An innovative method for identifying aquatic environmental characteristics: construction of water quality fingerprint spectrum and combination with two-dimensional correlation spectroscopy analysis

Abstract

With the deterioration of water quality and explosion of data scale, the diversification and complication of water quality analysis tasks are increasing, which require advanced analysis and visualization methods. Current multivariate statistical methods face challenges in processing multi-dimensional data and conducting causal analysis of fluctuations in water quality over temporal or spatial variations. In this study, a water quality fingerprint spectrum was constructed to perform multiscale and large-volume dataset. Based on the water quality fingerprint spectrum, a workflow of water quality was proposed for exploring the priority control regions, periods, and parameters were identified, and the key factors and effect pathways of response mechanisms, which consisted of three modules: data preparation, data interpretation, and causal analysis. In this study, we constructed a water quality fingerprint spectrum based on 792 water quality data from Lake Shahu as an example to present the methodology. The results showed that COD_Cr, TN, and TP were identified as priority control parameters of Lake Shahu, which were mainly affected by tourism activities and water replenishment. According to the results to conduct causal analysis, we found that the water replenishment could directly improve water quality (total effect: −0.464) by inhibition of algal growth and dilution of COD_Cr and TP concentrations, while tourism activities do the opposite (total effect: 0.562). The innovative analysis method conducted a thorough examination of fluctuations in water quality over spatio-temporal variations and paved the way between data and practical requirements in future works of water quality analysis and management.