Aggregate_area: userArea options

[andybeet]

The aggregate_area function gets passed a userAreas data frame which is used to proportion species landings and values attributed to statistical areas that straddle EPU definitions. This proportioning was based on the mskeyrun project.

The bundled comlandr::mskeyAreas data frame is the list of statistical AREAS and which EPUs they map to. This also contains the NESPP3 code and the proportion of the landings that should be attributed to each EPU. This object does ONLY contains statistical AREAS that straddle EPUs. Therefore when this is joined to the data all other Statistical Areas get assigned the EPU: "Other"

We should not use this object until we have fixed this.

Alternatively another data object (epuAreas) used to define EPUs by statistical AREAs is:

gom <- data.table(AREA = c(500, 510, 512:515), EPU = 'GOM')
gb  <- data.table(AREA = c(521:526, 551, 552, 561, 562), EPU = 'GB')
mab <- data.table(AREA = c(537, 539, 600, 612:616, 621, 622, 625, 626, 631, 632),
                  EPU = 'MAB')
ss  <- data.table(AREA = c(463:467, 511), EPU = 'SS')

epuAreas <- rbindlist(list(gom, gb, mab, ss))
epuAreas[, NESPP3 := 1]
epuAreas[, MeanProp := 1]

This also seems to have a flaw (when applyPropLand = T) in that the join by AREA and NESPP3 only assigns EPU to species with NESPP3 code = 1. The total landings will be correct for all of the species but they'll be assigned to EPU: "Other"

Looking into this

[andybeet]

Having said all of this. When we ran comlandr (for the SOE last year) to create the bennet index and comdat data set we used applyPropLand = F and the epuAreas object. Using these arguments the resulting joins and math are correct and valid (regardless of applyPropValue)

[andybeet]

I have included a check_argument_validation function that will prevent a user from using the wrong data sets in the worng manner. I have also disabled applyProp = T since we dont have the correct data sets created to pass as an argument.