BRIDGE-CH4: Code and Data for Global Soil Methane Uptake Prediction from Process-Based Simulations TEM and Chamber Observations from 1979 to 2025

This repository contains the source code and datasets for BRIDGE-CH4, a neural-process-based framework that jointly learns from coarse-resolution TEM simulations and sparse chamber observations for global soil methane uptake prediction. It includes model implementations, training and evaluation scripts, environmental drivers, and raw/processed methane datasets. Please refer to readme.md for instructions on getting started. Thank you for using our work!