Role of statistical remote sensing for Inland water quality parameters prediction

Abstract

Understanding the statistical relations among the Advanced Space borne Thermal Emission and Reflection Radiation (ASTER) data and observed water quality parameters, in order to develop a mathematical relation for the precise prediction of the missing data in a given area, is the main aim of the present study. This should enable to establish a spatial distribution map for each parameter of water quality for the area. The method was applied to Qaroun Lake in the Fayoum depression of Egypt.The water quality parameters obtained from ASTER data used in the present work are: Temperature, Turbidity, Hydrogen ion concentration (pH), Salinity, Total Dissolved Solids (TDS), Electrical Conductivity (EC), Total alkalinity, Total Organic Carbon (TOC) and Ortho-phosphorus.18 water sample data were used in the study: 15 sample data for mathematical model construction, giving the relation between the ASTER values and the water quality parameters, while 3 samples data were used to test the obtained model.The SPSS software of IBM was also used in the present research for the applied statistical analysis.The analysis showed a significant correlation between the observed values and the remotely sensed data with R2 > 0.94 and sig. < 0.01 in most cases. The calculated values resulting through the obtained equation showed a high accuracy: Root mean square error (RMSE) ranging from 0.8 to 0.014 and Standard Estimated Error (SEE) ranging from 0.9 to 0.0116.ERDAS Imagine and ArcGIS packages were used for applying the obtained mathematical model and spatial distribution map to the Qaroun Lake. Keywords: Remote sensing, Regression, Inland water quality, ASTER