1. Introduction
Water resources are basic natural resources and strategic economic resources related to the national economy and the people’s livelihood, and the controlling elements of ecological environment. Its sustainability is related to human development [
1]. The global climate change and the rapid development of urban society and economy have led to a series of water security problems such as water resource shortage, water environment pollution and water ecology deterioration, which have seriously affected the sustainable development of urban water resources and aggravated its vulnerability. Therefore, it is urgent to study the vulnerability of urban water resources, analyze its overall development trend and main influencing factors, and find its problems to propose adaptive control measures to reduce its vulnerability and promote its rational utilization and the sustainable and healthy development of urban social economy.
In terms of concept and research content, the research on water resource vulnerability originated from the field of groundwater. Albinet and Margat [
2] first proposed “groundwater vulnerability” in 1968. Later, scholars added studies on surface water and regional water environment [
3,
4,
5] and began to pay attention to the connection between human activities and water resources [
6,
7] and the influence of climate change on water resources [
8,
9,
10,
11].
In terms of vulnerability assessment indicator system, both DRASTIC method for groundwater vulnerability assessment [
12] proposed by the United States Environmental Protection Agency (USEPA) and pressure-state-response model based on composite model [
13] proposed by the Organization for Economic Co-operation and Development (OECD) have been widely applied in water resource vulnerability assessment. In terms of index system construction, indicators were initially selected from the aspect of natural characteristics, and natural factors such as buried depth of groundwater, vadose zone medium, overburden factor, runoff, rainfall, and evaporation coefficient were mainly considered [
14,
15]. Later, more attention was paid to the impact of human factors on water resource vulnerability [
16,
17,
18,
19,
20].
In terms of vulnerability evaluation methods, there were qualitative evaluation [
21] and quantitative evaluation. Qualitative evaluation means to find out the main influencing factors of water resource vulnerability by means of investigation and analysis and put forward measures to reduce water resource vulnerability. Quantitative evaluation means that researchers calculate the value of water resource vulnerability with data and models, which is more objective and accurate, mainly including function method [
22], set pair analysis method [
23], catastrophe progression method [
24], comprehensive index method [
25], fuzzy mathematical method [
26], TOPSIS method [
10], etc. In recent years, the combination of quantitative evaluation and GIS/RS technology has become a research hotspot [
27,
28,
29,
30].
For water resource vulnerability evaluation, it is the key problem to construct a comprehensive, systematic, scientific, and reasonable evaluation index system and select a practical evaluation model. Therefore, in this paper, we chose Jinan, Shandong province, as the representative of capital cities in North China that are short of water. The water resource vulnerability evaluation index system was constructed from two aspects of the natural endowment conditions and human activities. The improved analytic hierarchy process (hereinafter referred to as AHP) and entropy weight method (hereinafter referred to as EWM) were adopted to determine the combined index weights to achieve the unification of subjectivity and objectivity and make the calculation results more accurate and reliable. The comprehensive evaluation method was used to build the water resource vulnerability evaluation model. The data from 2008 to 2017 were selected to calculate the annual water resource vulnerability index of Jinan, and the evaluation results were analyzed. Combined with the water resource situation and water resource problems, the adjustment and control measures to reduce the water resource vulnerability of Jinan were explored.
2. Materials and Methods
2.1. Study Area
Jinan is the capital of Shandong Province in China (
Figure 1). It is located between latitudes of 36°01′ N and 37°32′ N and longitudes of 116°11′ E and 117°44′ E. By the end of 2017, the city covers an area of 7998 km
2, with a permanent population of 7.321 million and GDP of 720.2 billion yuan. From 1956 to 2015, the average precipitation was 671.1 mm. The water resources mainly come from atmospheric precipitation, which is characterized by insufficient total amount, great variation from year to year, uneven distribution in a certain year and uneven geographical distribution. With the rapid development of urbanization in Jinan, land use has changed significantly. The arable land, forest land and grassland that used to be natural have been converted into urban construction land. The natural landscape was replaced by industrial land, commercial land, residential land, etc. These changes have significant impacts on the temporal and spatial distribution of rainfall and runoff conditions. The comprehensive economy of Jinan remains in the forefront of Shandong Province, and it is an important comprehensive industrial city in the province. The water resources per capita are only 14.3 percent of the average of China. Therefore, it is urgent to analyze and evaluate the vulnerability degree of water resources in Jinan, to provide a basis for proposing adaptive countermeasures and promote the sustainable utilization of water resources and the sustainable development of economic society.
