Can statistical remote sensing aid in predicting the potential productivity of inland lakes? Case study: Lake Qaroun, Egypt

Abstract

AbstractThis study aimed at creating a sustainable and inexpensive Landsat-based electrical conductivity model that can easily notify fisheries managers of changes in electrical conductivity and hence the potential fish yield of Lake Qaroun in Egypt. The study integrated geospatial technology, field measurements, mathematical computations, and fish yield empirical model into the adopted methodology. Seventeen sampling sites covering the entire study area were selected to measure the electrical conductivity (EC; mS/cm) and water depths (D; m) of Lake Qaroun, Egypt, during November 2018. Spatial analysis tools within ArcGIS were used to extract EC data from non-surveyed sites. A high-resolution Sentinel-2B MSI and a cloud-free medium-resolution Landsat-8 OLI scenes for Lake Qaroun were used for morphometric and regression analyses, respectively. For regression, 75% of the dataset was used to build up the regression model, while the remaining 25% was used for validation. The study selected Landsat band ratios that correlated with the highest certainty (R > 0.80) with the examined EC. Stepwise regression model was then developed to predict EC from Landsat-8 data. In choosing the best regression model, the study selected the significant model (P < 0.05) with the highest coefficient of determination (R2) and the least error metrics. Finally, the developed model was applied in calculating the potential yield of Lake Qaroun. The innovative EC model derived in the current study using Landsat-8 OLI for Lake Qaroun showed a very good performance in estimating 95% of EC values significantly with high acceptable accuracy. In closure, the model can be used very efficiently as a decision support tool in assisting managers not only in monitoring the lake’s electrical conductivity regularly, during the month of November, but also in making preliminary estimates of the lake’s potential yield.