of how we could move from research to management for the sablefish example. So I think we will stop short of recommending any specific action on this, but look to the next round of updates. And perhaps pick up there where we left off.

We— there was a recommendation to consider perhaps two categories: stocks that don't have links to specific climate variables and those that do. And I think a strong recommendation that there's some value in developing general hard harvest control rules that will be applicable for most stocks most of the time, as opposed to things that are very specific, noting that there are options like HCR 7 where if there is more specific information about specific linkages, there could be a potential for a more detailed approach in those cases.

We will note that there is— could be some value in producing a prioritized list of objectives for comparison across harvest control rules. This is— we've noted that there's a lot of different objectives and a lot of different performance metrics, and I think a number of our comments are going to focus on prioritizing those and simplifying them, particularly with the goal of more transparency transparency for people trying to digest this information. We noted that various categories of performance metrics—status, yield—and we've suggested adding a category for transparency or simplicity, as this is a specific objective in the process, and to perhaps also add an objective related to the time and effort it would take to apply these approaches. How easy would it be to actually implement them?

There was a suggestion to consider whether it would make sense to select a key— one or more key metrics from each of the different categories or to combine those into some sort of an aggregated metric. I am sadly called back to our discussions of risk tables about whether there should be an average score or not, and I I fear that. However, I think we could at this point leave it up to the analyst to see what does and doesn't work, and maybe actually working with stakeholders to see what's most easily absorbed is better than us trying to guess.

And I guess the last, the last comment was just on realizing that our current— the current status quo approach of using the current HCRs plus a risk table, it's going to be important to understand whether a new HCR will do better than that, but we do recognize that the current system is difficult to simulate. And this, I guess, this is just— we're just flagging this as what is likely to be challenging moving forward in this process.

This. So I think those are the, the main points that we had discussion on.

We'll leave it there. Uh, it is after 3 o'clock, so we do still have our sablefish discussion to go, but let's take a 15-minute break, come back fresh, and hopefully move right through that.

Okay, if everybody could come on back to the table, about ready to get started.

Okay, we'll come back into session. We have finished everything but our discussion on sablefish from the Joint Groundfish Plan team, so we're almost there.

We will, so we will start out with our assigned leads. And first up is Martin. Go ahead, Martin.

Yeah, that's right. I promised brevity, but I'm going to— well, see how it goes.

Let's see. So the draft sablefish assessment was really comprehensive, and the SSC appreciated the efforts by the assessment authors. That went into developing alternative modeling approaches, including all the sensitivity runs. Um, there were 9 oral public comments and 15 written comments on the, on the agenda item, though some of them span different topics. Generally, the comments noted the large decline in ABC from when moving from Model 26.1 to 26.2 -FAA and highlighted that the change is due to a switch to a new modeling approach and not a change in the status of the stock.

There was a preference for a phased transition to the new modeling approach and support for the use of Model 26— or support for the use of Model 26.1 for harvest specifications.

So, um, backing up a bit, um, in November of— or October of 2025, the SSC endorsed Model 25.12 for management that was implemented in Spork, the new software package that could emulate the previous sablefish assessment and had the capacity to fit spatial models. This model was not used for harvest specification in 2025 due to the government shutdown, and since then a number of refinements have made— have been made to Model 25.12 to transition to a new reference model, these, uh, which is, uh, 26.1 and some references in the document indicate that it's a one-region model.

For the model— for model parameterization, there was an improved selectivity setup in Allure's survey CV, scaled— rescaled input sample sizes, and a change the logistics selectivity for the trawl fishery. For sablefish biology, there, there were updated length and weight parameters, an estimation of sex-specific natural mortality in the model, which I think was an SSC recommendation. Concerning data usage, the fishery CPUE index was dropped. Fixed gear and longline survey lengths were dropped when ages were available. And an early time fishery block was removed.

A carryover approach was implemented to construct a complete longline survey index. This was done by using the most recent survey for an area when it was not surveyed. Finally, extrapolated Western Aleutian stations were dropped. Overall, these changes are a fairly significant set of modifications to the model, that simplify, streamline, and modernize the assessment modeling approach. Changes had a relatively minor effect on the population trajectory and are well documented in the bridging analysis.

In addition to Model 26.1, two new models were available for consideration: If Fleets Is Area model, Model 26.2, and a true spatial model, Model 26.3. In June 2026, the ACIE review convened— was convened to review these new modeling approaches for sailfish. ACIE reviewers endorsed the FAA Fishing As Areas model as the best model, best scientific information available, and the plan team agreed with this recommendation.

Although this was a request in the terms of reference for CIE reviewers, BSIA has a rather precise definition in the context of fisheries management in the North Pacific. BSIA is considered the outcome of a full review process, including plan team and SSC review, followed by NOAA certification that BSIA was used. Consequently, the SSC approached its review of potential sablefish models as usual, taking the CIE review under consideration, under consideration, but not being bound by it.

The fleets as areas approach is commonly used in West Coast assessments and elsewhere, in in which spatial components of fisheries usually, or surveys, are modeled separately in a model without any actual spatial structure. For sablefish, the separate areas that were modeled were the Gulf of Alaska, the Bering Sea, and the Aleutian Islands. Several model configurations were considered that varied in the number of areas modeled by the fishery and the survey, and whether dome-shaped or asymptotic selectivity was modeled modeling. One advantage of the FAA approach was that it, it can incorporate information from the survey design without needing to interpolate or carry over survey estimates. The SSC has concern that using this modeling approach would reinforce the current survey schedule, but has— which has to be considered suboptimal from an assessment perspective.

Since the model is not spatial, each survey is interpreted by this model as an indicator of overall stock abundance through the lens of selectivity and catchability. This may be a concern for sablefish since trends on the edges of the stock, such as the Aleutian Islands and the Bering Sea, they show different trends than the core of the stock due to density dependence or range range extension.

This issue was not evaluated by the assessment authors, nor did the simulation framework consider this possibility. The second issue for the FAA approach is that selectivity for fisheries modeled this way often show dome-shaped selectivity patterns due to spatial differences in fishing mortality, recruitment, and migration. And there's a citation by Waterhouse et al., who was, I think, maybe a postdoc of Dave Sampson's, that looked at some simulation work on this issue.

A cross-testing simulation and estimation exercise using the 5-region spatial model— spatial model as an operating model evaluated various FAA model configurations and found that this— the model with the smallest median bias included 3 longline fleets, a trawl fishery fleet coupled with fishery and survey selectivity modeled as dome-shaped in the Bering Sea, but logistic elsewhere for both surveys and fisheries. The model recommended by the CEIE reviewers, the plan team, and the assessment author was a simpler model that modeled one longline fleet and one trawl fleet with both fisheries and surveys assumed to be asymptotic. This model also had a small medium bias. However, the SSC notes that FAA models with one longline fleet and one trawl fleet with asymptotic fisheries tended to be consistently biased high, and those with asymptotic surveys were consistently biased low, suggesting that the good performance of the author's recommended model was due, due to canceling biases rather than an appropriate set of assumptions.

The SSC recommended that the proposed models recommends that the proposed models going forward, the Model 25.12 as a continuity model, Model 26.1 and Model 26.2, which is the fleet as areas model. The, the author agrees that with the assessment author that the true spatial model 26.3 is not mature enough to use in management. While the SSC appreciates the assessment author's goal of parsimony, FAA models introduce complexity that needs to be explored. The SSC requests further evaluation of fleets as areas models that include dome-shaped selectivity for the Bering Sea fisheries and surveys noting that the closed-loop simulations with this configuration also showed good model performance, and that dome-shaped selectivity is an expected outcome of a fleet— fleet-as-areas models. Recommendations for this evaluation include dome-shaped selectivity, that dome-shaped selectivity should be evaluated using more flexible curves curves, such as the double normal that are already implemented in Spork.

The approach of using the gamma function is not often used for modeling selectivity because the ascending and descending portions of the curve are linked.

The SSC should request that an additional fleet as areas model be provided, the final assessment that allows for selectivity in for the Bering Sea fisheries and surveys to be dome-shaped. Um, also, the SSC also recommends use of the Five-Region Operating Model to evaluate emergent empirical selectivities for the regional fisheries and surveys by comparing survey and fishery age composition sampled regionally with abundance at the population level.

And then there, I have some further recommendations here.

If the FAA, Fleet as Areas approach remains the preferred approach for sablefish, the SSC recommends that process error and survey catchability be evaluated to account for density dependence and range expansion. Um, and I think we should also recommend consideration of spatial-temporal models for longline surveys that deal appropriately with missing observations. Um, I think we should support the plan team's recommendation for looking at REMO models, and I think we should strongly support continued exploration of the full spatial model is the most preferred approach that's likely to be most successful moving forward. And I don't think we're supposed to talk about allocation now, so I'm— I didn't include anything. So hopefully that's your understanding as well, Ian.