2.2. Construction of the Evaluation Index System
Based on the actual situation and data availability in Jinan, the evaluation indexes are divided into the natural endowment conditions and human activities in combination with the relationship between water resource reserves and demand. The natural endowment conditions (hereinafter referred to as NEC) reflect the reserves of water resources. The human activities (hereinafter referred to as HA) reflect the demand of human society for water resources. There is a certain restriction relationship between reserves and demand. When the available reserves are small, the demand cannot be fully satisfied, which means that water resources are more vulnerable. When the demand is small, there will be a surplus of available reserves, which means that water resources are less vulnerable.
The natural endowment conditions are divided into four evaluation indexes in two aspects: water resource condition and underlying surface condition (hereinafter referred to as WRC and USC). For the human activities, 14 evaluation indexes were chosen in terms of social driving force, integrated water, production water, domestic water, and ecological water (hereinafter referred to as SDF, IW, PW, DW, and EW).
There are 4 indexes in 2 aspects of NEC.
WRC. Precipitation per capita (X01) and precipitation per mu (X02) were chosen to measure the regional water resource reserves per capita and agricultural water resource reserves per mu.
USC. Land hardening area ratio (X03) and greening rate of built-up area (X04) reflect the influence on runoff of water cycle.
There are 14 indexes in 5 aspects of HA.
SDF. GDP per capita (X05) was chosen to reflect the regional economic scale and affluence degree. Per capita disposable income of urban residents (X06) was chosen to reflect the regional consumption ability from the angle of expenditure. GDP growth rate (X07) reflects economic growth rate. Labor productivity of the whole society (X08) reflects the vitality of economic growth.
IW. Water consumption per 10,000-yuan GDP (X09) and comprehensive water consumption per capita (X10) were chosen to measure the water consumption level of the whole industry and the regional per capita water consumption level.
PW. Water consumption for industrial added value of 10,000 yuan (X11) reflects the water consumption level of industry. Water consumption for agricultural added value of 10,000 yuan (X12) and irrigation water consumption per unit area (X13) reflect the water consumption level of agriculture.
DW. Per capita daily domestic water consumption (X14) and the proportion of domestic water consumption in total water consumption (X15) reflect the level of domestic water consumption.
EW. The sewage produced by industrial water and urban domestic water has a certain pressure on the ecological environment, which indirectly affects the water use of the whole industry. Therefore, three indexes are selected: wastewater discharge per 100-million-yuan GDP (X16), sewage treatment rate (X17) and the proportion of eco-environmental water consumption in the total water consumption (X18).
The evaluation index system of water resource vulnerability in Jinan is shown in
Table 1.
2.3. Sources of Data
The data comes from the water resources bulletin of Jinan, the national economy and social development statistical bulletin of Jinan, the statistical yearbook of Jinan, and relevant planning reports and statistical data of Jinan.
