Masoumi et al. (2026) Absolute validation of SWOT measurements over small reservoirs using legacy photogrammetric DEMs: A case study in Zanjan Province, Iran

Identification

• Journal: Journal of Hydrology Regional Studies
• Year: 2026
• Date: 2026-02-14
• Authors: Zohreh Masoumi, Mohammad J. Tourian
• DOI: 10.1016/j.ejrh.2026.103229

Research Groups

• Department of Earth Science, Institute for Advanced Studies in Basic Science (IASBS), Zanjan, Iran
• Center for Research in Climate Change and Global Warming (CRCC), Institute for Advanced Studies in Basic Sciences (IASBS), Zanjan, Iran
• Institute of Geodesy, University of Stuttgart, Stuttgart, Germany

Short Summary

This study validates SWOT satellite measurements of surface water extent and height over four small reservoirs in Zanjan Province, Iran, using high-resolution Digital Elevation Models (DEMs) derived from 1965 aerial photographs. It finds that an optimal PIXC classification variant effectively captures surface water dynamics, achieving a mean area RMSE of 0.15 km² and mean height RMSE of 2.16 m, though performance can degrade due to seasonal snow cover.

Objective

• To quantify the absolute accuracy of SWOT-derived surface water height and area in small reservoirs using pre-construction photogrammetric Digital Elevation Models (DEMs).
• To assess the impact of different surface-type classifications on the accuracy of the SWOT Pixel Cloud (PIXC) product.

Study Configuration

• Spatial Scale: Four small reservoirs (Taham, Kinevars, Boeen, Chargar) in Zanjan Province, northwestern Iran, with surface areas ranging from 0.09 km² to 2.5 km².
• Temporal Scale: SWOT data acquired from July 2023 to October 2025. Legacy aerial photographs from 1960–1965.

Methodology and Data

• Models used:
  • Structure-from-Motion (SfM) photogrammetry (for DEM generation).
  • Scale-Invariant Feature Transform (SIFT) algorithm (for feature identification in SfM).
  • RANdom-Sample Consensus (RANSAC) algorithm (for mitigating false matches in SfM).
  • Semi-Global Matching (SGM) algorithm (for dense point cloud creation in SfM).
  • Convex hull algorithm (for polygon construction from PIXC).
• Data sources:
  • Satellite: SWOT Pixel Cloud (PIXC) and Lake Surface Product (LakeSP) data (L2HRPIXC, L2HRLakeSP).
  • Observation: Legacy aerial stereo photographs (1960–1965, scale 1:20000, 1200 dpi scanning resolution). In-situ water level and limited surface area observations for Taham and Kinevars reservoirs (from Zanjan Regional Water Organization).
  • Ancillary: Shuttle Radar Topography Mission (SRTM) for Ground Control Points (GCPs). Sentinel-2 imagery for contextual analysis of snow cover.

Main Results

• The optimal PIXC classification variant (variant 6, combining open water and water-near-land pixels) consistently showed the strongest agreement with photogrammetric water extents.
• Spatial agreement, quantified by the Jaccard index for variant 6, ranged from 0.61 to 0.87 across the reservoirs.
• For variant 6, the mean area Root Mean Square Error (RMSE) was 0.15 km², with reservoir-specific area errors ranging from 5 % to 30 %.
• For variant 6, the mean height RMSE was 2.16 m, with reservoir-specific height RMSE values between 1.3 m and 3.17 m.
• The combined height-area normalized RMSE for variant 6 was 0.15 km²⋅m.
• The LakeSP product exhibited slightly degraded performance compared to variant 6, with an average area RMSE of 0.20 km² and height RMSE of 4.04 m.
• Seasonal snow cover significantly degrades SWOT PIXC classification performance by reducing coherence and increasing land–water misclassification, leading to lower Jaccard index values.
• Larger, more compact reservoirs (e.g., Taham) generally yielded higher-quality data and better spatial agreement, while smaller or complex-shaped reservoirs (Boeen, Chargar) showed greater variability and lower accuracy.
• Reservoir shape and orientation (e.g., elongated Kinevars) can influence SWOT detectability and data coverage.
• Inconsistencies and gaps were observed in SWOT PIXC and LakeSP data availability across epochs.

Contributions

• First study to simultaneously validate SWOT-derived water surface elevation and area using high-resolution pre-construction Digital Elevation Models (DEMs) derived from legacy aerial photographs.
• Demonstrates a novel approach for regional validation of SWOT in data-scarce semi-arid regions using historical photogrammetric data.
• Provides a systematic evaluation of different PIXC surface-type classification variants, identifying an optimal combination for small reservoirs.
• Quantifies the impact of seasonal snow cover on SWOT-based surface classification, highlighting a key environmental limitation for mountainous and cold-region environments.
• Offers an unbiased assessment of SWOT's native performance by avoiding post-processing or noise reduction techniques.

Funding

Not explicitly provided in the paper.

Citation

@article{Masoumi2026Absolute,
  author = {Masoumi, Zohreh and Tourian, Mohammad J.},
  title = {Absolute validation of SWOT measurements over small reservoirs using legacy photogrammetric DEMs: A case study in Zanjan Province, Iran},
  journal = {Journal of Hydrology Regional Studies},
  year = {2026},
  doi = {10.1016/j.ejrh.2026.103229},
  url = {https://doi.org/10.1016/j.ejrh.2026.103229}
}