The pollution of aquatic environments has had many negative repercussions. Water pollution not only leads to the deterioration of water quality, threatening human health and damaging ecosystems, but also causes serious economic losses [1
]. To solve the increasingly serious problem of water pollution, it is necessary to identify the risks of water environmental pollution and take effective risk control measures [2
]. In urban water supply systems, the protection of water resource environments is very important for urban water supply and can directly affect urban public security [3
]; therefore, it is necessary to establish an assessment level for pollution risk in important water resource areas. In addition, timely control measures should be taken to curb the deterioration of water resources in areas with high levels of pollution. As a major strategic issue that cannot be ignored [4
], the protection of water resources is not only related to urban security, but is also related to the ecological environment, which is also the reason why research on the identification and grade assessment of environmental water pollution risks is carried out. Among the different types of environmental water pollution, non-point source (NPS) pollution is difficult to effectively control and prevent due to its unclear effects and complexity, having become an important type of water pollution [5
The current research on NPS pollution is mostly focused on the pollution load model, which is used to estimate the load of NPS pollution and to analyze the influencing factors of pollutant output, the best management measures for NPS pollution, the temporal and spatial distribution of pollutants, and the risk of pollution occurrence [6
]. Load models mainly include mechanical models and empirical models. Mechanical models mainly simulate hydrological processes and are used to analyze the characteristics of rainfall runoff and confluence during nutrient migration to water [9
]. These models include CREAMS [11
], ANSWERS [12
], GLEAMS [13
], AGNPS [14
], WEPP [15
], EPIC [16
], AnnAGNPS [17
], BASINS [8
], and SWAT [18
] models. Although these models can provide accurate results, the values of model parameters cannot be obtained from field data and must be determined through model calibration. The complexity and low computational efficiency of these models limit their application to a certain extent [19
On the other hand, due to their relatively small amounts of data and numbers of parameters, empirical models such as the output coefficient model (ECM) are considered more reliable methods for simulating NPS pollution [21
]. The output coefficient model, first proposed by Omernik in 1976 [22
], uses multiple linear regression analysis to establish the relationships between N (Nitrogen) and P (Phosphorus) loads and land use types in order to predict eutrophication in water bodies. On this basis, Norvell developed a relatively simple output coefficient model in 1979 to simulate the effects of nitrogen and phosphorus output on the water quality of rivers and lakes to obtain more accurate research results [23
]. Since then, the output coefficient model has been continuously improved to obtain more accurate research results. As research increased, Johannes proposed the classical output coefficient model [24
] in 1996. Compared with those of the model proposed by Omernik and Norvell, the simulation results of this model are not only more accurate, but also do not rely on a large number of data points, which are difficult to collect, reducing the costs of monitoring and modeling. This model is an effective simulation model and is currently widely used for pollution load calculations [25
Although the output coefficients of different models reflect the uniqueness of the study area, the ECM uses the same output coefficient in different areas, while terrain heterogeneity is not considered when the model is used on a large scale, which limits the ECM [28
]. Many scholars have improved the ECM, for example by including a precipitation coefficient and terrain factors in the study. It has been shown based on application results that the improved model can optimize the estimation of non-point source pollution [30
]. Compared with the traditional empirical model (ECM), the improved empirical model (IECM) shows improvements regarding output coefficients such as terrain and precipitation, which enables the model to provide more accurate calculation results and extends its applications in terms of simulating nitrogen and phosphorus pollution [31
As nitrogen and phosphorus nutrients are the main causes of water quality deterioration and water eutrophication [32
], early NPS pollution risk assessments mainly used the loss of nitrogen and phosphorus to build an index system in order to study NPS pollution [7
]; however, it has been suggested that NPS pollution is caused by many factors, including land use, runoff, and distance [34
]. At present, most studies only assess the NPS pollution risk of a basin by estimating nitrogen and phosphorus loads [36
