Use of mathematical methods in the analysis and forecasting of nitrogen-phosphorus cycle indicators in the Pregolya river basin
Abstract
The analysis of multidimensional raw data on the state of aquatic ecosystems often requires the use of modern analytical tools capable of identifying hidden patterns. The purpose of this study is to analyze long-term water quality monitoring data for the Pregolya River (Kaliningrad) in order to identify the main factors determining variability in the hydrochemical regime and to assess their temporal dynamics. To analyze data on indicators of the nitrogen and phosphorus cycles, methods of multivariate statistics were applied, including hierarchical cluster analysis (Ward’s method), factor analysis (minimum residual method with varimax rotation), and time-series analysis (LOESS smoothing). As a result of the cluster analysis, three groups of observations were identified, reflecting different periods of nutrient input. Factor analysis made it possible to identify two latent factors: “nitrogen load from organic sources” and “the ratio of organic and mineral forms of nitrogen.” Time-series analysis of factor scores revealed nonlinear dynamics and long-term trends in changes in water quality. The study demonstrates that the integrated application of chemometric methods serves as an effective tool for diagnosing the condition of urban water bodies, identifying pollution sources, and providing a scientific basis for the development of management decisions in the field of water resource protection.