Adaptive Graph Learning with Transformer for Multi-Reservoir Inflow Prediction

TL;DR

AdaTrip is an adaptive graph learning framework using transformers that models dynamic spatial and temporal dependencies for multi-reservoir inflow prediction, improving accuracy and interpretability over existing methods.

Contribution

It introduces a novel adaptive, time-varying graph learning approach with attention mechanisms for multi-reservoir inflow forecasting, capturing hydrological dependencies dynamically.

Findings

Outperforms existing baselines on the Upper Colorado River Basin data

Enhances prediction for reservoirs with limited historical data

Provides interpretable attention maps revealing hydrological controls

Abstract

Reservoir inflow prediction is crucial for water resource management, yet existing approaches mainly focus on single-reservoir models that ignore spatial dependencies among interconnected reservoirs. We introduce AdaTrip as an adaptive, time-varying graph learning framework for multi-reservoir inflow forecasting. AdaTrip constructs dynamic graphs where reservoirs are nodes with directed edges reflecting hydrological connections, employing attention mechanisms to automatically identify crucial spatial and temporal dependencies. Evaluation on thirty reservoirs in the Upper Colorado River Basin demonstrates superiority over existing baselines, with improved performance for reservoirs with limited records through parameter sharing. Additionally, AdaTrip provides interpretable attention maps at edge and time-step levels, offering insights into hydrological controls to support operational decision-making. Our code is available at https://github.com/humphreyhuu/AdaTrip.