], which is not sufficient for analysis. For the assessment of basin pollution risk caused by NPS pollution, comprehensive impact factors need to be considered [37
]. By examining the literature and taking into consideration the process of pollutant production and reduction, in this study we select nitrogen, phosphorus, soil erosion intensity, distance, slope, and rainfall levels as the main factors contributing to NPS in order to evaluate the risk of NPS pollution in watersheds [38
The universal soil loss equation (USLE) is an empirical soil loss prediction equation [41
] that quantifies average annual soil loss. It is based on experimental observation data combined with statistical analysis and generalization of soil erosion impact factors. This equation is widely used mainly because it can be multiplied using a series of simplified variables to determine soil loss in a given area [43
]; however, the usefulness of the USLE is affected by survey data, which cannot be effectively combined with soil loss data, and its applications are limited to a certain extent [44
]. Conversely, the modified universal soil loss equation improves the process of expanding field data and combining soil erosion data to allow a wider range of applications [45
The loss amounts and spatial characteristics of non-point source pollutants in the Fuxian Lake Basin are analyzed in this study. The pollution risk and spatial distribution of non-point source pollution in the lake are then explored, and the risk levels for different regions are identified. By constructing the output coefficient model and soil erosion model, the pollution loads of nitrogen and phosphorus, the levels of sediment loss, and the spatial distribution of non-point sources are estimated. Then, the risk level of non-point source pollution is evaluated on this basis and the key areas of pollution are determined by using the multi-index comprehensive evaluation method. Finally, the risk levels for non-point source pollution are classified for different regions to analyze the levels and characteristics of non-point source pollution risk in different regions. Compared with previous studies, our study considering the influencing factors of NPS pollution is more comprehensive, while the regional risk level of the basin is further analyzed. In this way, the prevention and control measures of non-point source pollution in the basin can be better implemented and the ecological environment and water resources of Fuxian Lake can be effectively protected.
4.1. Analysis of Nitrogen and Phosphorus Pollution Load Values of NPS Pollution
The improved output coefficient model was used in this study to calculate the loads of nitrogen and phosphorus from NPS pollutants generated by different pollutant types into the lake. The obtained nitrogen and phosphorus load values are shown in the following table (Table 8
The pollution load of NPS nitrogen (1058.536 t) is much higher than that of phosphorus (100.845 t). The main source of nitrogen is land use. Due to the urbanization of the Fuxian Lake Basin, a large number of rural population areas are changing to urban population areas every year; however, due to the slow speed of economic development in Western China, rural populations still account for the majority of the population. Agricultural activities conducted by a large portion of the rural population related to land use and rural life cause a large amount of nitrogen loss. The total load is 617.025 t, accounting for 58.28% of the total and reaching up to 80.85% of the overall load. This is followed by livestock and poultry breeding (17.76%) and atmospheric deposition (1.39%). The main source of phosphorus is livestock and poultry breeding, with a total load of 51.52 t, accounting for 51.08% of the overall load. The phosphorus pollution loads from land use, rural population, and atmospheric sedimentation account for 23.03%, 21.25%, and 4.64% of the overall phosphorus pollution load from all sources in the Fuxian Lake Basin, respectively.
The main pollution source contribution rate of phosphorus is different from that of nitrogen, and the biggest pollution source of phosphorus load is the use of chemical fertilizers and other fertilizers in livestock and poultry breeding. The contribution rate of livestock and poultry breeding to the phosphorus load (51.08%) is much greater than its contribution rate to the nitrogen load (17.76%). This phenomenon is mainly because 87,000 pigs were produced in the basin by the end of 2016, accounting for 5% of the total number of pigs in Yuxi City. Due to the high output of phosphorus content in the excrement of each pig (0.279 kg), animal husbandry is a key factor related to NPS phosphorus pollution in the Fuxian Lake Basin. Second, land use and rural life are also important sources of phosphorus pollution load; however, the contribution rates of rural life to the phosphorus load (22.57%) and nitrogen load (21.25%) are not significantly different.