2.4. Comprehensive Evaluation of Urban Water Resource Vulnerability
2.4.1. The Improved Analytic Hierarchy Process
The AHP was proposed by the American Mathematician Satty T L in the 1980s. It better realizes the combination of qualitative and quantitative analysis. In the process of constructing the judgment matrix, the expert investigation method is used to make full use of the experience of the experts, and the data processing is supported by the strict mathematical theory. Therefore, this method can eliminate the influence of subjective factors to a certain degree and make the determination of weight more objective. However, in AHP, a consistency test is essential. In practical application, the judgment matrix is generally adjusted based on rough estimation. Although it is often effective, it is blind after all, and the possibility of having to go through multiple adjustments to pass the consistency test cannot be ruled out. However, in the improved AHP, the above problems are solved by using the concept of optimal transfer matrix. The consistency requirement is naturally met, and the weight values are directly calculated [
32]. The specific steps are as follows:
The system is analyzed with the improved AHP. The included factors are grouped. Each group is taken as a hierarchy. All hierarchies are arranged in the form of the highest level, a number of related intermediate levels and the lowest level. The arrange result is shown in
Table 1.
- 2.
Constructing the judgment matrix
The judgment matrix expresses the relative importance of each factor of each layer relative to a certain element of its upper layer. The constructed judgment matrix
A is as follows in Equation (1).
where
n is the total factor number of each layer relative to a certain element of its upper layer.
represents the value of relative importance of factor
to factor
in terms of element
at the adjacent upper level. It is namely the scale of importance. The “1~9” scale method is used. 1, 2, 3, …, 9 and their reciprocal can be taken as
.
equals 1 means that
is as important as
.
equals 3 means that
is slightly more important than
.
equals 5 means that
is significantly more important than
.
equals 7 means that
is strongly more important than
.
equals 9 means that
is absolutely more important than
. 2, 4, 6, 8 between them represent the median of the two adjacent judgments. The reciprocal of each number represents
is obtained when factor
is compared with factor
, and then 1/
is obtained when factor
is compared with factor
.
- 3.
Hierarchical single ordering
According to the judgment matrix A, the priority weights of factors in this hierarchy with respect to elements in the adjacent upper hierarchy are calculated, that is, hierarchical single ordering.
① According to matrix
A, matrix
B is obtained as Equation (2).
② According to matrix
B, matrix
C is obtained as Equation (3).
③ According to matrix
C, matrix
A* is obtained as Equation (4).
④ The eigenvectors of the matrix A* are calculated with square root method, and the vector W obtained by normalization is the result of hierarchical single ordering (no consistency test is required).
- 4.
Hierarchical total ordering
It is assumed that the hierarchical structure model is divided into three layers, namely, target layer
A, criterion layer
B and index layer
C. The hierarchical single ordering of all elements (
) of the criterion layer for target layer
A has been completed, and their values are respectively
. The result of hierarchical single ordering of each factor (
) at the index layer to a certain element
at the criterion layer is
, then the hierarchical total ordering, the weight of index
to target
A, can be expressed as Equation (5).
where
m is the total element number of the criterion layer.
The calculation results of the hierarchical total ordering of water resource vulnerability evaluation indexes in Jinan are shown in
Table 2:
2.4.2. Entropy Weight Method
EWM is an objective weight assignment method. When the weight is determined by EWM, the entropy value of an index is inversely correlated with the effective information amount and its weight. This method is hardly affected by subjective factors and effectively compensates for the inaccuracy of results caused by artificial reasons [
33], which can make the calculation results more authentic and reliable. The steps of EWM to determine index weight are as follows:
The raw data matrix
X can be constructed as Equation (6).
where
n and
m are the total number of samples and indexes, respectively.
The standardized judgment matrix
Y can be expressed as Equation (7). For the positive indexes,
can be obtained from Equation (8). For the negative indexes,
can be obtained from Equation (9).
where
and
are respectively the maximum and minimum values of the
ith index.
Then the entropy of the
ith index is according to Equation (10).
where
can be calculated as Equation (11).
The entropy weight
of the
ith index can be obtained from Equation (12).
The calculation results of the entropy weights of evaluation indexes of water resource vulnerability in Jinan are shown in
Table 3.
