Abstract
As a typical lake in the early stage of eutrophication, the non-point source pollution (NSP) in Erhai Lake was highly concerned. Since 2017, protection and rescue actions have been implemented in the Erhai Lake basin which significantly improved the water quality of Erhai Lake. But little attention has been paid to the interception effects of these actions on NSP. Based on high resolution datasets, including refinement land use/land cover (LU/LC) data, China Meteorological Assimilation Driving Database for the SWAT (CMADS), and Advanced Land Observing Satellite-1 (ALOS-12) data, the interception effects of ecological ponds newly built in 2018 on NSP was discussed with the support of the soil and water assessment (SWAT) model. These three high resolution datasets can meet the needs for simulating NSP by SWAT model, which was conductive to better reveal the interception effects of ecological ponds on NSP. Besides, the pollution load of shrubbery on Total nitrogen (TN)/Total phosphorous (TP), which has never been involved in similar research works in the Erhai Lake basin, was measured. Compared the temporal and spatial variations in TN/TP concentration before and after 2018, the interception effects of ecological ponds on NSP in the Erhai Lake basin were revealed by a sharp decrease in TN/TP concentration and a delayed presence of annual peak value about 1 month late in TN concentration before and after 2018. The interception intensity of NSP was determined by the number and volume of ecological ponds in each sub-basin and their corresponding upstream.
1 Introduction
Water environment security is the basis for social sustainable development. However, the increase in human population and worsening water environment pollution, including the point source pollution and non-point source pollution (NSP), have become more and more serious. With the rapid development of agricultural economy and the increase in food crop production for 12 successive years, China has become a large user of pesticides and fertilizers among the developing countries in the world [1]. The consequences of inappropriate behaviors in agricultural production, such as leaching of pesticides and chemical fertilizers, excretion of livestock and aquaculture industry, and the disordered discharge of rural domestic garbage, not only gradually eroded the agricultural ecological environment, but also had a huge impact on food and water security [2]. NSP has become a primary pollution source of water pollution and shows an increasing proportion in water pollution source [3,4]. More and more research efforts are being devoted to NSP, mostly on the patterns and its reduction strategies. However, it is a challenge to measure NSP and understand the mechanism of its emission due to the complexity in its formation, the uncertainty of emission direction, and the difficulty of NSP monitoring and control [5,6,7,8].
Erhai Lake, the second largest fresh water lake in the Yunnan-Guizhou Plateau, is one of the key protected lakes of the “new three lakes” in China, and it is the main drinking water source for Dali City and the surrounding areas [9]. In the last few years, the environmental problems of Erhai Lake have become increasingly serious, especially with the rapid expansion of economic crop planting and livestock breeding. Since 1990, the water quality of Erhai Lake has exhibited a downward trend [10,11] and has developed from an oligotrophic status to a mesotrophic status. Erhai Lake is considered to be in the representative preliminary eutrophic stage [9], and the NSP has become an important pollution source in Erhai Lake [3,12]. The nitrogen (N) and phosphorus (P) pollution loads of NSP (including precipitation and dust) are preliminarily estimated to be more than 80% of the total value in Erhai Lake [13].
To strengthen the protection of the Erhai Lake ecosystem further and to make Erhai Lake sustainable, schemes related to the Erhai Lake basin, lakeshore belt, and aquatic ecological protection area have been scientifically developed [14]. On April 1, 2017, the Erhai Lake basin was officially designated as the core area of a water ecological protection zone [15]. The protection and rescue actions, being named the “seven actions,” have been implemented in the Dali prefecture of Dali City and Eryuan County [16]. From 2017 to 2018, multi ecological ponds were built for the purpose of Erhai Lake water quality protection, and intercepting the rural and agricultural NSP [17]. From January to May 2018, the water quality of Erhai Lake was significantly improved and reached and stabilized in Class II. However, most research works in the Erhai Lake basin were mostly concentrated on the spatial and temporal pattern of NSP, how to improve the water quality, and protect the eco-environment in Erhai Lake before 2018 [18,19,20,21,22,23]. Little attention was paid on these treatment measures and their corresponding effects which were broadly carried out recently in the Erhai Lake basin since 2017. Only Li Dan et al. took 12 ecological ponds in three areas of the Erhai Lake basin as examples and studied the interception effects of ecological ponds on N by using spot measured data [17]. He revealed that the water quality of the effluent was significantly improved after flowing through the ecological ponds. In his research, the 12 ecological ponds being selected are located only in three towns of the Erhai Lake basin, which cannot represent and illustrate the overall characteristic of the whole basin. Similarly, the spot measured data of each ecological pond cannot explain the situation of the whole area. Besides, only the interception effect of ecological ponds on N was discussed, as to their functions on P were not explored. So, it is necessary to scientifically evaluate the treatment effects of all the protection and rescue actions on NSP intercepting, especially both N and P, that will be of great significance for the effective treatment of NSP in the Erhai Lake basin.
It comes to how to evaluate the treatment efficiency of the protection and rescue actions in the Erhai Lake basin? As too many different protection measures or schemes were carried out in the Erhai Lake basin, in our study, only the ecological ponds aiming at the NSP interception which have been finished in 2018 were taken into account. Owing to the high-altitude and hot-valley topography and a significant vertical differentiation in meteorology, limited measured site data cannot sufficiently reveal the spatial heterogeneity of NSP, so as the NSP interception efficiency of the ecological ponds all over the basin. The combination of remote sensing (RS), geographic information system (GIS), and the distributed hydrological model show prominent advantages in revealing the problems of prominent spatial heterogeneity, which had been confirmed by a number of NSP simulation studies in the Erhai Lake basin before the construction of ecological ponds in 2018 [24,25,26,27]. Among the numerous hydrologic models used for NSP simulation, the soil and water assessment (SWAT) model, developed by the Agricultural Research Bureau of the US Department of Agriculture, which comprehensively considers the impacts of meteorology, surface and underground runoff, soil types, plant growth, and agricultural management measures on the load of NSP, is the most widely applied and mature one [28,29,30,31,32].
The SWAT model is mainly driven by three datasets of land use/land cover (LU/LC), topography, and meteorology, which can directly determine the NSP simulation accuracy. The more refined and timely the driving datasets are, the more accurate the simulation results are. So far, the finest and timeliest datasets used in the SWAT model for the simulation of NSP in the Erhai Lake Basin are in the research done by Jin Chunling et al. in 2018 [33]. In her research, the LU/LC dataset from the Gaofen-1 satellite data with a spatial resolution of 16 m divided the LU/LC in the Erhai Lake basin into eight following categories: paddy field, dry land, irrigated land, rural residential area, urban land, woodland, grassland, and bare land. It is known that the pollution load of NSP varies from the changes in LU/LC. In the study of Hun-Taizi River basin, the pollution loads of Total N (TN) and Total P (TP) for arboreal forest was 1.913 and 1.389 t, while that for the shrubbery land arrived at 0.833 and 0.785 t, respectively [34]. A huge difference in the pollution load of TN and TP for these two kinds of forests could be seen. In the Erhai Lake basin, the shrubbery land occupying 36.68% of the forest land was not considered in Jin Chunling’s research on NSP, which could to some extent have some negative effects on the precision of her research on the NSP simulation in the Erhai Lake basin. In addition, the Erhai Lake basin, especially the plain areas around the lake, is extremely cut to pieces by the dense drainage network, the spatial resolution of 16 m for the LU/LC dataset cannot meet the quantification needs of the fragmented LU/LC. So, a more detailed LU/LC dataset of the Erhai Lake basin is needed to more precisely describe the causes for the spatial–temporal differentiation of NSP. Thus, LU/LC data with a more specified classification system and a higher spatial resolution of 6 m which was generated from the Gaofen-2 and Gaofen-6 satellite data were introduced into this study to more accurately evaluate the intercepting function of the ecological ponds built in 2018 on NSP in the Erhai Lake basin.
As to the meteorology dataset used to drive the SWAT model in Jin Chunling’s research, it was from the measured field data of the two meteorological stations, i.e., Dali City and Eryuan County [14]. Obviously, only two meteorological stations cannot give an accurate description of the regional climate heterogeneity in the Erhai Lake basin. The China Meteorological Assimilation Driving Database for the SWAT model (CMADS), established by Meng Xianyong in 2016 can better reflect the real meteorological field in China, especially in complex terrain regions that lack meteorological stations [35], has been widely used to drive the hydrological models of water resource studies on the Northwest China [36], Northeast China [37], and Middle China [38]. Liu Junlong et al. proved that the CMADS dataset could effectively drive the SWAT model by using it in the simulation of surface water cycle in the Erhai Lake basin where meteorological stations are scarce [39]. So, the CMADS dataset was chosen as the meteorology dataset to drive the SWAT model to accurately investigate the ecological ponds function on NSP, especially the TN and TP, in the Erhai Lake basin after 2018.
