This study retains the basic structure of DRASTIC model and obtains more specific evaluation results by adding land-use type and groundwater resource yield parameters, modifying the rating scale and weight of nine parameters. Comparison of the modified DRASTIC-LY vulnerability map with the map of the original DRASTIC-LY method revealed differences in 40.49% of the study area. The risk map shows that the very high vulnerability area decreased from 2.79 to 1.67%, while the high vulnerability area increased from 18.70 to 28.86%. Areas with low vulnerability increased by 10.15%, while areas with medium vulnerability decreased by 15.01%. The areas with very high groundwater vulnerability are mainly distributed in the Hanbin area on the north bank of the Han River, the areas with high are mainly concentrated on both sides of the Fujia River, while the areas with low are distributed in most areas in the west of the basin. The Pearson's correlation factor was 0.0583 in the original DRASTIC model, 0.1113 in the DRASTIC-LY method and 0.8291 in the modified DRASTIC-LY model, which indicated that the revised DRASTIC-LY model was more appropriate than the original model. The results can help the government with the protection of water resources.

  • Improve the DRASTIC model.

  • Determine the weight, grade and category of each parameter.

  • Based on the sensitivity analysis and model validation results, construct the best model suitable for groundwater vulnerability assessment.

  • The results of this study will help government managers improve the composition of water use and strengthen the protection of vulnerable aquifers.

In recent years, with the rapid growth of population, climate change and economic development, the pressure on the water environment in many countries in the world has gradually increased (Ahmed et al. 2018). As one of the important natural resources, groundwater resources are facing increasingly serious pollution and challenges (Tiwari et al. 2016). Groundwater pollution in developed and developing countries is very serious due to population growth and agricultural development (Maqsoom et al. 2020). In Argentina, due to the rise of international commodity prices and the support of new technologies, agricultural expansion has increased the pressure of land use on natural resources, especially the threat to groundwater quality (McLay et al. 2001). In Iran, the groundwater pollution problem is very serious due to population growth and agricultural development (Aller et al. 1987). China has inevitably encountered such problems. The low utilization rate of water resources and the large discharge of pollutants have led to the increasingly severe water environment in China, which has brought about a profound environmental and ecological crisis (Tian et al. 2020). Therefore, in order to ascertain the current situation of water environment quality, improve the quality of water environment and realize the sustainable development and utilization of groundwater resources, it is necessary to first determine the groundwater vulnerability of each region, which is vulnerable to groundwater pollution.

In 1968, Margat first proposed the concept of groundwater vulnerability. Then, in 1987, the International Conference on Soil and Groundwater Vulnerability defined that groundwater vulnerability refers to the sensitivity of groundwater to external pollution sources, which is the inherent characteristic of aquifers. Until 1993, the National Research Council of the United States defined groundwater vulnerability as the trend and possibility of pollutants reaching a specific location above the uppermost aquifer. Groundwater vulnerability is divided into two categories: one is special vulnerability, and the other is internal vulnerability. With the progress of the study, the model of groundwater vulnerability has expanded from DRASTIC to SINTACS, GOD, AVI, SYNTACS, SI and EPIK models (Sadat-Noori & Ebrahimi 2016; Ahada & Suthar 2018; Ahirwar & Shukla 2018), among which the DRASTIC model is the most widely used method in groundwater vulnerability assessment at present. The traditional DRASTIC model structure was proposed by the United States Environmental Protection Agency (EPA) in 1987 (Al-Mallah & Al-Qurnawi 2018). The model has been applied to assess groundwater vulnerability in many regions in the United States and has achieved good demonstration results, which has also been adopted by Canada, South Africa and European countries. Chinese researchers began to use this method in the 1990s. In recent years, more and more scholars have studied this method, and it has been improved and applied in many regions of the country.

In 2008, Wen et al. (2009) used professional model (DRASTIC model) and geographic information system (GIS) technology to evaluate the vulnerability of shallow groundwater in the Zhangye Basin. In 2012, Yin et al. (2013) used the DRASTIC model in the GIS environment to construct a zoning map of groundwater vulnerability in the Ordos Plateau. The results show that 24.8% of the study area has high pollution potential, 24.2% has medium pollution potential, 19.7% has low pollution potential and the remaining 31.3% of the area has no risk of groundwater pollution. In 2016, Wu et al. (2016) proposed the DRTILSQ model, based on DRASTIC and considering human factors, to assess the risk of groundwater pollution in the northern suburbs of Yinchuan City. In 2017, Yang et al. (2017) used a modified DRASTIC model to assess the vulnerability of groundwater in the Jianghan Plain. The results show that the improved DRASTIC model has a great improvement compared with the conventional model. After the amendment, the correlation coefficient was significantly increased from 41.07 to 75.31%. In 2017, Li et al. (2017) conducted a groundwater vulnerability assessment in the plain area of Tianjin City based on the DRASTIC model and GIS technology containing seven hydrogeological parameters. In 2018, Wu et al. (2018) used a modified DRASTIC model (AHP-DRASTLE model) to assess the vulnerability of groundwater to pollution in Beihai, China, to support the protection of groundwater resources in coastal areas of China. In 2018, He et al. used the DRACILM model to assess the vulnerability of nitrate pollution in the western Liaohe Plain. The correlation between vulnerability class and the concentration of NO3-N in the DRACILM model improved to 0.649, which was 40.6% higher than that obtained by DRASTIC (He et al. 2018).

In agricultural production areas, the use of chemical fertilizers and pesticides is one of the most important sources for the increase of nitrate and chloride in groundwater. Meanwhile, the main sources of nitrate in groundwater also include the discharge of industrial sewage, domestic sewage and livestock manure (Jafari & Nikoo 2019). However, the chloride in groundwater is partly from mineral fertilizers (potassium chloride in the mixture of nitrogen, phosphorus and potassium), and partly from industrial salt used in road maintenance (Garewal et al. 2019). Ankang City is located at the upstream of the Middle Route of the South-to-North Water Transfer Project in China and is an important water source conservation and conservation area (Liu et al. 2018). In the past 20 years, the amount of groundwater resources in Ankang Basin has decreased, and the water quality changes are mainly caused by mining activities, the conversion of large areas of dry land into paddy fields and the overuse of pesticides and fertilizers by human beings (Mondal et al. 2019).