Thanks, Martin. I think these are proposed model runs. Options, and we're not selecting a model and we're not selecting or recommending a model or recommending an apportionment. But if there are alternatives that we would like to see, I think it is within our purview to request a suite of options at this point.

Yeah, I guess I'd, you know, I did I didn't really have a huge issue with the carryover approach as being a reasonable approach, and which is what the author proposed. So I don't know if there needs to be a whole lot of additional discussion on that, but it is a little sure to be confusing to the public and, um, in the council as to why we would recommend a carryover approach be used for allocation but not as a preferred model approach. So if we— the recommendation in the end is to go with the fleets as areas approach, then we would have this sort of disconnect between the model structure and the advice on allocation. Although it, you know, there's just, you know, we're sort of super fortunate with this assessment in that the impact of allocation has been pretty carefully studied, and so we would be reasonable to conclude that it's probably not, you know, a stock concern to use something that's not 100% consistent with the assessment model.

I mean, I don't—. I'm not really— I didn't really sort of come down on whether, you know, the available choice, which of the available choices is preferred and so I don't think we need to have that discussion now. That's for December.

Okay, thanks, Martin. Fabio?

Thanks, Mr. Chairman. First, I concur with Dr. Dorn recommendations, all of them. That. And I want to thank you, the authors and the CA review, and acknowledge the substantial effort that went into this assessment.

And I have just a couple of comments, and I would like to— that SSC should recognize that the choice between the two assessment models models also has socioeconomic implications because of the models produce different ABCs and OFLs, and the revised survey treatment changed regional approach, man. So the SSC recommends that the economic and regional implications be clearly documented while emphasizing, emphasizing that economic consideration should inform decision-making but should not determine the biological model or modify the biological ABC. The SSC also should recognize the value of the Sablefish MSE as a framework for evaluating alternative harvest strategies under spatial varying population dynamics. The Joint Groundfish Plan team considered the MSE a strong example of alternative harvest control rule evaluation while noting remaining uncertainty in the movement components. The SSC should support continued development of the MSE, include evaluation of alternative HCRs and spatial assessment structures, and testing performance under possible spatial and recruitment scenarios.

And the SSC also supports using MSE to evaluate future spatial models before operational adoption, consistent with the CIA recommendation.

Also, the SSC should encourage continued development of socioeconomic components when feasible to support future evaluation of the economic consequences evidence of alternative management strategies. And finally, the SSC considers improved data collection to be the most important avenue for future assessment improvement. So should support maintaining or expanding the Longline Survey spatial coverage and increasing the OLS aging sampling, particularly to improve regional age composition information. And this improvement should provide more independent information for resolving recruitment and selectivity, catchability, and regional availability. I also have a couple of recommendations, but I don't need to go over that.

And finally, here was just a talk that I ask if the stock assessment authors could do a sensitivity fixing M using the model, the 26.1_1reg, using the same M in that model into the 26.2_faa model.

Because the assessment indicates that the downwards calling under FAA was driven primarily by lower M and R naught in addition to other structural changes. So that makes difficult to tell apart from the model comparison alone how much the difference is actually due to the redesign survey versus other assumptions. So My thought is to, for, to be retained the 26.1_1 reg as the coast-wide benchmark fallback model, retain the 26.2_FAA as the preferred current framework, and request this common M sensitivity that can clear the attribution of the ABC difference. That's all I have. Thank you.

Thanks, Fabio.

Other SSC comments?

Maybe while you're thinking, I have two clarifications. The— and I guess it's for Martin and Fabio. So the Groundfish Plan team recommended bringing forward three models in December. And you've both corroborated that we'd like to see those, that's 25.12, 26.1, and 26.2. The Groundfish Plant Team didn't recommend bringing full diagnostics on all three models.

Is that something that the SSC also supports, or would we like to see full diagnostics for all three of those?

I hadn't thought about that until now, Ian, but, um, thanks for bringing that up. Um, the—. I don't think there'd be a need for 25.12 because it's a continuity model, but I think that 26.1 and the 26.2 FAA are, are both— would— are both potentially, uh, would both potentially be adopted in December. And so for sure you'd want to see full diagnostics on both of those, as well as any other. I mean, the model that I'm proposing, you would also want to see the diagnostics.

Okay, thanks, Martin. And I guess that was my— going to be my follow-up was for the model that you're requesting the— or that we're discussing whether we're going to request, um, the polices areas model with dome-shaped selectivity in the Bering Sea fishery and survey, and potentially more flexible curves for the double normal and potentially also process error in survey catchability, would that be a model to be developed for December, or would that be a longer-term, see-it-next-year model? I guess the random error and catchability, I think, is more of a long-term research recommendation, But I mean, I do think it's just for us to understand the implications of a pretty strong assumption of asymptotic selectivity for everything where you really don't think that would be the case. Would be important to have that on the table to look at in December.

Okay, thanks for that clarification. Other SSC comments?

Okay, one more clarification, um, just to be sure that we're all on the same page. If we, we are commenting that we don't see anything unreasonable in the current suite of apportionment options, which I think therefore leaves it open to the author and the plan team to decide which ones to evaluate further in November.

My clarification is, do we want to be more specific and say we'd like to see the same set of options? Discussions in the fall or winnow that down or just be silent on it?

Bronze, could we just, um, note the potential discrepancy between the way the apportionment is done and what's, um, done in the model and kind of leave it to the discretion of the authors to consider that, or we could potentially recommend to, you know, use an approach that's more consistent.

Well, there's of course the REMA option is used a lot for apportionment. And that would have the advantage of sort of taking it out of the context of the model. And as they— as the PLAN team noted, not having age comps is an issue for using their ARIMA model in the assessment, but it's not an issue for apportionment. So maybe that would be a way to get around the awkward fact of the carryover— using the carryover approach. So maybe that's a new idea.

I mean, it would involve fitting RIMA models, but they, they're fast, and there's expertise at Akbay that— and that a lot of expertise. So if Dan or Matt needs some help, they certainly could find it.

So, so I guess, uh, in— that would be just another up, you know, we'd, I guess, say what Brown said is, and, and consider the ones that are on the table, and it's okay to consider the ones that are currently on the table, and then suggest their RIMA approach.

Other thoughts from the SSC on that or other topics?

Okay.

I will summarize where we've ended up then.

So we are going to thank the authors for their very comprehensive analysis and documentation. We're going to note the public testimony that we received, both written and oral. Uh, we will describe the history of where the— where we've been through this process, noting that it was a little atypical in that we saw the initial bridging or continuity analysis in October of 2025, but then it wasn't used last year due to the government shutdown. And so now we're essentially duplicating some of that effort and looking at some new changes on top of a model that was accepted by the SSC last October but never actually finally accepted in December. We will note that the— there are— there were a suite of changes and updates which we will list in the document.

I won't attempt to say them all again here between 25.12, the continuity model, model and 26.1, which was essentially the launching point for the new analyses we saw here.

We will note that there were two new models proposed after that extension to 26.2, including— sorry, 26.1, including 26.2, the fleets as areas models, which was brought in as the author and plan preferred model, as well as a spatial model that was evaluated as part of the CIE and has been, been brought along for use in a variety of other analyses. We will note the— that there was the CIE review in the spring, and that despite some comments in the CIE review about best scientific information available, the SSC approached this review as part of its normal review cycle in the the North Pacific management process.

We will have some summary of the general pros and cons of the fleets as areas models. We appreciated the authors presenting that. We had some additions of our own to that list. We will highlight that the current— that the change to fleets as areas model was necessitated by the change in the survey design, and that we'll note that that survey design seems suboptimal. We flag that differing trends in different areas can only be accounted for in a fleets-as-areas model through differences in selectivity and catchability, and not through actual different trends.

So this is the implicit part of a spatially implicit approach. And this leads us to conclude that there may be value in continuing to explore for the option for dome-shaped selectivity, which has been shown in other simulations to be a not unsurprising product of Flitz's ARIES models.

We will note that the models— the selection of 26.2 was driven by a variety of simulation analyses, which did seem to indicate some degree of canceling biases which isn't ideal and isn't the same thing as being completely unbiased. However, uh, we will recommend bringing forward the 3 models that were identified by the authors and the crab plan team: 25.12, the continuity model; 26.1, which is the updated single region model; and 26.2, the fleets as areas model. We will ask for full diagnostics for 26.1 and 26.2.