2.4.3. Comprehensive Evaluation Method
The final weights [
34] is determined with the method of combining subjective and objective weights, which can be expressed as Equation (13).
where
is the combined weight,
is the weight obtained by the improved AHP, and
is the weight obtained by EWM. Combined with the actual situation,
is taken as 0.5.
The calculation results of the combined weight of evaluation indexes of water resource vulnerability in Jinan are shown in
Table 3.
- 2.
The water resource vulnerability index
The state of the water resource system can be evaluated by its vulnerability index. The smaller the value, the better the state of the system. The water resource vulnerability index (hereinafter referred to as
VI) of the
jth sample can be calculated as Equation (14).
where
is the natural endowment condition vulnerability index and
is the human activity vulnerability index of the
jth sample. The calculation results of the water resource vulnerability indexes in Jinan from 2008 to 2017 are shown in
Table 4.
2.4.4. Water Resource Vulnerability Evaluation Threshold
Because the water resource system has the characteristics of complexity, regionality and variability, it is difficult to find the relevant fixed threshold for vulnerability classification in quantitative evaluation of regional water resource system vulnerability [
34]. In this paper, by referring to relevant literature [
25], and combining the actual situations of the study area, numerical distribution of calculated results and expert experiences, the urban water resource vulnerability is divided into five grades, as shown in
Table 5. According to
Table 5, the water resource vulnerability grade (hereinafter referred to as VG) of Jinan from 2008 to 2017 can be obtained, as shown in
Table 4.
3. Results
3.1. Index Weights Characteristics Analysis
The calculation results of the weights of water resource vulnerability evaluation indexes in Jinan by using the improved AHP, EWM and combined weight method are shown in
Figure 2.
As can be seen from
Figure 2, the maximum weight value (0.1299) and the minimum weight value (0.0106) of all indexes obtained from the improved AHP have the larger difference of 0.1193. It shows that under the subjective judgment, the influence of each index is very different. The difference between the maximum weight (0.0749) and the minimum weight (0.0370) obtained by using EWM is 0.0379, indicating that the influence of each index calculated by using the original data has little difference. The results calculated by the combined weight method are between the results of the above two methods, indicating that the combined weight method reduces the subjectivity generated by the expert scoring and the singleness of the calculation results of direct data, and increases the objectivity and accuracy of the evaluation results to a certain extent, and makes the calculation results more scientific and reasonable. Among the 18 indexes, the weight values of GDP growth rate (X07), water consumption per 10,000-yuan GDP (X09), precipitation per mu (X02) and precipitation per capita (X01) were 0.1024, 0.0875, 0.0849, and 0.0837, ranking the first, second, third, and fourth, respectively. This shows that these four indicators are the most important factors affecting the water resources in Jinan. X07 reflects the economic growth rate, and the faster the growth rate, the more serious the water resource vulnerability. X09 reflects the comprehensive water use efficiency of the whole businesses. The greater the water consumption, the more severe water resource vulnerability. X02 and X01 can represent the abundant and dry conditions in Jinan during the study period. The smaller the precipitation is, the more vulnerable the water resources are.
3.2. The Water Resource Vulnerability Characteristics Analysis
3.2.1. Characteristic Analysis of the Natural Endowment Condition Vulnerability Index
The calculation results of variation characteristics of the natural endowment condition vulnerability index in Jinan from 2008 to 2017 are shown in
Figure 3.