Therefore, it is necessary to accurately assess the vulnerability of groundwater to determine the health risk of groundwater and provide a reference for the implementation of South-to-North Water Transfer Project in the Middle Line and the improvement of water conservation function. The special purpose of this study is to (1) improve the DRASTIC model to include land-use type and model of groundwater resource yield, with emphasis on the impact of nitrate on groundwater vulnerability; (2) determine the weight, grade and category of each parameter, and establish the relationship between the parameter and the concentration of NO3-N; (3) based on the sensitivity analysis and model validation results, construct the best model suitable for groundwater vulnerability assessment in Ankang Basin. The results of this study will help government managers improve the composition of water use, strengthen the protection of vulnerable aquifers and build a green and environment-friendly new society.

Study area

Ankang Basin is located in the south of Shaanxi Province, extending from NW to SE. The study area is between 32° 40′ and 33° 00′ N, and 108° 16′ and 109° 06′ E (Figure 1). It is 88 km long from east to west, 3–8 km wide from north to south, with an area of 400 km2. Ankang Basin includes Hanbin District, Hanyin County and Shiquan County of Ankang City. The climate is hot in summer and wet and cold in winter, because it is distributed in the subtropical monsoon climate zone. The annual average temperature is 15–17 °C, the highest is 42.6 °C, the lowest is −16.4 °C and the annual precipitation is between 800 and 1,000 mm. Evaporation capacity is 736–1,238 mm/a. The highest temperature was 42.6 °C in July 1962, and the lowest temperature was −16.4 °C in December 1991. The basin lies between the Daba Mountains and Qinling Mountains, and the Yue River flows out from west to east. In the basin, floodplain, terrace, moraine and glacial erosion terrain, hills and alluvial fan terrain are widely distributed.
Figure 1

Location and sampling point distribution map in Ankang Basin of China.

Figure 1

Location and sampling point distribution map in Ankang Basin of China.

Close modal

In the basin, the primary river is the Han River, the secondary tributary is the Moon River and the tertiary rivers are the Fujia River, Ganges River, Donghe River and Guanyin River. The Yue River flows from the northwest to the southeast and flows into the Han River, while the Han River flows from the southwest to the northeast. The northern part of the basin is Qinling Mountains, with an altitude of between 480 and 1,000 m, while the southern part is Daba Mountains, with an altitude of between 1,220 and 2,120 m. The highest altitude of the basin is 390 m, the lowest is 245 m and the average is about 285 m. The soil of Ankang basin belongs to the yellow-brown soil zone of the bauxite region in China, with 8 soil categories, 17 subcategories, 33 soil genera and 71 soil species.

Controlled by topography, geomorphology and sedimentary environment, the groundwater in Ankang Basin is mainly divided into Pore water, Clastic rock pore water and Bedrock fissure water (Figure 2). Pore water is mainly divided into diving and confined water, which is distributed in Quaternary strata. Affected by the fact that the Quaternary strata are thick in the middle part and thin in the east and west sides, diving aquifers are widely distributed throughout the region. The lithology of the aquifer is mainly composed of sand and gravel of the Quaternary Holocene series, with a layer thickness of 2.5–7.5 m. The water level is shallow, and the depth to groundwater table is generally 3–5 m. The groundwater yield is generally between 100 and 300 m3/day in the west, 300 and 100 m3/day in the middle area and 50 and 100 m3/day in the east part. The confined water is only distributed between Heng River and Fujia River in the middle of the basin. The lithology of the aquifer is mainly composed of sand and gravel of the Quaternary Pliocene, often accompanied by argillaceous filling. The thickness of the aquifer is about 30–50 m, and the thickness in some areas is more than 100 m. The confined water head is about 10–20 m in the west and 20–30 m in the east of Wuli Town. The groundwater yield is generally between 620 and 1,670 m3/day, which has the significance of water supply. Groundwater in the basin is mainly recharged by surface water and atmospheric precipitation. The clastic rock pore water is distributed under the Neogene strata and belongs to confined water. The lithology of the aquifer is argillaceous sandstone, medium-coarse sandstone and conglomerate. The depth to groundwater table is 2.2–11.5 m and the water head is about 6.3–11.3 m above the ground. The groundwater yield is generally between 20 and 60 m3/day, which has the significance of water supply. Bedrock fissure water is mainly distributed in the strongly weathered zone of granite at the bottom of the basin, and also in a small area in the east and south of the basin. The groundwater yield is generally less than 50 m3/day, which does not have the significance of water supply. As the development and utilization of groundwater in the basin, 76.3% of groundwater is used for agricultural irrigation, 5.1% for industrial production and 18.6% for urban domestic water. Due to the hydraulic connection between surface water and groundwater, pesticides and fertilizers have become the main pollution sources affecting groundwater quality, especially the excessive use of nitrogen and organic pesticides.
Figure 2

Groundwater types in Ankang basin.

Figure 2

Groundwater types in Ankang basin.

Close modal

In order to ascertain the quality of shallow groundwater in the basin, 32 samples of shallow groundwater were collected from June 2021 to September 2022 (Figure 1). According to the principle of section control, the sampling density can reach 1 group/10 km2.

