Structure-Function Coherent Coarsening for Cross-Resolution Ecohydrological Modeling

TL;DR

This paper introduces a novel coarsening framework that preserves structural and functional details in ecohydrological models, improving cross-scale simulation accuracy and stability.

Contribution

The study develops the SFCC framework that maintains hydrological connectivity and functional heterogeneity during data coarsening, enhancing multi-scale ecohydrological modeling.

Findings

Hydro-aware coarsening preserves watershed morphology and improves runoff predictions.

Function-preserving methods reduce bias in categorical input aggregation.

Simulations show variables converge to steady states over time.

Abstract

Ecohydrological models are increasingly applied across multiple scenarios, yet their application remains constrained by high computational costs of fine-resolution simulations and structural inconsistencies in cross-scale modeling. This study develops a Structure-Function Coherent Coarsening (SFCC) framework that preserves both hydrological connectivity and functional heterogeneity during model input coarsening. We apply the VELMA model to 24 subbasins in the Salish Sea Basin, U.S. and examine three types of inputs: (i) DEM coarsened with a Hydro-aware approach that preserves drainage topology; (ii) land-use and soil-type datasets coarsened with function-preserving methods (Auto-weight and Auto-reassign) that retain small but process-dominant classes; and (iii) initial conditions coarsened with hydrology-, land-cover-, and soil-aware strategies to enhance temporal stability. Results show that the Hydro-aware method effectively preserves watershed morphology and yields more consistent runoff and nitrate predictions than mean-based coarsening across scales. For categorical inputs, the function-preserving methods alleviate the dominant-class bias of majority aggregation, particularly in basins where small high-impact patches drive nitrogen export. Long-term simulations further show that although hydrological variables equilibrate rapidly and biogeochemical processes adjust more gradually, deviations in both decrease over time and converge toward a steady state. These demonstrate that structural consistency and functional preservation together maintain dynamic stability through spatiotemporal feedback. Compared with existing work, the proposed SFCC framework operates directly at the data-input level, enabling more coherent integration of multi-source datasets and maximizing the retention of high-resolution information.