Projection units and survey points on land

[larogers123]

Currently in 7_rangeprojection.r, projections of CPUE (by wt) are made on a lat/lon grid. This is essentially the projected catch per unit of survey effort (~= survey haul) were a haul to be made in that particular place.

We then take the average of all projections within a larger grid box to calculate an average projected CPUE per cell (essentially accounting for the differences in rugosity within that cell).

For points that fall on land, rather than setting their CPUE to zero (as is currently in the code), I think we should just omit them from the average. Setting them to zero implies that our final output raster will be equivalent to a total catch within that cell (thus the catch within these coastal cells is lower in total), but I've rather been thinking of that raster as showing the relative density of fish. In this case, giving points on land a value of zero makes it look like the density of fish in the coastal areas is lower than otherwise.

It would be great to get from a density index (CPUE) to an absolute number, but I'm not sure how to do that. Possibly we could dig in to assessments (for assessed species), assume catchability for non-assessed species), and calculate area swept for a haul, etc. to get up to total abundance in an area, but would we gain much by doing this?

This should become clear at the next stage, so I'm flagging the points-on-land issue for now. To remove from average, just drop them from rugos object before merging with clim object.