26.2. We will also request further exploration, I guess, beginning with 26.2 of a flitsas area's model variation that would allow dome selectivity in the Bering Sea fishery and survey and may include more flexible curves, such as the double normal, to model that selectivity. And we would like to see that model for December, and we would like to see a full suite of diagnostics and information to go with that model.

In the longer term, we have some recommendations to continue to explore process error and survey catchability, which could— and also to continue to use the spatial model for evaluating the interaction between trends in a spatially dynamic population and missing data.

We will support continued development of the MSE and the various spatial and recruitment structures. We will also support development of socioeconomic indicators along with the MSE approach for alternative management strategies, and we will highlight that the socioeconomic implications of different choices of strategies are important, and the socioeconomic implications should be documented. However, we'll be clear that we're not using socioeconomic information as part of our scientific consideration of the models in this— during this process.

For apportionment, we will note that there were a wide range of apportionments presented to us and that the carryover approach seems generally reasonable but does highlight the discrepancy or inconsistency in between using that approach for apportionment but not in stock assessment model. We will suggest that the authors consider using a REMA approach because this could be applied without the issue of age composition, the inability to use it for age composition. It could be applied to the apportionment process rather than just a simple carryover, even though it may not be very different than a simple carryover.

We will have an overarching comment that the highlighting that the changes in the modeling this year were largely necessitated by changes in the data streams, and that in the bigger picture, improved data collection should be the highest priority for this stock.

And we will also ask for a sensitivity— sorry, this is a little bit out of order. We will ask for a sensitivity sensitivity using fixing the value for M from 26.1 and using that in 26.2 in order just to better understand the bridging between those models and the relative role of M versus the other changes in the models.

Is there anything I missed?

Jason.

I don't know that you missed it. I just I was catching up in my thoughts.

It probably goes under just the loss of data concerns, and I, I think we heard a lot of public testimony concerning the loss of the logbook, so I'm wondering if we shouldn't just highlight that and mention that as one of those data concerns.

Yeah, I think we touched on both of those in our discussion, so I would feel comfortable including both logs and surveys. Changes as part of that.

Thanks for catching that.

Okay, Martin. Yeah, I guess I'm not quite— I mean, I don't know what, what's going on really. I don't know why there's no longer any logbook data. And so, and it sounded like nobody else that we talked to knew either. So I'm just a little I'm just not sure what we would— I mean, we— I guess we can note it, but we don't have a solution and we don't know.

There's a lot, you know, it's like, well, I'm just a little struggling a little bit about why. Yeah, I guess it's okay, you know. I mean, I just— it seems like there's a lot of— there may be good reasons why that happened and it, and, or maybe there's bad reasons, but, you know, it's just we don't know what those were.

Go ahead and respond, Jason. Yeah, I don't, I don't know what the solution is either. I guess the intent is just to note that it was brought up and it was important to the testifiers and that it was one of the data streams that was formally in the assessment. And so I think I'm just sort of highlighting that in the minutes.

Katelyn.

Maybe we could also just add that CPUE is something that often would be a great fisheries indicator to include in an ESP. So even if it's not in the assessment for a particular reason, it is— it could still potentially be an indicator. So I would just differentiate between, um, something that has to necessarily be included in the assessment and something that— and maybe this was back in your other summary, Ian, about the social and economic kind of indicators, but it can be considered in that category, I think, as well.

Okay. Hopefully we can find some wording that keeps the concept but isn't too prescriptive because we didn't really dive in to a lot of detail on our discussion. Sherry?

I recognize I missed a good part of the discussion. But I was just thinking, I mean, it is possible that we could note that it came up in public testimony, say we're concerned if it rises to that, but also acknowledge that it, it wasn't presented to us why, why this index necessarily went away. Um, and that I don't know if you want to say concerned, but I think that we can acknowledge that we don't know what's behind it. But if there's concern, we note that, or we can also just summarize it in public testimony and acknowledge, you know, that— actually, we don't summarize public testimony because this isn't influential scientific information. So no, there are times that we've mentioned things that have come up.

I'll leave it— it was trying to help in phrasing for folks. You guys have to sort of figure out where you want to land on it though.

Does anybody have any concerns with it, with noting that we highlighted this loss of data? It— the topic includes both the survey and the logbooks, and we don't have a solution for either of them currently. We're just highlighting that that's what's happening.

Okay, I think we'll stick with that then.

All right, we are now going to transition, and that we've had a small change in schedule due to availability of staff and others. We are going to go over the GOA and straight to Bering Sea and Aleutian Islands Groundfish Plan Team report. Report. I realize it's already 4 o'clock, and I hate to say it, but we are going to go a little long today. We will absolutely stop by 5:45, but we will probably see how far we can get until then.

So bear, bear with us. We're going to have some catching up to do tomorrow afternoon as well, I'm sure. So over to you, Diana. Thank you. Through the chairs, I have Dr. Shotwell on the line, and she's sharing the slides from her computer.

So she is going to walk through the report. We do have two author presentations that are included in the Bering Sea Aleutian Islands Groundfish Planting Report. We have a separate author presentation from Dr. Ian Elliott on Bering Sea pollock, and then we have a presentation by Dr. Megan Bryan on Greenland turbot. So those— we'll switch to their slides as we get to that part of the report, and Dr. Schott will provide the rest of the report. And I believe that Kalei is back on and promoted again.

Great. And I'd just like to clarify for anybody waiting for public testimony, our plan here is to take the report, to take the initial parts of the plan team report. We're going to do the Bering Sea and Aleutian Islands harvest projections, and there will be a sign-up for testimony and we'll have our discussion and finish that item off. We'll do the same for Boguslav Polak, but then for the proposed models, we will go through and have all the presentation, get all the way through that, and then go and have our discussion afterwards. So there will be just one sign-up for all of the other proposed models.

Only the specifications models, specifically the harvest projections and Boguslav Polak, will have a separate public testimony.

You good to go, Kalei? Yes. Through the chair, can you hear me okay? We can. Thank you and welcome back.

I'm very excited to be back.

All right, members of the SSC, Kalei Shotwell, again, co-chair of the Bering Sea and Aleutian Islands Plan Team, presenting our planning report. I want to recognize and thank my co-chairs for— and council staff for their help doing— putting this together.

Okay, this is a summary table for the topics that we plan to present today. The team met in person on September 10th and 11th. Again, as before, you saw in the joint team, the link links to the individual topic presentations are in the headings in the minutes and in the presentation for your reference. The organization we'll have for today is I'll do a brief summary on the surveys, um, and then that's going to be followed by the harvest projections. I guess I'll pause after that, um, and see if there's any, um, comments, and then I'll move to Boguslav Polak, pause again, and then we can go into the 4 stocks with their proposed models.

And then at the end we just have a final— just one final slide going through some team discussions that we've had.

And then— oh, yes. And then we added a final summary slide for the models that you'll expect to see in November.

So we, we received a presentation from Duane Stevenson on the results of the Eastern Bering Sea Bottom Trawl Survey.

They completed all 350 standard stations. The Northern Bering Sea Survey was not conducted this year. That allows for field work associated with the bottom trawl survey modernization. Both the Eastern Bering Sea and Northern Bering Sea are scheduled for 2027. The surface and bottom temps were coldest, the coldest they were since 2013.

The cold pool was also the largest since 2012. Biomass for most of the species were stable or increasing. Pollock, pollock continues to be volatile and increased this year by 50%. Alaska skates continued their multi-year declined. There's also been some updated effort calculations associated with net, with net mensuration sensors, and the recalculation changed the 2024 and '25 biomass indices by about 3 to 4%.

So the survey lead said that it was important that the authors, when they update their time series, they update the whole time series and not just the most recent year. The teams discussed progress and timeline for modernization and then noted that initial fieldwork is started in 2026. The survey leads said they have an optimistic target of 2031, but their schedule is very uncertain. We also discussed incorporating more deepwater toes into the redesign, and the current intent is to go to about 400 meters, but they will be also looking at some 600-meter tows as they go.

And then Suzanne McDermott presented results of the Aleutian Islands Bottom Trawl Survey. They completed 373 stations of the 447 attempted tows. They collected over 89,000 lynx and 6,000 otoliths. They've also had new hardware software that was used for effort data collection. Their surface temperatures were cooler than average, but bottom temperatures increased to near average levels.

With— they also noted a distinct warming and localized hypoxia that was observed in the western Aleutians. I'll note that the O2 sensor was also in the Bering Sea survey as well, which is new information, and we discussed that as well. The general biomass trends with flatfish were below the mean, and rockfish were above the mean. The gadids were mixed, sablefish was above, and the skates were at low biomass levels. And the team discussed the new way of presenting presence-absence trends and their potential to examine future patterns in the future.

