kellijohnson-NOAA
commented
2 years ago @melissamonk-NOAA Text on VAST, though I am not sure that it is that informative
VASTWestCoast was explored for standardization of the \gls{s-wcgbt} data both north and south of Point Conception. Unfortunately, results were uncertain given the small number of tows that observed r spp. Truncating the spatial distribution of the survey to less than 300 m, which only eliminated a small handful of positive tows, did not decrease the uncertainty such that spatially-explicit parameters were estimable. Model convergence was more of an issue south of Point Conception rather than north of the break. Changing the distributional assumptions of the positive model or changing to a tweedie-like distribution that combines the two models did not increase the likelihood that the model could estimate spatially-explicit parameters. It was decided that a non-spatial model, which is more easily accomplished outside of the VAST framework would be best for all areas where the survey samples this species. Future research could investigate correlation structures between areas and if shared information across small regions of overlap would stabilize parameter estimation.
Writing assignments for shared sections with CA/OR/WA