Observation-Guided Neural Surrogate Learning for Scientific Simulation Emulation: A Single-Gauge Flood-Inundation Proof of Concept

TL;DR

This paper introduces an observation-guided neural surrogate model for urban flood simulation, combining hydrodynamic models with real gauge data to achieve high accuracy in emulating flood behavior.

Contribution

The novel framework integrates gauge observations with neural surrogates and uncertainty estimation for improved flood simulation emulation.

Findings

Emulator achieves R^2 approximately 0.99 outside the gauge pixel.

Mean absolute error below 0.01 meters in flood-depth predictions.

Strong pointwise consistency with gauge-derived local depth targets.

Abstract

We present an observation-guided neural surrogate-learning framework for scientific simulation emulation, demonstrated on urban flood-inundation mapping. The framework combines LISFLOOD-FP hydrodynamic simulations with a real Gauge L stage record that is mapped to the simulation grid and converted to a datum-consistent local water-depth target before being used as single-site supervision. Focusing on a 256 x 256 crop around Gauge L in the Chicago metropolitan area, the method first constructs an ensemble-approximated Gaussian-process/local analogue surrogate (EnsCGP) to obtain a coarse flood-depth estimate and an uncertainty proxy. A U-Net-ASPP neural corrector then refines the coarse map using only simulation-derived and geospatial inputs: EnsCGP depth, the uncertainty proxy, rainfall, and spatial coordinates. The converted gauge-derived local depth is used only as a pointwise training target at the mapped gauge pixel; simulation-based losses are evaluated away from that pixel. Across temporally held-out events from 2013-2019, the emulator closely reproduces LISFLOOD-FP simulation targets outside the gauge-constrained pixel, with R^2 approximately 0.99 and mean absolute error below 0.01 m, and shows strong pointwise consistency with the converted Gauge L local depth target under the stated rolling-year protocol. We interpret these results as strong simulator-emulation agreement with pointwise observation-guided correction, not as independent validation of real-world inundation accuracy or as a complete operational flood-forecasting system.