An integrated framework for actionable flood warnings using high resolution satellite imagery

Zhouyayan Li; Bekir Zahit Demiray; Marian Muste; Ibrahim Demir

doi:10.2166/hydro.2026.141

Back

An integrated framework for actionable flood warnings using high resolution satellite imagery

Journal article

Open access

Peer reviewed

An integrated framework for actionable flood warnings using high resolution satellite imagery

Zhouyayan Li, Bekir Zahit Demiray, Marian Muste and Ibrahim Demir

Journal of hydroinformatics, Vol.28(2), pp.196-210

02/01/2026

DOI: 10.2166/hydro.2026.141

Files and links (1)

url

https://doi.org/10.2166/hydro.2026.141View

Published (Version of record) Open Access

Abstract

Floods rank among the most devastating natural hazards globally, often affecting densely populated regions and causing significant social, economic, and environmental losses. Traditional flood models, while valuable, are limited by simplifying assumptions, numerical approximations, and insufficient data for accurate simulations. Recent advancements in data-efficient digital elevation–terrain model-based models offer improvements but remain sensitive to dynamic surface conditions. This study explores the use of high-resolution commercial PlanetScope imagery for flood inundation mapping in data-limited settings. To address early-stage reflectance issues stemming from the lack of onboard calibration in PlanetScope constellations, we introduce post-processing workflow called Quantile-based Filling and Refining. Initial flood maps generated using widely adopted normalized difference water index (NDWI) performed poorly, with F1 scores of 0.21 and 0.24 across the two study sites, and near-infrared (NIR)-based maps showed only moderate improvement (F1 up to 0.50 and 0.82). However, applying QFR to the NIR-based maps resulted in substantial gains, increasing F1 scores to 0.84 and 0.88, exceeding the accuracy of both NDWI- and NIR-derived maps as well as manual delineations. This study highlights the potential of commercial remote sensing imagery for accurate, small-scale flood mapping and demonstrates the effectiveness of QFR as a scalable method for enhancing flood prediction in challenging data environments.

decision-making

flood forecasting

flood inundation maps

image processing

map extraction

satellite imagery

Details

Title: Subtitle: An integrated framework for actionable flood warnings using high resolution satellite imagery
Creators: Zhouyayan Li - University of Iowa
Bekir Zahit Demiray - University of Iowa
Marian Muste - University of Iowa
Ibrahim Demir - Tulane University
Resource Type: Journal article
Publication Details: Journal of hydroinformatics, Vol.28(2), pp.196-210
DOI: 10.2166/hydro.2026.141
ISSN: 1464-7141
eISSN: 1465-1734
Publisher: IWA Publishing
Grant note: Iowa Department of Transportation: SPR-RE22(016)-8H-00
The study presented in this report was supported by the Iowa Department of Transportation-SPR-RE22(016)-8H-00.
Language: English
Date published: 02/01/2026
Academic Unit: Electrical and Computer Engineering; Civil and Environmental Engineering; IIHR--Hydroscience and Engineering; Injury Prevention Research Center; Geographical and Sustainability Sciences; Mechanical Engineering
Record Identifier: 9985139489402771

Metrics

1 Record Views