Surface water bodies mapping in Zimbabwe using landsat 8 OLI multispectral imagery: AA comparison of multiple water indices

Publication date: Available online 8 May 2018 Source:Physics and Chemistry of the Earth, Parts A/B/C Author(s): Mhosisi Masocha, Timothy Dube, Millicent Makore, Munyaradzi D. Shekede, Jacob Funani Water is integral for the survival of humans, animals, as well as vegetation and information on its spatial distribution is important for water resources management. Various methods have been applied to detect and map land surface water, especially due to the availability of a large range of satellite sensors. The main methods involved are either general feature classification techniques or thematic water body information extraction. Thematic water body extraction methods make use of band information and in this study; spectral indices are used to delineate land surface water for Zimbabwe using Landsat 8 OLI images. Multiple algorithms i.e. spectral indices were used to extract surface water bodies from Landsat 8 imagery. Then overall accuracies and Kappa coefficients and average reliability were then applied to assess the performance of the seven applied models. The results have shown that the LSWI plus VI can map surface water bodies with an overall accuracy of 86%, kappa statistic of 0.70 and a reliability accuracy of 89%. Comparatively, the index produced good results followed by the mNDWI and AWEIsh which had an overall classification accuracy of ±83% and an average reliability of 87%. Following on the list in terms of performance was the mNDWI, mNDWI, LSWI plus VI, and...
Source: Physics and Chemistry of the Earth, Parts ABC - Category: Science Source Type: research