The last key dataset for SWAT model is topography data which could obviously influence the NSP simulation precision [40,41,42]. Most NSP simulation studies tend to use the digital elevation model (DEM) datasets [43,44,45], which play an important role in various applications, such as hazard monitoring, flooding prediction, resource management, etc., [44,46], with the spatial resolution greater than or equal to 30 or 90 m to drive the SWAT model, which was generated from the Shuttle Radar Topography Mission (SRTM) [47]. Inskeep et al. concluded that model predictions based on input datasets with low spatial resolution may not accurately reflect transport processes occurring in situ [40]. In a similar study, Wilson et al. concluded that simulation results were affected by the resolution of input DEM [41]. A fine resolution input data would do better than a coarse DEM data. Every effort must be made to input DEM data at a finer resolution to minimize uncertainties in the model predictions. The model output errors significantly increased with a decrease in the spatial resolution of input soil data. The absolute vertical error of the free and in common use SRTM DEM data is suggested to be generally smaller than 16 m, and the absolute error of circular geolocation is less than 20 m [47]. Such an accuracy performance usually occurs in plain and low vegetated areas. However, in regard to mountainous areas and/or densely vegetated areas, the accuracy of the SRTM DEM is significantly degraded (e.g., dozens of or even hundreds of meters in elevation) [48,49]. So the SRTM DEM dataset cannot accurately describe the morphologies of the river systems in high-relief amplitude areas or estuary regions with lower-relief amplitudes [50]. The Erhai Lake basin is such an area surrounded by high-relief mountains with a flat dam area in the middle. The mountain area is densely covered by river network, while the dam area is densely distributed by the ditches instead. A total of 117 rivers and ditches flow into the Erhai Lake. The SRTM DEM dataset with a spatial resolution of 30 or 90 m cannot accurately describe the complex hydrological situation in the Erhai Lake basin. Thus, a DEM with a higher and more consistent accuracy (horizontal and vertical) is preferable. The Advanced Land Observing Satellite-1 (ALOS) dataset with a spatial resolution of 12 m (also called ALOS-12) obtainable from the PALSAR sensor of ALOS of the Japan Aerospace Research Institute shows a higher accuracy in areas with both large and small terrain fluctuations [51]. So, the ALOS-12 DEM dataset was preferred to input into the SWAT model to simulate the topography of the Erhai Lake basin for a more accurate prediction of NSP.
Based on the fine and timely datasets of LU/LC, CMADS, and ALOS-12 DEM, the SWAT model is driven to precisely simulate the NSP of TN and TP in the Erhai Lake basin before and after 2018, when a number of ecological ponds were built for the treatment of NSP of Erhai Lake. The objectives of this study are to (i) demonstrate the positive functions of the ecological ponds on intercepting NSP; (ii) explain the spatial and temporal variations in TN and TP related to ecological pond building and others; (iii) pinpoint the sub-basins that need further treatment for an effective improvement of the water quality in Erhai Lake; And (iv) provide theoretical references for the engineering treatment of NSP in the Erhai Lake basin.
2 Materials and methods
2.1 Study area
The Erhai Lake basin (Figure 1) includes the Erhai Lake, which is the second largest freshwater lake in the Yunnan–Guizhou Plateau, and the seven main rivers that flow into the Erhai Lake, namely, the rivers of Miju, Xizha, Yongan, Luoshi, Jinxi, Mangyong, and Boluo. The Erhai Lake basin covers 17 townships and 170 administrative villages in Dali City and Eryuan County, both belonging to the Bai Autonomous Prefecture of Dali in Yunnan Province, with a total area of 2,565 km2 and a population of over 900,000, of whom 70% are in the rural areas [52]. As a typical inland faulted freshwater lake, the area of Erhai Lake arrives at 250 km2, with 117 rivers and ditches discharging into it. The northern part of the Erhai Lake is the main water supply area, also called the “Bei Sanjiang” watershed, where the Miju, Luoshi, and Yongan rivers meet. This area accounts for 72% of the total area of the Erhai Lake basin. More than 70% of the annual total water from this area flows into Erhai Lake [33]. Furthermore, this part is the main farming area of the Erhai Lake basin, with a cultivated area of approximately 1.46 × 104 hm2, accounting for 58% of the total farmland area in the whole basin. The planting areas of rice and garlic account for approximately 60 and 30% of the total farmland area of this region, respectively [53]. Garlic planting is mainly distributed along the Luoshi river [54]. The sewage discharges from rural life, agricultural planting, and livestock breeding heavily pollute the regional rivers that flow into the Erhai Lake. The 18 streams of Cangshan Mountain in the western part of the Erhai Lake basin go straight into Erhai Lake. The runoff from Piedmont Dam area flows into the Erhai Lake by way of dense ditches. The dam area which is in the middle of the basin is characterized by dense villages and concentrated agricultural and tourism activities. All these dense human activities in this region significantly aggravate the basin NSP problem. The Boluo river located in the southern part of the basin originates from Dingxi and flows into the Erhai Lake through Fengyi valley, which is a high risk area with NSP [55].
Sketch map of the study area.
2.2 SWAT model principle
SWAT is physically based and semi-distributed continuous time, hydrological, long-term simulation, and lumped parameter deterministic model within the ArcGIS interface developed by United States Department of Agriculture Research service [56,57]. The physical processes of runoff, sediment transport, pesticide metabolism, crop growth, and nutrient cycling were linked by inputting the data of meteorology, soil properties, topography, vegetation, and agricultural management measures into the SWAT model. As SWAT model is integrated with RS, GIS, DEM, and other technologies, it can be used for the efficient hydrological simulation of different regions and different spatial and temporal scales to continuously predict hydrological and water resource responses in the long term. Additionally, the hydrological process revealed by SWAT model is close to that in the objective world, and most of the input data required by the model can be obtained from government departments. For areas that lack field measured data, the applicability of the SWAT model is also effective. Given the above mentioned advantages, SWAT model has been widely used by researchers and scholars since its development [4,44,58,59,60,61,62,63,64]. The SWAT model’s simulation process can be divided into two parts. First, the runoff module simulates and calculates the input and output of water, sediment, nutrients, and chemical substances in each sub-watershed. Second, the influx module simulates the transportation and gathering process of water, sediment, nutrients, and other elements through the river network to the outlet of the watershed.
2.3 Data and processing
Two kinds of data used in the SWAT model are spatial data and attribute data. The raster spatial data used in our study include the elevation data, soil type data, and LU/LC data. The attribute data comprise the meteorological data, soil physical and chemical attribute data, pond data, and the hydrological and water quality data used for model calibration and verification. The details of all data used in NSP simulation are shown in Table 1.
Data used for the NSP simulation in SWAT model
| Serial number | Data type | Data format | Spatial resolution | Data source |
|---|---|---|---|---|
| 1 | DEM | Grid | 12.5 m | JAXA ALOS website |
| 2 | LU/LC | Grid | 2 m | Gaofen-2 and Gaofen-6 |
| 3 | Soil type | Grid | 1:1,000,000 | Harmonized World Soil Database (HWSD) |
| 4 | Meteorology | Grid | 1/4° | CMADS model ver. 1.1 |
| 5 | Hydrology | .xls | Hydrological station of Diao Caogou in Dali County | |
| 6 | Water quality | .xls | PowerChina Kunming Engineering Corporation Limited | |
| 7 | Pond | .jpg | PowerChina Kunming Engineering Corporation Limited | |
| 8 | Soil physicochemical property | HWSD and results calculated from the SPAW software | ||
| 9 | River network | Vector |
The river and township data in a vector format are auxiliary data used for the analysis of simulation results. All the input data in the SWAT model must be unified in the same geographical coordinate of WGS-1984 and the same projection coordinate of UTM. For a more precise NSP simulation result, the datasets of LU/LC and soil type were both resampled to the same spatial resolution of 12 m as ALOS-12 DEM data with the nearest neighbor interpolation method.
The existing river network data was input into the SWAT model as a preset river network data to improve the accuracy and efficiency of the river network extracting work in the SWAT model. Based on the Gaofen-2 and Gaofen-6 satellite data, the LU/LC data including 16 categories (Table 2) were obtained by manual visual interpretation and then input into the SWAT model with a spatial resolution of 6 m and a total classification accuracy exceeding 90%. The results of the LU/LC classification in the Erhai Lake basin are shown in Figure 2. The CMADS dataset directly used in the SWAT model in our study proved to be of high accuracy with the two-tailed Pearson correlation coefficient between the CMADS data and the meteorological data of Dali’s national weather station arriving at 0.970 at the p-level of 0.01.