It can be seen from the spatial distribution of the nitrogen and phosphorus load intensity levels in the basin in Figure 4
that the spatial variation trend for NPS nitrogen load intensity is high on the east and west sides and low on the north and south sides. The western, eastern, and southeastern regions of the basin are high-intensity areas for NPS pollution loss and the NPS nitrogen load intensity in these areas is relatively high. The northern and southern sides of the lake belong to the plain area of the basin. Although a certain amount of agricultural land is present in this area, there is more nitrogen in the soil and the potential nitrogen loss rate is larger in this area than in other areas; however, the nitrogen load intensity is slightly lower than on the eastern and western sides of the basin because of the simultaneous influences of topography and rainfall.
The phosphorus load intensity of the NPS pollution is highest in the northeast of the basin and the largest source of phosphorus pollution is livestock breeding. In addition, the phosphorus pollution loads caused by land use and rural life are relatively high. The western part of the Fuxian Lake Basin also has a high phosphorus load intensity, which is not only related to the contribution of the main pollution sources to the NPS phosphorus load, but is also affected by topography. The slope range of this area is mostly above 30°, while the rainfall in this area is greater than the average rainfall in the basin; thus, this area is also affected by the influencing factors of rainfall.
4.2. Analysis of the Estimated Results of NPS Sediment Loss
The soil erosion map for the Fuxian Lake Basin in 2016 calculated according to the RUSLE is shown in Figure 5
. The soil erosion modulus of the Fuxian Lake Basin for the year 2016 ranged from 0 to 43,838.2 t·km−2
, while the average erosion modulus was 1158 t·km−2
. The estimated soil erosion results obtained in this study are basically consistent with the monitoring results for soil erosion in the Fuxian Lake Basin [59
], which indicates that the estimated results in this study are of high credibility.
According to the Standard for Classification and Gradation of Soil Erosion (SL190–2007) issued by the Ministry of Water Resources of China in 2007, the soil erosion intensity in the Fuxian Lake Basin is divided into five categories. The soil erosion classification table and intensity characteristics statistics are shown in Table 8
As shown in Figure 5
, the average erosion of the basin in 2016 was 1158 t·km−2
. Combining these data with the classification standard for soil erosion (Table 9
), we concluded that Fuxian Lake Basin shows mild erosion intensity. The spatial differences in soil erosion intensity in Fuxian Lake Basin are more obvious than in other areas, with slight and micro erosion being the most widely distributed classes. Among these, micro erosion in the basin mainly occurs on the north bank, south bank, and lake coast. The terrain in these areas is relatively flat, and most of the areas comprise cultivated land, residential areas, or areas of concentrated construction land. The area of light erosion is relatively large and is generally concentrated in planar form in the northern region.
The distribution of the moderate erosion is relatively scattered. Overall, the spatial distribution characteristics are similar to the characteristics for normal-intensity and extreme-intensity erosion areas, while the distribution is banded, which is greatly affected by the topographic characteristics. The areas of normal-intensity and extreme-intensity erosion in the Fuxian Lake Basin are relatively small, but they account for 21.94% of the total erosion, which is mainly due to the large soil erosion modulus. These areas are mainly distributed in grassland and bare land areas with steep slopes.
According to the soil erosion intensity characteristics of the Fuxian Lake Basin and the analysis shown in Figure 5
, we can see that the majority of soil erosion, i.e., 75.70% of the total erosion, is caused by mild and moderate erosion. From the perspective of the erosion area, the soil erosion in the Fuxian Lake Basin mainly involves micro and light erosion, with the erosion area accounting for 52.61% of the total area of the basin. The moderate erosion area covers 80.249 km2
, accounting for 11.84% of the total area of the basin, while the total area of extreme erosion is 24.497 km2
, accounting for 2.71% of the total area of the basin.
According to the analysis of the sediment loss volume corresponding to different sediment loss intensity levels (as shown in Figure 5
and Table 9
), the main sediment loss intensity in the Fuxian Lake Basin is slight loss, while the corresponding sediment loss volume accounts for 43.39% of the total loss volume. The second most prevalent loss intensity is moderate loss (Figure 6
), accounting for 32.31% of the total loss. Sediment losses caused by moderate-intensity and extreme-intensity zones account for 13.45% and 8.49% of the total sediment loss, respectively. The micro loss area is the smallest, accounting for only 2.36% of the total sediment loss.