We definitely like the new information they provided us. We discussed the dissolved oxygen information and whether hypoxic conditions might impact species distributions. And then we also suggested a comparison with other dissolved oxygen data that are available from different agencies. I think we might have mentioned the IPHC. And then there was noticed some species ID issues with skates that may affect that comparison with the long-term mean, that they were saying the skate was very low.

I guess Alaska skate and leopard skate were not distinct species prior to 2010.

We also discussed the potential to provide current biomass comparisons with the last survey and also the long-term mean as they showed in the future, since both of these metrics have utility. There were some size shifts in the ruffye blackspotted rockfish complex beginning after 2012, so we discussed that, but the change didn't really they seem to coincide with any changes in the survey methods. And we noted a very strong appreciation for the rapid processing of this data and the improved presentation materials.

Finally, we got the presentation from Abigail McCarthy and Nathan Laufenberger on the acoustic trawl survey, bottom trawl survey, acoustic trawl survey, and the acoustic vessels of opportunity. The acoustic trawl survey, ATS, is conducted biennially. The AVO is derived from the acoustic data collected during the Eastern Bering Sea bottom trawl survey, and that's collected annually.

For 2026, the ATS preliminary estimate was about 3.6 million tons and 13 1 billion pollock. They noted that that may change when bottom trawl information is included from the near sea floor— near the sea floor. They did observe more juvenile pollock than recent survey years with an age 2 component that was the largest since 2012— since the 2012 year class. ABO and ATS spatial patterns were similar in '26, and the team expressed appreciation for the substantial effort by the Mason GAP programs to conduct the survey and produce these rapid survey products for this meeting and for the stock assessments.

The team asked some questions regarding the temporal and spatial overlap, the seasonal, and they, they said that they had similar seasonal period, but they weren't sampled exactly at the same locations. The ATS midwater trawl methodology was also something we discussed. But it's not used to calculate CPE.

And then we suggested comparing ABO and ATS with numbers instead of biomass.

And that's it for the surveys. I guess I'll keep going into the harvest projections. Dr. Melissa Heltuk and several lead authors that were in the meeting presented the harvest projections projections for these 7 stocks that are listed on the slide. The table in the slide shows the summary of those harvest projections, including the stock, the lead author, the ABCOFL, and the assessment schedule.

The team supported the authors' ABCOFL and the biologically informed recommended distributions, or the BRDs, for all 7 stocks. The team also discussed ways to improve readability for future reports, including specifying no reduction from max permissible ABC instead of what we had said, no reduction from ABC, that, and then distinguishing recent spawning biomass trends from projected biomass and including a brief explanation of the harvest projection process. We also discussed the shift in Kamchatka flounder from even to odd year biennial assessment schedule and that they were, they planned to keep the odd year assessment schedule moving forward.

That's all I have for harvest projections, so I guess I'll stop if you'd like me to. Yeah, this is good. It's going to be a little awkward. We may not— it may not take us very long to move through some of these, but we will do a full finish on harvest projections before we continue on. So are there any additional questions on the harvest.

Harvest projections.

Okay, we will— I'll check for public comment on the harvest projection specifically.

Seeing none, we can go right to our leads for harvest projections. So we'll start with Andrew.

All right. Well, thank you. Thank you for the overview and summary of the the plan team recommendations for the harvest projections and for the authors for producing them. I'll just note that there was no public testimony for these specific stocks.

The stocks, Bering Sea stocks that we received harvest projections for are Northern Rock Sole, Aleutian Island Pollack, Kamchatka flounder, northern rockfish, Alaska plaice, hourtooth flounder, and flathead sole. Just note that full assessment for some of these stocks are scheduled for 2027, except for Alaska plaice and flathead sole, which are scheduled for 2028.

And I just note that specifications changed slightly for all stocks compared to 2026. 6, Slightly higher for Aleutian Island Pollock, Kamchatka Flounder, Alaska Place Flathead Sole, and they're slightly lower for Northern Rockfish, Hairtooth Flounder, and Northern Rock Sole.

So I think the SSC should concur with the plan team and author recommended AFC and OFLs for these 7 stocks for 2027-2028 with no reduction from maximum ABC. I'll note there isn't any area apportionment for any of these stocks. And finally, just note that none of the authors or plan team highlighted any major concerns related to these stocks. And finally, that none of these stocks are subjected to overfishing and, and are not overfished or approaching an overfished condition.

And that's all I got. Thanks.

Thank you. We'll go next to Sarah.

Um, yeah, thanks. Uh, I think, I mean, I agree with, um, with everything so far. I would just add that, um, there's The continued multi-year declining trend, um, and that's flagged as a potential concern. So additional tracking of that, um, may be considered. And, um, there are continued concerns around capacity constraints and data gaps.

That seems to be a common theme throughout. Today, but worth mentioning that there are data limitations that are affecting some of the assessment processes, and some of them are pretty clearly explained and documented, and some— there remain some questions around that, and we may We may want to clarify, clarify that for transparency's sake in, as we go through this, the discussion and the stock assessment process.

Um, let's see.

The ecosystem status report. I guess that's a bit later on, so I'll come back to that. Yeah, we can hold comments on the remainder of the report till we get to the end. We'll have a catch-all in case we need to loop back to anything. We really just need to cover the harvest projections for now.

Yep. That's it. Hi, thanks. It just— it was brought to my attention that in the— on the slide there for Flathead Soul, it looks like ABC and OFL have been reversed.

That's—. I think it's just a typo, but the numbers are correct in the— in the Harwich projection report. Thanks.

Thanks, and we'll just clarify the table that we have is correct. It's just a slide issue. That's correct. And I, through the chair, I should have noted in the beginning also, you do have a table that has only the specs for which you are taking up separately at this meeting, as well as your proposed specifications posted to your agenda. Thanks.

This is somewhat new for us to be getting through more than just a little bit of specs in October, so. We're doing our best to make this process smooth. Did you have another comment, Andrew? I guess just, um, just the other, other thing I wanted to mention, I forgot to mention, is just, you know, appreciating the discussion that the plan team had on ways to make future harvest projection presentations clear to readers, the reports, and just, uh, just want to ask, I think the minutes should note The SSC appreciates the efforts that have been made by authors to standardize these reports and address some of the recommendations that we've— that the SSC has given in the past, such as standardizing methods used to estimate remaining harvest for the current year and some other things which I can highlight in specific examples of what we've brought forward in the past that they've been making efforts to, to standardize. Thanks.

Great, thanks. And Sarah, sorry, I didn't mean to cut you off. Did you have anything else to add to the harvest projection specifically? No, that was it. I caught myself going too far.

Thanks. No worries, thank you. Okay, I, I will just very briefly summarize that we are going to support— we're going to agree with the plan team on all of the OFLs and ABCs these accepting this issue in the slide here with no reduction from Max ABC. We'll note that overfishing did not occur. None of these stocks are overfished or appear to be approaching an overfished condition.

We will note that some of these specs, the specs change slightly from those in 2026. We'll thank the plan team for their overview and summary. And our note that we appreciated the discussion on ways to make these reports easier for people to understand. We support the recommendations, those recommendations, and we'll reiterate a few of our previous suggestions. We'll also note that we will be seeing full assessments for many of these species coming in 2027 and a few in 2028.

Sheri. Sure. I'm not fast enough to do this. I don't know that we asked for public testimony on harvest projections. I'm guessing we did.

Yay. Thank you. Sorry, we did, and we received none. So we will summarize them in depth.

Okay, that finishes us off on harvest projections. So please continue.

Okay, sorry about that typo on the ABC for lettuce. That's what I get for doing these things late. Okay, moving on to Boguslav Pollock. Dr. Jimmy Anelli provided the review of the operational update assessment for Boguslav Pollock. This is a Tier 5 assessment with no new survey data.

Oh, I'll slow down. Looks like there's a delay. David Bryan provided an informal overview of ongoing research efforts, including some explorations of size-based spawning timing and survey timing. And the team discussed potential for changes in the sex ratio on the spawning grounds due to that size-based spawning timing and how that may impact the sex ratio observed by the survey. The survey group also noted that there's been a request out to fund a 2028 survey for Boguslaw.

And the team agreed with the author's proposed Tier 5 model and harvest recommendation.

And that is all we have for Bogusław Pollak.

Thank you. That was efficient. It was. So we will check for public comment on Bogusław Pollak.

Seeing none, we can go to our discussion and back to Andrew.

