ISPRS-Archives The International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences ISPRS-Archives Int. Arch. Photogramm. Remote Sens. Spatial Inf. Sci. 2194-9034 Copernicus Publications Göttingen, Germany 10.5194/isprs-archives-XLII-4-W18-1097-2019 A NOVEL WATER INDEX (SWI) FOR SALTY WATER FROM LANDSAT 8 OLI/TIRS Yousefian F. 1 Sahebi M. 2 Shokri M. 1 Moradi M. 1 K. N. Toosi University of Technology, Mirdamad, Tehran, Iran Remote Sensing Research Center, K. N. Toosi University of Technology, Mirdamad, Tehran, Iran 19 10 2019 XLII-4/W18 1097 1105 Copyright: © 2019 F. Yousefian et al. 2019 This work is licensed under the Creative Commons Attribution 4.0 International License. To view a copy of this licence, visit https://creativecommons.org/licenses/by/4.0/ This article is available from https://isprs-archives.copernicus.org/articles/XLII-4-W18/1097/2019/isprs-archives-XLII-4-W18-1097-2019.html The full text article is available as a PDF file from https://isprs-archives.copernicus.org/articles/XLII-4-W18/1097/2019/isprs-archives-XLII-4-W18-1097-2019.pdf

Monitoring natural resources is one of the most important tasks in earth observation and remote sensing satellites. Water resources play a crucial role in the life of human on the planet. Among the water resources, salty lakes are of particular importance in biological, physical and environmental issues. In this study, a new Salty Water Index (SWI) for Landsat 8 Operational Land Imager (OLI) images is proposed based on salty lakes by particle swarm optimization (PSO), where water doesn’t combine by cloud, shadow, and salty areas. SWI is implemented on four famous and important salty lakes with the proper distribution of the whole world and different Salinity, including Lake Assal, Great Salt Lake, Eyre Lake, and Lake Urmia. The performance of SWI is compared with other water indices by overall accuracy, f-score, kappa coefficient, and standard deviation to mean ratio. Results show the efficiency of SWI on all cases due to 0.0055 Standard deviation to mean for SWI compared to 0.0395, 0.0255, 0.0873, 0.0214, 0.0524, 0.0408 and 0.0375 for NDWI, MNDWI, AWEIsh, AWEI, WRI, MOWI, and MBWI, respectively. Also, Effectiveness criteria (E) determines the efficiency of each band of Landsat 8. In this regard, results show the high performance of Green and Near IR band in all conditions and relatively proper performance of some other bands based on a special condition of each case study. The proposed method is also suggested to readers to obtain novel spectral indices of other classes and other sensors.