4.3. A Risk Assessment of NPS Pollution
The final NPS pollution risk distribution map was obtained by superimposing the six factors, i.e., nitrogen, phosphorus, sediment, distance, slope, and rainfall, as shown in Figure 7
. The distribution of the NPS risk level in Fuxian Lake as calculated using six factors is shown in Figure 6
. It can be seen from Figure 6
that the higher the risk value of the non-point source pollution, the higher the risk of non-point source pollution is. According to the spatial distribution map of non-point source pollution risk in the Fuxian Lake Basin, it can be seen that the risk values range from 0 to 1.03, while the average risk value is 0.5015.
In terms of the overall spatial distribution of NPS pollution, the risk is highest in the northern plain area, followed by the southern area. The higher risk areas are distributed in narrow strips, which are mainly related to the topography of those parts of the basin. The spatial differences in NPS pollution risk in the Fuxian Lake Basin are determined by various influencing factors.
In terms of pollution source factors, the spatial distributions of nitrogen and phosphorus factors are consistent with the characteristics of comprehensive risk distribution. The overall sediment factor value is small. Areas with larger values are mainly distributed in parts of the north and south, and their distribution is scattered. The spatial distribution of the distance factor shows a trend of higher values in the northern region of the basin than in the southern region. This is mainly because the water network in the region has a higher distance factor value and is more concentrated, while the distance from the lake is shorter, so the migration distance is smaller and the distance factor is larger. The influence characteristics of annual rainfall factors are relatively unclear. The spatial distribution presented by the slope factor is higher in the northern part of the basin than in the southern part and also higher in the western part than in the eastern part. Although the terrain in the northern region is relatively flat, its risk value is relatively high due to the higher loss intensity of nitrogen and phosphorus factors, the denser river network, and the shorter migration distance.
The risk of NPS pollution is affected not only by nitrogen and phosphorus but also by the sediment, migration distance, slope, and rainfall. Basins with high-risk characteristics for NPS pollution are close to rivers and lakes and the migration distances of their pollutants are relatively short. During the rainstorm season, the degree of pollutant loss in runoff is reduced and the ratio of pollutants entering rivers and lakes increases, thereby increasing the risk of water pollution. In the northernmost low-risk watershed area, the migration distance is relatively long, which reduces the risk of pollutants entering rivers and lakes to a certain extent. Unlike nitrogen and phosphorus pollution sources, the sediment factor, slope factor, and rainfall factor have no obvious influence on the risk of NPS pollution.
As the main form of water pollution, NPS pollution is more difficult to monitor and quantify than point source pollution, and is also more difficult to research, prevent, and control. In this study, IECM and RUSLE models were used to estimate the non-point source nitrogen and phosphorus pollution loads, the sediment loss, and their spatial distribution in the basin. On this basis, the NPS risk level and characteristics of Fuxian Lake Basin were obtained, which makes this study more valuable for reference than previous studies in terms of water source protection.
Although the export coefficient model is an important model for estimating nitrogen and phosphorus pollution [11
], rainfall has an important impact on non-point source pollution; therefore, under the premise of considering the temporal and spatial heterogeneity of rainfall factors, in this study we introduced rainfall factors and topographic factors to improve the export coefficient model, which has higher credibility for the average annual pollution load of the basin [14
]. Previous NPS risk studies focused more on the generation and transportation of nitrogen and phosphorus pollutants and less on the impact of factors such as soil loss in the NPS pollution study area; however, as is known to all that the differences in topographic slopes and soil erosion resistance levels in different basins, as well as the large-scale soil and water losses in these basins caused by human activities, can cause huge damage to the water resources [8
]. As such, based on previous studies, in this study we comprehensively considered the nitrogen and phosphorus loads, as well as the amounts of soil and water loss, and estimated the NPS risk within the spatial scope, not only making the estimated results more accurate [33
], but also providing practical guidance for water resource protection.