LU/LC classification system of the Erhai Lake basin in the SWAT model
| Name | Four-digit character code of LU/LC classification in the SWAT model | Character |
|---|---|---|
| Paddy field | RICE | In the standard false color image, the paddy field is pink with a fine texture; mainly located in the western and northern parts of the Erhai Lake basin; two crops a year |
| Slope farmland | AGRR | Cultivated land with slopes >25°; one crop a year |
| Cultivated land | CORN | Dryland located in dam areas of the Erhai Lake basin with slopes <25°; excludes paddy fields |
| Forest | FRST | Mostly evergreen coniferous forests with a small number of deciduous broad-leaved forests |
| Shrubbery | FRSD | Appears orange or light yellow with a coarse texture in the standard false color image; mainly located in the eastern mountainous area of the Erhai Lake basin where water is poor |
| Orchard | ORCD | Characterized by rows in the dam area of the Erhai Lake basin |
| Grassland | HAY | Appears yellow or grey with a fine texture; located in steep slopes; has poor water conditions in the mountainous area |
| Waterbody | WATR | River, lake, ditch, reservoir, or pond |
| High density urban land | URHD | County area larger than 6,000 m2 |
| Medium density urban land | URMD | Residential land with an area of 2,000–6,000 m2 |
| Low density rural land | URLD | Rural and suburban areas with an area of less than 2,000 m2 |
| Industrial and mining land | UIDU | Appears bright white with a coarse texture; located in urban-developed areas or steep-slope mountainous area |
| Parkland | UINS | Tourist attractions and scenic spots; mainly located around the ancient town of Dali |
| Land for transportation | UTRN | Roads at all levels |
| Commercial land | UCOM | Large accommodation and restaurant areas, mainly located around the ancient town of Dali |
| Bare land | BARR | Appears white with a coarse texture |
LU/LC map of the Erhai Lake basin.
The soil database of the SWAT model constitutes the soil type’s distribution map, soil type index table, and soil attribute file. The soil type distribution map was obtained by clipping the HWSD at the scale of 1:1,000,000. The parameters needed in the soil database of the SWAT model included 24 fields that could be separated into the following five categories: irrelevant, default, conventional, software calculation, and empirical formula calculation parameters. The conventional parameters, including soil texture, occurrence layer name, occurrence layer serial number, layer relative thickness (cm), top layer depth (cm), bottom layer depth (cm), 2–0.2 mm (%), 0.2–0.02 mm (%), 0.02–0.002 mm (%), <0.002 mm (%), and organic matter (Cg/k), could be acquired directly from the HWSD. The software calculation parameters, including soil wet density, effective water holding capacity, and saturated hydraulic conductivity, were calculated from the SPAW hydrology software developed by the Washington State University. The other key parameters were soil erodibility (K factor) and soil hydrology grouping. Soil erodibility was obtained by using the Erosion Productivity Impact Calculator (EPIC) model developed by Williams in 1990 [65]. Soil hydrology grouping was determined according to the minimum infiltration rate.
Four kinds of water bodies are defined in the SWAT model: ponds, wetlands, depressions or potholes, and reservoirs. The dynamic variations in the inflows, outflows, and water volumes of the four kinds of water bodies were clearly defined to reflect their differences and to ensure that their changes could conform to their actual hydrological processes. All ecological ponds finished in 2018 are shown in Table 3, and the related parameters used to drive the SWAT model are listed below in Table 4. The ecological ponds can store sewage in rainy season and when into dry season, the stored and purified water by ecological pond wetland will be reused in the field to realize water conservation [17]. All ecological ponds and their parameters were written into the SWAT model for the NSP simulation in the period of 2018–2020 to explore the interception efficiency of these ecological ponds on NSP in the Erhai Lake basin.
Information about fifteen ecological ponds newly built in 2018
| No. for ecological ponds | Longitude | Latitude | Area (m2) |
|---|---|---|---|
| 1 | 100.13 | 25.95 | 210 |
| 2 | 100.12 | 25.98 | 1,530 |
| 3 | 100.10 | 25.96 | 490 |
| 4 | 100.27 | 25.73 | 1,200 |
| 5 | 100.26 | 25.85 | 1,170 |
| 6 | 100.32 | 25.59 | 930 |
| 7 | 100.20 | 25.69 | 520 |
| 8 | 100.13 | 25.75 | 1,040 |
| 9 | 100.16 | 25.74 | 870 |
| 10 | 100.15 | 25.84 | 1,360 |
| 11 | 100.12 | 25.89 | 1,150 |
| 12 | 100.10 | 25.91 | 530 |
| 13 | 100.11 | 25.85 | 780 |
| 14 | 100.11 | 25.94 | 920 |
| 15 | 100.13 | 25.94 | 790 |
Parameters of the ecological ponds in the Erhai Lake basin
| Variable name | Definition | Value | Unit |
|---|---|---|---|
| PND_FR | Fraction of sub-basin area that drains into ponds | Value of PND_FR should be between 0.0 and 1.0. | 0–1 |
| PND_PSA | Surface area of ponds when filled to the principal spillway | Required if PND_FR >0.0 | Hectare |
| PND_PVOL | Volume of water stored in ponds when filled to the principal spillway | Required if PND_FR >0.0 | 104 m3 |
| PND_ESA | Surface area of ponds when filled to the emergency spillway | Required if PND_FR >0.0 | Hectare |
| PND_EVOL | Volume of water stored in ponds when filled to the emergency spillway | Required if PND_FR >0.0 | 104 m3 |
| PND_VOL | Initial volume of water in ponds | Required if PND_FR >0.0 | 104 m3 |
| IFLOD1 | Beginning month of non-flood season | Required if PND_FR >0.0 | Month |
| IFLOD2 | Ending month of non-flood season | Required if PND_FR >0.0 | Month |
| NDTARG | Number of days needed to reach target storage from current pond storage | 15 (default value) | Day |
| SECCIP | Water clarity coefficient for ponds | The default value for SECCIP is 1.00, which uses the original equation | 0–1 |
Since 2003, the closure and relocation of livestock and poultry activities had been broadly implemented in the Erhai Lake basin, which resulted in a sharp decreasing number of livestock and stabilized at a low value in 2017 [66]. Therefore, the contribution of livestock and poultry breeding to NSP in the Erhai Lake basin was not considered in this study. Owing to little change in the amount of fertilization utilization in farming land before and after 2018, the fertilizer activities were not considered either in the NSP simulation in the Erhai Lake basin.
2.4 Simulation with the SWAT model
Related studies have shown that climate changes will cause corresponding changes in the runoff and water quality [53,67]. For precisely evaluating the function of ecological ponds on NSP, three simulation schemes were designed to exclude the impact of climate changes on them during the study period. In the simulation Scheme 1 (Scheme 1), no ponds were written into the SWAT model in 2015–2017. In the simulation Scheme 2 (Scheme 2), 15 newly constructed ecological ponds were written into the SWAT model in 2018–2020. In the simulation Scheme 3 (Scheme 3), none of ecological ponds were written into the SWAT model in the same period of 2018–2020 as Scheme 2. By comparing the simulation results of Scheme 1 with Scheme 3, the impact of climate changes on NSP could be determined. Comparing the simulation results of schemes 1 and 2, the effects of ecological ponds on NSP could be eventually measured.
2.4.1 Sub-basin division and simulation
According to the research on Erhai Lake basin [43], the number of sub-basins is a key parameter for ensuring a good simulation precision of the runoff in SWAT. For the Erhai Lake basin, if the number of divided sub-basins is less than 85, then the runoff simulation results will be greatly affected by the number of divisions. If the number is too high than 85, the runoff depth will be stable, but some false river networks will be generated due to the high number of narrow and long streams. So, in this study, the Erhai Lake basin was divided into 115 sub-basins and 13,813 hydrological response units, by setting the classification threshold value of LU/LC data to 10% and that of soil data to 20%.
2.4.2 Calibration and sensitivity analysis of hydrological and water quality parameters
SWAT calibration and uncertainty program (SWAT-CUP) links sequential uncertainty fitting 2 (SUFI-2), generalized likelihood uncertainty estimation (GLUE), parameter solution (ParaSol), Markov Chain Monte Carlo (MCMC), and particle swarm optimization (PSO) algorithms to SWAT [68,69], and has been widely used for calibration and sensitivity analysis [70,71]. As the SUFI-2 algorithm has gained the most popularity in parameterization, sensitivity analysis, calibration, validation, and uncertainty analysis of hydrological parameters [72], it was chosen as the calibration and sensitivity analysis algorithm in SWAT-CUP software in this study. Supported by all available datasets, the SWAT model was run first-round, respectively, in the period of 2015–2017 and 2017–2019 to get the initial values of 35 parameters, runoff and water quality items being used in the first-round iteration of SWAT-CUP. By setting the year of 2015 and 2018 as the preheating period, and the year 2016 and 2019 as the verification period, the sensitivity analysis of the parameters was iterated several times with the SUFI-2 algorithm in SWAT-CUP. In each round of running of SWAT-CUP, all the parameters will get a new value and its corresponding t-test value and p-value. The t-test value is used to assess the sensitivity level of each parameter, while the p-value is used to measure the significance level of the sensitivity of each parameter. The greater the absolute value of the t-test is, the more sensitive the parameter is, while for the p-value, it is the opposite. Parameters with t-test value greater than 1 and p-value less than 0.5 are the most sensitive parameters which will significantly influence the NSP simulation accuracy with the SWAT model [73]. In this study, nine parameters of hydrology and six parameters of water quality were determined as the most sensitive parameters to the NSP simulation with the SWAT model in the Erhai Lake basin. With all 15 parameters being assigned to the optimal values corresponding to a maximum of t-test value and a minimum of P-value shown in Table 5, the SWAT model was last run, respectively, in the period of 2015–2017 and 2018–2020 to get the final value of hydrology and water quality items in the Erhai Lake basin.
