inst/scripts/03_temperature_carstm.R

In jae0/ecofishmod: Ecosystem-based fishery stock assessment models using conditional autoregressive models to represent spatial and temporal variabilty.

#   !!! WARNING, this uses a lot of RAM !!! 96 GB

if (!exists("year.assessment")) {
  year.assessment=lubridate::year(Sys.Date())
  year.assessment=lubridate::year(Sys.Date()) -1
  year.assessment = 2018
}


# construct basic parameter list defining the main characteristics of the study
# and some plotting parameters (bounding box, projection, bathymetry layout, coastline)
p = aegis.carstm::temperature_carstm(
  DS = "parameters",
  project_name = "temperature",
  variabletomodel = "t",
  carstm_model_label = "production",
  inputdata_spatial_discretization_planar_km = 1,  # km controls resolution of data prior to modelling to reduce data set and speed up modelling
  inputdata_temporal_discretization_yr = 24/365,  # ie., every 2 weeks .. controls resolution of data prior to modelling to reduce data set and speed up modelling
  yrs = 1999:year.assessment,
  spatial_domain = "SSE",  # defines spatial area, currenty: "snowcrab" or "SSE"
  areal_units_resolution_km = 25, # km dim of lattice ~ 1 hr
  areal_units_proj4string_planar_km = aegis::projection_proj4string("utm20"),  # coord system to use for areal estimation and gridding for carstm
  areal_units_source = "lattice", # "stmv_fields" to use ageis fields instead of carstm fields ... note variables are not the same
  areal_units_overlay = "none"
)


  # to recreate the underlying data
  sppoly = areal_units( p=p, redo=TRUE )  # this has already been done in aegis.polygons::01 polygons.R .. should nto have to redo
  spplot( sppoly, "AUID", main="AUID", sp.layout=p$coastLayout )


  M = temperature.db( p=p, DS="aggregated_data", redo=TRUE )  # will redo if not found .. not used here but used for data matching/lookup in other aegis projects that use bathymetry
  M = temperature_carstm( p=p, DS="carstm_inputs", redo=TRUE )  # will redo if not found
  # to extract fits and predictions

  fit = carstm_model( p=p, M=M )

  # extract results
  fit = carstm_model( p=p, DS="carstm_modelled_fit" )  # extract currently saved model fit
  res = carstm_summary( p=p, operation="load"  ) # to load currently saved results

  plot(fit)
  plot(fit, plot.prior=TRUE, plot.hyperparameters=TRUE, plot.fixed.effects=FALSE )
  s = summary(fit)
  s$dic$dic
  s$dic$p.eff

  # maps of some of the results
  vn = paste(p$variabletomodel, "predicted", sep=".")
  carstm_plot( p=p, res=res, vn=vn )

  vn = paste(p$variabletomodel, "random_sample_iid", sep=".")
  if (exists(vn, res)) carstm_plot( p=p, res=res, vn=vn, time_match=list(year="1950", dyear="0") )

  vn = paste(p$variabletomodel, "random_auid_nonspatial", sep=".")
  if (exists(vn, res)) {
    res_dim = dim( res[[vn]] )
    if (res_dim == 1 ) time_match = NULL
    if (res_dim == 2 ) time_match = list(year="2000")
    if (res_dim == 3 ) time_match = list(year="2000", dyear="0.8" )
    carstm_plot( p=p, res=res, vn=vn, time_match=time_match )
  }

  vn = paste(p$variabletomodel, "random_auid_spatial", sep=".")
  if (exists(vn, res)) {
    res_dim = dim( res[[vn]] )
    if (res_dim == 1 ) time_match = NULL
    if (res_dim == 2 ) time_match = list(year="2000")
    if (res_dim == 3 ) time_match = list(year="2000", dyear="0.8" )
    carstm_plot( p=p, res=res, vn=vn, time_match=time_match )
  }

}


# end