All right, well, just thank the plan team and the authors for their work on this update assessment. Boguslav Pollock is assessed every 2 years, and the last full assessment was in 2024, and we got an operational update assessment presented this year. There's no public testimony. As we heard, this stock is assessed under Tier 5. and it's based on random effects autoregressive model of acoustic survey trends.

The acoustic trawl survey is conducted every 4 years, the last survey taking place in 2024, so there's no new data for this.

The area has been closed to directed pollock fishing since 1992, and bycatch is primarily in the non-pelagic trawl, harpeth flounder. Fishery and has been low in recent years. Uh, I think that our minutes will reflect that the SSC agree with the plan team and authors' recommendations for the 2027 and 2028 OFL and maximum permissible ABC for this stock. I'll just note that there are no elevated risks for this stock in the risk table, and so no buffer was recommended to reduce the ABC from maximum, maximum ABC.

So for the stock, because it's Tier 5, an overfished status determination cannot be made, but this stock's not subject to overfishing. I'll just note that the stock is caught at levels well below the FOFL of 0.313.

And yeah, and just appreciate the summary on the update on the exploration of size-based spawn timing and survey timing, and look forward to updates on that as that moves forward.

Thanks. Okay, next is Sarah. I'm sorry, next is Sherri. And I've got no additional comments. Thanks.

Thank you. Comments from other SSC members?

I had one, and just to note that this is the first update or full that we're adopting in this cycle. And we were expecting, I think, 5-year projections on all of the assessments this year, but I'd note it's not in this one, and that's probably just because it's an update. So we may want to ask for it to be added next time around. It appears to— those 5-year projections appear to be in all the full assessments this year. So we may just want to add that to our list for this one for next time.

Other comments from the SSC?

Dana, um, I don't think we do projections for Tier 4 and 5 assessments, right?

That could be the case. Yeah, I mean, it's just survey biomass estimates, so it'd be less informative. Okay. Yeah, good, good point. We'll scratch that one then.

Okay, I hardly need to summarize. We're going to support the plan team recommendations for '27 and '28 OFL and ABC with no buffer from maximum permissible. We have no overfished status determination and this stock is not subject to overfishing.

We'll summarize a little bit of the background on the stock. Biennial assessment, uh, we received no public testimony and we had no new survey data. So this is about as straightforward as it gets.

Back to you for the next item, and I guess this next piece will include all of the rest of the plan team report on all of the proposed models. And then we will take public testimony on everything at the end and go through all of our discussions on each piece. Uh, would start this section by asking that where we are going to hear directly from authors. Because of how far behind we are, and to no fault of your own, we are really going to hold you to a 10-minute presentation. So please do your best to give us your best shot in 10 minutes.

Thank you.

Thank you. We'd be moving to Jimmy and Ellie's presentation now if Jim's ready to share his screen.

Sorry, I'm there, I'm there.

Can you hear me?

Can you hear me? I'm sorry, we can't hear you, Jim. That was a yes? Yes. Okay.

And you're seeing the screen? Great. Yeah. This, like all the other ones that you've heard today, or some of them, is a fill in the gaps for the furlough, the government shutdown, what happened. And just wanted to acknowledge the authors and remind you of last year's activities.

So all of this you saw last September, so nothing super new there. We provided a discussion paper, and this is all in available in the link under 2025 discussion, the discussion paper that went to the CIE in 2025. And then also after the CIE, the analysis that I presented to you, which were pretty extensive. You can see in the table of contents there and go through it. And for future sablefish, you should consider my three double 3-parameter double logistic as a promotional pick there.

So here's my little DAG, if you will, of the model numbering and where we are, where we were. And so last year we had our normal 2024 model that was Model 23, that we had a RTMB transition, and it developed this Model 25, which began being a phantom, but it was an exact match from the ADMB. And what you'll see in a minute is where we've gone since then. There's a couple of important data product issues that are happening that we started last year. One is the movement from traditional microscopic ages to— with this Fourier transform near-infrared approach that is— was used last year for bottom trawl survey, and now for the fishery data from 2025, it'll be included along with 2026 bottom trawl survey ages.

Another big change is this estimator that we've been using for a couple of years expanding age samples, and you saw a glimpse of that last year, and there's details in the report, but I won't go into that for brevity here. But those are just kind of the before-modeling type approaches. But what I really want to make sure you understand or complete— this is about a model check and about a pushed to move to the next generation of Life, at least for next year. And we— here I just highlighted the old model, how it matched with the 2023, and we did checks. But we're also taking all of the same data, obviously making the bridge, but also featuring all of the things that what the new platform can do.

And so that was for the 2025 model that never got used, but could have, but we developed all of that and it's there since then. As you've seen in other assessments, the development of RSeattle, soon to be rebranded apparently, has been done and This is what I recommended to the Commission— sorry, the Commission— the planning team. And these are comparing kind of the bridge model from AAD Model Builder, changing a few things that it needed to make the 23.2 model. And that was what I was going forward with. And that's just the check.

So sorry. Going backwards a little bit again, sorry, that, that's not backwards.

This was just the model changes to make the RSeattle changes that Grant Adams had done. This other one was what I did last year. And so the new model will develop from this translation check, and that'll be called model 26. 0 To start out, and that's, you can see that the changes between the translation check and the development one is insignificant. And just to convince you that some of the selectivities look alike, this is just one figure on the fishery changes between 23.2 and 26 side by side.

And along with this figure, at the plan team and in the document you can see online, I looked at the OSAs, the self-tests simulating data. I looked at the removing one at a time indices, also looking at retrospectives and projections, all working fine. And even added the feature of developing DSIM, and hopefully we'll be able to have a stepwise transition from an ESP appendix in future so that the DSIM aspect can be applied. We also did MCMCs that performed quite well and also had some new kind of data leverage figures for seeing what parts of the model— what parts of the data most are most influential in the model. So those are all in the main document, but I did just remind you of this is the just through 2024 data, the fits and kind of a new way of showing what the correlation is like between the observed and predicted, quite good fits around the board.

And then doing the same demonstration of whole index fits, which ones are most sensitive. For example, if you drop off the bottom trawl survey, it really can change. Some of the estimates in the near term. And so just again reflecting the importance of the data and keeping that consistency, we did find in presentation and through the CIE that I had done 20 appeals on my retrospectives, and the moans bro for lots of appeals looked really good, but the CIE said you should do shorter ones, and so they look less good, and these are reproduced here. And so we're working on some of the reasons of that, and I did present some aspects that seem to be related to the data, which it usually is, on why the retrospectives are bad.

And so—. Sorry—. And so, yeah, that's kind of the bottom line of kind of analytical things that need to happen, but I just wanted to conclude that this, you know, RSeattle, what I'm calling Model 26, you know, aligns well with the 80 Model Builder bridge, has all the features that we need, and I think is a good way forward. And I highly encourage you to look at stuff in the HTML because I think there's some really cool stuff in there. And that's all I have.

Did I make it 10 minutes?

I think you might have beaten 10 minutes. Well done.

We have a few questions for you. I see Chris Sidden first. Yep, Ian, in the plan team document summary, the report, so just so you're clear, so the groundfish plan team report, it says that you're going to be bringing, they're recommending bringing 4th model 23.0. So is that the plan? Is that the original non-improved model that you're bringing along with 26.0?

I just want to be clear. Thanks. Yeah, that's the 80 Model Builder one. Yep. Okay, I just want to make sure.

So you're bringing 23.0, the 80 MB one, and then 26.0, the RSeattle. Okay, yes, that's the current plan. I'll be able to compile that 10 years from now too, I guarantee you. Thank you.

Next question from Martin.

Um, so I'm wondering about 26.0 and how it's set up. It's basically, it's using the same sort of for selectivity, and you showed really close, really, really close patterns between— yeah, so I'm thinking that that model uses the same sort of framework that you've used historically for modeling selectivity, which is kind of a sort of a, um, sort of a penalized sort of smoothing approach. And is that correct, Jim? It's just— it's— he's just copied over the, the, the work that you had done previously on the smoothing selectivity, um, uh, yes. Okay.

Pretty closely in the feature I haven't really highlighted is that those can now be treated as random effects, and so my arbitrary penalization in the past can be actually estimated in the future, and I'm sure future people will be gung-ho at doing that instead of double normal, in my opinion. Sorry. No, I mean, that was going to be The one— oh, sorry.

I guess the other question is, where does 25.0 fit in the grand scheme of things now? Is that just depreciated, or is that still sort of on the table as a potentially useful model? And my understanding is that that's a conversion straight into RTMV rather than through our Seattle, which is, of course, programmed in our TMB. So, that's slightly different structure. Sorry, our Seattle is still TMB.