Results of parameter sensitivity analysis of the hydrology and water quality items in SWAT-CUP
| Parameters sensitive to hydrology | Parameters sensitive to water quality | |||||
|---|---|---|---|---|---|---|
| Parameter name | |t| | p-value | Parameter name | |t| | p-value | |
| GW_DELAY | 6.271 | 0.001 | GW_DELAY | 5.765 | 0.001 | |
| SOL_BD (1) | 4.035 | 0.007 | CN2 | 2.077 | 0.083 | |
| SOL_K (1) | 3.274 | 0.017 | SFTMP | 1.884 | 0.109 | |
| CH_K2 | 2.174 | 0.073 | GW_REVAP | 1.466 | 0.193 | |
| ALPHA_BNK | 1.910 | 0.105 | ESCO | 1.442 | 0.199 | |
| CH_N2 | 1.736 | 0.133 | CH_N2 | 1.374 | 0.219 | |
| CN2 | 1.490 | 0.187 | ||||
| ESCO | 1.215 | 0.270 | ||||
| ALPHA_BF | 1.119 | 0.306 | ||||
2.4.3 Evaluation of the simulation results
The determination coefficient of R 2 and the NASH coefficient named ENS were selected to evaluate the NSP simulation results. R 2 reflects the degree of agreement between the simulation results and the field measured data. When R 2 = 1, the two kinds of data are completely correlated. When R 2 < 1, the closer it is to 0, the lower the degree of agreement is. When ENS = 1, the simulated results are in good agreement with the measured data. In case R 2 > 0.6 and ENS > 0.5, the simulation results will be considered credible [73]. The NSP simulation results of monthly runoff and TN/TP from SWAT model in 2017 and 2020, and the corresponding field measured data in the same year, were input into SWAT-CUP to get R 2 and ENS. For the runoff simulation in 2017, R 2 = 0.82 and ENS = 0.73, while in 2020, R 2 = 0.83 and ENS = 0.77. For the simulation accuracy of water quality, R 2 = 0.80 and ENS = 0.64 in 2017, while in 2020, R 2 = 0.81 and ENS = 0.66. The total simulation accuracy could meet our requirements for the NSP study in the Erhai Lake basin. From Figure 3 below, the runoff and water quality field measured data and the corresponding simulation results of No. 65 sub-basin in the Erhai Lake basin could be seen.
Simulation results of runoff and water quality of the sub-basin No. 65 in the Erhai Lake basin, respectively, in 2017 and 2020: (a) runoff simulation results in 2017, (b) water quality simulation results in 2017, (c) runoff simulation results in 2020, and (d) water quality simulation results in 2020.
3 Results
From Figure 4b, it can be seen that the differences in TN/TP concentration between Scheme 1 and Scheme 3 was minute, which only accounted for 0.91 and 0.16% of the corresponding average in Scheme 1, respectively. While the differences in TN/TP concentration between Scheme 1 and Scheme 2 arrived at 26391.87 and 9354.65 kg/km2, which were 80.65 and 476.19 times of the differences between Scheme 1 and Scheme 3 (Figure 4a). It could be concluded that the climate of the Erhai Lake basin varied slightly before and after the construction of ecological ponds in 2018 which could be seen from the research of Guoying and Pei [11]. The minute changes in TN/TP concentration caused by the climate variation in this period could be completely ignored. That is to say, by comparing the TN/TP simulation results between Scheme 1 and Scheme 2, the interception effects of the ecological ponds on NSP in the Erhai Lake basin could be thoroughly determined.
Variations in TN/TP concentrations in the schemes of 2 and 3 in comparison to the scheme 1. (a) TN and (b) TP.
3.1 Temporal variations in TN/TP concentrations before and after 2018
3.1.1 Temporal variation in TN/TP concentration before and after 2018 in the overall basin
The temporal variation in TN/TP in the overall basin was measured by the TN/TP concentration at the main outlet of the Erhai Lake basin. The multi-year average of annual total amount of TN per unit area at the main outlet of the Erhai Lake basin was of 12820.42 kg/km2 in the whole of 2015–2017 when the fifteen ecological ponds had not been built. From June to September, it showed a relatively high value in TN concentration with a period average of 2435.12 kg/km2, which accounted for 19.00% of the annual total amount (Figure 5a). A double-peak morphology was seen in June and August with the two peak values being extremely close to 2973.69 kg/km2. From December to April in the next year, a low value was observed with a period average of TN concentration only arriving at 131.9 kg/km2. While in the period of 2018–2020, when 15 ecological ponds had been built and written into SWAT model, a sharp decrease by 3399.62 kg/km2 could be seen in TN concentration, which occupied 26.52% of the corresponding value before the ecological ponds being built. A unimodal morphology in September was found in TN annual variation which was obviously different from the double-peak morphology before the ecological ponds were built. The TN concentration in September of this period was only 813.83 kg/km2, which accounted for 83.2% of that in September before 2018. A low value period also appeared from December to April in the next year in TN concentration with a period average of 16.03 kg/km2 before and after 2018.
Variation in TN/TP concentration before and after 2018 (kg/km2): (a) TN and (b) TP.
The multi-year average of annual total concentration of TP per unit area at the main outlet of the Erhai Lake basin (Figure 5b) was only 31.87% of that of TN. The temporal variation in TP in the Erhai Lake basin was consistent with that of TN with their correlation coefficient arriving at 0.91 which indicated a significant correlation at the p-level of 0.01. A notable difference of a highest value (926.43 kg/km2) in June was found in TP concentration per unit area. The decrease in the amount of TP concentration after 2018, when ecological ponds were built, was 49.19% of that before 2018. That decrease proportion of TP before and after 2018 was greater than that of TN.
3.1.2 Temporal variation in TN/TP before and after 2018 in a typical sub-basin
As the variations in TN/TP concentration before and after 2018 in the overall basin were mainly embodied in the period from June to September, the analysis of TN/TP temporal variation for a typical sub-basin was conducted only in the high-value period. The No. 65 sub-basin was chosen as the typical one mainly for the following four reasons: (1) it is located in the middle of the Erhai Lake basin; (2) its area is in the middle level; (3) the number of newly built ecological ponds in this sub-basin is the maximum (six reservoirs); and (4) we collected field measured data of water quality in this sub-basin which was meaningful to our study. By comparing the TN/TP concentration per unit area in monthly scale in 2016 (2 years before the ecological ponds were built) and 2020 (2 years after the ecological ponds were built) in the No. 65 sub-basin (Figure 6), a significant decrease was found with the TN/TP concentration per unit area in 2020 only arriving at 75.12%/68.32% of that in 2016. In 2016, the peak value of the average TN/TP concentration per unit area in monthly scale appeared in June, reaching 1367.21/357.37 kg/km2, and then decreased to 726.56/222.62 kg/km2 in July, and it slightly increased from August to October with an increment of 100 kg/km2 per month. In contrast, in 2020, the peak value of average TN/TP concentration per unit area in monthly scale was not apparent until July, with a peak value of only 59.97%/61.90% of that in 2016.
Annual variation in TN/TP concentration in No. 65 sub-basin in 2016 and 2020: (a) TN and (b) TP.
3.2 Spatial variation in TN/TP concentration before and after 2018
From Figures 7–9, a similar spatial distribution of TN/TP concentration could be seen in the Erhai Lake basin before and after 2018. Thus, the spatial distribution characteristic of TN/TP concentration was only explored in the period before 2018 when the 15 ecological ponds were not yet built. As there was a sum of 115 sub-basins in SWAT simulation in this study, only sub-basins with relatively higher and lower values of TN/TP concentration were discussed. The highest value of TN/TP concentration of 2534.33/775.54 kg/km2, which was 3.56/3.65 times of the watershed’s average corresponding value (Figure 10), could be found in the No. 26 sub-basin located in the town of Sanying in Eryuan County (Figure 9). The area of the No. 26 sub-basin was somewhat small of only 14.77 km2, and it accounts for only 4.13% of the average area of all of the sub-basins. The input of TN/TP into this sub-basin was 112342.52 kg/34387.03 kg, in the same time an output of TN/TP, respectively, of 112304.21 kg/34364.93 kg was found, thus a reduction ratio of TN/TP was shown to have arrived at 0.03%/0.06% in No.26 sub-basin owing to deposition or absorption. The concentration of TN/TP in No. 44, 46, and 56 sub-basins was followed with an average concentration of TN/TP in these three sub-basins arriving at 2074.09 kg/km2/637.44 kg/km2. The TN/TP concentration in No. 37, 57, 19, and 98 sub-basins was in an similar level of 1619.49/507.89 kg/km2. The lowest annual average concentration of TN/TP was 66.92 kg/km2/0.58 kg/km2, only 9.40%/0.27% of the overall average value, which was in No. 110 sub-basin in Xiaguan Town, Dali City. A relatively low value of TN/TP concentration was derived for the No. 32, 115, 113, 114, and 38 sub-basins, with an average TN/TP concentration of 92.15 kg/km2/10.34 kg/km2.
Spatial distribution map of TN/TP concentration before 2018 (kg/km2): (a) TN and (b) TP.
Concentration of TN/TP in each sub-basin before 2018 (kg/km2).
Spatial distribution map of TN/TP concentration after the engineering implementation (kg/km2): (a) TP and (b) TN.