Methodology

The DRASTIC model was developed by the U.S. Environmental Protection Agency (EPA) to evaluate groundwater pollution potential for the entire United States (Moghaddam et al. 2017). The acronym DRASTIC stands for the seven parameters used to calculate the DRASTIC index (DI) value: the depth to groundwater table (D), net recharge (R), aquifer properties (A), soil properties (S), topography (T), impact of the vadose zone (I) and the hydraulic conductivity of the aquifer (C). Each factor is mainly rated on a scale of 1–10, which indicates the relative pollution potential of a given factor (Table 1) (Hosseini & Saremi 2018). The seven parameters are then assigned with weights ranging from 1 to 5, reflecting their relative importance (Table 1). The DI or vulnerability rating is then computed by applying a linear combination of all factors (Babiker et al. 2005):
formula
where D, R, A, S, T, I and C are the seven parameters and the subscripts r and w are the corresponding rating and weights, respectively. The higher the value of DI index, the greater the groundwater vulnerability to pollution.
Table 1

Data source and format for the seven parameter data layers

ParametersData sourcesScale of map prepared
61 long-term observation wells and 10 monitoring wells, 2021–2022 1:100,000 
precipitation and irrigation data from 1985 to 2020 1:100,000 
27 hydrogeological drill-hole data, 2021–2022 1:100,000 
10 water penetration test, 2021–2022; World Soil Database (HWSD), 2012 1:100,000 
27 hydrogeological drill-hole data, 2021–2022 1:100,000 
27 hydrogeological drill-hole data, 2021–2022 1:100,000 
33 pumping tests and 27 hydrogeological drill-hole data, 2021–2022 1:100,000 
ParametersData sourcesScale of map prepared
61 long-term observation wells and 10 monitoring wells, 2021–2022 1:100,000 
precipitation and irrigation data from 1985 to 2020 1:100,000 
27 hydrogeological drill-hole data, 2021–2022 1:100,000 
10 water penetration test, 2021–2022; World Soil Database (HWSD), 2012 1:100,000 
27 hydrogeological drill-hole data, 2021–2022 1:100,000 
27 hydrogeological drill-hole data, 2021–2022 1:100,000 
33 pumping tests and 27 hydrogeological drill-hole data, 2021–2022 1:100,000 
Based on the multi-source basic data of the study area, the thematic layer of the seven parameters of the model was constructed (Figure 4). All data were based on GIS format, and ArcGIS 10.2 software was used for overlay analysis and calculation. The depth to groundwater table (D) was based on 61 long-term observation wells and 10 monitoring wells, and the shallow groundwater level measurement data were obtained in August 2021 and May 2022. Based on these scattered data, a raster image with a pixel size of 10 m was created using Kriging interpolation (Figure 3(b)). Then, according to the definition of DRASTIC model, the depth to groundwater table obtained by the difference is given and assigned a rate of 4–10. Net recharge (R) refers to the amount of water reaching the underground aquifer after passing through the aeration zone. In order to accurately calculate the net recharge parameters, we used the FEFLOW7.4 model. The R parameter was calculated by comprehensively considering hydrological data, meteorological data, characteristics of aeration zone and hydrogeological conditions (Figure 3(c)). The obtained values of net recharge were grouped and rated from 1 to 4. The aquifer properties (A) refer to aquifer characteristics that affect solute migration and transformation process. The A parameter was obtained by using geological map (1:100,000), hydrogeological map (1:100,000) and 27 borehole data (Figure 3(d)). Then, the A factor was differed to create a parameter map of different hydrogeological units. Finally, the hydrogeological unit was rated from 4 to 9. The soil properties (S) represent the soil characteristics of the seepage zone and the ability of surface water to penetrate the ground. The soil medium map was mainly based on the global soil database and the results of the third survey of land quality in China (Figure 3(e)). According to the soil texture, the grade of soil properties was from 3 to 10. Topography (T) refers to the use of the global digital elevation model (DEM) with a precision of 10 m to determine the percentage slope within a specific range (Figure 3(f)). Then, according to the slope map, multiple slope equivalent areas were divided and assigned a score from 1 to 10. The influence of vadose zone (I) can be defined as the influence of unsaturated zone characteristics. The ecological geological map (1:50,000), hydrogeological map (1:100,000) and geological map (100,000) were used to obtain the contour map of the vadose area (Figure 3(g)). These data enabled us to accurately describe the profile characteristics of the seepage zone and then compiled them into parameters that can be recognized by the DRASTIC model. The hydraulic conductivity (C) of the aquifer represents the capacity of the aquifer to output and store groundwater per unit volume. The hydraulic conductivity of the aquifer in the study area was determined by analyzing the hydrogeological data, using the data of 27 boreholes and the data of 33 pumping tests. According to the definition of the DRASTIC model, different hydraulic conductivity zones in the study area were assigned with ratings from 1 to 5 (Figure 3(h)).
Figure 3

Seven layers of the DRASTIC-LY model, (a, depth to water, b, concentration, c, net recharge, d, aquifer media, e, soil, f, topography (slope), g, impact of the vadose zone, h, hydraulic conductivity, i, land-use type and j, groundwater resource yield).

Figure 3

Seven layers of the DRASTIC-LY model, (a, depth to water, b, concentration, c, net recharge, d, aquifer media, e, soil, f, topography (slope), g, impact of the vadose zone, h, hydraulic conductivity, i, land-use type and j, groundwater resource yield).

Close modal
Figure 4

The map of intrinsic vulnerability based on DRASTIC model.

Figure 4

The map of intrinsic vulnerability based on DRASTIC model.

Close modal

After creating seven layers with attribute structure, the DI value was obtained by weighted summation of multiple parameters and finally the groundwater vulnerability assessment map was obtained by overlaying a thematic map using ArcGIS platform (Rahman 2008). Generally speaking, the value of DI is between 56 and 187. The lower the DI value, the weaker the potential vulnerability risk of groundwater, and the higher the DI value, the greater the risk of groundwater pollution in this area (Al-Abadi et al. 2017). The DI was divided into five categories, namely very low vulnerability (<75), low vulnerability (100–75), medium vulnerability (125–100), high vulnerability (150–125) and very high vulnerability (>150) (Lad et al. 2019).

DRASTIC-LY model

Based on the original DRASTIC model, between Qinling and Daba Mountains, the potential risks of the model of groundwater resource yield and land-use type for pollution identification should also be considered (Kozlowski & Sojka 2019). In the water conservation function area, the change of groundwater quality is mainly caused by human activities and agricultural fertilization and the high concentration of nitrate is the main factor affecting the quality of groundwater (Figure 3(b)). The model of groundwater resource yield (Y) in a specific area shows the self-purification ability and renewal rate of groundwater. Based on the systematic analysis of geological and hydrogeological conditions, the threshold value of the Y parameter in the study area was determined through pumping test and water injection test of a large number of boreholes (Figure 3(j)). The initial weight of Y parameters in the DRASTIC-LY model is set to 4. Based on the data of the third land-use survey in China, a land-use map was prepared to assess the potential risk of groundwater pollution (Figure 3(i)). The following types of land use are considered: woodland, grassland, water and reservoir, cultivated land and architecture land, with grades of 2, 4, 5, 7 and 10, respectively (Table 2) (Sarkar & Pal 2021). The initial weight of land-use type parameters in the DRASTIC-LY model is set to 5. The values of DI in the DRASTIC-LY model range from 54 to 260 in theory (Aslam et al. 2020).

