Estimation of Nitrate Trends in the Groundwater of the Zagreb Aquifer

Geosciences

Zoran Kovač¹  Zoran Nakić¹  Andrea Bačani¹  Drago Špoljarić²  David Stanek² 

[1] Faculty of Mining, Geology and Petroleum Engineering, University of Zagreb, 10000 Zagreb, Croatia;Intellomics Ltd., 10000 Zagreb, Croatia;

关键词: linear regression;    LOESS non-parametric regression;    confidence and prediction intervals;    data aggregation;    nitrate contamination;    Zagreb aquifer;    trend’s evaluation at different scales;   

DOI  :  10.3390/geosciences8050159

来源: DOAJ

【 摘 要 】

Nitrates present one of the main groundwater contaminants in the world and in the Zagreb aquifer. The Zagreb aquifer presents the main source of potable water for the inhabitants of the City of Zagreb and it is protected by the Republic of Croatia. The determination of contaminants trends presents one of the main tools in groundwater body status and risk assessment. In this paper, the use of regression analysis on the aggregated data, together with confidence and prediction intervals, at different observation scales has been evaluated. Nitrate concentrations are generally decreasing in almost all areas, observed at different observation scales. It has been shown that linear regression can be efficiently used in the estimation of nitrates trends. Results showed that the calculation of confidence and prediction intervals can provide more useful conclusions than the calculation of the trend’s statistical significance. Also, the results suggest that confidence and prediction intervals can be used in groundwater body chemical status and risk assessment, respectively. Data smoothing and data aggregation are generally desirable, but have certain limitations. If too much data is aggregated, trend estimation by regression analysis can point to false conclusions. Evaluation of trends at different observational scales can provide more realistic trend estimation, as well as more precise identification of areas where groundwater protection measures should be implemented.

【 授权许可】

Unknown