As can be seen from
Figure 3, from 2008 to 2017, the natural endowment condition vulnerability index of Jinan increased and decreased with fluctuations, with the minimum value being 0.057 in 2013 and the maximum value being 0.215 in 2014, and the difference between them being 0.158, which showed an increasing trend overall. Among them, it showed a decreasing trend from 2008 to 2010, 2012 to 2013 and 2014 to 2016. This is because the natural endowment condition vulnerability index is the sum of the water resource condition vulnerability index and underlying surface condition vulnerability index. From 2008 to 2017, the water resource condition vulnerability indexes in Jinan were large, with great fluctuation, and generally showed an increasing trend. The underlying surface condition vulnerability indexes were small, and the variation fluctuation was small, and showed a slightly decreasing trend overall. As a result, the variation characteristics of the natural endowment condition vulnerability index is consistent with that of the water resource condition vulnerability index. In 2014, the values of both X01 and X02 were the minimum in the series, and their weight values were the largest, so the water resource condition vulnerability index was the largest, leading to the maximum natural endowment condition vulnerability index. In 2013, the values of both X01 and X02 were the second largest in the series, and their weight values were the largest. Therefore, the water resource condition vulnerability index was the second smallest, and the natural endowment condition vulnerability index reached the minimum value when combined with the underlying surface condition vulnerability index, which was small in that year.
3.2.2. Characteristic Analysis of the Human Activity Vulnerability Index
The calculation results of variation characteristics of the human activity vulnerability index in Jinan from 2008 to 2017 are shown in
Figure 4.
As can be seen from
Figure 4, the fluctuation of the human activity vulnerability index in Jinan from 2008 to 2017 was small and showed a decreasing trend overall, except for slight increases in 2011 and 2017, with the minimum value of 0.205 in 2016 and the maximum value of 0.520 in 2008, with a difference of 0.315. The human activity vulnerability index is the sum of the vulnerability index of SDF, IW, PW, DW, and EW. From 2008 to 2017, the vulnerability indexes of SDF, IW, PW, and EW all showed a decreasing trend overall, and the values of the former three were large. Only the vulnerability index of DW showed an increasing trend overall, and its value was small, so human activity vulnerability index showed a decreasing trend overall. The slight increase in the human activity vulnerability index in 2011 was due to the increase in X09, X13 and X14 and the decrease in X17. The human activity vulnerability index increased slightly in 2017 as both X05 and X06 reached their maximum values in the series.
3.2.3. Characteristic Analysis of the Water Resource Vulnerability Index
The calculation results of variation characteristics of the water resource vulnerability index in Jinan from 2008 to 2017 are shown in
Figure 5.
As can be seen from
Figure 5, the water resource vulnerability index of Jinan increased and decreased with fluctuations from 2008 to 2017, with the minimum value being 0.287 in 2016 and the maximum value being 0.660 in 2008, and the difference value being 0.373, which showed a decreasing trend overall. However, it increased in 2011, 2014, and 2017. This is because the water resource vulnerability index is the sum of the natural endowment condition vulnerability index and the human activity vulnerability index.
From 2008 to 2017, the natural endowment condition vulnerability indexes of Jinan were small, with changes of increase and decrease and fluctuations, and generally showed an increasing trend. Among them, it showed a significant increase in 2011, 2014, and 2017.
From 2008 to 2017, the human activity vulnerability indexes of Jinan were relatively large, showing a decreasing trend overall, except for a slight increase in 2011 and 2017. In other words, the natural endowment condition vulnerability index contributes the change characteristics of ups and downs to the water resource vulnerability index, and the human activity vulnerability index contributes the overall trend to water resource vulnerability index.
The maximum value in 2008 was achieved because GDP growth rate, labor productivity of the whole society, production water and ecological water were all in the worst state.
The minimum value obtained in 2016 was because precipitation, GDP growth rate, labor productivity of the whole society, production water and ecological water were all close to the best state.
4. Discussion
According to the combined weight analysis of water resource vulnerability evaluation indexes in Jinan, the weights of X07 and X09 ranked the first and second respectively in the human activities, and the weights of X02 and X01 ranked the third and fourth respectively in the natural endowment conditions. This shows that these four indicators are the most important factors affecting the water resource vulnerability in Jinan.
