Issues with N. America coastline and EEZ objects

[andrew-edwards]

@remi-daigle

I accepted your pull request but had meant to first pull my dev-andy changes into main, and so reverted the pull request. Else I'm getting tangled up!

Also, the plotting didn't quite work. I think some issues are:

1. coastline_north_america could be simplified, as it includes provinces and states that I don't think we really need. And goes out to Hawaii. plot(coastline_north_america) gives

Though the actual .rda file is not too big, so maybe this isn't such a problem?

2. north_america_eez.rda is 16 Mb, which is the second biggest object in the package (but has no time component), so I think it would be worth getting it smaller, as it will make the loading of pacea take longer.
   Also, it has a mr_df class; not sure what that is. Looks different to the bc one:

> class(north_america_eez)
[1] "sf"         "mr_df"      "tbl_df"     "tbl"        "data.frame"
> class(bc_eez)
[1] "sf"         "data.frame"
> 

Think we need to have a bit of wrangling like for the BC objects in data-raw/coastline, though I'm no expert on this stuff (I can ask someone to help this week maybe).

And plot(north_america_eez) gives this:

which is probably okay (you wouldn't just plot this), but seems to be missing Alaska coast?

3. I tried this, which I thought would work, but doesn't quite:

plot.pacea_st(glorys_bottom_temperature,
                      bc = FALSE,
                      eez = FALSE, 
                      north_america = TRUE, 
                      restrict_plot = TRUE)

The restrict_plot should automatically redo (crop) the axes, but I wonder if it's due to the projections being used (and the option is being ignored). Maybe removing Hawaii will help, not sure.

4. Given the plotting uses ggplot, I wonder if we could just add the north america coast etc. to marea, and just add new options to the end of:

plot.marea_st <- function(x,
                          ...){
  pacea:::plot.pacea_st(x,
                        bc = FALSE,
                        eez = FALSE,
                        ...)
}

So something like:

plot.marea_st <- function(x,
                          ...){
  tplot <- pacea:::plot.pacea_st(x,
                        bc = FALSE,
                        eez = FALSE,
                        ...)
if(north_america == TRUE){
    tplot <- tplot +
      geom_sf(data = coastline_north_america, fill = NA, lty = "dotted")
  }
}

Then also try the restrict_plot option like in plot.pacea_st().

[jamiectam]

Hi! Just chiming in here. I think we should retain the provinces/states in the coastline file, if possible. I they are relevant in in the Maritimes regions for certain fisheries, and I don't think the coastline file was that big (1Mb?). I'm fine with removing Hawaii, though.

We did forget about the Alaska coast (MRGID 8463)! We should add that, but agree that we should try to reduce the file size. I wonder if there is a simple solution with st_simplify in the sf package?

[andrew-edwards]

Yep, strangely it's the EEZ that is the big one. It is a MULTIPOLYGON (though so is pacea::bc_eez) - I'm thinking it can be simplified with some st magic like we have in pacea\data-raw\coastline.