And so, we kind of made a decision. I mean, I didn't really care which way to go. They both gave pretty much the same answer. The reason for going to RSeattle, and you'll hear more tomorrow in Cole's presentation, is that, for example, the one-step-ahead residuals are better under TMB models, and that's what RSeattle is. I mean, RTMB can do them as well, but I'm pretty sure this is still true.

It might that RTMB— well, I shouldn't overspeak, but there are still advantages for pure TMB models, and that's what rSeattle is. Even though it's got the R in front of it, it's still written in TMB. And yeah, it's a bit faster for MSE-type applications compared to the RTMB. So that The model 25, still there on my computer, will be handed down, but is pretty much a one-off mirror of the other model with not all the features, but enough of them.

Okay, thanks. So then 25.0 is, is not sort of being carried forward into the future. Yeah, okay. I don't know if there's a way to make that clear because it had me confused a bit because there's a lot of arrows here. And yeah, so one thing that might help is that if you go look at the 2025 previous work, that's where the bridging really happened.

Okay, I didn't repeat the same bridging exercise this year.

Okay, don't see any more questions here in the room. Thanks, Jim. You betcha.

Okay, I think this is me, and we— I can go over the planning team's discussion and recommendations for Francine Pollock unless there's further questions for Jim.

No, go ahead, we're done with questions. Okay. The team did discuss the strong retrospective pattern that was similar in all the models Jim showed us and its linkage to the absence of the 2020 bottom trawl survey. But as Jim mentioned, he considered this a data-related issue rather than anything due to model robustness concerns. We also discussed the treatment of the age-1 selectivity in the bottom trawl survey, as it has substantial flexibility with the lightly penalized selectivity parameter.

And we suggested the authors could evaluate alternative treatments for that H1 selectivity. They could make it time invariant or, or change the penalty. The team also discussed whether changes in the ecosystem structure could influence the retrospective pattern and the observed volatility of the bottom trawl survey biomass survey index.

We also discussed OSA residuals, that they were showing some patterns indicating some potential data weighting and fit issues. But the team also discussed the transition from ADMB to RSeattle, as Jim just went through, and considered the bridge very thorough with similar population trajectories, and noted the many advantages of continuing development in RSeattle, which included the integrated DSM.

Option. So for recommendations, the team agreed with the authors' proposed models, bringing forward Model 23.0, which is the base, and Model 26.0, which is the RSEattle model for November, as we just discussed. We also suggested evaluating alternative treatments of H1, bottom trawl surface selectivity, performing some self-tests on some simulated datasets, looking into the bottom trawl survey index volatility, and considering any ecosystem-related causes of that retrospective pattern, and doing that for both Model 23 and 26.

We also talked a little bit about clearly documenting the age composition weighting and prioritizing some further evaluation there. And then supported continued development of the Eastern Bering Sea Pollock ESP as resources allowed that. And that's all we had for the plan team review.

Great. Looks like we don't have any questions. Please continue.

Okay, next I'll go through the yellowfin sole proposed models. Dr. Ingrid Spies presented her proposed assessment models in June this year. The assessment went through a CIE review, which focused on the transition from ADMB to SS3. They also reviewed the appropriate assessment tier, going from Tier 1 to Tier 3, treatment of growth and survey data, and the environmental covariates that were used in survey catchability. The CIE supported the transition to SS3 and using Tier 3, and the authors proposed moving forward with the Tier 3 SS3 model for November.

And then the authors proposed 3 total models to be reviewed: the updated SS3 base model, another model that builds on the base and uses the sample ISS approach to determine input sample sizes for the survey age compositions, And then a final model that builds off that ISS model and uses the Eastern Bering Sea survey bottom temperature rather than the survey start date as the environmental covariate for survey catchability.

Next slide.

The team discussed the potential to evaluate the cold pool as a covariate and to continue consideration of using environmental covariates in the survey. And catchability components for future assessments. We also discussed the empirical weight at age and the increase in weight at older ages in recent years that, that might be due to low sample sizes at the oldest ages. Don't have as many sample sizes up there. The team suggested future exploration of growth modeled as a function of temperature.

And then the team agreed with the authors' proposed Tier 3 models the 3 I mentioned before for November. And that's all we had for Yellowfin Sole.

We have a question from Franz. Uh, yeah, thanks, Kaleigh. So the, the stock is still in Tier 1 at this point, and, uh, is, you know, is, uh, the— are the authors bringing forward both the Tier 1 and Tier 3 calculations? Because here the focus is pretty much entirely on Tier 3 in the assessment itself and in the GPT report.

We did not request the Tier 1 model, but we have Megan Bryan online. I'm wondering if she has any thoughts on that. I don't think we requested the Tier 1 model.

Hi, yep, this is Megan Bryan. Correct, I don't think we discussed that, and we were not planning to bring Tier 1 calculations forward in November because there was agreement for Tier 3.

Is the SSC asking us to bring forward Tier 1?

Yeah, thanks. I mean, typically we I mean, we can take that up in discussion, but maybe Wisconsin staff here, we can get clarity on that. I mean, we have— we're officially in Tier 1, so our status quo approach would be a Tier 1 version of the model.

Sherry? So I had thought the same. What I— When I went back to the plan team review in October of 2025, um, they presented a comparison. So when we did this for Pollock, it came forward in— we reviewed it in October and we made the final decision in December. When I go back to our October minutes and the process that happened at the plan team, we actually said the SSC supports the authors and Groundfish Plan Team's recommendation to move forward with these models which have changed and move the stock Tier 1 to Tier 3.

So I guess my question is, and this is sort of our process, can we make that decision? It looks like we made the decision in October, but we didn't have a December meeting, and, and so I'm wondering, it's the same question sort of that France has, do we, do we need to bring this forward in final specs, or do we, can we go back to our recommendation in October that we accept it and move on with Tier 3.

I see that Megan's online. If you want to chime in with anything— I'm in. —The assessment perspective, go ahead.

Okay, I'm sorry. Um, yes, chiming in from the assessment perspective. Um, if we're moving to use SS3 like we have shown this year, and I think that was the agreement last year, our ability to do the Tier 1 calculations, we don't have the same ability to do the Tier 1 calculations using SS3 like we did with the bespoke model. I think we'd have to estimate the Ricker curve outside of the assessment model to do that. So our ability to get the tier Tier 1 calculations from the tier— from the SS3 model is not as straightforward as with the bespoke AD&B model.

And that's something we brought up last year. Excuse me.

Okay, thanks for that clarification. I guess this is sort of uncharted territory because we've never reviewed the same assessment 2 Octobers in a row without something in between.

Through the chair, that was kind of the nature of the plan team's discussion of it as well. We really assumed that because of how different last year was, but that, that had come up, um, already for your October and our September meeting, um, that was kind of where we all were landing on it too. We didn't even discuss Tier 1. We just made the assumption that it was going forward under Tier 3 as it would have done most likely last year. So, um, from the plan team's perspective, we didn't have that discussion.

So I guess for right now, are there any clarifying questions that will help us? And we can have more discussion on the topic we need to when we get there.

Yeah, I guess the only question I have is, yeah, I mean, because our It's literally— we've got about 7 words there, and it's the order of those 7 words that are the issues, because we supported bringing forward 2 different models for comparison, and we said move the stock from Tier 1 to Tier 3. So I recognize, um, yeah, I recognize what that is. Do you know procedurally, do we need to— um, we've gotten to the point now where if we've got bridging models, we've accepted bridging to a new platform in October. That can happen on its own. We've—.

Models have to be in December. We've moved apportionments, have to be in December. Is there any information on moving tiers and whether that can be done in October or December?

Through the chair, I don't think we've ever faced that situation before, and we wouldn't have last year had we not had we had the assessments coming forward, although I don't believe that you did ask for the Tier 1 comparison last year before we knew that the shutdown would affect that.

That's my recollection as well, that we were just moving forward with Tier 3. Okay, I think we can continue.

Okay, and now I think we'll switch over to Meagan Bryan for Greenland Turbot.

Kolei, I think you have to stop sharing so I can share.

Okay, good afternoon everyone. Can you see my screen? Because I'm not seeing it on my end. Yes, we can. You can— oh, thank you.

Okay, so I will try to get this through this very quickly. Today I'll provide a brief summary of the preliminary assessment for Greenland turbot that I presented to the groundfish plant— or the BSAI Great Groundfish Plant Team in September.

