Morphological Change and Quantification of Uncertainty in Remotely-Sensed Measurement of Indus River

The increasing accessibility and resolution of remote sensing (RS) images of the Earth’s images has significantly improved the accuracy and spatio-temporal range with which we can analyse the channel movement and width variations. We summarized and evaluated current best practices, tested novel techniques to deal with the error and uncertainty, and present recommendations for futuristic study utilizing RS observations of Indus River migration and width variations in this paper. This study also deals with the morphological changes in the Indus River. It was observed that mean root mean square error - RMSE (Landsat 8) value for difference between actual centerline and remote sensing delineated centerline across all years was found to be 0.127 km in the study reach. The error found in Landsat 7 delineated centerline was more as compared to Landsat 8 and Landsat 5, due to the scan line error. In addition, Remote sensing delineated river widths were compared with the actual cross-section surveyed on or near to image acquisition date. The coefficient of correlation between actual and remote sensing delineated river widths comes out (R2 = 0.8665), and RMSE was found to be 115.4 m. Lastly, we advocate concentration on enhancing the simplicity, generalization, and open-source prospects of programs and tools used to calculate the channel planform variations and spatially varying error and uncertainty, resulting in single platform for measuring and comparing outcomes.