Spatial variation map of TN/TP concentration after the engineering implementation (kg/km2): (a) TN and (b) TP.
The average reduction in TN/TP concentration per unit area in the Erhai Lake basin was about 233.98 kg/km2/83.43 kg/km2 after 2018 compared to that before 2018. The number of sub-basins exceeding the average of TN/TP concentration was 60, accounting for 82.36% of the total number of 115. Furthermore, more than half (55%) of the 60 sub-basins were constructed with ecological ponds, indicating a remarkable interception effect of ecological ponds on NSP in the Erhai Lake basin. The maximum variation in TN/TP concentration before and after 2018 could be seen in No. 46 sub-basin from Figures 10 and 11. Compared to that before 2018, there was a notable reduction in 1289.72 kg/km2/438.47 kg/km2 in TN/TP concentration per unit area in No. 46 sub-basin in the Erhai Lake basin after 5 ecological ponds with a total area of 0.42 km2 and a capacity of 0.005 km3 were newly built in this sub-basin in 2018. Besides, numerous ecological ponds with a total area of 32.76 km2 were built in its upstream, which makes the total area of the ecological ponds in No. 46 sub-basin and its upstream arriving at 33.18 km2. In sub-basins of No. 44, 57, and 56, compared to that before 2018, there was an average reduction of 587.85 kg/km2/216.58 kg/km2 in TN/TP concentration per unit area after 2018, because only 3 ecological ponds with a total area of 1.2 km2 were newly built in these three sub-basins and their upstream. The minimum variation in TN/TP concentration before and after 2018 was found in No. 110 sub-basin, with a value of 22.26 kg/km2/0.27 kg/km2, respectively, which was only 3.49%/0.07% of that in No. 46 sub-basin. This sub-basin is located in the mountainous area of the upper reaches of the Erhai Lake basin where no new ecological reservoir was newly built in 2018.
Concentration of TN/TP in each sub-basin after engineering implementation (kg/km2).
4 Discussion
4.1 Advantages of high-resolution RS data in NSP simulation in the Erhai Lake basin
Although the Erhai Lake basin is surrounded by mountain areas, 34.75% of the total area belongs to the piedmont plain where dense river networks are developed with a total of 117 rivers [74]. Upon the precise ALOS-12 DEM data with a spatial resolution of 12 m, a total of 115 sub-basins and about 196 rivers and ditches were generated with the SWAT model. The resulted average area of sub-basins was 25.03 km2, only occupying 0.98% of the total Erhai Lake basin area. Compared to similar NSP simulation studies with the SWAT model in Erhai Lake basin, a similar average sub-basin area was found, which was around 28 km2 [33]. The sub-basin division scheme in our study cannot only take full advantages of the high spatial resolution data of ALOS-12 DEM but also avoid false water systems owing to over division [43]. The widely used CMADS dataset had also been used in SWAT model for runoff simulation in the research of Erhai Lake basin, and was proved to be capable of effectively driving the SWAT model, for its probability distribution correlation with the observed data in Dali meteorological station, arriving at more than 0.95 in Skillscore value for daily temperature and at 0.8 in Skillscore value for daily precipitation [39]. By using CMADS datasets, the NSP simulation precision with the SWAT model in the Erhai Lake-a poorly gauged region with a significant climate vertical zonality-could also be absolutely improved compared to the status of only three gauged stations [75].
For studies about LU/LC on NSP, most are concentrated either on the dominant LU/LC type, such as forest [76], grassland [77], cultivated land [78,79], etc., or on all LU/LC changes in the study areas [34,80,81]. The types of LU/LC related in these studies are mostly cultivated land, paddy land, grassland, forest land, residential land, bare land, and waterbody, while the shrubbery is rarely concerned in NSP studies so far. Based on the simulation results of NSP, the contribution of shrubbery to TN/TP concentration occupied about 6.5%/5.4% of the total value of all 9 LU/LC categories which were concerned in the SWAT model in the Erhai Lake basin. The contribution proportion of arboreal forest to TN/TP in the Erhai Lake basin was about 8.94%/8.30% of the total value of all 9 LU/LC categories. All these indicated that the contribution of shrubbery to TN/TP NSP pollution should not be ignored. Based on these three types of high precision data which could better describe the high heterogeneity of morphology, climate, and LU/LC in the Erhai Lake basin, the simulation accuracy of NSP with SWAT model can be improved, which had already been proved by a value of ENS = 0.64 and the corresponding value of R 2 = 0.80.
4.2 Temporal variation in TN/TP concentration before and after 2018
By the three simulation schemes designed in this study, the temporal variation of NSP in the Erhai Lake basin could be determined before and after 2018 when additional 15 ecological ponds were built. Nitrogen was the main pollution load in the whole basin scale or a sub-basin scale, with the annual total concentration of TN per unit area being 3.74 times of that of TP. A similar conclusion, with the equal-standard discharge volume of TN and TP arriving at 1,371.29 m3/a and 420.10 m3/a, was found in the research carried by Lu et al. (2017) in the Erhai Lake basin [82]. This is because the source of nitrogen is more than that of phosphorus. The source of TN originates not only from the transformation of organic matter, but from the transformation of nitrite, nitrate, and ammonia nitrogen. While that for TP is only from the transformation of organic matter. The multi-year average of annual total amount of TN and TP per unit area at the main outlet of the Erhai Lake basin declined, respectively, from 12820.42 and 4273.47 kg/km2 to 7541.42 and 2374.15 kg/km2 by comparing that in 2015–2017 to that in 2018–2020 when 15 ecological ponds were newly built. Thus, the intercepting effects of the ecological ponds on NSP, especially on TN/TP could be determined. The sewage flowing into the ecological ponds was purified obviously with the TN concentration varied from 1.35–23.54 mg/L in the ecological pond inlets to 0.26–10.62 mg/L in the outlets [17]. This is because the nitrogen and phosphorus in the sewage could be degraded through the biochemical and non-biochemical processes in the ecological ponds, including plant absorption, natural degradation, microbial degradation, substrate absorption, biological assimilation, etc., [83]. The intercepting effect of the fifteen ecological ponds on NSP in the Erhai Lake basin was enlarged by the peak value of TN/TP concentration per unit area being shown nearly 1–3 months late. A similar phenomenon was found in sub-basin No. 65 for 6 ecological ponds which were newly built in 2018.
Climate change is one of the main factors affecting NSP. Precipitation and temperature are the two dominant factors impacting streamflow and NSP loads [84]. Precipitation in particular has the greatest influence on stream and sediment, whereas temperature influences nitrogen and phosphorus concentrations in surface water [85]. Although, the climate variation before and after 2018 was too slight to be neglected on the variations in NSP during the same period in the Erhai Lake basin, owing to the obvious inter-annual variability in precipitation and temperature in monthly scale, a relatively high value of TN/TP concentration per unit area could be seen from May to October which is the rainy season of the Erhai Lake basin. The peak values of TN concentration shown in June and September was mainly affected by high values of precipitation and temperature at 168 mm and 35°C, respectively, which were 2.3 and 1.4 times of their corresponding annual averages.
4.3 Spatial variation in TN/TP concentration before and after 2018
From Figure 9, the spatial variation in TN/TP before and after 2018 showed a similar performance with a relatively high value of TN/TP concentration per unit area mainly concentrated in the northern areas of the Erhai Lake basin, which was consistent with the conclusions drawn by Xiang Song et al. in 2020 [86]. Besides the precipitation discussed above, the spatial variation in NSP in the Erhai Lake basin was mainly dominated by agriculture NSP which could be measured by different LU/LC types [87]. The contribution of different LU/LC types to N/P pollution load is as follows: cultivated land > paddy land > orchard > construction land > traffic land > grassland > shrubbery > forest [88,89]. According to the study of Tong et al., the construction land including the URHD/URMD/URLD and UTRN increases the coverage percentage of impervious surface in the Erhai Lake basin, which can aggravate the pollution load of nitrogen and phosphorus into the Erhai Lake [88]. The topography and slope are two additional factors that will influence the pollutant deposition in a watershed [90]. For the individual sub-basins mentioned in Section 3.2, a relatively high ratio of CORN with 35.57% of the total sub-basin area being covered by CORN could supply a large number of pollution source of nitrogen and phosphorus according to the NSP research of Tong et al. in the Erhai Lake basin [88]. Besides, the terrain of this sub-basin was relatively flat with a proportion of 86.2% below 20° which was adverse to the export of nitrogen and phosphorus pollutions in this sub-basin. As to the three sub-basins with a relatively higher concentration of TN/TP, namely the No. 44, 46, and 56 sub-basins, a similar phenomenon could be seen with a relatively high proportion of farmland arriving at 37.23%, among which, 21.19% was CORN and 15.15% was RICE. In sub-basins of No. 37, 57, 19, and 94, a high total proportion of farmland and residential land arriving at 44.69%, and a flat terrain with 71.23% of the area below 20° in slope were found which could well explain the higher concentration of TN/TP in these sub-basins. In adverse, these sub-basins with relatively low value of TN/TP concentration per unit area were mostly covered by forest land and showed steep terrain. 64.16% of the sub-basin of No. 110 with the lowest concentration of TN/TP was covered by forest and grassland. And the terrain of this basin was so steep with 89.32% of the area above 20°. The sub-basins of No.44, 57, and 56 with a relatively low concentration of TN/TP showed a steep terrain with nearly 85% of areas greater than 20°, and a high proportion of more than 75% being covered by shrubbery, evergreen forest, and grassland. It could be concluded that the spatial variation in TN/TP concentration in the Erhai Lake basin was mainly decided by LU/LC and topography.
