Evaluation of the Threshold for an Improved Surface Water Extraction Index Using Optical Remote Sensing Data

Abstract

In this study, we proposed an automatic water extraction index (AWEI) threshold improvement model that can be used to detect lake surface water based on optical remote sensing data. An annual Landsat 8 mosaic was created using the Google Earth Engine (GEE) platform to obtain cloud-free satellite image data. The challenge of this study was to determine the threshold value, which is essential to show the boundary between water and nonwater. The AWEI was selected for the study to address this challenge. The AWEI approach was developed by adding a threshold water value based on the split-based approach (SBA) calculation analysis for Landsat 8 satellite images. The SBA was used to determine local threshold variations in data scenes that were used to classify water and nonwater. The class threshold between water and nonwater in each selected subscene image can be determined based on the calculation of class intervals generated by geostatistical analysis, initially referred to as smart quantiles. It was used to determine the class separation between water and nonwater in the resulting subscene images. The objectives of this study were (a) to increase the accuracy of automatic lake surface water detection by improvising the determination of threshold values based on analysis and calculations using the SBA and (b) to conduct a test case study of AWEI threshold improvement on several lakes’ surface water, which has a variety of different or heterogeneous characteristics. The results show that the threshold value obtained based on the smart quantile calculation from the natural break approach (AWEI ≥ −0.23) gave an overall accuracy of close to 100%. Those results were better than the normal threshold (AWEI ≥ 0.00), with an overall accuracy of 98%. It shows that there has been an increase of 2% in the accuracy based on the confusion matrix calculation. In addition to that, the results obtained when classifying water and nonwater classes for the different national priority lakes in Indonesia vary in overall accuracy from 94% to 100%.

Anzali Wetland Crisis: Unraveling the Decline of Iran's Ecological Gem

“…Consequently, the automatic water detection index (AWEI shadow ) (Feyisa et al., 2014) was employed for delineating multi‐temporal water pixels from other forms of LUC in the study area. This method is robust for water detection in complex aquatic ecosystems such as wetlands, addresses challenges posed by the presence of shadow and low albedo, which may hinder other methods such as NDWI and MNDWI (Doña et al., 2021; Laonamsai et al., 2023; Nguyen et al., 2019; Yulianto et al., 2022; Zhai et al., 2015). To enhance the robustness of the derived AWSA, Otsu dynamic thresholding (Otsu, 1979) coupled with a Canny edge filter was applied, overcoming the challenges of dynamic thresholding and constraining the number of input pixels for those located near water bodies (Donchyts et al., 2016).…”

“…Consequently, the automatic water detection index (AWEI shadow ) (Feyisa et al, 2014) was employed for delineating multi-temporal water pixels from other forms of LUC in the study area. This method is robust for water detection in complex aquatic ecosystems such as wetlands, addresses challenges posed by the presence of shadow and low albedo, which may hinder other methods such as NDWI and MNDWI (Doña et al, 2021;Laonamsai et al, 2023;Nguyen et al, 2019;Yulianto et al, 2022;Zhai et al, 2015).…”

Zagros Grass Index—A New Vegetation Index to Enhance Fire Fuel Mapping: A Case Study in the Zagros Mountains

“…Figure 4 illustrates the variatio NDVI changes between areas above 1000 m AMSL and those below 1000 m AMSL. Consequently, the elevation affects tree and grass species' SOS, MAX, and EOS [63,64]. Accordingly, the study area was classified into two major classes: (i) the area over 1000 m AMSL and (ii) the area under 1000 m AMSL.…”

“…The SRTM DEM was additionally employed to delineate the study area based on elevation, as it significantly influences the phenology of both tree and grass species. Consequently, the elevation affects tree and grass species' SOS, MAX, and EOS [63,64]. Accordingly, the study area was classified into two major classes: i) the area over 1000 m AMSL and ii) the area under 1000 m AMSL.…”