Reproduce “Advancing Mass-Conserving Deep Learning for Distributed Hydrologic Modeling: A Case Study in a Snow-Dominated Watershed” on OpenResearch
Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.
Universal app for Apple silicon and Intel. Requires macOS 11 or later.
About this paper
Yihan Wang, Lujun Zhang, N. Benjamin Erichson, Yang Hong, Tiantian Yang
PECAN extends mass-conserving neural hydrology to grid-cell storage and runoff routing, producing spatial diagnostics while training only on gauge streamflow. Its relaxed variant, PECAN-MCR, permits water injection or removal. In one Colorado watershed it reaches outlet KGE 0.89 versus VIC’s 0.88, but has worse bias in the highest 2% of flows; its latent states do not validate snow storage physically.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.