From the average reduction amount of TN/TP concentration per unit area, we could conclude that the fifteen ecological ponds newly built in the Erhai Lake basin in 2018, indeed, played a significant role in intercepting TN/TP and purifying water quality. The spatial variation in the interception efficiency of these ecological ponds on NSP depended on the number of ecological ponds newly built in each sub-basin and its upstream sub-basins. The greater the number and area of ecological ponds that are being newly built, the greater is the reduction efficiency on TN/TP pollution load.
5 Conclusion
As a typical lake in the early stage of eutrophication, the NSP problem in Erhai Lake is extremely complicated. Since 2017, protection and rescue actions have been taken in the Erhai Lake basin, especially in 2018 when numerous ecological ponds were built. All these actions have significantly improved the water quality of Erhai Lake. Given that most studies on Erhai Lake were being focused on NSP simulation and its variation, few studies paid attention to the interception effects of protect and rescue actions on NSP in the Erhai Lake basin. On the basis of high-resolution datasets, the NSP was simulated with the SWAT model in the Erhai Lake basin. By comparing the temporal and spatial variations in the TN/TP concentrations before and after 2018, the effectiveness of the ecological ponds on pollution interception around the lake was revealed. This study can provide a scientific theoretical basis for the ecological engineering construction for NSP treatment in the Erhai Lake basin. The following conclusions can be drawn from this work:
High precision of dataset will absolutely improve the simulation accuracy of NSP in the Erhai Lake basin. With a high resolution of DEM data, the complex river network in the Erhai Lake basin could be precisely represented. The advantages of CMADS datasets in NSP simulation was obvious in poor-gauged region. With a high spatial resolution and a refinement LU/LC classification system, the contribution of LU/LC on NSP will be accurately measured.
Owing to more sources of TN than that of TP, a relatively large amount of difference could be seen in TN and TP concentration per unit area, no matter in basin scale or sub-basin scale. But the temporal variations in TN and TP in monthly scale were notably similar and both showed a double-peak morphology in June and August because of abundant precipitation and high temperature in the same period before 2018. After 2018, when 15 ecological ponds were newly built in the Erhai Lake basin, not only the concentration of TN/TP pre unit area decreased significantly, but a unimodal morphology of temporal variation in TN/TP was shown with the peak value of TN delayed by 1 month. The sewage interception effects of the ecological ponds on NSP were proved in the Erhai Lake basin.
The spatial variation of the annual average TN/TP concentration before and after 2018 both showed a highest value in the No. 26 sub-basin and a lowest value in the No. 110 sub-basin, which were owing to great differences in LU/LC and topography between these two sub-basins. A flat terrain coupled with a high proportion of “source” type of LU/LC will absolutely aggravate NSP in the Erhai Lake basin, and vice versa.
The interception effects of ecological ponds on TN/TP concentration per unit area were determined by the number of ecological ponds being built in each sub-basin and their upstream. The six new ecological ponds built in the No. 46 sub-basin and other ponds in its upstream resulted in the largest reduction in TN/TP concentration in it after 2018. On the contrary, the absence of ecological ponds in the No. 110 sub-basin and even in its upstream deduced the lowest reduction in TN/TP concentration in the No. 110 sub-basin.
Because the main purpose of this article was to discuss the effects of ecological ponds on NSP in the Erhai Lake basin with the support of SWAT model, the influence of high-resolution datasets on NSP simulation results in the Erhai Lake basin was not thoroughly elaborated. With more and more RS data and assimilation data arising, the research on NSP simulation upon SWAT model driven by different precision datasets should be carried out to thoroughly reveal the effects of different datasets in NSP simulation with SWAT model, especially in regions with significant spatial heterogeneity of topography and meteorology. Along with the implementation of the protection and rescue actions named the “seven actions,” great changes have taken place in the Erhai Lake basin, especially in aspects of LU/LC which could remarkably influence the NSP variation in Erhai Lake basin. With the support of high resolution RS data, especially GF data, a precision study of LU/LC variation in NSP should be carried out. Constrained by limited ecological ponds information, only fifteen newly built ecological ponds were discussed in this article. In the future, more and more information about the ecological ponds, not only related to their numbers, location and volume but the planted vegetation and their corresponding monitoring data, will also be added in the SWAT model to thoroughly discuss the ecological ponds effect on NSP interception in the Erhai Lake basin
-
Funding information: 1. Key Technology Project from Power Construction Corporation of China (DJ-ZDXM-2018-38); 2. Key Research Program of Yunnan Province (2019BC002); 3. The Technology Innovation Research and Development Project of Chengdu Science and Technology Bureau，2019-YF05 -02641-SN; 4. Key Research Program of Yunnan Province (2019BC002); 5. Key Technology Project from Power Construction Corporation of China (DJ-ZDXM-2018-38).
-
Author contributions: data curation: W.C. and Y.Q.; formal analysis: W.C., Y.Q., L.XN., C.J., F.R., and L.S.; funding acquisition: W.C., F.R., and L.S.; methodology, Y.Q., L.XN., and C.J.; writing – original draft: L.XN., Y.Q., and W.C; writing – review and editing: C.J., F.R., L.S., and L.J. All authors have read and agreed to the published version of the manuscript.
-
Conflict of interest: The authors declare that they have no conflict of interest.
-
Institutional review board statement: Not applicable.
-
Informed consent statement: Not applicable.
-
Data availability statement: All data that supports the findings of this study could be seen from Tab1 in the article. Data types of DEM, Soil type, Meterology, Soil physicochemical property are openly availabile in websits which were listed in the data source column in Tab1. As to data types of LULC, water quality, pond and river network, they are available from Power China Kunming Engineering Corporation Limited. That kinds of data are not publicly available due to privacy restrictions.
References
[1] He X. Modernization of national food security governance system and governance capacity. Chin Rural Econ. 2020;6:12–5.Search in Google Scholar
[2] Wittman H, Chappell MJ, Abson DJ, Kerr RB, Blesh J, Hanspach J, et al. A social-ecological perspective on harmonizing food security and biodiversity conservation. Regional Environ Change. 2017;17:1291–301.10.1007/s10113-016-1045-9Search in Google Scholar PubMed PubMed Central
[3] Cao L. Changing port governance model: Port spatial structure and trade efficiency. J Coast Res. 2020;95:963–8.10.2112/SI95-187.1Search in Google Scholar
[4] Liu HT, Chen JC, Zhang L, Sun KX, Cao WZ. Simulation effects of clean water corridor technology on the control of non-point source pollution in the Paihe River basin, Chaohu Lake. Env Sci Pollut R. 2021;28:23534–46.10.1007/s11356-020-12274-xSearch in Google Scholar PubMed
[5] Zhang W, Wu S, Ji H-J, Kolbe H. Estimation of agricultural non-point source pollution in China and the alleviating strategies I. Estimation of agricultural non-point source pollution in china in early 21 century. Sci Agric Sin. 2004;37:1008–17.Search in Google Scholar
[6] Collins AL, Zhang YS, Winter M, Inman A, Jones JI, Johnes PJ, et al. Tackling agricultural diffuse pollution: What might uptake of farmer-preferred measures deliver for emissions to water and air? Sci Total Env. 2016;547:269–81.10.1016/j.scitotenv.2015.12.130Search in Google Scholar PubMed
[7] Hao FH, Yang ST, Cheng HG, Bu QS, Zheng LF. A method for estimation of non-point source pollution load in the large-scale basins of China. Acta Sci Circum. 2006;26:375–83.Search in Google Scholar
[8] Sidemo-Holm W, Smith HG, Brady MV. Improving agricultural pollution abatement through result-based payment schemes. Land Use Policy. 2018;77:209–19.10.1016/j.landusepol.2018.05.017Search in Google Scholar
