Extracting association rules from changes in aquifer drawdown in irrigation areas of Qazvin plain, Iran

Abstract

Proper planning for better management of water resources, especially groundwater, is of particular importance due to water deficits and reductions in aquifer drawdown depth. The Qazvin plain has been exposed to variations of water levels in the aquifer drawdown during recent years. The present study aims to evaluate the dependency among the variations of aquifer drawdown groundwater levels and different effective factors using data mining association rules and the Apriori algorithm. Information on changes in aquifer drawdown depth and the affecting factors (2004–2018) were considered to extract the association rules. The results of association rules were evaluated by confidence, support, and lift indices. Based on the results, increasing the volume of the discharged water from agricultural wells with more than 0.8 million cubic meters had the highest correlation with the aquifer drawdown with more than 1 m. In addition, the extracted rules demonstrated that the aquifer drawdown was less than 0.3 m when the volume of water delivered to the irrigation network and discharged from agricultural wells was more than 0.5 and less than 0.2 million cubic meters, respectively. Further, the air temperature and water demand of plants had a direct relationship with the volume of the aquifer drawdown in most extracted rules while the precipitation volume and percentage of air humidity had reverse relations with the aquifer drawdown. Finally, due to the unfavorable conditions of the Qazvin plain, further use of irrigation networks and suitable cropping patterns is strongly recommended regarding improving aquifer conditions.