Machine learning-based seasonal SMAP soil moisture retrieval integrating MODIS drought indices: A case study of the Wujiang River Basin
Ju Zhao, Hanyu Lu, Pengfei Qu, Yongyi Yuan · 2026
Read Machine learning-based seasonal SMAP soil moisture retrieval integrating MODIS drought indices: A case study of the Wujiang River Basin by Ju Zhao, Hanyu Lu, Pengfei Qu, Yongyi Yuan (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/plos-pone-0351643.txt.
Description
Ju Zhao and colleagues sharpened satellite estimates of soil moisture over China's Wujiang River Basin, a rugged karst landscape where fields and rock fracture the ground. A NASA satellite reads moisture coarsely, and one model for all seasons fits none of them well there. The team trained separate seasonal models, mixing satellite drought indicators with terrain data, and compared five learning algorithms.
The seasonal CatBoost models beat single-model baselines, doing best in spring and autumn. Reading the model's reasoning showed its logic shifting with the year: winter estimates leaned on terrain, summer ones on heat and water signals. That rotation matches how moisture actually behaves in the basin. Better seasonal soil moisture maps feed drought warnings and water planning in mountain country where ground sensors are sparse.
Catalog record
| CREATOR | Ju Zhao, Hanyu Lu, Pengfei Qu, Yongyi Yuan |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 74592 |
| LICENSE | CC_BY |
| RIGHTS | CC BY |
| SOURCE | PLOS |