Table 2

Rating scales for land-use types and model of groundwater resource yield

Land-use type
Model of groundwater resource yield
TypesValueRange (m3/day)Value
Woodland <100 
Grassland 100–300 
water,reservoir 300–1,000 
Cultivated land 1,000–3,000 
Architecture 10 >3,000 
Land-use type
Model of groundwater resource yield
TypesValueRange (m3/day)Value
Woodland <100 
Grassland 100–300 
water,reservoir 300–1,000 
Cultivated land 1,000–3,000 
Architecture 10 >3,000 

Applying the DRASTIC model to intrinsic vulnerability

The internal vulnerability assessment map of Ankang Basin was established by using the DRASTIC model. The groundwater vulnerability index was calculated on the ArcGIS platform, and the data sources were shown in Table 1. The weights and evaluation levels of vulnerability indicators were shown in Table 3.

Table 3

Original and modified values of the DRASTIC-LY parameters

Depth to groundwater
Net recharge
Aquifer media
Range (m)ABCRange (mm)ABCAquifer mediaABC
>9.0 4.66 3.22 <165 2.63 1.12 Silty sand and pebbles 6.52 3.58 
6.0–9.0 14.47 10.00 165–185 11.75 5.00 Gravel and pebbles 9.17 5.03 
3.0–6.0 6.6 4.56 185–205 6.25 2.66 Gravel and mud 16.40 9.00 
<3.0 10 3.24 2.24 205–225 7.72 3.29 Slate and sandstone 11.31 6.21 
    >225 1.90 0.81 Gravelly sandstone 0.92 0.50 
Soil type
Topography
Influence of the vadose zone
Soil typeABCRange slope (%)ABCGeological formationABC
Paddy soil 8.54 2.99 >20 – – Loam 5.34 2.40 
Submergenic paddy soil 4.94 1.73 15–20 4.70 1.66 Sandy soil 10.69 4.81 
Yellow-brown soil 4.03 1.41 10–15 7.64 2.69 Glutenite 1.13 0.51 
Yellow-brown earth 1.66 0.58 5–10 14.19 5.00 Gritstone 20.00 9.00 
Yellow cinnamon soil 14.28 5.00 <5 10 6.60 2.33 Igneous rock 1.67 0.75 
Skeleton soils 10 13.80 4.83         
Hydraulic conductivity
Land-use type
Model of groundwater resource yield
Range (m/day)ABCTypeABCRange (m3/h)ABC
<10 10.35 5.00 Woodland 2.92 1.76 <100 7.78 3.61 
10–20 8.53 4.12 Grassland 16.55 10.00 100–300 1.12 0.52 
20–30 9.83 4.75 water,reservoir 7.63 4.61 300–1,000 19.39 9.00 
30–40 6.35 3.07 Cultivated land 8.16 4.93 1,000–3,000 6.07 2.82 
>40 8.65 4.18 Architecture 10 6.95 4.20 >3,000 7.36 3.42 
Depth to groundwater
Net recharge
Aquifer media
Range (m)ABCRange (mm)ABCAquifer mediaABC
>9.0 4.66 3.22 <165 2.63 1.12 Silty sand and pebbles 6.52 3.58 
6.0–9.0 14.47 10.00 165–185 11.75 5.00 Gravel and pebbles 9.17 5.03 
3.0–6.0 6.6 4.56 185–205 6.25 2.66 Gravel and mud 16.40 9.00 
<3.0 10 3.24 2.24 205–225 7.72 3.29 Slate and sandstone 11.31 6.21 
    >225 1.90 0.81 Gravelly sandstone 0.92 0.50 
Soil type
Topography
Influence of the vadose zone
Soil typeABCRange slope (%)ABCGeological formationABC
Paddy soil 8.54 2.99 >20 – – Loam 5.34 2.40 
Submergenic paddy soil 4.94 1.73 15–20 4.70 1.66 Sandy soil 10.69 4.81 
Yellow-brown soil 4.03 1.41 10–15 7.64 2.69 Glutenite 1.13 0.51 
Yellow-brown earth 1.66 0.58 5–10 14.19 5.00 Gritstone 20.00 9.00 
Yellow cinnamon soil 14.28 5.00 <5 10 6.60 2.33 Igneous rock 1.67 0.75 
Skeleton soils 10 13.80 4.83         
Hydraulic conductivity
Land-use type
Model of groundwater resource yield
Range (m/day)ABCTypeABCRange (m3/h)ABC
<10 10.35 5.00 Woodland 2.92 1.76 <100 7.78 3.61 
10–20 8.53 4.12 Grassland 16.55 10.00 100–300 1.12 0.52 
20–30 9.83 4.75 water,reservoir 7.63 4.61 300–1,000 19.39 9.00 
30–40 6.35 3.07 Cultivated land 8.16 4.93 1,000–3,000 6.07 2.82 
>40 8.65 4.18 Architecture 10 6.95 4.20 >3,000 7.36 3.42 

A: Original rating; B: average concentration of NO3-N (mg/L); C: Modified rating value.

