Comparative Evaluation of Spectral Indices and Sensors for Mapping of Urban Surface Water Bodies in Jodhpur Area : Smart & Sustainable Growth

Abstract

Remote sensing data increasingly used to identify and map open water bodies at comprehensive scales. The Jodhpur city experienced high urbanization in last one decade with population about 13Lakhs, and therefore needs effective public amenities alongwith safe water. In this research, compared accessible surface water bodies mapping approaches using seven spectral indices, viz. normalized difference water index (NDWI), tasseled cap wetness index (TCW), automated water extraction index (AWEIsh and AWEInsh), modified normalized difference water index (MNDWI), Water ratio index (WRI) and normalized difference vegetation index (NDVI) as well as two medium resolution sensors (Sentinel-2A and Landsat 8 OLI). The combinations of different water algorithms and satellite sensors were used to evaluate accuracy of the open water body. The results confirmed that water algorithms have high accuracies with Kappa Coefficients ranging from 0.12 to 0.98. The MNDWI water algorithms performed better than other water indices algorithms, and could be associated with pure water dominance in study area.The resultant water mapping from Sentinel-2A (10m) data has superior accuracies than Sentinel-2 (20m) and Landsat 8 OLI (30m). This research illustrates the enhanced performance in Sentinel-2A (10m) and Sentinel-2A (20m) for mapping of water body. The present study shows the availability of alternate water resources, which shall also be useful during frequent maintenance work of Rajeev Gandhi canal which supply water to whole city. The nearest neighbour technique is used to resample Sentibel-2A (10m) data of the visible and near IR bands to 20 m resolution bands to perform further analysis and comparisons.