[Feature]: Add plots of spatial trends in abundance

[iantaylor-NOAA]

Describe the problem your feature request is related to.

We look at spatial patterns of densities and residuals but our focus with resulting indices is on changes in abundance on a coastwide scale, and sometimes on the scale of WA, OR, and CA, but trends in abundance may be changing at a finer scale and maps are more compelling than lines showing time series by area.

Describe the solution you'd like

I got to see some cool plots of spatially variable trends for Pacific spiny dogfish in a presentation by @LindsayDavidson yesterday. I think these would be useful to add to the set of default diagnostics for U.S. west coast indices created by this package. My ideal would be to see the trends from start to end of the WCGBTS survey plus decadal changes (e.g. 2003-2013 and 2013-2023).

This is a low priority, but @kellijohnson-NOAA suggested that it's worth posting an issue about.

Describe alternatives you have considered

Stick with the status-quo diagnostics.

Additional context

No response

[ericward-noaa]

This is something we did for this paper:

https://onlinelibrary.wiley.com/doi/abs/10.1111/ecog.05176

Fundamentally it's a slightly different model from the conventional ones we use, because of the spatial trend (so involves fitting a 2nd round of models). Because of that I don't know that bundling it in indexwc makes the most sense? But maybe we split it off into its own repo? It'd be really easy to us GH actions to fit the models annually and produce the kinds of trend maps you're looking for. Open to other ways of dealing with this

[iantaylor-NOAA]

Thanks @ericward-noaa. I look forward to reading the paper during a less busy time. I had naively assumed that you could just subtract the estimated distribution of density at the end of the time series from that at the beginning, but of course that would not have the statistical validity of a model that explicitly incorporates the trend.

I'm also not imagining that {indexwc} needs to reinvent any of this stuff, but posted the issue in the hopes that we can benefit from some of the tools that others are exploring with {sdmTMB} elsewhere. But that's not a high priority right now regardless.