The intrinsic vulnerability index (DI) of Ankang Basin was between 56 and 187, which can be divided into five classes (Mondal et al. 2019; Aslam et al. 2020): very high, high, medium, low and very low (Figure 4). The vulnerability classification in the northwest, central and southeast parts of the basin was relatively high, which means that this area was more vulnerable to external pollution than other areas; this result was mainly due to the shallow depth of the groundwater, the high hydraulic conductivity of the aquifer or the large proportion of sand and gravel in the aquifer and the aerated zone and the groundwater is mainly recharged by surface water (Karan et al. 2018). The assessment results were basically consistent with the contour map of the initial concentration of nitrate (Figure 3(a)), while nitrate in groundwater was mainly brought in by human activities, which indicated that the groundwater in this area was vulnerable to human activities. In the central and eastern parts of the basin, due to the deep groundwater depth or low hydraulic conductivity, the vulnerability level was medium. Low vulnerability areas were concentrated, mainly distributed in the central and western parts of the basin. Due to the vadose zone (loam), hydrogeological characteristics of deep groundwater level and aquifer medium (fine sand), groundwater vulnerability was low. In summary, in the western, central and southeastern parts of the basin, shallow groundwater depth, rapid groundwater runoff and semi-open environment were the main reasons for the high vulnerability level. However, in most of the western and central parts of the basin, the main reasons for the low vulnerability level were the deep groundwater level, slow flow and closed environment.

Assessment of nitrate vulnerability using DRASTIC-LY model

The accurate vulnerability map of the Ankang Basin was obtained by using the DRASTIC-LY model. Due to the importance of land-use type and groundwater resource yield model in reflecting human activities, the weight should be given priority consideration. As shown in Table 2, the weights of land-use type and groundwater resource yield parameters were set to 5 and 4, respectively.

The accurate vulnerability index (DI) in the Ankang Basin was between 74 and 273 and could be divided into five classes (Wu et al. 2016; Shakoor et al. 2020): very high, high, medium, low and very low (Figure 5). The southeast of Ankang City, Wuli Town, the southeast of Hanyin County and the northeast of Shiquan County had higher groundwater vulnerability levels, which meant that this area was more vulnerable to external pollution than other areas. In the southeast of Ankang City, pumping groundwater along the Han River and the accumulation of pollutant groundwater along the groundwater flow are the main reasons for the high vulnerability level; in Wuli Town, groundwater vulnerability is mainly affected by groundwater resource output (Figure 3(d)) and aquifer medium (Figure 3(j)); in the southeast of Hanyin County and the northeast of Shiquan County, land-use type is the main reason for high vulnerability. In the northern part of Yue River and the northern part of Ankang City, the nitrate concentration is high, but due to the slow flow rate of groundwater, the low development and utilization potential and the aquifer medium, the groundwater vulnerability classification is medium. The groundwater vulnerability classification in the central, northern and western areas of Ankang Basin is relatively low, which is basically consistent with the assessment results using the DRASTIC model.
Figure 5

The map of nitrate vulnerability based on DRASTIC-LY model.

Figure 5

The map of nitrate vulnerability based on DRASTIC-LY model.

Close modal

Revising the rating scale of each parameter

Adjust the rationality of the rating scale to improve the accuracy of the specific vulnerability assessment results (Sinha et al. 2016; Hao et al. 2017). Using statistical methods, the classification of nine parameters defined in the original DRASTIC-LY model was correlated with the concentration of NO3-N to optimize the rating scale. For no-continuous parameters such as aquifer media, influence of the vadose zone, hydraulic conductivity, land-use type, model of groundwater resource yield, it was necessary to maintain all existing rating scales in the area. The original value, modified value and corresponding average nitrate concentration in groundwater of each parameter in the DRASTIC-LY model were shown in Table 3. According to the average nitrate concentration, the rating of each category parameter had been modified, and the rating was controlled within 5 or 10 grades. The rating of soil type and topography parameters was modified from 10 to 5. The original rating of net recharge (4) was lighter than the other 9 parameters (5, 9 and 10), and the revised rating had been appropriately modified to 5 according to the importance of the parameters. The revised parameter ratings had three categories (including 5, 9 and 10) to ensure that the results were more reasonable and reliable.

Revising the parameter weights

The greater the weight of the parameter, the more likely it is to affect the groundwater vulnerability in the model than other parameters. Different regions and different parameters also have different impacts on groundwater vulnerability. For example, the impact of the vadose zone and topography that significantly affect NO3-N concentrations were assigned high and low weights, respectively. However, whether the theoretical research can be applied to the practice of the study area needs further discussion. In this study, it is necessary to revise the weights applicable to the model of this study area. In order to optimize the DRASTIC-LY model, the weight of each parameter needed to be recalculated. Using Pearson's (r) correlation, the correlation between each parameter and the average NO3-N concentration was calculated to achieve the optimized parameters. According to the maximum weight value (5) specified by the DRASTIC model, the new weighting factor was recalculated. If a parameter is not statistically significant, it will be excluded from the vulnerability equation. As shown in Table 4, Pearson's (r) correlation and revised weights it can clearly be concluded that the ‘Topography’ and ‘Soil type’ parameters were not statistically significant and should be weakened from the vulnerability model. The non-significant correlation between ‘NO3-N’ concentration and ‘Topography’ indicated that the runoff velocity of groundwater had little impact on nitrate concentration in groundwater (Rezaei et al. 2018; Zenebe et al. 2020). The non-significant correlation between ‘NO3-N’ concentration and ‘Soil type’ means that the adsorption and chemical reaction of soil to NO3-N can be ignored (Mondal et al. 2017; Jhariya 2019). It can also be concluded that the weights for the depth to groundwater table, land-use type, hydraulic conductivity and model of groundwater resource yield changed very little. However, the weights of the net recharge and the impact of the vadose zone parameters had decreased, although they were still relatively high. In addition, the weights of groundwater resource yield and aquifer media in the model were slightly increased. The modified weights showed their importance in the assessment of groundwater vulnerability in the study area. Finally, the revised weights for depth to groundwater table (D), net recharge (R), aquifer media (A), soil type (S), topography (T), impact of the vadose zone (I), hydraulic conductivity (C), land-use type (L) and model of groundwater resource yield (Y) are 5, 3, 4, 1, 1, 4, 3, 5 and 5, respectively (Hu et al. 2018; Joshi & Gupta 2018).

