adding kriging fusion method, several modifications in raster predictions functions

master script, temp predictions, for Oregon with 7 hold out prop, 10 samples and 12 dates

master script, multisampling first predictions with new code using prop hold out 10% to 70%

initial commit, analyses paper part 1 with revised results

master script, test run for NASA

master script, preparing data for Oregon for the test run

raster prediction, modification for screening of training, runs by stage

covariates production script, fixing bug without stage 2

gam day, predictions for paper testing covariates combination 4

kriging day, predictions for paper, testing covariates combination 3