From Soil Moisture Spatial Patterns to Catchment Nitrate Dynamics using Explainable AI
• Deep learning-XAI approach to predict daily nitrate concentrations in 8 catchments using streamflow and SMAP-Hydroblocks soil moisture maps. • Spatial patterns of soil moisture account for 30% of model feature importance for nitrate predictions. • A
Authors
- Ralf Merz (ORCID: https://orcid.org/0000-0002-3698-2339)
- Noemi Vergopolan (ORCID: https://orcid.org/0000-0002-7298-0509)
- Andréas Musolff (ORCID: https://orcid.org/0000-0002-0115-1359)
- Larisa Tarasova (ORCID: https://orcid.org/0000-0002-7675-0751)
- Felipe Saavedra
- Carolin Winter (ORCID: https://orcid.org/0000-0002-4238-6816)
- Zhenyu Wang
Institutions
- Helmholtz Centre for Environmental Research (DE)
- University of Freiburg (DE)
Publication Details
- Published
- 2025-09-22
- DOI
- https://doi.org/10.22541/essoar.175855522.20150676/v1
- Primary Topic
- Hydrology and Watershed Management Studies
- Type
- article
- Field-Weighted Citation Impact
- 0.00