This code reproduces the results for XGBoost presented in 'A widely-used eddy covariance gap-filling method creates systematic bias in carbon balance estimates'
The code has been succesfully run in the Windows environment using Python 3.8.13
matplotlib==3.5.2, pandas==1.4.2, numpy==1.22.4, sklearn==0.0, tensorflow==2.9.1, keras==2.9.0, xgboost==1.6.1, openpyxl==3.0.10, geopandas==0.11.1, seaborn==0.12.1
run pip install -r requirements.txt to install all above packages.
analyze_FLUXNET.py analyze_north.py