Hmm, okay, why isn't this moving forward? Okay. Um, and so in September, presented several Tier 3 models to the plan team, um, and a lot of the discussion centered on data conflicts identified through their prior exploratory modeling or through the Tier 3 assessment model diagnostics that you'll see in a bit. Um, really, there's considerable tension between the survey length composition data and the survey indices, and in general, Francis reweighting leads to strong downweighting of the length composition data, especially survey length compositions, to better fit the survey indices. We also noticed this year, we have noticed in the past and this year, that the deepwater surveys, there's some conflict between the deepwater surveys and the shelf survey.

If you were to remove the time blocks that are on selectivity for the shelf the Shelf Bottom Trawl Survey. Basically, the other, the deepwater surveys drive an overestimation of that, the early observations from the Shelf Bottom Trawl Survey. Also, we noticed that an improved fit to the longline survey, which at this current moment is our only survey, that samples the adult population, that can result in poorer fits to the other surveys.

And along these lines, in 2025, we demonstrated that the current model that's used for management that starts in 1945, the— these early estimates that you— of recruitment that you can see in the top left panel is mainly driven by the catch observations, which is shown— which are shown on the bottom left.

But— and this corresponds to a time period when we don't have a lot of composition data, but we also are noticing that different data sources are in conflict about the initial recruitment estimate and population size. So, you know, there's considerable uncertainty about whether the stock is incredibly productive and small or large and less productive. And so, given the data conflicts and these uncertainties, um, in general, we think that at some point we'll need to discuss how do we approach this assessment moving forward, what data are needed to alleviate some of these conflicts, do we continue with a Tier 3 model, as an update assessment moving forward until we have answers to how to improve the data that is available for this assessment, or do we use a different approach to manage the stock? So now I'm going to summarize the models that were presented in September. So I'm going to go really quickly through these models.

The models listed here don't represent all the models that were presented in September, but they're the ones that I'm going to highlight when I go over the results. All of the models build iteratively, so 16.4 is the last accepted model. 25.0 Is the same model, but is using the recommended RPNs for— from the AFSC longline survey, using the recommended interpolation method. And then the remaining models all built iteratively from there. After 25.0, the input sample size for the length data were updated to be the number of hauls for the fishery fleets and the out— and using the output from the AFSC ISSR package for the bottom trawl surveys.

And so this was done to allow for interannual variability in the fishery input sample size and the survey sample size. And they— all of the models also implement Francis reweighting, which is an objective approach to weight the data that was used. And then some other specifics, 25.2, in that mod, in 25.2a, the autocorrelation parameter and the staff recruitment relationship was fixed. Next because it's not estimated confidently in the model. I mean, the values used is based on published values for flatfish.

25.3A is building on 25.2a, and there we use the float option and stock sensitive— sensitive SS3 to get survey catchability, and this allows the model to derive the MLE for catchability. And then 25.4a, We removed the time block on the slope bottom trawl survey selectivity that starts in 2011 because that only includes 2 years of data, and that was meant to stabilize the estimation of the slope survey selectivity.

Okay, we also had presented a few models where we have a later start date, again, to kind of address this uncertainty about initial model conditions. They're both based— they're— 26.2, I think it should be A, is based on 25.3A. The model starts in 1977. We include an estimate of equilibrium catch, which is just the mean of the observed catch between 1977 1960 and 1976, and we have early recruitment deviations starting in 1967. And that's based on what we're seeing here in the bottom right corner for the growth curves, given the variability in length at age.

We're assuming that length data is probably only informative about early recruitment deviations, maybe 10 years in the past, given where the growth curves asymptote and giving that variability in length at age.

Okay, yes. Okay, so jumping into some results, this is just showing the model fits to the survey indices. You have the fits to the bottom, the shelf bottom trawl survey at the top left, the slope bottom trawl survey the bottom left, and the AFSC longline survey in the bottom right. What I want you to focus on here are certain colors. So royal blue is Model 25.0, and the warmer colors are the models that I briefly highlighted moments ago.

I'm also including the likelihood table for— to show you the change in the likelihoods for these models. And so just looking at this table here, you can see that when implementing Francis reweighting, you get an overall improvement in the fit, the model fit to the survey likelihood, to the surveys, but that's really being driven by the improvement in the fit to the AFSC Longline Survey RPNs. And again, that's at the bottom right, and, and that improvement is really driven by this kind of more recent time period, the fit to the more recent time period of the observations.

Okay. And so I just wanted to show what the impact of data weighting in terms of the input sample size. So this table just summarizes what is used, the status quo status quo, and the averages over all years are shown in this table. And so the first column is the status quo input sample size that's used in the base model. And then the second column is showing you the adjusted sample size using the variance adjustment that's used in the last accepted assessment.

And so really, the— there's higher— the input sample size is reduced in all cases, but the model is weighted towards, you know, the bottom trawl survey length composition data. Now, when I update the input sample size using the AFSC longline— not the AFSC longline, the AFSCISS R package— you can see there's a fairly large jump in the average numbers across or the AFSC across the two bottom trawl surveys. If the same is true when you use the number of trawls from, for the fisheries. And then when you use Francis reweighting, there's a direct, the model Francis reweighting really wants to downweight all of the data sources, length composition data sources, but really downweights the length composition the information coming from the length composition from the bottom trawl surveys. So there's a drastic change in the data weighting using the Francis reweighting method.

So that generally, that generally leads to some loss of the fit to the survey length composition data, and I'm not going to go through that because it's a lot of information, but you can see that in Figures 9 through 13 in the report. But again, you know, Essentially downweighting the, the length composition data in the model allowed for better fit to the survey indices, especially the longline RPNs. And so it does have some impacts on our estimates of recruitment, this, the data weighting. So the, the top left figure is showing the estimates of age zero recruits across Across the models, again, you know, you want to concentrate on the royal blue and the warmer colors. And then on the bottom right, I'm showing you the recruitment deviations.

So essentially, you know, there's, there's some similarity in terms of the estimates of the recruit deviations, the trends in the recruitment deviations are the same. But using Francis reweighting, we essentially kind of expect fewer recruits, especially in some of the years where we have data than the base model. And this becomes more apparent where you look at the dark red and orange where you start the model in 1977. The— our expectation of recruitment is much lower than, Models 25.3 and 25.4 in the base model.

So just looking at the estimates of spawning stock biomass across— not, not, sorry, it's just showing spawning stock biomass, not fishing mortality, but across the models. And so here you can see that all the models essentially estimate the same general trend, especially starting 1978 forward. But starting the model in 1977, our perception of the stock changes. And you can see here that, you know, if we start the model later, the initial population size is— or SSB is much higher than if we include those early recruitment deviations and start the model earlier in 1945, where the model estimates the population to be much smaller. And much more productive.

Looking at the more recent time period for our estimates of SSB, you know, the models that are implementing the Francis reweighting estimate SSB to be somewhat smaller than the base model.

So just going back to the idea of some of the data conflicts, I'm just showing you the likelihood profiles for, um, unfished recruitment. The length data are informative about unfished recruitment, so this profile is showing the length likelihood components. So black is all of them together, royal blue is the trawl fishery, bright green is the longline fishery, neon green is the shelf bottom trawl survey, orange is a slope, and red is the AFSC longline survey, and you can see in general across the models, the shelf survey wants the initial recruitment, unfish recruitment, to be much smaller than, say, the AFSC longline survey or the slope survey. It's more apparent when you start the model in 1977, you can see the estimates of R0 much larger. It seems to be driven by the longline survey and the slope survey, whereas all of the other likelihood components are indicating a smaller R0.

Okay, in terms of retrospective patterns, I'm just showing Model 25.0, which is essentially the base model. With 25.3a and 25.4a. And you can see the retrospective pattern does increase in those models where one, catchability is not fixed for the bottom trawl surveys and we're using the Francis reweighting approach.

The top, the top figures are showing the, the retrospective pattern in SSB and then just in the bottom row is showing you the survey catchability estimates from the model, given the different peels. So this is the full model going to the 10th peel here. And so in the base model, you know, the bottom trawl survey catchability is fixed, and so you do see some changes in the longline survey. When you free up, the model to get the MLE for the bottom trawl surveys, you can see that the Q estimate for the shelf bottom trawl survey decreases as you remove data. And that seems to be driving one of the main drivers of this retrospective pattern.

Okay, so just to quickly summarize those Tier 3 models, the impacts of data weighting and the likelihood profiles and R0 show that there are several data conflicts in the model. Yes, there is, there are stronger retrospective patterns in some of the newer alternative models, but we recommend bringing forward models 25.0, 25.3a, and 25.4a in November. The advantage of the, the latter two is that they incorporate decision-making following best practices. In a lot of cases, and I think a more defensible approach to catchability.

