Akkimi et al. (2026) FlDepth: A New Method for Estimating Fluvial and Pluvial Flood Depths from Near Real-Time Satellite-Derived Inundation Map and Topography

Identification

• Journal: Water Resources Management
• Year: 2026
• Date: 2026-02-16
• Authors: Anjaneyulu Akkimi, Srinivasulu Jella, Radha Manasa Peesapati, Amanpreet Singh, Durga Rao Khv, Prakash Chauhan
• DOI: 10.1007/s11269-025-04405-1

Research Groups

National Remote Sensing Centre (NRSC), Indian Space Research Organisation (ISRO), Hyderabad, Telangana, India

Short Summary

This study introduces FlDepth, a novel GIS-based method for estimating fluvial and pluvial flood depths using near real-time satellite inundation maps and topographic data. The method demonstrated high accuracy (errors < 20 cm, NSE ~ 1) compared to hydrodynamic models and ICESat-2 data, outperforming existing tools like FwDET in diverse terrains.

Objective

• To propose and evaluate a novel GIS-based methodological approach (FlDepth) for estimating floodwater depths by leveraging near real-time flood extents derived from multi-sensor satellites and topographic data.

Study Configuration

• Spatial Scale: Regional to basin scale, covering specific river stretches (e.g., Godavari–Sabari section, 200 km length) and entire states (Assam: 8.95 × 10^3 km^2; Bihar: 9.82 × 10^3 km^2). Flood depth estimations were performed at 10 meter and 30 meter spatial resolutions.
• Temporal Scale: Near real-time flood events, including the 2024 monsoon season (June–October) in Assam and Bihar, and specific flood events on 17 August 2020 (Godavari) and 04 July 2024 (Brahmaputra), and 13 August 2024 (Ganga–Kosi plains).

Methodology and Data

• Models used:
  • FlDepth: Proposed GIS-based method for flood depth estimation, relying on flood extent boundaries and terrain elevations to derive water surface elevation (WSE) through cross-section interpolation.
  • MIKE 21FM: 2D hydrodynamic model used as a benchmark for evaluating FlDepth.
  • FwDET (Floodwater Depth Estimation Tool): Widely used tool for comparison with FlDepth.
• Data sources:
  • Satellite-derived flood inundation maps: Near real-time extents from multi-sensor satellites (RISAT-1 A SAR, Sentinel-1 A SAR, Resourcesat-2 A, Landsat-8, Landsat-9, EOS4 RISAT-1 A MRS SAR).
  • Topographic data: ALTM DTM (1 meter resolution LiDAR-based Digital Terrain Model) for Godavari–Sabari, and FABDEM (Forest and Buildings removed DEM, 30 meter resolution) for Assam and Bihar.
  • Satellite altimetric data: ICESat-2 ATL13 (inland water surface heights) for validation of water levels.
  • Land Use/Land Cover (LULC) data: Used for assigning bed resistance in the MIKE 21FM model.

Main Results

• FlDepth demonstrated strong alignment with hydrodynamic model simulations, achieving an accuracy greater than 87%, errors less than 20 centimeters, and a Nash-Sutcliffe Efficiency (NSE) of approximately 1.
• FlDepth consistently outperformed FwDET, producing smoother, more continuous, and realistic flood depth patterns without the abrupt discontinuities, triangulations, or missing values often seen in FwDET results.
• Water levels estimated by FlDepth and the hydrodynamic model showed near-perfect correlations (R² ≈ 0.99) with ICESat-2 ATL13 inland water surface heights in the Godavari–Sabari region, significantly stronger than FwDET (R² ≈ 0.85).
• Operational applicability was successfully demonstrated in flood-prone Assam and Bihar during the 2024 monsoon, highlighting the method's resilience and predictive accuracy for both fluvial and pluvial floods over large and complex terrains.
• A consistent offset of approximately 2.5 meters was observed, with model-estimated water levels systematically overestimating ICESat-2 values, indicating potential discrepancies in the underlying DTM.

Contributions

• Introduces FlDepth, a novel, automated, and computationally efficient GIS-based methodology for near real-time flood depth estimation from satellite-derived inundation maps and topographic data.
• Provides a practical and scalable solution for flood depth mapping that can be applied across diverse flood types (fluvial, pluvial) and spatial scales, with minimal input requirements.
• Significantly improves upon existing DEM-based methods like FwDET by generating smoother, more hydrologically consistent, and realistic flood depth fields, particularly in large, low-slope regions.
• Validates the method using a dual framework of hydrodynamic model benchmarks and independent satellite altimetric observations (ICESat-2 ATL13), enhancing reliability.
• Highlights the capability of ICESat-2 ATL13 data for calibrating and validating hydrological and hydrodynamic models, contributing to improved model performance.
• Addresses a critical gap in flood risk management by enabling rapid translation of satellite flood extents into actionable depth estimates for emergency response, damage assessment, and policy-making in data-scarce environments.

Funding

The authors declare that no funds, grants, or other support were received during the preparation of this manuscript.

Citation

@article{Akkimi2026FlDepth,
  author = {Akkimi, Anjaneyulu and Jella, Srinivasulu and Peesapati, Radha Manasa and Singh, Amanpreet and Khv, Durga Rao and Chauhan, Prakash},
  title = {FlDepth: A New Method for Estimating Fluvial and Pluvial Flood Depths from Near Real-Time Satellite-Derived Inundation Map and Topography},
  journal = {Water Resources Management},
  year = {2026},
  doi = {10.1007/s11269-025-04405-1},
  url = {https://doi.org/10.1007/s11269-025-04405-1}
}