Treffer: Code for 'Tree species explain only half of explained spatial variability in plant water sensitivity'

This repository contains the analysis code that led to the results presented in the article in Global Change Biology titled “Tree species explain only half of explained spatial variability in plant water sensitivity" (Konings et al., 2024). This description is written assuming the reader has read that article. A copy of this repository is also stored at Github at https://github.com/agkonings/PWSSpeciesAnalysis . Most of the data files associated with this analysis (including the input plant water sensitivity map, the subsetting of theis map to FIA sites, the biogeographic predictor values at the FIA sites, their dominant species, and the predicted plant water sensitivity values from the species-based and random forest models) are uploaded to an associated Dryad repository. The Dryad repository can be found at doi:10.5061/dryad.g1jwstr05 This may be helpful for running some of the code in this repository or for further analysis. Much of the core analysis code is contained in the file rf_regression.py, using an HDF5 file of input maps for the random forest that is created in make_data.py. The file pullFIADominantSpecies.py identifies the FIA sites where a single species covers more than 75% of the basal area. The file pws_calculation.py calculations the initial PWS maps, as described in the article. The file checkCrossCorrs.py is used to determine which potentially inputs have sufficiently low cross-correlated to be useable for the random forest model, as described in the article’s methods file. It also is used to calculate some of the supplementary figures. Lastly, the file writeDataToNETCDF.py is used to write the output of the scripts to the main netCDF output file stored in the Dryad repository. It uses as input a python dill file (packaging the entire contents/state of the Python console session at the end of running rf_regression.py). The main figures of the article are derived from the following scripts:Figure 1: plotPWSForCommonSpecies.pyFigure 2: plotPWSForCommonSpecies.pyFigure 3: rf_regression.pyFigure ...