Han Eunjin, Ines Amor V M, Koo Jawoo
International Research Institute for Climate and Society, Columbia University, NY, 10964, USA.
Department of Plant, Soil and Microbial Sciences, Michigan State University, Michigan State University, MI, 48824, USA.
Environ Model Softw. 2019 Sep;119:70-83. doi: 10.1016/j.envsoft.2019.05.012.
One major challenge in applying crop simulation models at the regional or global scale is the lack of available global gridded soil profile data. We developed a 10-km resolution global soil profile dataset, at 2 m depth, compatible with DSSAT using SoilGrids1km. Several soil physical and chemical properties required by DSSAT were directly extracted from SoilGrids1km. Pedo-transfer functions were used to derive soil hydraulic properties. Other soil parameters not available from SoilGrids1km were estimated from HarvestChoice HC27 generic soil profiles. The newly developed soil profile dataset was evaluated in different regions of the globe using independent soil databases from other sources. In general, we found that the derived soil properties matched well with data from other soil data sources. An ex-ante assessment for maize intensification in Tanzania is provided to show the potential regional to global uses of the new gridded soil profile dataset.
在区域或全球尺度上应用作物模拟模型的一个主要挑战是缺乏可用的全球网格化土壤剖面数据。我们利用SoilGrids1km开发了一个分辨率为10公里、深度为2米的全球土壤剖面数据集，该数据集与DSSAT兼容。DSSAT所需的几种土壤物理和化学性质直接从SoilGrids1km中提取。利用土壤传递函数推导土壤水力性质。其他无法从SoilGrids1km获得的土壤参数则根据HarvestChoice HC27通用土壤剖面进行估算。利用来自其他来源的独立土壤数据库，在全球不同地区对新开发的土壤剖面数据集进行了评估。总体而言，我们发现推导得到的土壤性质与其他土壤数据源的数据匹配良好。通过对坦桑尼亚玉米集约化的事前评估，展示了新的网格化土壤剖面数据集在区域到全球层面的潜在用途。