And then just quickly, I think if I have a few more minutes, future steps. This model is lacking age data. Age data are available from the Shelf Bottom Trawl Survey and the Slope Bottom Trawl Survey, but we don't really have any of that yet from the commercial fisheries, and the Age and Growth Program are working on that, so that I can work with them to see if we can develop informative age distributions from the fishery otolith collections. The hope there would be that this incorporating age data could stabilize the model, better inform recruitment, maybe alleviate some of the data conflicts that we're seeing, but this will take several years before it's available for use in the model, given aging priorities, but the Aging Growth Program has— is starting to age some of the fishery otoliths as we speak, and I received 1 year of data already. Another limitation is that we don't have otoliths from the longline survey, so that it would be something that we would need to look into and see if that would even be a possibility to collect.

Another— and this has come up, but this is not a point on this slide, but this has come up a few times, you know, we also used the slope survey data in this assessment, and that hasn't occurred since 2016. And so that does seem to be a data gap that would be— we see as an advantage to fill if possible through the survey modernization activity that is going on here at the center. So I guess one of the main questions that at some point we will have to kind of confront is whether we, you know, depending upon decisions made this year, do we conduct, you know, this assessment as an update assessment until age data are available.

The plan team also asked me to provide a Tier 3 model. And so I did that this year. And so that's kind of to kind of get this question, you know, do we explore other management options for the stock, given some of the data limitations? Do we move towards a Tier 4 or Tier 5 assessment? You know, Tier 4, you know, you do need to have selectivity and mature— and a maturity ODrive, which in this case is fairly old, but that's the same for the assessment model.

And then the Tier 5 model, that's pretty straightforward to implement. And you see the results of that here on the slide. Slide, the REMA model was fit to basically all of the survey data that's available, the Aleutian Islands bottom trawl survey, the Aleutian Islands longline survey RPNs, the shelf bottom trawl survey biomass, slope biomass, and then the longline survey slope RPNs as well. And you can see here, you know, given the fits, The overall estimate of biomass basically tracks the, the fits to the survey, the survey time series. So, that will be brought forward in November.

And again, that will be part of the discussion about whether we move forward with a Tier 5 as an interim approach while we wait for data sources to be, you know, data gaps to be filled, or do we hold off and think of some other options in the meantime? Okay, I think that's all I have. Tried to get through that kind of quickly. Great, thank you. We have at least one question here, Franz.

Yeah, thanks, Megan, for that overview. Appreciate this is kind of a tough one, data limits. Becoming a data-limited stock, I guess. I have a couple questions, one about the time blocks and the recommendation on the models, and the other about the Tier 5 approach.

So the models that you're proposing both include the time blocks that have been used historically for the slope survey, selectivity. Um, yes, just looking back through that, it doesn't seem like they particularly well support it, um, in general. And then you propose two models, 25.3A and .4A, um, where 4A is understood, um, combines the two selectivity periods to provide a little more stability. And if I recall, the results don't really change all that much, but I was wondering what your kind of thinking was behind bringing forward these two models, given that the differences are pretty minor, and from the, you know, the 25-3A seems really sensitive to, you know, just having those two surveys and the last period, like the selectivity, for example, was all over the place.

So I guess my question is, are those, you know, sufficiently distinct to be informative? In particular, is 25-3A, you know, a good model given that it also had poor convergence properties, I think, and a poor retrospective?

So is there a good reason to bring both of those forward?

Well, the— yes, they both have a larger retrospective, and so that is a disadvantage of those. I think the reason why I wanted to bring those forward is because One, we have catchability right now and the base model is fixed. And that has been, that comes up, this has come up a couple of times when this stock has been reviewed. Essentially, those estimates come from a model run where basically the, I think the shelf bottom trawl survey data from 2007 to 2016. I think 2007 to 2014 were removed from the model because that was a period— or maybe 2012— a period of relatively high recruitment.

And to get those estimates, and then they were fixed, and then the model was reran using those fixed estimates. And so I know that came up in the— all of the CIE reviewers were questioned if that was actually a valid, valid approach to come up with a catchability estimates for the bottom trawl surveys to fix them because it was kind of double using the data. And so it seemed like using— allowing the model to derive the MLE of catchability seemed to be maybe a more defensible approach, and so that's why I wanted to bring that forward.

Yeah, thanks, that helps some. I mean, I guess the question was really, is there enough benefit to maintaining those 3 selectivity periods to— and given the poorer performance of 25-3. Okay, okay, so for the bottom trawl survey? Yes. For the slope?

Yeah, it has the slope survey. Okay, sorry.

Yeah, so the first time period is basically to account so we have some early length composition data that is coming from a cooperative survey that was done prior to 2002, and so the first time block is representing a possible difference in selectivity. And then the, yeah, so then the, the remain, remaining time blocks, the two time blocks, were basically to account for changing availability for that strong recruitment class coming between coming onto the slope.

And so, yeah, that's why I— and it doesn't seem like— so 25.4 is to remove— only have one time block to represent the more modern survey.

And so that's why I wanted to bring 25.4. 25.3, Yeah, agreed. I don't know if that necessarily necessarily needs to come forward. We didn't have a long discussion about time blocks this time, so it seemed like a reasonable one to, to bring forward because that's been kind of the status quo.

Okay, thanks. I appreciate it. My other question was about the Tier 5 scaling in particular. I was trying to understand what that's scaled to. Is that the biomass estimate that comes out of that, is that reflecting, you know, kind of a scaling to the slope survey?

So, the kind of the expected biomass on the slope as informed by the other indices, or how to interpret that? Okay. So, the Tier 5 model, we have two scaling parameters. One basically is scaling the Aleutian Islands RPNs to the Aleutian Islands bottom trawl survey. And then the other scaling parameter is scaling the RPNs from the slope, from the longline survey to the— yeah, to basically the slope survey, the bottom trawl survey biomass.

Okay, thank you. And so then, yeah, so then the— but given that there's a short time series for the slope, you can see that, you know, you see this, like, drop after 19— not 19, 2016, I mean. That's being driven by the RPNs.

We have another question from Dana.

Hi, Megan. Thanks. I kind of picked up similar to what Franz was getting at there and don't really see the utility in 25.3A. To be brought forward, but I guess I wasn't ready to give up on the 1977 start. So I was just wondering if you could kind of elaborate on what the obstacles were for 26.3, which I think would be a good analog to 26.4A.

Mm-hmm. Yes. So as I was running that model, I did run into some— it seemed to— I was running into bounding issues on selectivity parameters once I started it in 19— that particular model in 1977, and just kind of ran out of time to try to troubleshoot that. And so did not suggest moving forward with the earlier— the later start models.

I think in your first slide, you talked about how the deepwater surveys changed the trajectory quite a bit. In the original, like, 16 5-year-old model, have you done like a leave one out kind of analysis where you take each survey out one at a time?

Yes, I think I did that last year, and I cannot remember exactly how that all went. I was planning on doing that for the November report. Perfect. Including that in the November report. Okay, thanks.

Okay, I'm not seeing other questions. Thanks for the presentation, Megan.

Thanks for your time.

Okay, uh, I guess we will continue with The plan team discussion recommendations for Greenland turbot, as long as people still have interest. The team discussed the persistent conflict between— Okay. The team discussed the persistent conflict between survey indices and the length compositions and the continued retrospective patterns that Megan mentioned. They suggested some sensitivity activity runs of recruitment autocorrelation, later start year, and Tier 4 applications to do that for future investigations. We also discussed the importance of additional age information for the slope survey and from fishery collections, and potentially looking into an otolith collection to start up on the longline survey if that was possible.

And then the authors noted that an ESP has been initiated for this stock to develop a more comprehensive view of available information that's not used in the assessment and include data, um, and then that this information could be used in the risk table. And then for recommendations, the team agreed with the author's proposed Tier 3 models, as Meagan mentioned, 2500, 253A, 254A, uh, for November, and to also include an updated Tier 5 REMA. To use as potential interim advice. This does not constitute a permanent change in tier, but just to have that available. We also recommended to clearly document how the AFSC longline survey was used in both Tier 3 and 5 approaches and to include how the revised design might impact the use of that survey.

And we recommend to include information needed to improve the assessment in the research data gap section, also to include priorities for additional age data and criteria for selecting.

And— oh, and also to include the importance of deepwater survey information.

And then we also had some additional recommendations for future research, which are included in the minutes. Concerning reweighting, autocorrelation, and Tier 4 approaches.

And that's all we have.

Great. Do we have any questions on the plan team comments on Greenland Turbot?

Okay, it looks like no. I know I said we're going to 5:45, but it's 5:30. If anybody objects, I'm sorry, we're going to quit for the day and start on COD proposed models in the morning.

Okay, we'll be there.