At present, the risk assessment of non-point source pollution tends to focus on changes of water quality [23
], the analysis of water pollution characteristics [17
], and the protection of ecological environments [41
], but less on the calculation and assessment of the risk levels of non-point source pollution loads [2
]. In addition, in the study of non-point source pollution loads, most scholars mainly focus on the study of nitrogen and phosphorus nutrients as pollution sources [13
]. Under the premise of comprehensively considering the influence of nitrogen and phosphorus nutrients and sediment as major pollutants on non-point source pollution [33
], this study establishes a comprehensive risk assessment index to assess the risk of non-point source pollution, which makes the obtained research results more reliable.
There is no doubt that this study still has some shortcomings. First, this study only discussed the ecological risk assessment of NPS pollution in the Fuxian Lake Basin in 2016 and did not study changes on a long-term scale, meaning that the results of the study are insufficient to influence environmental protection and ecological restoration; therefore, in subsequent research, data from more years will be obtained and NPS pollution in multiple years will be analyzed and simulated to determine the trend for the ecological risk of NPS pollution and its influencing factors.
Second, the migration process for NPS pollutants is extremely complex. It involves many natural and human factors, including the topography, climate, soil, rainfall, and runoff. Although this study comprehensively considered six factors and was very accurate, there is still room for improvement. In future studies, the influences of different factors on NPS pollution should be considered more comprehensively to provide further evidence for the evaluation factors of NPS pollution.
Finally, in the risk assessment and analysis of NPS pollution in this study, the analysis of water characteristics was not sufficient, as pollutants in lakes may also cause differences in NPS pollution to some extent. In future studies, relevant research should be carried out on the sensitivity of lakes to pollutants.
Considering the above deficiencies and prospects [61
], future research should focus on China’s territorial spatial planning and integrate the risk assessment of the NPS pollution of river basins into territorial spatial planning to better address ecological pollution and environmental remediation [62
This study evaluated the ecological risk of NPS pollution in the Fuxian Lake Basin to improve the empirical model and comprehensively consider the factors that affect NPS pollution. The following main conclusions were reached:
(1) The nitrogen load intensity of the Fuxian Lake Basin is greater than the phosphorus load intensity, while the distributions of NPS nitrogen and phosphorus pollution load intensities have certain spatial differences. The regions with higher nitrogen NPS load intensity are distributed in the west, east, and southeast of the basin. The phosphorus pollution load intensity of the NPS is higher in the northeast and west of the basin;
(2) In 2016, the annual average erosion modulus of the Fuxian Lake Basin was 1158 t·km−2·a−1. The average intensity of the Fuxian Lake Basin erosion is classified as mild. The intensity of the sediment loss in the Fuxian Lake Basin is mainly classified as slight, and is mostly distributed on the north bank of the lake and the construction land area along the lake. The intensity of the sediment loss is higher in grasslands and bare land area in the north, west, and southeast of the basin in areas with high topographic relief;
(3) The risk values for NPS pollution in the Fuxian Lake Basin range from 0 to 0.916, with an average risk value of 0.407. Spatially, the risk is higher in the plain area to the north of the lake, mainly due to the joint actions of nitrogen, phosphorus, and distance. Moreover, through the analysis of the risk of NPS pollution in the basin, it was found that the whole basin is primarily medium risk. The risk of NPS pollution is greatly affected by the loss of nitrogen and phosphorus and is affected by other factors to a certain extent.
Through the assessment and analysis of different NPS risks, high-risk areas should be managed and pollution control planning projects and ecological restoration work should be strengthened. Reasonable pollution control planning should be carried out in medium-risk areas to reduce the possibility of increasing pollution. Protection and preventative measures should be taken in low-risk areas and the development intensity of these areas should be rationally planned. The planning and treatment of different risk levels is conducive to NPS risk control in the Fuxian Lake Basin and has a positive guiding role in local ecological restoration and environmental protection.