WANG Jiyng , MU Xioqin CHEN Shouji LIU Wei WANG ZhuoDONG Zhnfeng

a Department of Resources & Environment, Chengdu University of Information Technology, Chengdu, 610225, China b Institute of environmental planning, Ministry of Ecology and Environment, Beijing, 100012, China

Keywords：

Water resources carrying capacity

DSPERM framework model

Annealing-projection pursuit model

Evaluation indicator

Dianchi Lake Basin

ABSTRACT

With the intensified impact of human activities, most lakes have been severely disturbed and the lake ecosystem has been seriously damaged, which exerted a great impact on the living environment of human beings in the lake basins. The health of the lake ecosystem has gradually become one of the hot issues in recent years. In this study, the water resources carrying capacity(WRCC) was used to reveal the chain relationship between human activities and water environment in the economic development of the Dianchi Lake Basin in Kunming City of China during 2005–2015. Specifically, we chose 25 evaluation indicators related to the water environment and socialeconomic activities, classified them into six subsystems, i.e., the driving force subsystem (D), the water resources situation and consumption subsystem (S),the water resources pressure subsystem (P), the water environmental situation subsystem (E), the response subsystem (R), and the management subsystem (M), and built a comprehensive assessment system-DSPERM framework model. Simulated annealing-projection pursuit model which reflects the structure or feature of high-dimensional data was adopted to calculate the WRCC of the Dianchi Lake Basin during 2005–2015 by weighting each evaluation indicator and each subsystem of the DSPERM framework model. The results show that the WRCC of the Dianchi Lake Basin was in level III (medium carrying capacity) from 2005 to 2012. Since 2013, the WRCC has been at level II (strong carrying capacity), and from 2005 to 2015, it showed a gradual upward trend. The evaluation indicators of each subsystem varied greatly and exhibited different development trends.The indicators of the water resources pressure subsystem had the greatest impact on the WRCC, followed by the indicators of the water environmental situation subsystem and the water resources situation and consumption subsystem. We recommend that the DSPERM framework model and the simulated annealing-projection pursuit model constructed in this work can be used to analyze the dynamic changes of the WRCC over the years. They have the advantages of practicability and feasibility, and can provide the basis for the scientific decision-making and comprehensive management of regional water environment planning.

1. Introduction

With the increasing impact of human activities, most lakes are in a state of strong disturbance, and the lake ecosystem is seriously damaged, which has a great impact on human living environment. Lake ecosystem health has gradually become one of the hot issues in recent years. Water resources carrying capacity (WRCC) affects the development scale and structure of social economy, resources and population. Study on the WRCC has important theoretical guiding significance for eco-economic construction (Ouyang et al., 2015; Guo et al., 2018;Jia et al., 2019), which can provide decision-making basis for lake planning and management.

Previous studies on the WRCC were mainly focused on a single component, such as available water resources or water assimilative capacity (Jiang et al., 2001). In addition, the previous research methods were relatively single, most of these studies lack of comprehensive and quantitative research results. At present, there is no unified research method to evaluate the carrying capacity of water resources at home and abroad. Most foreign researches classified it as the field of sustainable development (Walmsley, 2002).

The WRCC is a complex system with both nature and human social attributes. Many factors, such as population,economy, society, resources availability and technology, act as positive or negative feedbacks to the WRCC(Giupponi et al., 2004; Gilmour et al., 2005). These factors have complicate nonlinear relationships with the degree of the WRCC (Forte et al., 2017; Jia et al., 2018). The contribution of each factor to the WRCC cannot be substituted and compensated by a single pressure-dimension. Therefore, how to construct an index system for evaluating the WRCC is the key to study regional WRCC.

