Code & Data for cross-validation of LPI based predictive models.

This repository contains the Code and Data needed to conduct spatial cross-validation of LPI based predictive models. - lag_cv_analysis.R This is the main script that loads the trained models, compares their predtive utility, calculates dissimilarity matricies to produce main manuscript figures. Corresponding files are given for analysing models from Spooner et al (fi_cv_analysis.R) - spat_proj_analysis.R This script analyses the LPI spatial projection map data "Data Prep" contain scripts used to prepare data for analysis, including extracting climate and land-use data and working with WDPA and IUCN range maps. lpi_fi_data_prep.R R code to combine Spooner et al (2018) separate data on lambda, mass, luc and cc. Separate datasets for populations whose modeled trend has an rsquared > 0.5 and all populations are generated. cru_extract.R R code to extract temperature data corresponding to LPI populations. hyde_extract_local.R R code to extract land-use data relating to LPI populations. The 'models' directory contains codes to contr Processed data used to make the projections is here: * bird_spat_dat.rds * mam_spat_dat.rds Code to collate the species distribution data and extract associated environmental rates of change can be found in: * lpi_spat_map_dataprep.R, cc_rate_calc_hpc.R, and luc_rate_calc_hpc.R are used Code to take the data and make the above rds files, to be used in the projections: * lpi_spat_map_model_analysis.R