Eric Ward

This is referring to this issue, [https://github.com/pbs-assess/sdmTMB/issues/146](https://github.com/pbs-assess/sdmTMB/issues/146) The major changes are: - adding new family for the standard curve models - allowing a 2nd dataframe to be passed in (currently...

Add std normal

4

- include CDF for normal - possibly with / without estimated detection threshold - possibly with / without variance mean linear relationship - add all to delta models

Update residual-checking.Rmd

2

Added diagnostics for delta/presence-absence models: AUC, TSS TSS addition motivated by discussion here: https://github.com/pbs-assess/sdmTMB/issues/84 It'd be easy to turn the TSS code into a function if it's more broadly useful

The time-varying model applies to global effects, and spatial_trend represents a single field. It'd be nice to have the spatial trend modeled as a RW / AR(1) process

Add variable meshes by time step

3

This is a niche problem, but there's analysts who would like to use sdmTMB to model data that includes closed areas that are variable over time. Time-varying meshes would also...

This is the fix for the quadratics roots code -- using this to finish the spatial indicators paper / ECCWO project. Not sure whether this is better left on its...

I had someone request this recently, but was hoping to have a high level overview on when to standardize or scale covariate and response data, vs when to not do...

I think the solution here is just modifying `coef()`to take `model` as an argument ``` r library(dplyr) library(tidyr) library(sdmTMB) set.seed(1) # Build a mesh to implement the SPDE approach: mesh...

I talked to Lewis last week about this. Currently includes examples of: - main effects by species - spatial fields by species, with and without shared variance - spatiotemporal fields...