In this study, annealing-projection pursuit model was introduced to calculate the WRCC. Annealing-projection pursuit model is a kind of statistical method for processing and analyzing high-dimensional data, where highdimensional data were projected into low dimensional subspace to find the projection reflecting the structure or characteristics of the original high-dimensional data, so as to facilitate the study and analysis of high-dimensional data. The model has the advantages of robustness, anti-interference and high accuracy, and has been widely used in many fields (Liu and Liu, 2021; Tang et al., 2021). Up to now, a series of multi-dimensional evaluation models has been established, such as three-dimensional evaluation model (Wang et al., 2017a), vigor organization resilience (VOR) for ecology evaluation (Tait et al., 2001; Niyogi et al., 2003), pressure state response (PSR)(Gentry-Shields and Bartram, 2014), driving force-state-response (DSR) (Smeets and Weterings, 1999) and driving force-pressure-state-impact-response-management (DPSIRM) (Sun et al., 2018).

However, water environment is a complex system with many influencing factors. The DSR or PSR framework model is too simple to establish a causal chain to fully reflect the causal relationship between various influencing factors and water environment. Therefore, we attempted to construct a regional WRCC evaluation index system to reflect the causal relationship between human social development and water resources.DSPERM framework model, which can separate the driving force and pressure system, the influence and state system, and the management and response system, is more in-depth than DSR and PSR models. In this study,DSPERM framework model that includes the driving force subsystem (D), the water resources situation and consumption subsystem (S), the water resources pressure subsystem (P), the water environmental situation subsystem (E), the response subsystem (R) and the management subsystem (M) was used to calculate the WRCC.DSPERM model can reflect the development of population and economy and is the vigor to promote the development of water resources. The development of vigor subsystem brings about resources pressure, changes the organization, state and elasticity of water resources, and enhances people’s management of water resources.Eventually, it promotes the coordinated development of each subsystem through strengthening management.The indicator system based on the DSPERM framework model reflects the relationship between influencing factors and water resources system. Therefore, we constructed an evaluation index system using the DSPERM framework model based on the understanding of the connotation of the WRCC and the mechanism of interaction between various elements.

In the present work, we chose the Dianchi Lake Basin in Yunnan Province of China as the research object. The Dianchi Lake is the sixth largest freshwater lake in China and the largest highland freshwater lake in Yunnan Province. It is an important water source for the survival and development of Kunming City in history, known as the ‘Plateau Pearl’. In the past 30 years, the rapid economic development of the Dianchi Lake Basin has brought tremendous pressure on the basin’s water environment. The water body is polluted and the lake eutrophication is serious, resulting in the dual dilemma of water environment pollution and water resources shortage. In this study,we used the DSPERM framework model and the annealing-projection pursuit model to calculate the WRCC of the Dianchi Lake Basin to determine the main factors affecting the lake ecosystem and analyze the effects of various factors. The study will provide scientific reference for economic development planning, ecological environment protection and sustainable utilization of water resources in the lake basin.

2. Study area and data sources

2.1. Study area

The Dianchi Lake Basin (24°40′–25°02′N, 102°36′–102°47′E) is located in Kunming City, Yunnan Province, China(Fig. 1). The altitude of the Dianchi Lake Basin is 1887.40 m; the total area is 309.50 km2, the lake volume is 1.56×109m3and the average water depth is 5.30 m. The Dianchi Lake Basin is now in a state of mild eutrophication.

Fig. 1. Geographical location of the Dianchi Lake Basin.

2.2. Data sources

The research data mainly come from the 2006–2016 statistical yearbooks of Yunnan Province (Yunnan Provincial Bureau of Statistics, 2006–2016), 2006–2016 statistical yearbooks of Kunming City (Kunming Bureau of Statistics,2006–2016), 2005–2015 Water Resources Bulletins of Yunnan Province (Yunnan Provincial Water Resources Department, 2005–2015) and 2005–2015 Statistical Communiques of Kunming City on National Economic and Social Development (Kunming Bureau of Statistics, 2005–2015). Monitoring data of total phosphorus, total nitrogen,ammonia nitrogen, permanganate index (CODMn) and five days’ biochemical oxygen demand (BOD5) are from the literature of Ouyang et al. (2015).

