A denoising–classification–retracking method to improve spaceborne estimates of the water level–surface–volume relation over the Urmia Lake in Iran

Abstract

ABSTRACT Decreasing the volume of the Urmia Lake, as the largest inland water body in Iran, is one of the current environmental and water resource management concerns. This study obtains a reliable spaceborne water level (WL)–area–volume relationship for the Urmia Lake using terrestrial, aerial and satellite-based data. The aim of this study is to improve Urmia Lake’s WL derived from satellite altimetry and, consequently, to more accurately estimate the volume of the lake for the last decade. To this end, improved WL is obtained from the Satellite with Argos and Altika (SARAL/AltiKa) and Jason-2 altimetry missions by performing a post-processing method. The post-processing method includes a denoising, a classification and appropriate retracking algorithms. The results are validated against in situ gauge data and also compared with results from Prototype Innovant de Système de Traitement pour les Applications Côtières et l’Hydrologie (PISTACH) and Prototype on AltiKa for Coastal, Hydrology and Ice (PEACHI) products. The Denoising–Classification–Retracking (DCR) method improves the root mean square error (RMSE) of WL with respect to those of PISTACH and PEACHI by 54% and 24%, respectively. The surface area of the lake is determined from Landsat 7 Enhanced Thematic Mapper Plus (ETM+) images based on calculating normalized difference water index (NDWI). The results are validated against the surface area obtained from aerial photogrammetry and Cartosat high resolution image. Moreover, based on bathymetric map a Look-up table including surface area and volume of the lake at specific levels are formed. The obtained surface area is then compared with the values of the Look-up table. The normalized root mean square error between surface extent obtained from proposed method and corresponding values is about 11%. The estimated lake’s volume is compared with the level-volume curve from the bathymetric data. The result showed the RMSE of this comparison is about 0.12 km3. Our validated results show that the lake has lost 75% of its volume from late 2008 to early 2016 but continued with an increase in its volume in May 2017 twice as much as in early 2016. Our results support urgent or long-term restoration plan of Lake Urmia and highlight the important role of spaceborne sensors for hydrological applications.