Study region: Four small reservoirs in northwestern Iran varying in size and shape-Taham, Kinevars, Boeen, and Chargar-serve as major drinking and irrigation water sources for semi-arid Zanjan Province. Study focus: Because in situ monitoring of small reservoirs is often irregular or incomplete, satellite observations such as SWOT offer an alternative, but require validation. We generate highresolution Digital Elevation Models (DEMs) from 1965 aerial stereo photographs using Structurefrom-Motion photogrammetry to reconstruct pre-reservoir topography. This enables regional validation of SWOT-derived surface water extents and heights using DEMs and, where available, in situ observations. New hydrological insights: SWOT Pixel Cloud (PIXC) and Lake Surface Product (LakeSP) data (July 2023-October 2025) are used to evaluate PIXC classification. Six classification variants are generated by combining different PIXC surface-type classes and are evaluated against the photogrammetric DEM. The optimal variant, combining open water and water-near-land classes, yields Jaccard indices of 0.61-0.87. For this variant, mean area RMSE is 0.15 km2 and mean height RMSE is 2.16 m. Reservoir-specific area errors range from 5 % to 30 %, while height RMSE ranges from 1.57 to 3.17 m, slightly outperforming LakeSP. Our results further indicate that seasonal snow cover can degrade SWOT-based surface classification by reducing coherence and increasing land-water misclassification. Overall, results show that SWOT effectively captures surface water dynamics in small reservoirs, supporting monitoring in data-scarce semi-arid regions.