priors significantly affect results and should be explicitly defined up front in masterScript. values of priors have also been made more reasonable.
covariates are scaled based on values in their respective layers, however, some have data outside of norway (eg corine), some convert it all to 0 (eg road density), some is all NA (eg met weather), leading to inconsistent standardisation.
all spatial processing used res and crs defined in masterScript, except for prediction, this now uses the parameters as defined in masterScript.
What changes have been made?
masterScript.R - define prior.range and prior.sigma
pipeline/import/initialiseRepository.R - add prior.range and prior.sigma to controlPars.RDS
pipeline/import/environmentalImport.R - convert regionGeometry to raster, and use it to set all rasters outside of regionGeometry to NA
pipeline/integration/utils/metadataProduction.Rmd - bug: split region with comma, in cases where multiple regions were used to define regionGeometry.
pipeline/models/speciesModelRuns.R - use res and projCRS as defined yn masterScript for predictionData. use prior.range & prior.sigma from controsPars.RDS. bug: if modelRun = "allSpecies", redList should be NULL
Why have changes been made?
What changes have been made?