3. Methodology

The general methodology used in this study is displayed in Figure 2. First of all, it is necessary to find all relevant indicators that reflect the WRCC, including socio-economic indicators, water resources indicators, water quality indicators, investment management indicators, etc. Second, according to the impact of social and economic activities on water resources, we identified and adjusted the commonness and difference of the indicators based on the relevant literature. Third, by searching for the representative data of statistical yearbooks, following the principle of availability, we eliminated the indicators that cannot be obtained. Fourth, we constructed the DSPERM framework and divided the indicators into six evaluation categories to establish the evaluation index system. Last, we used the simulated annealing-projection pursuit model to evaluate the WRCC of the Dianchi Lake Basin in 2005–2015.

Fig. 2. Demonstration of the research flow chart. WRCC, water resources carrying capacity. DSPERM is a framework model that includes the driving force subsystem (D), the water resources situation and consumption subsystem (S), the water resources pressure subsystem (P), the water environmental situation subsystem (E), the response subsystem (R) and the management subsystem (M).

3.1. DSPERM framework model

The lake basin ecosystem is a system with multiple indicators. The impact of human social and economic activities on river basin is becoming more and more serious, including positive and negative impacts on water resources. There are two kinds of pressures on water resources. The first is the supply and demand of water resources, and the second is the enrichment and discharge of pollutants, which directly affect the water quality and quantity of the basin. The relationship between social and economic activities and water resources is bidirectional.Usually, human social and economic activities have negative impacts on the quality and quantity of water resources,meanwhile, when water resources are scarce and water pollution reaches a certain degree, some management measures should be taken to improve the water quality and quantity. Considering the two-way effect, an ideal index system should fully reflect the impact of socio-economic factors on water resources. So, we constructed the DSPERM framework model, as shown in Figure 2.

Based on the data from the statistical yearbooks, we identified the relevant indicators of the WRCC, which can reflect the subsystems’ attributes, including socio-economic indicators, water resources indicators, water quality indicators and investment management indicators (Kunming Bureau of Statistics, 2006–2016; Yunnan Provincial Bureau of Statistics, 2006–2016). According to the commonalities and differences of the indicators involved in this study, we adjusted, merged and eliminated the indicators to maintain the comprehensiveness,representativeness and independence of the data as much as possible (Table 1). We also explored the impacts of these indicators on the WRCC. Finally, following the principles of comprehensiveness, representativeness and independence of these indicators, we established the WRCC indicator system of the Dianchi Lake Basin based on the DSPERM framework model.

Table 1Indicator data of the Dianchi Lake Basin in 2005–2015.

For the driving force subsystem, urbanization rate is an important indicator to measure the level of economic development; per capita GDP is an important indicator to display the overall level of economic development; and population density reflects the density of urban population.

For the water resources pressure subsystem, the proportion of irrigated area to common cultivated land area and the amount of chemical fertilizer per unit cultivated land area indirectly reflect the impact of chemical fertilizer on water body, which are typical indicators to describe the non-point source pollution. Industrial wastewater discharge per 10,000 CNY industry output value, total volume of industrial wastewater discharge and consumption wastewater discharge reflect the pressure of industrial production and domestic consumption on water resources.

For the water resources situation and consumption subsystem, water resources per capita and water resources per area directly or indirectly reflect the utilization capacity of regional water resources; water consumption per 10,000 CNY GDP, industrial water consumption per 10,000 CNY industrial output value and agricultural water consumption per 10,000 CNY industrial output value all reflect the amount of water resources consumed by industrial and agricultural productions.

For the water environmental situation subsystem, total phosphorus, total nitrogen, ammonia nitrogen, CODMn and BOD5 are commonly used indicators to characterize the level of water pollution.

For the response subsystem, ratio of city sewage disposal, rate of treated industrial wastewater up to the discharge standard and recycling rate of water for industrial use represent the urban and industry wastewater discharge and reflect the city’s ecological damage degree and water resources situation. Recycling rate of water for industrial use characterizes industrial water saving efficiency; it is generally considered that 100% is the best value and 60% is the middle value for the developing countries, and for the developed countries, the value is greater than 85%. In China, recycling rate of water for industrial use in developed cities such as Beijing is higher than 90% (Beijing Municipal Bureau of Statistics, 2006–2016).

