Trend analysis on water quality index using the least squares regression models

Abstract

River water pollution requires continuous water quality monitoring that promotes the improvement of water resources. Therefore, the trend analysis on water quality data using mathematical model is an important task to determine whether the measured data increase or decrease during the time period. This paper is intended to highlight the applicability of the least squares regression models to fit the WQI data of the Skudai River, Tebrau River and Segget River located in Johor, Malaysia. As per the 12 years of trend analysis, the data of WQI are collected from the Environmental Quality Reports 2009-2020. The least squares method is utilized to estimate the unknown constants of the linear, quadratic, cubic, polynomial of degree four and degree five regression models. The advantage of using proposed models is that it can be implemented easily even on relatively low computational power systems. The results show that the higher degree polynomial model fits the data reasonably well, in which the polynomials of degree 4 and 5 have lowest average error. Assessment of actual and predictable values of WQI shows that the trends in WQI for all study areas are downward year after year.