Statistical approaches to explore the linkages between physicochemical parameters and BQEs, and set river nutrient threshold concentrations in Hungary

Abstract

The objective of the Water Framework Directive (WFD) is to achieve good ecological status in surface waters by 2027. To make a proper evaluation of the ecological status of watercourses, it is necessary to harmonize class boundaries for chemical and biological quality elements (BQEs). This paper aims to explore the linkages between physicochemical parameters and BQEs and set river nutrient threshold concentrations that support good ecological status. Regression and mismatch methods were applied to find the relationship between phytoplankton (PP) and phytobenthos (PB) ecological quality ratio and mean total phosphorus (TP) and total nitrogen (TN) concentrations. Nutrient thresholds have been suggested for several water types, which are varied in the case of highland rivers 1.8–6.2 mg TN/l, 180–400 μg TP/l; in the case of lowland rivers 1.4–5.0 mg TN/l, and 100–350 μg TP/l. These values are similar to what other studies found, but the relationship between biology and nutrients was weaker. Besides nutrients, additional data of measured dissolved organic carbon, 5-day biochemical oxygen demand, chemical oxygen demand with potassium permanganate method, and information about hydromorphological features were involved in the analysis. The research demonstrates that random forest can be used as a nonlinear, multiparametric model for predicting biological class from five variables with 35–81% error for PP and with 18–47% error for PB. HIGHLIGHTS Statistical analysis demonstrated in the paper will be the first publication of Hungarian ecological monitoring data processing.; The research pointed out some failures in the implementation of the Water Framework Directive, namely that considering every biological quality element throughout criteria development might be a mistake.; The paper demonstrates how random forests can be used for biological classification.;

Keywords

• ecological status
• nutrient thresholds
• random forest
• statistical methods
• water framework directive