A hybrid deep learning framework for estimating urban tree transpiration
for two species, while maintaining strong performance with limited training data. SHAP identifies model-derived species-specific response patterns consistent with known differences in hydraulic architecture, and diagnoses regimes where Jarvis assumptions underperform. This framework offers a practical pathway for embedding process knowledge into broader urban evapotranspiration and microclimate models.
Authors
- Thomas Rötzer (ORCID: https://orcid.org/0000-0003-3780-7206)
- Lingye Yao (ORCID: https://orcid.org/0000-0003-3588-8225)
- Mohammad A. Rahman (ORCID: https://orcid.org/0000-0001-9872-010X)
- Astrid Moser-Reischl (ORCID: https://orcid.org/0000-0002-1288-411X)
- Stephan Pauleit (ORCID: https://orcid.org/0000-0002-0056-6720)
- Longhao Zhang (ORCID: https://orcid.org/0000-0002-1875-0059)
- Xue Zhong (ORCID: https://orcid.org/0000-0002-1958-1373)
- Qi Li (ORCID: https://orcid.org/0000-0002-0060-3376)
- Kai-Xin Liu
- Jun-Ru Yan
- Xiang Zhang (ORCID: https://orcid.org/0000-0002-8736-4979)
Institutions
- Karlsruhe Institute of Technology (DE)
- Southwest University (CN)
- Tianjin University (CN)
- Chinese University of Hong Kong (HK)
- Ecosystem Sciences (AU)
- Southwest Jiaotong University (CN)
- Technical University of Munich (DE)
- South China University of Technology (CN)
Publication Details
- Journal
- Journal of Environmental Management
- Published
- 2026-08-28
- DOI
- https://doi.org/10.1016/j.jenvman.2026.130780
- Primary Topic
- Plant Water Relations and Carbon Dynamics
- Type
- article
- Field-Weighted Citation Impact
- 0.00
Funders
- Alexander von Humboldt-Stiftung
- Deutsche Forschungsgemeinschaft
- National Natural Science Foundation of China
- China Postdoctoral Science Foundation
- China Scholarship Council
- Bayerisches Staatsministerium für Umwelt und Verbraucherschutz