Risk assessment of shallow groundwater contamination under irrigation and fertilization conditions

Abstract

Due to extensive agricultural irrigation and fertilization, China now faces a severe problem of non-point source groundwater contamination. In this paper, the Daxing district of Beijing, a typical agricultural area with high fertilizer application, is chosen as the study area. With the consideration of intrinsic vulnerability and nitrate contamination load of groundwater, DRASTIC-based methodology was used to establish a GIS-based groundwater contamination risk assessment model, namely DRSIN model. The DRSIN model contains five parameters: depth to groundwater (D), net recharge (R), soil type (S), impact of the vadose zone (I) and nitrogen (N). By employing cluster analysis and fuzzy synthetic evaluation, the influence of agricultural non-point source contamination of groundwater on the formation of nitrate contamination was discussed under the existing conditions and 11 different irrigation and fertilization scenarios. The results show that the groundwater contamination risk in the north and east of Daxing District is higher than that in the south and the west, which conforms to the monitoring results of nitrate nitrogen content in groundwater. This indicates the need for reasonable groundwater utilization and protection planning to reduce agricultural non-point source contamination. Groundwater contamination risk decreases most significantly under the scenario of irrigation amount reduced by 25 %, nitrogen application reduced by 25 % and groundwater depth increased by 5 m. The findings provide data for reasonable groundwater development and utilization in this area.

Appendix

Model construction

Inverse distance-weighted (IDW) interpolation in ArcGIS was employed for spatial interpolation of the five factors. Cluster analysis was performed with SPSS to calculate the weights, and fuzzy synthetic evaluation was used for risk classification. Finally DRSIN model was constructed.

For groundwater contamination risk assessment by fuzzy synthetic evaluation method, the procedures are listed as follows:

• (a) Sample values were collected for 31 sites, and 5 indicators were included to reflect groundwater contamination risk. A score (0–10) was assigned to the characteristic value of each factor at each sampling point according to measurements. Thus indicator set X _i of the ith sampling point was constructed:

  $$X_{i} = \left[ {X_{i1} , \, X_{i2} ,{ \ldots }X_{i5} } \right],\quad{i = 1, \, 2, \, \ldots p}$$

  (4)

• (b) Construction of evaluation set

  $$V = \, \left[ {v_{1} , \, v_{2} , \ldots ,v_{5} } \right] ,$$

  (5)

where v ₁ : value assigned as 1 for score of 0–2.

v ₂ : value assigned as 2 for score of 3–4;

…

v ₅ : value assigned as 5 for score of 9–10.

• (c) For each factor X _ij , single factor evaluation matrix [r _ij1,r _ij2,…,r _ijm ] was constructed. That is, r _ijk (0 ≤ r _ijk  ≤ 1) represents the evaluation on factor x _ij using v _k , where j = 1,2,…,5 and k = 1,2,…,5. Thus the single factor evaluation matrix R = (r _ijk )5 × 5 was obtained.

• (d) Synthetic evaluation was performed according to the weight of each factor A = (a ₁,a ₂,…,a ₅): B = A O R = (b ₁,b ₂ ,…,b ₅), which is a fuzzy subset on V. Depending on the definition of operation O, M ($\times ,\;{ \oplus }$) was chosen. Weighted averaging model was used to calculate b _l = ∑(a _l ·r _ls ), (s = 1,2,…,5), where b _l is the probability of reaching the corresponding degree of contamination. Finally, the synthetic evaluation matrices of each sampling point were solved.

• (e) The evaluation results obtained using the above-mentioned method was then used for inverse distance-weighted (IDW) interpolation in MAPGIS. The groundwater contamination zoning map of Daxing District was obtained. Synthetic evaluation was carried out using C programming language. IDW interpolation in MAPGIS was applied to obtain groundwater contamination risk zoning map of Daxing District.