From 2008 to 2017, the natural endowment condition vulnerability index of Jinan was in the range of [0.057, 0.215], showing a decreasing-increasing-decreasing-increasing-decreasing-increasing fluctuation, and generally showing an increasing trend, indicating that the natural endowment condition vulnerability degree of Jinan gradually increased. The reasons may be that X03 and X04 were relatively stable from 2008 to 2017, and their weights were relatively small, and X01 and X02 were the main factors affecting the natural endowment conditions, and the values of the two showed a trend of increasing-decreasing-increasing-decreasing-increasing-decreasing, and generally showed a decreasing trend. Therefore, the lack of precipitation resources was the key cause of the natural endowment condition vulnerability in Jinan.
From 2008 to 2017, the human activity vulnerability index of Jinan was in the range of [0.205, 0.520], and showed a decreasing trend overall. From 2008 to 2017, the strictest water resource management system was gradually implemented, and with the continuous optimization and upgrading of the industrial structure, the labor productivity of the whole society was improved. Except for the domestic water consumption, water consumptions of other industries in the city had a downward trend, and water-use efficiency was significantly improved. The discharge of wastewater was gradually decreased, the sewage treatment capacity was constantly improved, and the proportion of ecological water consumption in the total water consumption was constantly increased, which gradually reduced the human activity vulnerability in Jinan.
From 2008 to 2017, the water resource vulnerability index of Jinan was in the range of [0.287, 0.660], and generally showed a decreasing trend. The change could be divided into three stages: severe vulnerability in 2008, moderate vulnerability from 2009 to 2012 and in 2014, and mild vulnerability in 2013 and from 2015 to 2017. In Jinan, the great attention was paid to the construction of long-term water-saving management mechanism, and remarkable achievements were made in water-saving work. The economic restructuring, the level of agricultural industrialization and industrial economic returns were improved. At the same time, the water use structure was improved, and the water-use efficiency and economic development grew synchronously and promoted each other. All these made the water resource vulnerability of Jinan gradually reduce. In 2014, the precipitation was the minimum value, and the per capita daily domestic water consumption was the maximum value in the series, resulting in high water resource vulnerability in this year.
5. Conclusions
Based on the natural endowment conditions and human activities, the evaluation index system of water resource vulnerability in Jinan was constructed, and the weights were combined with the improved analytic hierarchy process and entropy weight method. Then, the comprehensive evaluation method was adopted to calculate the water resource vulnerability index, and the evaluation results were analyzed.
The GDP growth rate, water consumption per 10,000-yuan GDP, precipitation per mu and precipitation per capita are the most important factors affecting the water resource vulnerability in Jinan. It experienced a transition process from severe vulnerability to moderate vulnerability to mild vulnerability, and gradually decreased overall. These conclusions are basically consistent with the water resources situation of Jinan during the study period, which shows that the method selected is reasonable and applicable, and the calculation results can provide guidance for the sustainable management of water resources in Jinan. In the future, if the industrial structure can be further reformed and the allocation of water resources can be further optimized according to the relationship between the amount of water resources and the economic development of various industries, the water-use efficiency will be greatly improved. The level of water resource management can be improved by strengthening the construction of smart water systems. Relying on the improvement of scientific and technological level, the control and management of sewage discharge can be strengthened, the sewage treatment capacity can be improved, and the utilization of reclaimed water can be strengthened. The above measures have important practical significance for improving the ecological environment, improving the utilization rate of water resources, promoting economic development, and reducing the water resource vulnerability.
The evaluation index system of urban water resource vulnerability established in this paper contains 18 indexes mainly related to water quantity, which is still not perfect. Therefore, the actual situation of the region should be studied more extensively and deeply, water quality and water ecology and other related indicators should be further considered, the evaluation index system should be further optimized and perfected to make it more comprehensive and effective. Due to data availability, dates of 2008–2017 are chosen in this paper. The result would have been more robust if it was evaluated for longer period.
Jinan is selected as the study area in this paper. The research results can provide reference for cities with similar natural endowment conditions and human activity levels. In the future, the comparative analysis of multiple cities should be carried out to explore the impact of the natural endowment conditions, human activity levels, and other factors on the urban water resource vulnerability.