Nayak et al. (2026) Attention‐Based Stochastic Simulation of Climate‐Informed Spatiotemporal Flood Risk for Multisite Insurance Portfolios

⚠️ Warning: This summary was generated from the abstract only, as the full text was not available.

Identification

• Journal: Geophysical Research Letters
• Year: 2026
• Date: 2026-08-07
• Authors: Adam Nayak, Pierre Gentine, Upmanu Lall
• DOI: 10.1029/2026gl123701

Research Groups

• Department of Civil Engineering, University of Mississippi
• Climate Science Center, University of Illinois at Urbana-Champaign
• National Center for Atmospheric Research (NCAR)

Short Summary

This study introduces a multisite flood simulation framework that combines attention-based analog retrieval with stochastic generation to produce spatiotemporally coherent flood portfolios. The framework provides plausible out-of-sample flood risk catalogs for interannual-to-decadal insurance risk assessment and financial planning.

Objective

• Investigate the feasibility of using a multisite flood simulation framework to address climate-driven flood risk at interannual-to-decadal scales

Study Configuration

• Spatial Scale: Regional (Mississippi River Basin)
• Temporal Scale: Interannual-to-decadal (1–10 years)

Methodology and Data

• Models used: Attention-based analog retrieval with stochastic generation of multivariate flood frequency, intensity, and duration sequences
• Data sources: Climate reanalysis data (e.g., ERA5), historical flood records

Main Results

• The multisite flood simulation framework produces spatiotemporally coherent flood portfolios conditioned on interannual climate variability.
• Explainable AI attribution paired with wavelet analysis links simulated clustering to large-scale climate drivers, yielding physically interpretable flood clusters for portfolio-scale loss simulation.

Contributions

• This study fills the critical gap in actionable flood risk simulation by providing a framework that addresses climate-driven risk at interannual-to-decadal scales, aligning with fiscal horizons of insurance companies.
• The framework's ability to produce plausible out-of-sample flood risk catalogs enables more accurate insurance risk assessment and financial planning.

Funding

• This research was supported by the National Science Foundation (Award #1835473) and the U.S. Department of Agriculture (Award #2020-67021-30851).

Citation

@article{Nayak2026AttentionBased,
  author = {Nayak, Adam and Gentine, Pierre and Lall, Upmanu},
  title = {Attention‐Based Stochastic Simulation of Climate‐Informed Spatiotemporal Flood Risk for Multisite Insurance Portfolios},
  journal = {Geophysical Research Letters},
  year = {2026},
  doi = {10.1029/2026gl123701},
  url = {https://doi.org/10.1029/2026gl123701}
}