S&M1869 Research Paper of Special Issue
Published: May 16, 2019
Application of Water Indices in Surface Water Change Detection Using Landsat Imagery in Nepal [PDF]
Tri Dev Acharya, Anoj Subedi, He Huang, and Dong Ha Lee
(Received April 16, 2018; Accepted March 13, 2019)
Keywords: Landsat, water index, NDVI, NDWI, MNDWI, WRI, SWI, change detection, Sheyphoksundo, Phewa, Koshi
Surface water change is a very important indicator for environmental, climatic, and anthropogenic activities. Remotes sensors, such as Landsat, have been providing data since the last four decades, which are useful for extracting land cover types such as forest and water. Researchers have proposed many surface water extraction techniques, among which index-based methods are popular owing to their simplicity and cost effectiveness. By using the standard and Otsu threshold methods, water features can be mapped, thus changes can be detected. On the basis of the results of this study, the following water indices were applied, i.e., normalized difference vegetation index (NDVI), normalized difference water index (NDWI), modified normalized difference water index (MNDWI), water ratio index (WRI), and simple water index (SWI) for surface water extraction, and their changes using Landsat images. Three unique test sites from Nepal, which represent overall types of water features found in the country, were selected for this study. A model was developed in ArcGIS to differentiate the water features based on index methods and changes were calculated in terms of positive and negative. The results show that most of the case water index methods based on standard and Otsu thresholds cannot accurately separate water from its nearby backgrounds, such as melting ice, shadows of hills with or without vegetation, grasslands, and wet barren sand. From the current cases, it is not recommended to use the water indices in Nepal without any appropriate expert opinion or sitewise calibrated thresholding for automated water detection. However, they can be used for the overall change detection. Most of the change maps of the selected lakes showed good accuracies in some unique cases, such as the dark shadow and forest near a water body. SWI is good for stagnant lakes but not for water bodies with shifting features, such as the Koshi River. The model can be very useful in quickly understanding the changes in water bodies and taking the necessary measures to planners, but not much in accurate mapping. Furthermore, studies of different seasons, sensor data, and sites are necessary in the use of standard water indices for accurate change detection.
Corresponding author: Dong Ha Lee
This work is licensed under a Creative Commons Attribution 4.0 International License.
Cite this article
Tri Dev Acharya, Anoj Subedi, He Huang, and Dong Ha Lee, Application of Water Indices in Surface Water Change Detection Using Landsat Imagery in Nepal, Sens. Mater., Vol. 31, No. 5, 2019, p. 1429-1447.