Are coastal categories determined by trajectories of pelagic origin?

[rBatt]

The current trajectory method look are based on following a surface temperature over space. Just like land-water could be a boundary, deep-shallow could be a boundary.

I see 2 main ways to address this issue:

1. find which temperature grids ever had a haul in them, and limit the universe to those areas. This creates "edge effects" where hauls may or may not have occurred. Advantage is that the shallow-deep boundary is already defined by in-hand data (trawl). The disadvantage is that there are on-shelf locations that were not part of the haul survey, and also that if you wanted to know what strictly deepwater fishes were doing, you would be excluding their trajectories (i.e., no more looking at deep water)
2. Get gridded data of depth, and turn that into a shelf-deep definition. Conduct the trajectory simulation/ analysis twice: once for an on-shelf universe, once for an off-shelf universe. Advantage: thorough, and applies pre-existing algorithm (what to do at a land-water boundary, or at a universe boundary) to the new problem, so not much new code needed. Disadvantage is that the new data are needed, and those need to be turned into a shelf definition.

@mpinsky Know of an easy way to get 1º data that is binary for 0 = off-shelf 1 = on-shelf? Or at least 1º depth? I can look too.

[mpinsky]

ETOPO2: http://www.ngdc.noaa.gov/mgg/fliers/01mgg04.html

[rBatt]

Awesome! @mpinsky 's link is global 2 minute relief data. This corresponds to the 2nd approach I suggested in opening this issue.

[rBatt]

Read-in the data for this in my last commit, once I figure out how to best align with 1º grid data, I should be able to use depth as a boundary the same way I used land as a boundary.

[rBatt]

@mpinsky Took you suggestion, used the ETOPOS data.

View the depth categories here

View the result of the constrained trajectories here