Search papers, labs, and topics across Lattice.
This paper introduces Physics-Informed Environmental Retrieval (PIER), a model-agnostic framework that enhances embedding-based retrieval by incorporating a physics-aware stream to ensure consistency with underlying physical processes. By leveraging local verifiers trained on physics-derived flux features, PIER adapts retrieval weights based on scenario-specific diagnostic features, allowing for a more reliable integration of environmental data. Experiments across 356 lakes over 41 years demonstrate that PIER significantly improves predictions for water temperature and dissolved oxygen compared to traditional methods.
PIER outperforms standard retrieval methods by ensuring that environmental predictions are consistent with physical dynamics, leading to more accurate modeling of complex systems.
Accurate modeling of environmental systems is fundamental to scientific understanding and decision-making, yet remains challenging because observations are limited and physical dynamics vary across systems. Retrieval-augmented approaches offer a natural path to transfer knowledge across systems, but standard embedding-based retrieval does not guarantee consistency of underlying physical processes, since scenarios with similar embeddings may arise from different underlying mechanisms. We propose Physics-Informed Environmental Retrieval (PIER), a model-agnostic framework that augments embedding-based retrieval with a physics-aware stream that scores candidates by flux-response consistency with the target, using local verifiers trained on physics-derived flux features. A weight adjustment mechanism then learns per-scenario weights that adaptively balance the two retrieval streams based on diagnostic features summarizing physics-stream reliability. Experiments on 356 lakes across the Midwestern United States spanning 41 years show that PIER consistently outperforms baselines for water temperature and dissolved oxygen prediction, and serves as a general augmentation strategy across diverse backbones.
