Kazemi Elham, Karyab Hamid, Emamjome Mohammad-Mehdi
Department of Environmental Health Engineering, School of Health, Qazvin University of Medical Sciences, Qazvin, Iran.
Bahonar Blvd, College of Medical Sciences, Qazvin, Iran.
J Environ Health Sci Eng. 2017 Nov 21;15:23. doi: 10.1186/s40201-017-0287-x. eCollection 2017.
The presence of nitrate is one of the factors limiting the quality of groundwater resources, particularly in arid and semi-arid climates. Therefore, the knowledge about the distribution of nitrate in groundwater and its source has an effective role in protecting health. The study aimed to optimize an interpolation method to predict the nitrate concentration and assessment of aquifer vulnerability in Qazvin plain.
One hundred sixty-two deep wells in Qazvin plain aquifer were randomly selected and nitrate concentration was analyzed in four different lands including agricultural, residential, steppe and mixed-use areas. Interpolation was done by IDW, Spline, Kriging and National neighbor methods using ArcGIS software. To select the best interpolation method, errors of predicted values were determined by Mean Relative Error (RME) and Root Mean Square Error (RMSE). For analysis of potential vulnerability of aquifer to nitrate pollution due to agricultural activity and sewage leaks, hazard factors and control factors were used for identification of hazard indexes (HI) using IPNOA and IPNOC model.
The results showed that in 8.82% and 18.52% of samples in agricultural and residential areas, the detected nitrate was above the acceptable level at 50 mg/L. National neighbor method with the lowest RME and Spline method with the lowest RMSE were provided the most accurate estimates of nitrates in the aquifer. The highest hazard was obtained in agricultural areas (HI = 6.11). Also, the most influential parameters on aquifer vulnerability were mineral fertilizer (HF = 3), organic fertilizers (HF = 3), irrigation systems (CF = 1.04) and tillage patterns (CF = 1.04).
According to the results, National neighbor with the lowest RME was preferable than the other spatial interpolation methods for prediction of nitrate concentration in the aquifer. This method provided similar spatial distribution maps of nitrate in groundwater and that was an efficient method for assessing water quality. Hazard index as a result of agricultural activities (IPNOA) was ranged from "very low" to "low" which was in accordance with detected and predicted nitrate concentration in the aquifer. In addition he hazard of nitrate contamination from household (IPNOC) was in very low (class 2).
硝酸盐的存在是限制地下水资源质量的因素之一，尤其是在干旱和半干旱气候地区。因此，了解地下水中硝酸盐的分布及其来源对保护健康具有重要作用。本研究旨在优化一种插值方法，以预测加兹温平原地下水中硝酸盐浓度并评估含水层脆弱性。
在加兹温平原含水层随机选取162口深井，分析了农业、住宅、草原和混合用途四个不同区域的硝酸盐浓度。使用ArcGIS软件通过反距离权重法（IDW）、样条插值法、克里金法和最近邻点法进行插值。为选择最佳插值方法，通过平均相对误差（RME）和均方根误差（RMSE）确定预测值的误差。为分析农业活动和污水泄漏导致含水层对硝酸盐污染的潜在脆弱性，使用危害因素和控制因素，通过IPNOA和IPNOC模型确定危害指数（HI）。
结果表明，在农业和住宅区域，分别有8.82%和18.52%的样本检测到的硝酸盐含量高于50mg/L的可接受水平。最近邻点法的RME最低，样条插值法的RMSE最低，这两种方法对含水层中硝酸盐的估计最为准确。农业区域的危害最高（HI = 6.11）。此外，对含水层脆弱性影响最大的参数是矿物肥料（HF = 3）、有机肥料（HF = 3）、灌溉系统（CF = 1.04）和耕作方式（CF = 1.04）。
根据结果，RME最低的最近邻点法比其他空间插值方法更适合预测含水层中的硝酸盐浓度。该方法提供了与地下水中硝酸盐相似的空间分布图，是评估水质的有效方法。农业活动导致的危害指数（IPNOA）范围为“非常低”到“低”，这与含水层中检测到的和预测的硝酸盐浓度一致。此外，家庭来源的硝酸盐污染危害（IPNOC）为非常低（2类）。