Table 4

Original and modified rating values of the DRASTIC-LY parameters

ParameterOriginal weightsPearson's (r) correlationRevised weights
Depth to groundwater 0.486 
Net recharge 0.295 
Aquifer media 0.411 
Soil type 0.117 
Topography 0.104 
Impact of the vadose zone 0.402 
Hydraulic conductivity 0.323 
Land-use type 0.479 
Model of groundwater resource yield 0.451 
ParameterOriginal weightsPearson's (r) correlationRevised weights
Depth to groundwater 0.486 
Net recharge 0.295 
Aquifer media 0.411 
Soil type 0.117 
Topography 0.104 
Impact of the vadose zone 0.402 
Hydraulic conductivity 0.323 
Land-use type 0.479 
Model of groundwater resource yield 0.451 

Utilities of vulnerability maps for groundwater protection and management

The scatter diagram of nitrate concentration in groundwater and the optimized DRASTIC-LY model evaluation map were shown in Figure 6. The results indicated that the optimized DRASTIC-LY model had a high correlation with the actual nitrate concentration and was most suitable for understanding the specific assessment of groundwater pollution vulnerability in different regions. The nitrate concentration (NO3-N) values in the shallow groundwater were classified into five classes such as 0.96–2.70, 2.70–5.40, 5.40–10.00, 10.00–20.00 and 20.00–32.00 mg/L. Comparison of the modified DRASTIC-LY vulnerability map with the map of the original DRASTIC-LY method revealed differences in 40.63% of the study area. The risk map shows that the very high vulnerability area decreased from 2.79% (original model) to 1.67% (modified model), while the high vulnerability area increased from 18.70 to 28.86%. Areas with low vulnerability increased by 10.15% compared to those predicted by the original DRASTIC-LY map. Areas with medium vulnerability decreased by 15.01% compared to those predicted by the original DRASTIC-LY map.
Figure 6

Map of specific vulnerability using modified DRASTIC-LY model.

Figure 6

Map of specific vulnerability using modified DRASTIC-LY model.

Close modal

On the north bank of the Hanjiang River in Ankang City, the vulnerability of groundwater is very high, which means that the leakage of leachate from the landfill site is the main reason for the high vulnerability. Compared with the prediction of the original DRASTIC-LY model, there are still high vulnerability areas on the east and west sides of the Fujia River, and this area is slightly larger than the area predicted by the original model. The main reason is that the area is located in an area with intensive human activities, and groundwater is largely exploited while receiving rainfall and river recharge (Moghaddam et al. 2018). It is worth noting that the low vulnerability area in the western part of the basin has increased significantly, mainly because the groundwater in this area is dominated by lateral runoff and receives little recharge from precipitation and river (Mogaji & San Lim 2017).

Groundwater is the core element of natural resources, and its moderate development, utilization and effective protection ensure the sustainable use of human beings (Gemail et al. 2017; Tian et al. 2021). The assessment of groundwater vulnerability is crucial for the protection of groundwater resources in Ankang basin and the implementation of the water transfer project in the middle route of China's South-to-North Water Transfer Project. The groundwater vulnerability assessment map helps the government to make wise decisions to prevent the impact of industrial production, agricultural activities and domestic sewage discharge on groundwater resources. The optimized DRASTIC-LY model is the most suitable model to evaluate the specific vulnerability of groundwater in Ankang Basin to nitrate pollution.

Sensitivity analysis of the modified DRASTIC-LY model

As shown in Table 5, sensitivity analysis is a single-parameter sensitivity analysis by comparing the revised weight and the effective weight. The effective weight of the parameter reflects the function of the relationship between its theoretical weight and the nine parameters in the optimized DRASTIC-LY model. In this study, there is a slight deviation between the theoretical weight and the effective weight calculated by the optimized DRASTIC-LY model. The calculation results of the single-parameter sensitivity analysis show that the average effective weight is between 6.37 and 49.84%, indicating that the nine indicators in the vulnerability assessment model are quite different. The effective weights of the R, S, T, I, H and Y parameters (7.54, 2.66, 2.26, 10.34, 9.57 and 14.25%, respectively) are less than their theoretical weights (9.68, 3.23, 3.23, 12.90, 9.68 and 16.13%, respectively, which have standard deviations of 3.12, 1.54, 1.08, 4.69, 3.84 and 6.27%, respectively). Soil type and topography parameters had little effect on groundwater vulnerability compared to the other seven parameters (Wei et al. 2021). The effective weights for D, A and L (21.25, 13.22 and 14.25%, respectively) are higher than their theoretical weights (16.13, 12.90 and 16.13%, respectively). Depth to groundwater, aquifer media, model of groundwater resource yield and land-use type are the four parameters most affected by human activities, which play a key role in groundwater vulnerability assessment (Oroji & Karimi 2018). Net recharge, hydraulic conductivity and impact of the vadose zone are the naturally determined parameters, which also play a certain role in groundwater vulnerability assessment (Baghapour et al. 2016; Saida et al. 2017; Khan & Jhariya 2019).

Table 5

Statistics of the single-parameter sensitivity analysis

ParameterRevised weightRevised weight (%)Effective weighting (%)
MinMaxAverageStandard deviation
Depth to groundwater 16.12 16.13 9.36 34.02 21.25 
Net recharge 9.67 9.68 2.38 13.03 7.54 
Aquifer media 12.92 12.90 1.99 24.80 13.22 
Soil type 3.23 3.23 0.90 7.21 2.66 
Topography 3.23 3.23 0.87 6.37 2.26 
Impact of the vadose zone 12.92 12.90 2.03 23.85 10.34 
Hydraulic conductivity 9.67 9.68 4.91 22.39 9.57 
Land-use type 16.12 16.13 5.83 49.84 18.88 
Model of groundwater resource yield 16.12 16.13 2.59 26.95 14.25 
ParameterRevised weightRevised weight (%)Effective weighting (%)
MinMaxAverageStandard deviation
Depth to groundwater 16.12 16.13 9.36 34.02 21.25 
Net recharge 9.67 9.68 2.38 13.03 7.54 
Aquifer media 12.92 12.90 1.99 24.80 13.22 
Soil type 3.23 3.23 0.90 7.21 2.66 
Topography 3.23 3.23 0.87 6.37 2.26 
Impact of the vadose zone 12.92 12.90 2.03 23.85 10.34 
Hydraulic conductivity 9.67 9.68 4.91 22.39 9.57 
Land-use type 16.12 16.13 5.83 49.84 18.88 
Model of groundwater resource yield 16.12 16.13 2.59 26.95 14.25 