For the management subsystem, percentage of investment in environmental protection of GDP, per capita public park area and percentage of coverage by greenery and plants in developed area represent the scale of investment for water resources and environmental protection, and reflect the strength of environmental governance and the government’s emphasis on water resources management. Further, forest cover rate reflects the ecological measures of water resources protection.

Considering the present situation of serious water pollution in the Dianchi Lake Basin, and combining the experts’opinions and the existing literature on the evaluation criteria of the WRCC (Jin et al., 2006), we adjusted the evaluation criteria appropriately. The threshold of relevant water pollution assessment indicators and standards issued by local governments and planning objectives were also taken into account. According to above mentioned information, we classified the evaluation criteria of the WRCC into five levels: level I (very strong carrying capacity),level II (strong carrying capacity), level III (medium carrying capacity), level IV (weak carrying capacity) and level V (very weak carrying capacity), as shown in Table 2.

Table 2Evaluation level criteria for each indicator of the WRCC.

3.2. Simulated annealing-projection pursuit model

3.2.1. Normalization of indicators

We adopt the extreme value normalization method here, so that each indicator can be directly compared with each other.

For positive indicators shown in Table 2, normalization can be expressed as:

For negative indicators shown in Table 2, normalization can be expressed as:

where of the five evaluation levels of thejthindicator; andxjminis the minimum value of the five evaluation levels of thejthindicator.is the value of thejthindicator of theithsample;xijis the normalized indicator;xjmaxis the maximum value

3.2.2 Projected characteristic value

The projected characteristic valueZiis defined as:

whereZiis the projected characteristic value of theithsample; andajis the projection vector ofthejthindicator.

3.2.3. Optimal projection vector

The determination of the projected characteristic valueZiis the key to find the optimal projection vectora,which reflects the structure of high-dimensional data.

wheres(a) is the distance between inter classes, which was defined as the variance of the projected characteristic valueZiandZk(Zkis the projected characteristic value of thekthsample);d(a) is the inner class density;Zis the average value ofZi;rikis the distance between the projected characteristic valuesZiandZk;rmaxis the maximum value ofrik;andRis the window width of the density, which is related to data characteristics. The general range ofRis fromrmax+0.5 m tormax+2.0 m. In Equation 7,f(R–rik) is a monotone density function, which decreases with the increase ofrik.

If the distance between inter classes is large, the sample is more fragmented. The larger the inner class density,the more the significant the sample classification. Consequently, we constructed a projection objective function to get the optimal projection vector. When the projection objective function reaches the maximum, the optimal projection vector can be obtained. The projection objective functionQ(a) is calculated by the following formula:

3.2.4. Influence weight of each indicator

Using the law of data itself, the projection vectorajwas directly obtained by the above mentioned equations and optimized by simulated annealing (Yu et al., 2017; Assad and Deep, 2018). Proportion of projection vectorreflects the influence weight of each indicator on the WRCC. The influence weight of each indicator was calculated by Equation 9 on the basis of projection vectoraj.

wherewjis the influence weight of thejthindicator on the WRCC.

3.2.5. Simulated annealing

Kirkpatrick et al. (1983) successfully introduced simulated annealing method into the field of combinatorial optimization. At present, simulated annealing method is widely used in optimization problem of resources allocation,machine learning, neural networks and signal processing (Yu et al., 2017; Assad and Deep, 2018; Dhagat and Jujjavarapu, 2021; Sheik Sulaiman et al., 2021). Cooling was used to control the process of algorithm. When the value of control parameter approaches to zero, the relative optimal solution can be obtained. The optimal projection vectoris obtained by simulated annealing. The above operations were conducted using Matlab 7.0 software(MathWorks, USA; https://www.mathworks.com/).

whereTis the control parameter (°C); andtis the number of iterations. The initial temperatureT0is set as 1×1010°C,the ending temperatureTfis 0°C, and the acceptance probability formula is expressed as: exp (∆f/T)>rand, in which∆fis the increment of the optimization objective of the evaluation function, andrandis a random number between 0 and 1.