[9] Zhang J, Zhi M. Effects of basin nutrient discharge variations coupled with climate change on water quality in Lake Erhai, China. Env Sci Pollut R. 2020;27:43700–10.10.1007/s11356-020-09179-0Search in Google Scholar PubMed
[10] Zheng G, Yu X, Jiang N, He L. Variation process and trend of Erhai Lake water quality. J Northeast Forestry Univ. 2004;32:99–102.Search in Google Scholar
[11] An G, Guo Z, Ye P. Climate impacts and its effects on water quality of Erhai Lake during the period of 1989–2019 in Dali area, Yunnan. Geoscience. 2022;36:406–17.Search in Google Scholar
[12] Luo P, Cao H. Temporal and spatial variations of TN and TP in different water periods of Erhai Lake. J Yangtze River Sci Res Inst. 2021;38:14–9.Search in Google Scholar
[13] Husain A, Muneer MA, Wu F, Yin G, Shen S, Wang F, et al. Application of optimum N through different fertilizers alleviate NH4+ -N, NO3-N and total nitrogen losses in the surface runoff and leached water and improve nitrogen use efficiency of rice crop in Erhai Lake basin, China. Commun Soil Sci Plan. 2019;50:716–38.10.1080/00103624.2019.1589482Search in Google Scholar
[14] Lin S, Shen S, Zhou A, Lyu H. Sustainable development and environmental restoration in Lake Erhai, China. J Clean Prod. 2020;258:120758.10.1016/j.jclepro.2020.120758Search in Google Scholar
[15] Li W, Ma L, Zang Z, Gao J, Li J. Construction of ecological security patterns based on ecological red line in Erhai Lake basin of Southwest China. J Beijing For Univ. 2018;40:85–95.Search in Google Scholar
[16] Cao Z. Influence of “seven acivities” of Erhai treament on the development of passenger station industry. China Mark. 2019;3:72–3.Search in Google Scholar
[17] Li D, Chu Z, Liu Y, Huang M, Du B, Zheng B. Characteristics and effects of nitrogen interception in the ecological multi-pond wetlands in Erhai Lake sub-basin. Res Environ Sci. 2019;32:212–8.Search in Google Scholar
[18] Peng W, Wang S, Liu X. Review on evolution of pollutant control measures in the Erhai Lake basin. J China Inst Water Resour Hydropower Res. 2005;3:95–9.Search in Google Scholar
[19] Guo HC, Liu L, Huang GH, Fuller GA, Zou R, Yin YY. A system dynamics approach for regional environmental planning and management: A study for the Lake Erhai basin. J Env Manage. 2001;61:93–111.10.1006/jema.2000.0400Search in Google Scholar PubMed
[20] Wang X, Deng Y, Tuo Y, Cao R, Zhou Z, Xiao Y. Study on the temporal and spatial distribution of chlorophyll a in Erhai Lake based on multispectral data from environmental satellites. Ecol Inform. 2021;61:101201.10.1016/j.ecoinf.2020.101201Search in Google Scholar
[21] Tang Q, Ren T, Wilko S, Liu H, Lei B, Lin T, et al. Study on environmental risk and economic benefits of rotation systems in farmland of Erhai Lake basin. J Integr Agric. 2012;11:1038–47.10.1016/S2095-3119(12)60096-3Search in Google Scholar
[22] Zhu R, Wang H, Yu D, Zeng C, Shen H, Chen J. Dynamic changes of microcystins and phytoplankton during the cyanobacterial bloom in Lake Erhai in 2013. J Lake Sci. 2015;27:378–84.10.18307/2015.0302Search in Google Scholar
[23] Pan X, Lin L, Huang Z, Liu M, Dong L, Chen J, et al. Distribution characteristics and pollution risk evaluation of the nitrogen and phosphorus species in the sediments of Lake Erhai, Southwest China. Env Sci Pollut R. 2019;26:22295–304.10.1007/s11356-019-05489-0Search in Google Scholar PubMed
[24] Li X, Dong L. Study on agricultural non-point source pollution load and control measures of Erhai Lake basin. Hubei Agric Sci. 2011;50:3535–9.Search in Google Scholar
[25] Zhai Y, Shang X, Shen J, Wang X. Application of SWAT model in agricultural non-point source pollution investigation in Lake Erhai watershed. Res Environ Sci. 2012;25:666–71.Search in Google Scholar
[26] Xu Q, Yang K, Li J, Yang Y. Agent-based modeling of land use change and simulation of response on nonpoint source pollution for Erhai Lake watershed. Shuili Xuebao. 2014;45:1272–83.Search in Google Scholar
[27] Zhang C, Lu J, Chen X. Analysis of the economic circle trade work of article shipping routes based on structural holes theory. J Cent China Norm Univ. 2017;51:108–14.Search in Google Scholar
[28] Zhang P, Liu Y, Pan Y, Yu Z. Land use pattern optimization based on clue-s and SWAT models for agricultural non-point source pollution control. Math Computer Model. 2013;58:588–95.10.1016/j.mcm.2011.10.061Search in Google Scholar
[29] Zhou CW, Yu LF, Zhou Y, Yan LB. Hydrological and ecological effect of Caohai watershed regulation project based on SWAT model. Appl Ecol Environ Res. 2019;17:427–36.10.15666/aeer/1701_427436Search in Google Scholar
[30] Wang XM, Yang F, Yang S, Chen X, Yuan Q, Ye J. Study on non-point source nitrogen and phosphorus pollution and its law of accumulation in villages and towns. Agro Food Ind Hi Tec. 2017;28:2005–9.Search in Google Scholar
[31] Knapen A, Poesen J, Galindo-Morales P, De Baets S, Pals A. Effects of microbiotic crusts under cropland in temperate environments on soil erodibility during concentrated flow. Earth Surf Proc Land. 2007;32:1884–901.10.1002/esp.1504Search in Google Scholar
[32] Khalid K, Ali MF, Abd Rahman NF. The development and application of Malaysian soil taxonomy in SWAT watershed model. 2015;79:77–88.10.1007/978-981-287-365-1_7Search in Google Scholar
[33] Jin C. Study on non-point source pollution in the western and northern regions of Erhai basin using SWAT model. Master dissertation. Beijing, China: Chinese Research Academy of Environment Sciences; 2018.Search in Google Scholar
[34] Heng Z, Miao L, Chun L, Sun L, Ling J. Non-point source pollution simulation under land use change scenarios in Hun-Taizi River watershed. Ying Yong Sheng Tai Xue Bao = J Appl Ecol. 2016;27:2891–8.Search in Google Scholar
[35] Meng X, Wang H, Chen J. Profound impacts of the China meteorological assimilation dataset for SWAT model (CMADS). Water-Sui. 2019;11:832.10.3390/w11040832Search in Google Scholar
[36] Liu J, Liu S, Shangguan D, Xu J. Applicability evaluation of precipitation datasets from CMADS, ITPCAS and TRMM 3B42 in Yurungkax river basin. J North China Univ Water Resour Electr Power(Natural Sci Ed). 2017;38:28–37.Search in Google Scholar
[37] Zhang L, Wang H, Meng X. Application of SWAT model driven by CMADS in Hunhe river basin in Liaoning province. J North China Univ Water Res Electr Power (Natural Sci Ed). 2017;38:1–9.Search in Google Scholar
[38] Gao X, Zhu Q, Yang Z, Wang H. Evaluation and hydrological application of CMADS against TRMM 3B42V7, PERSIANN-CDR, NCEP-CFSR, and gauge-based datasets in Xiang River basin of China. Water-Sui. 2018;10:1225.10.3390/w10091225Search in Google Scholar
[39] Liu J, Xu J, Yuan Z, Zhou M. Application of SWAT model driven by CMADS in water cycle simulation of Erhai basin, Yunnan province. Yangtze River. 2020;51:65–72.Search in Google Scholar
[40] Inskeep WP, Wilson WJ, Snyder JP, Macur RD, Gaber RE, Input HM. Parameter and model resolution effects on predictions of solute transport. J Environ Qual. 1996;25:453–62.10.2134/jeq1996.00472425002500030011xSearch in Google Scholar
[41] Wilson JP, Wraith IW, Snyder JM, Gis-based RD. Solute transport modeling applications: Scale effects of soil and climate data input. J Environ Qual. 1996;25:445–53.10.2134/jeq1996.00472425002500030010xSearch in Google Scholar
[42] Chaubey, Cotter AS, Costello TA, Soerens TS. Effect of DEM data resolution on SWAT output uncertainty. Hydrol Process. 2005;19:621–8.10.1002/hyp.5607Search in Google Scholar
[43] Li M, Han H, Liu X, Hao W. Study on optimal sub-division scheme of watershed using SWAT model—a case study in Erhai basin, Yunnan province. Chin J Agrometeorol. 2012;33:185–9.Search in Google Scholar
[44] Nantasaksiri K, Charoen-Amornkitt P, Machimura T. Land potential assessment of Napier grass plantation for power generation in Thailand using SWAT model. Model validation and parameter calibration. Energies. 2021;14:1326.10.3390/en14051326Search in Google Scholar
[45] Nasab MT, Grimm K, Bazrkar MH, Zeng L, Shabani A, Zhang X, et al. SWAT modeling of non-point source pollution in depression-dominated basins under varying hydroclimatic conditions. Int J Env Res Pub He. 2018;15:2492.10.3390/ijerph15112492Search in Google Scholar PubMed PubMed Central
[46] Singh P, Gupta A, Singh M. Hydrological inferences from watershed analysis for water resource management using remote sensing and GIS techniques. Egypt J Remote Sens Space Sci. 2014;17:111–21.10.1016/j.ejrs.2014.09.003Search in Google Scholar
