don't renumber years in speciesEcoregion

[CeresBarros]

At the start of Biomass_core, there is a chunk of code that renumbers the years in speciesEcoregion so that they start at 0. This chunk has two components. The first checks whether there are any years <= successionTimestep and takes the last available. The it binds traits for this year with any traits for years > successionTimestep and renumbers all years as year - min(year):

  speciesEcoregion[, identifier := year > P(sim)$successionTimestep]
  speciesEcoregion_True <- speciesEcoregion[identifier == TRUE, ]
  speciesEcoregion_False <- speciesEcoregion[identifier == FALSE, ]
  if (NROW(speciesEcoregion_False)) {
    speciesEcoregion_True_addon <- speciesEcoregion_False[year == max(year), ]
    speciesEcoregion_True <- rbindlist(list(speciesEcoregion_True_addon, speciesEcoregion_True))
  }
  sim$speciesEcoregion <- speciesEcoregion_True[, ':='(year = year - min(year), identifier = NULL)]

I believe this can be problematic when the user/another module provides a table of trait values for different years for several reasons:

Assuming that speciesEcoregion starts with trait values for every year of the simulation:

1. if start(sim) and end(sim) are both <= successionTimestep, then only the last year of trait values will be used.
2. if start(sim) and end(sim) are both > successionTimestep, years in speciesEcoregion will be renumbered from 0 to end(sim)-start(sim) and later on, no trait values will be found at time(sim)
3. if start(sim) is <= successionTimestep and end(sim) > successionTimestep, the trait values used at start(sim) will correspond to whatever year closest to successionTimestep in the range [start(sim), successionTimestep) and the remaining years in speciesEcoregion will be subtracted this year, meaning that the trait values used at each year will be shifted "backwards"

According to the LANDIS-II documentation, speciesEcoregion should have values for year 0 , which is presumably why this "year rescaling" is happening -- although I couldn't find a place where we use speciesEcoregion[, year == 0] (but I may have missed it). So I propose that, instead of renumbering the years in this table, we check for trait values at year == 0 and if they don't exist we take the trait values at min(year) and "duplicate them" for year == 0.

Could others please have a look and see if I missed anything?

[CeresBarros]

I've converted this to a draft for the time being as this will need testing before ready to be evaluated for a merge.

[ianmseddy]

Investigating this now... the code chunk is unfamiliar to me.