FLUXCOM-X monthly net ecosystem exchange on global 0.05 degree grid for 2003

X-BASE NEE is based on the FLUXCOM-X framework which trains machine learning models on in-situ eddy covariance data and uses them to produce this global product. The X-BASE experiment is a basic configuration to serve as a baseline for the FLUXCOM-X framework and includes as predictors the core meteorlogical data, plant functional type classification as well as MODIS based vegitation indicies and land surface temperature. XGBoost was used as the machine learning algorithm. The GPP estimates from the eddy covariance data was based on the Nighttime Partitioning method.

Published paper: https://egusphere.copernicus.org/preprints/2024/egusphere-2024-165/
Nelson, J.A., Walther, S., Jung, M., Gans, F., Kraft, B., Weber, U., Hamdi, Z., Duveiller, G., Zhang, W., 2023. FLUXCOM-X-BASE. https://doi.org/10.18160/5NZG-JMJE

X-BASE NEE 数据集基于 FLUXCOM-X 框架构建，该框架通过对原位涡度相关数据进行机器学习模型训练，并利用训练得到的模型生成全球产品。X-BASE 实验作为一个基本的配置，旨在为 FLUXCOM-X 框架提供基准，其中包含核心气象数据、植物功能型分类以及基于 MODIS 的植被指数和地表温度作为预测因子。在本次研究中，XGBoost 机器学习算法被选用。涡度相关数据估算的 GPP（总初级生产力）是基于夜间分配法进行的。