The nitrate concentration of shallow groundwater in 32 different villages was tested and analyzed. The concentration of nitrate in groundwater was accurately determined by spectrophotometry. The measured nitrate concentration and the parameters affecting the recharge, runoff and discharge of groundwater are comprehensively considered and used to modify the hydrogeological parameters in the original model to obtain the modified DRASTIC-LY model. Pearson's correlation factor was 0.0583 in the original DRASTIC model, 0.1113 in the DRASTIC-LY method and 0.8291 in the modified DRASTIC-LY model (Figure 7). The conclusion shows that the vulnerability assessment map constructed by the modified DRASTIC-LY model is more accurate than that constructed by the original model.
Figure 7

Relationship between nitrate concentration and DI values: (a) original DRASTIC model, (b) DRASTIC-LY model, (c) modified DRASTIC-LY model.

Figure 7

Relationship between nitrate concentration and DI values: (a) original DRASTIC model, (b) DRASTIC-LY model, (c) modified DRASTIC-LY model.

Close modal

In order to effectively address the impact of human activities on groundwater vulnerability assessment results, this paper is based on the structure of the DRASTIC model, but (1) in Qinling Mountains-Daba Mountains, an important water conservation area, the impact of land-use types and groundwater resource yield parameters on groundwater vulnerability was considered, (2) the weight of net recharge, soil type and impact of the vadose zone parameters in the model was reduced, (3) according to the correlation between the nine parameters in the model and the NO3-N concentration, the parameter weight was modified.

  • 1.

    DRASTIC, DRASTIC-LY and optimized DRASTIC-LY models were used to evaluate the vulnerability of groundwater to nitrate pollution. Comparison of the modified DRASTIC-LY vulnerability map with the map of the original DRASTIC-LY method revealed differences in 40.49% of the study area. The risk map showed that the very high vulnerability area decreased from 2.79 to 1.67%, while the high vulnerability area increased from 18.70 to 28.86%. Areas with medium vulnerability decreased by 15.01% compared to those predicted by the original DRASTIC-LY map. Areas with low vulnerability increased by 10.15% compared to those predicted by the original DRASTIC-LY map.

  • 2.

    The evaluation results indicate that the areas with high groundwater vulnerability are mainly distributed on the north bank of the Han River, the areas with high groundwater vulnerability are mainly concentrated on both sides of the Fujia River and the areas with low groundwater vulnerability are distributed in most areas in the west of the basin. Natural conditions and human production activities jointly affect the vulnerability of groundwater, among which industrial production, agricultural activities and domestic sewage are the main reasons for the increase of groundwater vulnerability.

  • 3.

    The Pearson's correlation factor was used to determine the statistical relationship between nitrate concentrations in groundwater and groundwater vulnerability maps. The Pearson's correlation factor was 0.0583 in the original DRASTIC model, 0.1113 in the DRASTIC-LY method and 0.8291 in the modified DRASTIC-LY model, which indicated that the revised DRASTIC-LY model was more appropriate than that constructed by the original model.

Due to the reduction of project funds and the limited sample size, the accuracy of this study was not particularly accurate. In addition, microbial samples and N isotope samples had not been analyzed and tested, so it is impossible to determine the exact source of nitrogen and the impact of nitrification and denitrification on vulnerability. Therefore, the limitation of this study was that it could not accurately characterize the source, migration and transformation of nitrogen. This study comprehensively described the inherent and specific vulnerability of Ankang Basin and laid a foundation for nitrogen monitoring and dynamic evaluation in the future.

This work was supported by the Project of Water Ecological Restoration to Support Investigation in Shaanxi Section of Hanjiang River of South-to-North Water Diversion Project (No. ZD20220207) and Ecological Restoration Survey Project of Important Ecological Areas in Southern China (CGS). The author thanks the editors and reviewers of the journal for their valuable comments, which were of great help to the improvement of the paper. We give thanks to Changlai Xiao, Xiujuan Liang, Xu Honggen, Xinhua Bao, Bo Zhang and Shanghai Du for helping us with data processing skills and writing skills during the writing process.

All relevant data are included in the paper or its Supplementary Information.

The authors declare there is no conflict.

