PLSTM-Reg v1.0: a regional physics-encoded LSTM model for simulating reservoir operations without site-specific training

Representing reservoir operations in large-scale hydrological models remains difficult due to complex release decisions and scarce operational records. Here, we develop PLSTM-Reg v1.0, a regional physics-encoded deep learning framework that explicitly enforces mass balance and operational constraints to simulate reservoir operations across diverse systems. The framework is evaluated using 256 representative reservoirs across the Contiguous United States, focusing on three core capabilities: temporal generalization to unseen periods, spatial transfer to unseen reservoirs, and historical data reconstruction in data-scarce reservoirs. Under temporal testing, the regional model improves 1 d-ahead release forecasts from a median Kling–Gupta Efficiency (KGE) of 0.83 to 0.96 relative to local counterparts (i.e., models trained exclusively on site-specific records), and reduces poorly simulated cases (KGE<0.8) from 41.8 % to 2.3 %. For long-term simulation, storage performance reaches a median KGE of 0.79, a modest gain over local models (0.76) but with notable robustness for reservoirs with large capacity. When transferred to unseen reservoirs, the model substantially outperforms widely used rule-based schemes: median KGE rises from 0.55 (best benchmark) to 0.73 for release and from 0.22 to 0.59 for storage, and the proportion of usable simulations (KGE>0.5) increases from 56.6 % to 89.8 % for release and 14.4 % to 61.7 % for storage. In historical storage reconstruction, incorporating monthly satellite-derived surface area as a proxy for missing ground-based records enables reconstruction accuracy comparable to local models. These results demonstrate that cross-reservoir deep learning combined with physical knowledge provides a scalable scheme for representing human water management within large-scale hydrological and land surface models.