- R 57.7%
- Python 42.3%
| Filename | Latest commit message | Latest commit date |
|---|---|---|
| analysis | ||
| dnn | ||
| dnn_predictions | ||
| functions | ||
| preprocessing | ||
| LICENSE | ||
| README.md | ||
europe_tree_competition_publi
This repository contains code for the publication "Loss of competitive strength in European conifer species under climate change"
Data that goes along with the code can be downloaded here: 10.5281/zenodo.19161115
There are several folders:
-
preprocessing: This folder contains scripts that were used for data preprocessing. Hereby we used the data from the harmonized simulation database: https://zenodo.org/records/12750180 The first steps are done in R (version 4.2.2) to extract forest state transitions from the simulation database. In a second part, this information is then used to extract simulation years with corresponding daily climate information from the harmonized simulation database. This second part is done in Python (version 3.9), where a climate compressor DNN is applied to reduce the dimensions in the training data. See supplementary material for detailed description
-
dnn: This folder contains the code to train the deep neural network
-
dnn_predictions: This folder contains the code to run dnn predictions with the trained dnn for different forest states.
-
analysis: This folder contains all code to run the statistical analysis and produce the figures.
-
functions: This folder contains functions for the analysis