3.2.6. Sensitivity analysis

Parameters are the basis of mathematical models; they are determined using multiple sets of input or output data of the training samples. The input or output data are subject to some errors due to various factors, so the results of optimization by annealing-projection pursuit model will produce some errors. This will inevitably lead to some uncertainties in the estimated parameters of annealing-projection pursuit model, and has a certain impact on the reliability and stability of the model evaluation results. Therefore, it is necessary to analyze the sensitivity of annealingprojection pursuit model, which can be verified by performing a sensitivity analysis on the optimal projection vector.

The relationship of the relative error of annealing-projection pursuit model with the relative error of projection vector and its sensitivity can be described using Equation11.

where ΔZi/Ziisthe relative error of annealing-projection pursuit model (ΔZi=Zi–Zi0, in whichZi0is the initial value ofZi); Δajisthe relative error of projection vectoraj(Δaj=aj–aj0, in whichaj0is the initial value ofaj); andSajis the sensitivity of projection vectoraj. The value ofajcan be calculated by Equations 4–8.

When Δajis nearing to 0, Equation 11 can be expressed as:

4. Results and discussion

4.1. Optimization results of sensitivity analysis

The normalized values of the samples were input into the simulated annealing-projection pursuit model. The relative error and absolute error of projection vector ajin Equation 13 were obtained using least square regression fitting calculations, as shown in Table 3. The absolute value of relative error is less than 10% for projection vectoraj. The results suggest that the optimized annealing-projection pursuit model is reliable.

Table 3Optimal projection vector (aj) and its relative error range and absolute error range.

4.2. Influence weights of each indicator and subsystem

We calculated the sum of the influence weights of each indicator, which reflects the impact weight of each indicator on the WRCC (Fig. 3). The top 10 indicators with the greatest impact on the regional WRCC were as follows: ammonia nitrogen>total volume of industrial waste water discharge>water consumption per 10,000 CNY GDP>consumption waste water discharge>percentage of investment in environmental protection of GDP>water resources per area>population density>industrial wastewater discharge per 10,000 CNY industry output value and total phosphorus>proportion of irrigated area to common cultivated land area.

Fig. 3. Influence weights of all indicators in the six subsystems. x1, per capita GDP; x2, population density; x3,urbanization rate; x4, industrial wastewater discharged per industry output value; x5, total volume of industrial waste water discharged; x6, consumption waste water discharged; x7, amount of chemical fertilizer per unit cultivated land area; x8, proportion of irrigated area to common cultivated land area; x9, water resources per capita; x10, water resources per area; x11, water consumption per 10,000 CNY GDP; x12, industrial water level (industrial water consumption per 10,000 CNY industrial output value); x13, agricultural water level (agricultural water consumption per 10,000 CNY industrial output value); x14, total phosphorus; x15, total nitrogen; x16, ammonia nitrogen; x17,permanganate index (CODMn); x18, five days’ biochemical oxygen demand (BOD5,); x19, ratio of city sewage disposal; x20, standard treatment rate of industrial wastewater; x21, recycling rate of water for industrial use; x22,forest cover rate; x23, percentage of investment in environmental protection of GDP; x24, per capita public park area;x25, percentage of coverage by greenery and plants in developed area.

We also calculated the sum of the influence weights of each subsystem (Fig. 4), which reflects the impact weight of each subsystem on the WRCC. The influence weight value of the water resources pressure subsystem is the largest,followed by the water environmental situation subsystem and the water resources situation and consumption subsystem, indicating that these three subsystems have great influences on the integrated system (Fig. 4). To be specific, the indicators with high values of influence weight, such as industrial wastewater discharged per 10,000 CNY industry output value, total volume of industrial waste water discharge, consumption waste water discharge and proportion of irrigated area to common cultivated land area, belonged to the water resources situation and consumption subsystem. Therefore, the water resources situation and consumption subsystem had a great influence on the comprehensive evaluation system of the WRCC.

Fig. 4. Influence weights of all subsystems.

