Statistical and Visual Comparison of Water Quality Changes Caused by a Large River Restoration Project

Abstract

This article presents a statistical and visual comparison of water quality changes caused by a large river restoration project. Since water quality data are often shown as a non-normal distribution with high seasonal variations, appropriate statistical methods should be selected according to data characteristics for accurate scientific decision-making. In this study, a normality test was first performed using the Shapiro-Wilk test and two statistical comparison tests were then performed, including the paired T-test and the sign-test. Seasonality was considered by comparing monthly data pairs. In addition, a diagonal pair comparison plot was proposed as a visual comparison method. This plot is a graphic data display, where monthly paired water quality is represented by X-Y coordinates. From this study, it was concluded that the series of statistical and visual methods would be suggested for comparison of non-normally distributed water quality with high seasonal variations.