arXiv cs.AI· Xilin Xia·· 11 小时前AI 评分38
从水文定律到流域模型：基于过程的模型在三大洲达到深度学习精度
From hydrological laws to catchment models: A process-based model matches deep-learning accuracy on three continents
AI 导读
研究者提出受三条物理定律约束的低维流域模型 CaRT+，每流域 35 个参数、14 个状态，不含逐日神经网络，在英国 602 个和美国 593 个实测流域的留出年份中位 NSE 分别达 0.863 和 0.765，与有状态 LSTM 的 0.869 和 0.751 相当。
正文 · 原文
Abstract:Dooge asked whether hydrology has laws at the catchment scale. We ask whether a low-dimensional system built under physical laws can match the accuracy of neural networks that outperform process-based models on the same inputs. Three laws constrain a lumped catchment model. Conservation makes it a flux-connected storage network, dissipation fixes the sign of every flux and aggregation determines how the unresolved composition enters the fluxes, through effective laws over threshold distributions and a memory of the forcing. These laws admit four kinds of refinement. A parsimonious model under them, CaRT (11 parameters per catchment), is refined in a loop assisted by AI agents. Each refinement addresses an observed deficit through a hypothesis from the literature. The result, CaRT+, has fourteen states, 35 parameters per catchment and no daily neural network. Most refinements follow from aggregation over thresholds or from variables that carry no water but modulate fluxes. On 602 British and 593 US gauged catchments, single CaRT+ models reach median Nash-Sutcliffe efficiencies of 0.863 and 0.765 in held-out years. A stateful LSTM reaches 0.869 and 0.751 and an entity-aware LSTM 0.862 and 0.754, while three-seed LSTM ensembles remain ahead of the CaRT+ ensemble by 0.003-0.012 in median paired NSE. On 505 drier Australian gauged catchments unused in development, single CaRT+ models reach 0.778 against 0.762 and 0.753 for the two LSTMs. The results suggest that catchment-scale laws exist as laws of form, with the constitutive families shared and only their parameters learned.
| Comments: | 61 pages, 12 figures, 13 tables, plus 19 pages of Supporting Information |
| Subjects: | Machine Learning (cs.LG); Artificial Intelligence (cs.AI); Dynamical Systems (math.DS); Geophysics (physics.geo-ph) |
| Cite as: | arXiv:2611.06382 [cs.LG] |
| (or arXiv:2611.06382v1 [cs.LG] for this version) | |
| https://doi.org/10.48550/arXiv.2611.06382 arXiv-issued DOI via DataCite (pending registration) |
Submission history
From: Xilin Xia Dr [view email]
[v1]
Wed, 4 Nov 2026 11:45:47 UTC (3,479 KB)
来源：arXiv cs.AI · arxiv.org