4.3. Comprehensive evaluation results of the WRCC in the Dianchi Lake Basin

Based on the grading criteria in Table 2, we calculated the assessment criteria of the WRCC, and the results are as follows: level I, WRCC value>4.3774 (very strong carrying capacity); level II, 2.3565

Fig. 5. Comprehensive evaluation results of each subsystem of the WRCC in the Dianchi Lake Basin from 2005 to 2015. I, WRCC value>4.3774 (very strong carrying capacity); II, 2.3565

The evaluation result of the WRCC in the Dianchi Lake Basin from 2005 to 2012 was in the range of level III.Since 2013, the WRCC of the Dianchi Lake Basin was in the range of level II. The comprehensive evaluation index of the WRCC of the Dianchi Lake Basin showed a gradual upward trend after 2006. The overall trend of the WRCC indicated that the WRCC of the Dianchi Lake Basin has been improved in the past 10 years, especially after 2006.

From Figure 5 it can be seen that the evaluation indicators of each subsystem of the WRCC in the Dianchi Lake Basin varied greatly and exhibited different development trends. Among the six subsystems, the water resources pressure subsystem showed the largest increase in the WRCC values (Fig. 5). That is because the evaluation indicators of this subsystem showed relatively large increases in their values (Table 1), contributing more to the combined value of this subsystem.

Among the indicators of the water resources situation and consumption subsystem, both industrial water consumption per 10,000 CNY industrial output value and agricultural water consumption per 10,000 CNY agricultural output value decreased significantly (Table 1). Industrial water consumption per 10,000 CNY industrial output value decreased from 174.01 m3/10,000 CNY in 2005 to 35.55 m3/10,000 CNY in 2015. Agricultural water consumption per 10,000 CNY agricultural output value decreased from 1310.34 m3/10,000 CNY in 2005 to 444.65 m3/10,000 CNY in 2015. In summary, the values of water consumption per 10,000 CNY GDP, industrial water consumption per 10,000 CNY industrial output value and agricultural water consumption per 10,000 CNY agricultural output value have decreased year by year (Table 1), indicating that the efficiency of water resources utilization has gradually improved.

From 2005 to 2015, the evaluation indicators of the water environmental situation subsystem showed a stable trend with slight fluctuations, which is mainly due to the increase in the concentrations of CODMnand total nitrogen, and the decrease in the concentration of total phosphorus (Table 1). Ammonia nitrogen and BOD5did not change much,and the concentration of ammonia nitrogen in 2005 was in the range of level II of the surface quality standard(Ministry of Environmental Protection of the People’s Republic of China, 2002). From 2005, the concentration of total phosphorus fluctuated between 0.12 and 0.20 mg/L, within the level III (Ministry of Environmental Protection of the People’s Republic of China, 2002). In 2005, 2007, 2014 and 2015, BOD5concentrations were better than level I, while the rest of the years were little worse than level I, indicating that BOD5is an indicator that does not need to be focused on prevention and control. From 2005 to 2015, CODMnconcentrations were better than level I. As for total nitrogen, only in 2005 and 2012, the concentrations reached the level V. This implies that total nitrogen is an indicator that needs to be controlled as a priority. The reason for the total nitrogen exceeding the standard of the Dianchi Lake Basin is that the basin has been dumped a large amount of domestic, industrial and agriculture wastewater for a long time, and the sources of water pollution were mainly domestic sewage, industrial wastewater and agricultural pollution (Gao et al., 2013). Therefore, it is very urgent to prevent and control the water pollution of total nitrogen.

The response subsystem showed an increasing trend from 2005 to 2013 with a relatively low increase rate, and thereafter a slight decrease in 2014 and 2015 (Fig. 5). The standard treatment rate of industrial wastewater increased from 98.06% in 2005 to 99.98% in 2015, and the rate of city sewage disposal rose from 67.61% in 2005 to 94.36%in 2015.

The evaluation indicator values of the management subsystem showed an increasing trend during 2005–2015.The percentage of investment in environmental protection of GDP rose from 2008 to 2013 and slightly decreased in 2014 and 2015 (Table 1). This indicates that the government, enterprises and social organizations in Kunming City have paid more attention to environmental protection (He et al., 2010; Kunming Dianchi Administration Bureau,2018), which contributed to the improvement of the water environment.

In order to determine the socio-economic regulation strategies in the Dianchi Lake Basin, it is necessary to analyze the representative indicators that contribute significantly to the changes of socio-economy. The direct impact of socio-economic development on the water environment can be reflected by the water resources pressure subsystem and the driving force subsystem.

