The Perfect Assessment

[StardustSB]

What would the perfect stock assessment need? After thinking about how I would answer this question, I felt that starting with the end product and working backward was best. This is how I visualized it. I know I have oversimplified, but maybe that is what we need to do. Simplify the question first, go through the exercise and then dig into the details.

If Data & Information were available like groceries at Costco, what would you put in your shopping cart? Now take a step back. What work needed to be done for that item to exist on the shelf in the first place?
Someone at the store had to pull it from the back inventory and put it on the shelf (Database). Someone before that, packaged it up ready for consumption (data). Before that, someone cooked/baked it (aging otoliths. running genetics, fecundity etc). Someone prepared (maybe same person) the ingredients (dissection, break & burn, dna). And, before that someone grew & harvested the raw ingredients (collection of specimens). Take another step back. Was that last person the farmer/rancher (fisheries dependent)? Or was it a company owned farm that hires staff and follows strict procedures that cannot be deviated from no matter what the product is (fisheries independent)?

[shcaba]

Terrific framing of this issue. I can take this to one of our weekly stock assessment meetings and start to answer your questions. I can post those answers to this space so we can capture the ideas. Hope to report back some ideas in the next couple of weeks (we don't have a meeting next week).

[StardustSB]

Thanks Jason, I know the analogy is a bit primitive but that’s how I had to break it down to really try and figure out. What are we asking for, what do we need. The idea of starting with the finished product and then stripping it back from there each step, I think would be very helpful.

[shcaba]

Some initial responses to the above questions:

_If Data & Information were available like groceries at Costco, what would you put in your shopping cart?__

DATA (4 major types)

• Catches (landings AND discards) - accurate, with uncertainty estimates if based on estimates. This helps us count how many/how much fish is out there. This should be complete for every removal (catch + discard) source, so every fleet. Leaving out major sources of catches and discards will bias how many fish we think was and is out there.
• Length compositions (randomly and representatively (if we can't measure every fish, which we usually can't) sampled from all sectors). This allows us to estimate fleet or survey selectivity (how the fish are removed from the population), recruitments, and the reduction of the stock status from unfished (what we sometimes call depletion). It can also help estimate natural mortality.
• Age compositions (randomly and representatively sampled from all sectors, with corresponding lengths). Because population age structure is very important to know, but lengths don't usually give us the best resolution of the oldest individuals. This can also help with estimating natural mortality (see below).
• Age and length samples that cover both small and large individuals in order to estimate growth curves. Having this over time to look at time-varying growth would be even better. This is not necessarily the same as having fishery-dependent age and length data, as those may not cover the smallest and biggest sizes needed to get growth parameters estimated well.
• Abundance indices from fishery-independent surveys to give us trends in stock dynamics. Scientifically designed surveys can give better signal on the stock dynamics if designed specifically for a species. If it is a larger survey that samples many species, not all species may be sampled equally well. PARAMETERS
• Natural mortality. One of the most influential life history parameters that is hard to directly measure. We often rely on maximum age or age and growth estimates to estimate this value. But the above data (length and ages) can help us estimate this in the stock assessment, so another reason having that type of data is key.
• Maturity samples from fish in the modeled area. Understanding how this changes over time may be important too.
• Fecundity samples from fish in the modeled area.
• Weight, in paired with lengths, helps understand the biomass connection of the lengths being sampled.
• DNA. Can be helpful to understand stocks and even identify species (e.g., vermilion vs sunset rockfishes; blue vs deacon rockfishes)

Now take a step back. What work is needed to be done for that item to exist on the shelf in the first place? Fishery-dependent

• Catches have to be tracked and enumerated. Samples of the size, weight, ageing structures, gonads, DNA are all potential, but each take time. We are limited by the number of samplers and the processing time for sampling each fish. Ideally, we'd sample every fish, but not an opportunity. Live fish fishery makes acquiring many of these data sources even more challenging. The recreational fishery is VERY hard to sample. The recreational fishery (especially private boats and shore modes) remains the most challenging place to sample.

Fishery-independent

• Designed sampling programs can either be opportunistic (getting whatever you can) or planned out (surveys).
• Limited resources (time, personnel), the large coastline, and highly diverse ecosystem of our groundfishes makes getting the variable types and places of data very hard.
• The use of commercial boats to carry out designed surveys has shown to be a cost-effective way to get larger coverage for less cost, though it is still expensive.

Other complications to data collection

• No-take areas- figuring out how to measure within no-take areas and what to do with that information needs consideration, especially as those areas either grow and mature over time. -Rebuilding plans- rebuilding plans often shutdown all ability to acquire any data sources, which in turn makes it difficult to understand how the population is reacting to the rebuilding implementation.

[StardustSB]

Great start. I believe the exercise of building the framework for assessments in a "dream scenario" is important. This will help us better understand what has been missing, or done in a less effective way because of policy/funding/timing/access etc. We (the royal we, industry, state, fed, ngo) can then work to deal with those hurdles so you (science teams) aren't working with one arm tied behind your back. I look forward to hearing more feedback.

[shcaba]

I look forward to feedback and more discussion on this. Let me know if we can clarify anything, or provide more details.