mcolvin
commented
5 years ago @sreynolds2 That seems reasonable but does make an assumption that that relationship is similar among segments. It also might be a poor assumption as cpue might be correlated with bend length (i.e., more gear in shorter bends and therefore higher cpue)-seems like somebody should have looked at that already. A different approach but related might be to take all bends over the last 4 years or so and come up with some sort of relative 'density' from cpue and then use a multinomial to allocate fish out.
Neither way is great to extrapolate to both basins so I think going with whatever has the fewest or most robust assumptions.
We could use the approach from Steffensen et al. 2017 (linear relationship between CPUE and RD abundance estimates) to provide presumably more accurate initial abundances by segment and origin given the CPUE data is available by origin. Then, basin abundance = sum of segment abundances and relative density (for assigning bend location) could also be assigned assuming similar densities of bends within segment.
Additionally, we could use CPUE by segment and origin as a default input that is update each year. Then we could have a code that generates the relative density values and the initial abundance inputs from this input. This would be a nice option (more automation of the updating process) given we can obtain CPUE by segment and origin each year and assume the same linear relationship in Steffensen et al. 2017 or easily update that relationship.
Reference: Steffensen, K.D., L.A. Powell, and M.A. Pegg. 2017. Using the robust design framework and relative abundance to predict the population size of pallid sturgeon Scaphirhynchus albus in the lower Missouri River: Journal of Fish Biology, v. 91, no. 5, pp. 1378-1391, 10.1111/jfb.13457.