The WRCC values of the water pressure subsystem showed an increasing trend, from 0.07 in 2005 to 0.25 in 2014(Fig. 5). Among the evaluation indicators of the water pressure subsystem, the amount of chemical fertilizer per unit cultivated land area increased during 2005–2015. As a percentage of commonly used in arable land area, its increase indicated a high pressure of agricultural surface source discharge. Industrial wastewater discharge per 10,000 CNY industry output value reached the level V. The result implies that the discharge intensity of industrial wastewater in the Dianchi Lake Basin is relatively high and the pressure on the water environment is relatively great.

For the driving force subsystem, its main constraints during 2005–2015 were population density, urbanization rate and per capita GDP, as there was a large gap between the values of these indicators and the optimal values. The values of these indicators in the driving force subsystem had steadily increased year by year (Table 1). The indicator with the largest increase was the per capita GDP, which increased from 1.77×104CNY/person in 2005 to 5.97×104CNY/person in 2015; it has exceeded the national average value after 2010 (3.00×104CNY/(person•a)), indicating a faster economic development. The ratio of city sewage disposal increased from 67.61% in 2005 to 94.36% in 2015.The percentage of investment in environmental protection of GDP was on the rise, indicating that the environmental protection efforts are increasing.

5. Conclusions

The impacts of human social and economic activities on lakes is becoming more and more serious, including positive and negative influences. Considering such a two-way effect, and in order to fully reflect the impacts of socio-economic factors on the water environment, we constructed an evaluation indicator system of DSPERM (the driving force subsystem, the water resources situation and consumption subsystem, the water resources pressure subsystem, the water environmental situation subsystem, the response subsystem and the management subsystem).The DSPERM framework model comprehensively simulated water environmental-related pressures, driving forces and other aspects; it provides new evaluation concepts for the lake ecological security assessment and lake governance in China. Annealing-projection pursuit model was employed to calculate the WRCC of the Dianchi Lake Basin. The result indicates that the WRCC of the Dianchi Lake Basin was in the level III (medium carrying capacity)from 2005 to 2012. Since 2013, the WRCC of the Dianchi Lake Basin was in the level II (strong carrying capacity),and after 2006, it showed a gradual upward trend. The overall trend suggests that the WRCC has improved in the Dianchi Lake Basin in the past 10 years, especially after 2006. The evaluation indicators of the WRCC in the Dianchi Lake Basin were quite different, showing different development trends. The Dianchi Lake Basin has received a large amount of domestic sewage and industrial wastewater for many years. The main sources of water pollution are domestic sewage, industrial wastewater and agricultural pollution. Indicators of industrial wastewater discharge per 10,000 CNY industry output value, total volume of industrial waste water discharge, consumption wastewater discharge and proportion of irrigated area to common cultivated land area all belonged to the water resources pressure subsystem; therefore, the water resources pressure subsystem had the most significant influence on the WRCC. The sum of the influence weights of each subsystem indicators can reflect the impact weight of the subsystem on the WRCC. The water resources situation and consumption subsystem, the water resources pressure subsystem and the water environmental situation subsystem have great influences on the integrated water system.The study provides scientific suggestions for economic development planning, ecological environment protection and sustainable utilization of water resources in the Dianchi Lake Basin.

In this study, we used the DSPERM model to simulate the mutual response between economic society, water resources and water environment, and reveal the interactions between human activities and water resources in the economic development of the Dianchi Lake Basin. The evaluation indicator system based on the DSPERM model is comprehensive, practical, convenient and accurate. It combined with annealing-projection pursuit model can effectively reflect the changes in the WRCC of the Dianchi Lake Basin and also provide a scientific way to evaluate the WRCC of other basins.

Declaration of competing interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Acknowledgments

This research was supported by the National Social Science Foundation of China (20&ZD091), the Science and Technology Department Project of Sichuan Province, China (21RKX0358, 2019JDJQ0006), and the Social Science Planning Project of Sichuan Province, China (SC18B027).