Developing a reliable model for aquifer vulnerability

• Published in: Stochastic environmental research and risk assessment Vol. 30; no. 1; pp. 175 - 187
• Main Authors: Jang, Cheng-Shin, Lin, Chien-Wen, Liang, Ching-Ping, Chen, Jui-Sheng
• Format: Journal Article
• Language: English
• Published: Berlin/Heidelberg Springer Berlin Heidelberg 01.01.2016; Springer Nature B.V
• Subjects: Agricultural land; Agricultural management; Alluvial fans; Aquatic Pollution; Aquifers; Assessments; Chemistry and Earth Sciences; Computational Intelligence; Computer Science; Discriminant analysis; Earth and Environmental Science; Earth Sciences; Ecological risk assessment; Environment; Ground-water quality; Groundwater; Groundwater management; Math. Appl. in Environmental Science; Mathematical models; Nitrates; Original Paper; Physics; Pollution; Pollution abatement; Pollution control; Probability Theory and Stochastic Processes; Reliability; Statistics for Engineering; Vadose water; Waste Water Technology; Water Management; Water Pollution Control; Water protection; Water quality; ISEW, Taiwan; Aquifer vulnerability; Discriminant analysis; Groundwater; DRASTIC; Nitrate-N
• ISSN: 1436-3240; 1436-3259
• DOI: 10.1007/s00477-015-1063-z

The assessment of aquifer vulnerability to pollution is crucial for planning a sound management strategy of groundwater quality protection and farmland fertilizer use. This study establishes a reliable model for aquifer vulnerability assessment with an excellent performance for predicting groundwater nitrate-N contamination in the Choushui River alluvial fan, Taiwan based on the DRASTIC method. To promote the prediction performance of aquifer vulnerability assessment, discriminant analysis (DA) was applied to determine the weights of factors in the DRASTIC model by comparing the model results with the observed nitrate-N data. Key factors influencing the presence of groundwater nitrate-N pollution were characterized for different concentration thresholds. The results of analysis reveal that the modified DRASTIC model using DA significantly improves prediction performance for aquifer vulnerability assessment, and groundwater protection zones can be determined correctly based on the modified DRASTIC index. Furthermore, the sensitivity of the factors in the modified DRASTIC model indicates that the depth to the groundwater and aquifer media are critical when the nitrate-N concentration is less than 3 mg/L, while the impact of the vadose zone plays a vital role in controlling nitrate-N pollution of over 5 mg/L.