Balcerowska-Czerniak Grażyna, Westad Frank
Institute of Mathematics and Physics, Bydgoszcz University, Poland.
Department of Engineering Cybernetics, Norwegian University of Science and Technology, Norway.
J Environ Manage. 2025 Aug;389:126104. doi: 10.1016/j.jenvman.2025.126104. Epub 2025 Jun 9.
The new possibility of using statistical water quality index for characterization of the harmful algal bloom's distribution is presented. A model is formulated based on the mathematical formula that integrates two statistical measures of abnormal behavior in a process or system, Hotelling T and Q residuals, as additional model outputs for the basic Principal Component Analysis (PCA). In this paper, initially available physico-chemical water quality parameters were transformed into a smaller set to build the model. The final quantity from a PCA model, the TQ_PCA index, has shown to be good measure for the growth of the toxic haptophyte Prymnesium Parvum present in the Oder River/Poland in 2024. Here, we also study how multifactorial short-term changes of the water quality are summarized in a few parameters, including temperature, dissolved oxygen, pH, TN/TP ratio and a proposed salinity index, and how they can be viewed in simple comparative plots. We make a quantitative assessment and show that the effect of pH has become one of the key factors. It has been suggested that the specific range of the salinity index favors the golden algae growth in both low and high salinity conditions. However, the most important result of this study is the demonstration of the TQ_PCA index's potential as a new indicator, to detect changes in the aquatic environment prior to the harmful algal blooms occurrence. The objective of the proposed on-line monitoring method is to help authorities to improve seasonal prediction of harmful algal blooms (HABs). This study could also lead to greater understanding of water systems in response to other environmental drivers.
本文提出了利用统计水质指数来表征有害藻华分布的新可能性。基于一个数学公式构建了一个模型，该公式整合了过程或系统中异常行为的两种统计量度——霍特林T统计量和Q残差，作为基本主成分分析（PCA）的额外模型输出。在本文中，最初可用的物理化学水质参数被转换为一个较小的集合以构建模型。PCA模型的最终量，即TQ_PCA指数，已被证明是2024年波兰奥得河中存在的有毒定鞭藻微小原甲藻生长的良好量度。在这里，我们还研究了水质的多因素短期变化如何在几个参数中得到概括，包括温度、溶解氧、pH值、总氮/总磷比和一个提议的盐度指数，以及如何在简单的比较图中查看这些参数。我们进行了定量评估，并表明pH值的影响已成为关键因素之一。有人提出，盐度指数的特定范围有利于金藻在低盐度和高盐度条件下生长。然而，这项研究最重要的结果是证明了TQ_PCA指数作为一种新指标的潜力，能够在有害藻华发生之前检测水生环境的变化。所提出的在线监测方法的目的是帮助当局改进对有害藻华（HABs）的季节性预测。这项研究还可能有助于更深入地理解水系统对其他环境驱动因素的响应。
