Supplementary files for "Data-driven estimation of nitric oxide emissions from global soils based on dominant vegetation covers"

In-situ observations collected from publications, data-driven model codes, DNDC simulation files, and the supporting data for all figures of this study are uploaded. In-situ observations included 1,356 observations of soil nitric oxide (NO) emissions from 192 sites, including 1,032 for cropland soils from 70 sites, 114 for grassland soils from 36 sites, and 208 for forest soils from 86 sites. Data-driven models provided three machine learning methods, including random forest (RF), generalized boosted regression model (GBM), and radial basis function (RBF). The DNDC simulation files included simulation files of 51 selected sites.本研究已上传取自已发表文献的原位观测数据、数据驱动模型代码、DNDC模拟文件，以及所有配图的支撑数据。其中原位观测数据包含来自192个站点的1356条土壤一氧化氮（NO）排放观测记录，涵盖70个农田站点的1032条记录、36个草地站点的114条记录，以及86个森林站点的208条记录。数据驱动模型涵盖三种机器学习方法，分别为随机森林（random forest, RF）、广义提升回归模型（generalized boosted regression model, GBM）以及径向基函数（radial basis function, RBF）。DNDC模拟文件包含51个选定站点的模拟数据文件。