نوع مقاله : مقاله پژوهشی
موضوعات
عنوان مقاله English
نویسندگان English
Recent advances in remote sensing have provided effective solutions to address data scarcity. This study estimates river discharge using the remote sensing-based geoBAM (geomorphologically-enhanced variant of BAM) algorithm, applied in both expert and unsupervised classification frameworks. The algorithm follows the McFLI approach and incorporates river geomorphological features. To obtain the initial dataset, river width information was extracted from satellite imagery, while the remaining hydraulic parameters were obtained from the HEC-RAS model. A total of 17 images of Landsat 8 as well as 78 images of Sentinel-2 from the 2017–2018 water year were used to analyze a non-braided and non-meandering section of the Karun River between Mollasani and Ahvaz. Time series validation of the estimated discharge result against observed data showed that Sentinel-2-based discharge estimates outperformed those from Landsat 8 in both classification modes (NSE values of 0.53 vs. 0.20 and 0.74 vs. 0.14 for expert and unsupervised modes, respectively). The improved spatial and temporal resolution of Sentinel-2 led to more accurate discharge estimation. Interestingly, the unsupervised mode yielded better results than the expert mode when using Sentinel-2 data, which may be due to a mismatch between predefined expert priors and the actual hydraulic characteristics of the study area.
کلیدواژهها English
