Monitoring of Total Dissolved Solids Using Remote Sensing Band Reflectance and Salinity Indices: A Case Study of the Imperial County Section, AZ-CA, of the Colorado River
Godson Ebenezer Adjovu, Haroon Stephen, Sajjad Ahmad
Abstract
Water quality-related issues are becoming a global concern due to the fast pace at which water bodies are deteriorating. The deterioration continues to worsen due to climate change, urbanization, and human activities. Several parameters affect the quality of water. One of such parameters is total dissolved solids (TDS) also known as salinity. It is estimated that salinity in the Colorado River causes close to $400 million in economic damages to the United States every year. This work aims to employ remote sensing techniques, which unlike in situ measurements have been proven to be economical and effective means of water quality monitoring. The overarching goal of the study is to detect salinity in the Colorado River using Landsat 8 Operational Land Imager (OLI) satellite data. The objective of this research was therefore to develop a mathematical relation between remote sensing spectral indices and TDS for the prediction of TDS concentration at the Imperial County Section, AZ-CA, of the Colorado River. Spectral indices were extracted from the satellite data using different 30 m resolution reflectance combinations within the ArcGIS Pro software. The study developed univariate regression for the spectral reflectance and indices and ground truth data of TDS. Both data sets were acquired from the United States Geological Survey (USGS). Results produced R2 of 0.62 to 0.74 for different bands and indices. It was established through sensitivity analysis that one salinity index offers the best model for salinity estimations for the study area with an absolute percentage error of 1.76%. Results offer a possibility of using Landsat 8 OLI for the estimation of TDS in the Colorado River Basin (CRB).