Ahada
C. P. S.
&
Suthar
S.
2018
A GIS based DRASTIC model for assessing aquifer vulnerability in Southern Punjab, India
.
Modeling Earth Systems and Environment
4
(
2
),
635
645
.
Ahirwar
S.
&
Shukla
J. P.
2018
Assessment of groundwater vulnerability in Upper Betwa River Watershed using GIS based DRASTIC model
.
Journal of the Geological Society of India
91
(
3
),
334
340
.
Ahmed
I.
,
Nazzal
Y.
&
Zaidi
F.
2018
Groundwater pollution risk mapping using modified DRASTIC model in parts of hail region of Saudi Arabia
.
Environmental Engineering Research
23
(
1
),
84
91
.
Aller
L.
,
Bennet
T.
,
Lehr
J. H.
,
Petty
R. J.
&
Hackett
G.
1987
DRASTIC: Standardized system for evaluating groundwater pollution potential using hydrogeologic settings
.
Journal of the Geological Society of India
29
(
1
).
Babiker
I. S.
,
Mohamed
M. A. A.
,
Hiyama
T.
&
Kato
K.
2005
A GIS-based DRASTIC model for assessing aquifer vulnerability in Kakamigahara Heights, Gifu Prefecture, central Japan
.
Science of the Total Environment
345
(
1–3
),
127
140
.
Baghapour
M. A.
,
Nobandegani
A. F.
,
Talebbeydokhti
N.
,
Bagherzadeh
S.
,
Nadiri
A. A.
,
Gharekhani
M.
&
Chitsazan
N.
2016
Optimization of DRASTIC method by artificial neural network, nitrate vulnerability index, and composite DRASTIC models to assess groundwater vulnerability for unconfined aquifer of Shiraz Plain, Iran
.
Journal of Environmental Health Science and Engineering
14
(
1
),
13
.
Garewal
S. K.
,
Vasudeo
A. D.
,
Landge
V. S.
&
Ghare
A. D.
2019
Groundwater vulnerability mapping using modified DRASTIC ANP
.
GRADEVINAR
71
(
4
),
283
296
.
Gemail
K. S.
,
El Alfy
M.
,
Ghoneim
M. F.
,
Shishtawy
A. M.
&
Abd El-Bary
M.
2017
Comparison of DRASTIC and DC resistivity modeling for assessing aquifer vulnerability in the central Nile Delta, Egypt
.
Environment Earth Science
76
(
9
),
350
.
He
H. Y.
,
Li
X. G.
,
Li
X.
,
Cui
J.
,
Zhang
W. J.
&
Xu
W.
2018
Optimizing the DRASTIC method for nitrate pollution in groundwater vulnerability assessments: a case study in China
.
Polish Journal of Environmental Studies
27
(
1
),
95
107
.
Hu
X. J.
,
Ma
C. M.
,
Qi
H. H.
&
Guo
X.
2018
Groundwater vulnerability assessment using the GALDIT model and the improved DRASTIC model: a case in Weibei Plain, China
.
Environmental Science and Pollution Research
25
(
32
),
32524
32539
.
Karan
S. K.
,
Samadder
S. R.
&
Singh
V.
2018
Groundwater vulnerability assessment in degraded coal mining areas using the AHP-Modified DRASTIC model
.
Land Degradation and Development
29
(
8
),
2351
2365
.
Maqsoom
A.
,
Aslam
B.
,
Khalil
U.
,
Ghorbanzadeh
O.
,
Ashraf
H.
,
Tufail
R. F.
,
Farooq
D.
&
Blaschke
T.
2020
A GIS-based DRASTIC model and an adjusted DRASTIC model (DRASTICA) for groundwater susceptibility assessment along the China–Pakistan economic corridor (CPEC) route
.
ISPRS International Journal of Geo-Information
9
(
5
),
332
.
McLay
C. D. A.
,
Dragten
R.
,
Sparling
G.
&
Selvarajah
N.
2001
Predicting groundwater nitrate concentrations in a region of mixed agricultural land use: a comparison of three approaches
.
Environment Pollution
115
(
2
),
191
204
.
Moghaddam
H. K.
,
Jafari
F.
&
Javadi
S.
2017
Vulnerability evaluation of a coastal aquifer via GALDIT model and comparison with DRASTIC index using quality parameters
.
Hydrological Sciences Journal-Journal Des Sciences Hydrologicaques
62
(
1
),
137
146
.
Mondal
N. C.
,
Adike
S.
,
Singh
V. S.
,
Ahmed
S.
&
Jayakumar
K. V.
2017
Determining shallow aquifer vulnerability by the DRASTIC model and hydrochemistry in Granitic Terrain, southern India
.
Journal of Earth System Science
126
(
6
),
1
23
.
Mondal
I.
,
Bandyopadhyay
J.
&
Chowdhury
P.
2019
A GIS based DRASTIC model for assessing groundwater vulnerability in Jangalmahal area, West Bengal, India
.
Sustainable Water Resources Management
5
(
2
),
557
573
.
Sadat-Noori
M.
&
Ebrahimi
K.
2016
Groundwater vulnerability assessment in agricultural areas using a modified DRASTIC model
.
Environmental Monitoring and Assessment
188
(
1
),
1
18
.
Shakoor
A.
,
Khan
Z. M.
,
Farid
H. U.
,
Sultan
M.
,
Ahmad
I.
,
Ahmad
N.
,
Mahmood
M. H.
&
Ali
M. U.
2020
Delineation of regional groundwater vulnerability using DRASTIC model for agricultural application in Pakistan
.
Arabian Journal of Geosciences
13
(
4
),
195
.
Sinha
M. K.
,
Verma
M. K.
,
Ahmad
I.
,
Baier
K.
,
Jha
R.
&
Azzam
R.
2016
Assessment of groundwater vulnerability using modified DRASTIC model in Kharun Basin, Chhattisgarh, India
.
Arabian Journal of Geosciences
9
(
2
),
98
.
Tian
H.
,
Liang
X. J.
,
Gong
Y.
,
Qi
L. L.
,
Liu
Q.
,
Kang
Z.
,
Sun
Q. F.
&
Jin
H. T.
2020
Health risk assessment of nitrate pollution in shallow groundwater: a case study in China
.
Polish Journal of Environmental Studies
29
(
1
),
827
839
.
Tian
H.
,
Liang
X. J.
,
Sun
Q. F.
,
Liu
Q.
,
Kang
Z.
&
Gong
Y.
2021
Evaluation of drinking water quality using the water quality index (WQI), the synthetic pollution index (SPI) and geospatial tools in Lianhuashan District, China
.
Polish Journal of Environmental Studies
30
(
1
),
141
153
.
Tiwari
A. K.
,
Singh
P. K.
&
De Maio
M.
2016
Evaluation of aquifer vulnerability in a coal mining of India by using GIS-based DRASTIC model
.
Arabian Journal of Geosciences
9
(
6
),
438
.
Wu
X. Y.
,
Li
B.
&
Ma
C. M.
2018
Assessment of groundwater vulnerability by applying the modified DRASTIC model in Beihai City, China
.
Environmental Science and Pollution Research
25
(
13
),
12713
12727
.
Yin
L. H.
,
Zhang
E. Y.
,
Wang
X. Y.
,
Wenninger
J.
,
Dong
J. Q.
,
Guo
L.
&
Huang
J. T.
2013
A GIS-based DRASTIC model for assessing groundwater vulnerability in the Ordos Plateau, China
.
Environment Earth Sciences
69
(
1
),
171
185
.
This is an Open Access article distributed under the terms of the Creative Commons Attribution Licence (CC BY 4.0), which permits copying, adaptation and redistribution, provided the original work is properly cited (http://creativecommons.org/licenses/by/4.0/).