search

brianlangseth-NOAA / Spatial-Workshop-SPASAM

Spatially stratified simulation-estimation framework incorporating multistage stock-recruit relationships and incorporating larval IBM outputs
1 stars 0 forks source link

Initial abundance issues? #32

Closed brianlangseth-NOAA closed 2 years ago

brianlangseth-NOAA commented 2 years ago

Our initial abundance (as established by decaying from R0) shows higher abundance around age 6 than for age 5. How can this be if the age composition is based on using mortality to decay from R0. Do we have a non-zero mortality rate? The same is true for ages 19-22 (but not easily seen in figure)

image

AmySchueller-NOAA commented 2 years ago

Does the initialization allow for deviations from a decay from R0? Are there deviations being estimated for the initial age structure?

Amy

On Fri, May 13, 2022 at 12:41 PM Brian Langseth @.***> wrote:

Our initial abundance (as established by decaying from R0) shows higher abundance for some ages compared to earlier ages. How can this be if the age composition is based on using mortality to decay from R0. Do we have a non-zero mortality rate?

[image: image] https://user-images.githubusercontent.com/27824606/168329046-18b483d3-c121-4eba-a9de-80250e878e86.png

— Reply to this email directly, view it on GitHub https://github.com/brianlangseth-NOAA/Spatial-Workshop-SPASAM/issues/32, or unsubscribe https://github.com/notifications/unsubscribe-auth/AGKZMOQ4W65FL2JBYGNEQFDVJ2A4JANCNFSM5V346JBQ . You are receiving this because you are subscribed to this thread.Message ID: @.***>

--

Amy M. Schueller, PhD

Research Fish Biologist, SEFSC

NOAA Fisheries | U.S. Department of Commerce

Office: (252) 666-7408

www.fisheries.noaa.gov

brianlangseth-NOAA commented 2 years ago

Although the process is a tad confusing deviations in Ninit are allowed. I haven't yet tried to include Ndevs. Right now we decay from R0 based on natural mortality by age, which doesn't use any deviations.

If each age is decayed based on the M from previous ages, the only way for initial abundance to be higher than a previous age would be to have negative mortality. Thus, I wonder if a constant M is assumed for all ages before, or whether there is some other minor error.

brianlangseth-NOAA commented 2 years ago

M is assumed to the be the same for the age at which M is defined AND for all ages before. For example, for M at age = 0.3, 0.27, 0.24, 0.21..., 0.27 is applied for all years (1) before. Consequently, one can obtain increase amounts of decay because as M linearly declines, exp(-M) is decreasing less and less, and so raising it to an increasing power results in a parabolic curve that starts to increase.

Is applying M to all years before it appropriate? Seems like M should change for each year going back in the past.

JDeroba commented 2 years ago

We could just average among across ages and use an age-invariant M. This will produce an identical fit to an age-varying M that is not being estimated. Deroba and Schueller 2013, Performance of stock assessments with misspecified age- and time-varying natural mortality, Fisheries Research 146 27-40.

brianlangseth-NOAA commented 2 years ago

This issue is not as relevant given the finding that estimating Ndevs is doable in issue #37 with other changes to speed up the runtime. With our decision to use a shortened time series, Ndevs is even more appropriate to estimate, thus setting to a decay based on mortality is not as relevant. If however we do ever set to a decay function, the issue raised here should be revisted.