[47] Farr TG, Rosen PA, Caro EA, Crippen R, Duren R, Hensley S, et al. The shuttle radar topography mission. Rev Geophys. 2007;45:RG2004/200710.1029/2005RG000183Search in Google Scholar
[48] Berthier E, Arnaud Y, Vincent C, Remy F. Biases of SRTM in high-mountain areas: Implications for the monitoring of glacier volume changes. Geophys Res Lett. 2006;33:L08502.10.1029/2006GL025862Search in Google Scholar
[49] Weydahl DJ, Sagstuen J, Dick OB, Ronning H. SRTM DEM accuracy assessment over vegetated areas in Norway. Int J Remote Sens. 2007;28:3513–27.10.1080/01431160600993447Search in Google Scholar
[50] Wang H, Liu J, Xie Z, Wang W. Hydrological application-based analysis on precision of multi-source digital elevation model. Water Resour Hydropower Eng. 2018;49:30–8.Search in Google Scholar
[51] Courty LG, Soriano‐Monzalvo JC, Pedrozo‐Acuña A. Evaluation of open‐access global digital elevation models (AW3D30, SRTM, and ASTER) for flood modelling purposes. J Flood Risk Manag. 2019;12:1–27.10.1111/jfr3.12550Search in Google Scholar
[52] Pang Y, Xiang S, Zhaosheng C, Hu X, Xue L, Teng Q. Impacts of urbanization on rural domestic sewage emissions in Erhai Lake basin. Res Environ Sci. 2015;28:1246–52.Search in Google Scholar
[53] Hu W, Duan Zy, Lu Y, Zhai L-M, Chen X, Liu P. The distribution of farmlands soil nitrogen and carbon and crop yield relationship in northern Erhai lake. Chin J Soil Sci. 2011;42:1138–42.Search in Google Scholar
[54] Wang Z, Xie J, Fang D, Wu D, Kong H, Liu A, et al. Nitrogen pollution load in farmlands under two typical farming systems in the north of Lake Erhai region. J Ecol Rural Environ. 2013;29:625–9.Search in Google Scholar
[55] Chen X, Li X, Qian W, Hu S. Response characteristics of algal chlorophyll-a to nitrogen，phosphorus and water temperature in Lake Erhai based on quantile regression. Environ Sci. 2017;38:113–20.Search in Google Scholar
[56] Arnold JG, Srinivasan R, Muttiah RS, Williams JR. Large area hydrology modelling and assessment part I: Model development. J Am water Resour Assoc Modeling Earth Syst. 1998;34:73–89.10.1111/j.1752-1688.1998.tb05961.xSearch in Google Scholar
[57] Arnold JG, Fohrer N. SWAT2000: Current capabilities and research opportunities in applied watershed modelling. Hydrol Process. 2005;19:563–72.10.1002/hyp.5611Search in Google Scholar
[58] Zhang Z, Chen S, Wan L, Cao J, Zhang Q, Yang C. The effects of landscape pattern evolution on runoff and sediment based on SWAT model. Env Earth Sci. 2021;80:2–466.10.1007/s12665-020-09315-6Search in Google Scholar
[59] Li CY, Fang HY. Assessment of climate change impacts on the streamflow for the Mun river in the Mekong Basin, southeast Asia: Using SWAT model. Catena. 2021;201:201.10.1016/j.catena.2021.105199Search in Google Scholar
[60] Hoghooghi N, Bosch DD, Bledsoe BP. Assessing hydrologic and water quality effects of land use conversion to Brassica carinata as a winter biofuel crop in the southeastern coastal plain of Georgia, USA, using the swat model. Gcb Bioenergy. 2021;13:473–92.10.1111/gcbb.12792Search in Google Scholar
[61] Duan Y, Luo M, Guo X, Cai P, Li F. Study on the relationship between snowmelt runoff for different latitudes and vegetation growth based on an improved SWAT model in Xinjiang, China. Sustain-Basel. 2021;13:1189.10.3390/su13031189Search in Google Scholar
[62] Cheng J, Gong Y, Zhu DZ, Xiao M, Zhang Z, Bi J, et al. Modeling the sources and retention of phosphorus nutrient in a coastal river system in China using SWAT. J Env Manage. 2021;278:111556.10.1016/j.jenvman.2020.111556Search in Google Scholar PubMed
[63] Chen Y, Marek GW, Marek TH, Porter DO, Brauer DK, Srinivasan R. Simulating the effects of agricultural production practices on water conservation and crop yields using an improved SWAT model in the Texas high plains, USA. Agr Water Manage. 2021;244:244.10.1016/j.agwat.2020.106574Search in Google Scholar
[64] Apostel A, Kalcic M, Dagnew A, Evenson G, Kast J, King K, et al. Simulating internal watershed processes using multiple SWAT models. Sci Total Env. 2021;759:1–14.10.1016/j.scitotenv.2020.143920Search in Google Scholar PubMed
[65] Williams JR. Epic-erosion/productivity impact calculator: 1. Model documentation. Techn Bull - U S Dep Agric. 1990;4:206–7.Search in Google Scholar
[66] Jing WU, Xin AG, Zhao YG, Wang J, Sun LM, Sun Q. Estimation of pollutants production from dairy farms and analysis of corresponding environmental capacity in Erhai Lake basin of Yunnan. Anim Husb Feed Sci. 2017;2016:1564–7.Search in Google Scholar
[67] Rostami S, He J, Hassan Q. Riverine water quality response to precipitation and its change. Environments. 2018;5:8.10.3390/environments5010008Search in Google Scholar
[68] Abbaspour KC. SWAT-CUP-2012: Swat calibration and uncertainty programs - a user manual. Swiss Federal Institute of Aquatic Science and Technology, Dübendorf, 2012.Search in Google Scholar
[69] Deng W, Shang S, Cai X, Zhao H, Song Y, Xu J. An improved differential evolution algorithm and its application in optimization problem. Soft Comput. 2021;25:5277–98.10.1007/s00500-020-05527-xSearch in Google Scholar
[70] Roth V, Nigussie TK, Lemann T. Model parameter transfer for streamflow and sediment loss prediction with SWAT in a tropical watershed. Env Earth Sci. 2016;75:1321.10.1007/s12665-016-6129-9Search in Google Scholar
[71] Ha LT, Bastiaanssen W, Griensven AV, Dijk A, Senay GB. SWAT-CUP for calibration of spatially distributed hydrological processes and ecosystem services in a Vietnamese river basin using remote sensing. Remote Sens Hydrol Earth Syst Sci Discuss. 2017;251:1–35.10.5194/hess-2017-251Search in Google Scholar
[72] Abbaspour KC, Vejdani M, Haghighat S. SWAT-CUP calibration and uncertainty programs for SWAT. Modsim International congress on modelling & simulation land water & environmental management integrated systems for sustainability; 2007.Search in Google Scholar
[73] Shang X, Wang X, Zhang D, Chen W, Chen X, Kong H. An improved SWAT-based computational framework for identifying critical source areas for agricultural pollution at the lake basin scale. Ecol Model. 2012;226:1–10.10.1016/j.ecolmodel.2011.11.030Search in Google Scholar
[74] Xiang S, Wan L, Pang Y. Spatial-temporal variation of inflow river water quality under land use effect. J ofAgro⁃Environment Sci. 2020;33:2474–83.Search in Google Scholar
[75] Chen X. Refined simulation of non-point-source pollution and identifying multilevel priority management areas at the watershed scale-a case study of Lake Erhai. Beijing: China Institute of Water Resources and Hydropower Research; 2019.Search in Google Scholar
[76] Zhang, Fukushima, Onda, Mizugaki, Gomi, Kosugi, et al. Characterisation of diffuse pollutions from forested watersheds in Japan during storm events - its association with rainfall and watershed features. Sci Total Envir. 2008;390(1):215–26.10.1016/j.scitotenv.2007.09.045Search in Google Scholar PubMed
[77] Wood FL, Heathwaite AL, Haygarth PM. Evaluating diffuse and point phosphorus contributions to river transfers at different scales in the Taw Catchment, Devon, UK. J Hydrol. 2005;304:118–38.10.1016/j.jhydrol.2004.07.026Search in Google Scholar
[78] Reza A, Eum J, Jung S, Choi Y, Owen JS, Kim B. Export of non-point source suspended sediment, nitrogen, and phosphorus from sloping highland agricultural fields in the East Asian monsoon region. Env Monit Assess. 2016;188:188.10.1007/s10661-016-5681-9Search in Google Scholar PubMed
[79] Ahn SR, Kim SJ. The effect of rice straw mulching and no-tillage practice in upland crop areas on non-point source pollution loads based on HSPF. Water-Sui. 2016;41:8–73.10.3390/w8030106Search in Google Scholar
[80] Bai XY, Shen W, Wang P, Chen XH, He YH. Response of non-point source pollution loads to land use change under different precipitation scenarios from a future perspective. Water Resour Manag. 2020;34:3987–4002.10.1007/s11269-020-02626-0Search in Google Scholar
[81] Gossweiler B, Wesstrom I, Messing I, Villazon M, Joel A. Impact of land use change on non-point source pollution in a semi-arid catchment under rapid urbanisation in Bolivia. Water-Sui. 2021;13:13.10.3390/w13040410Search in Google Scholar
[82] Lu Z, Yu B, Zhang H, Yang Z. Correlation analysis of agricultral non-point source pollution and water environment change in Erhai Lake watershed. J Cent China Norm Univ. 2017;51:215–23.Search in Google Scholar
[83] Gorito AM, Ribeiro AR, Almeida CMR, Silva AMT. A review on the application of constructed wetlands for the removal of priority substances and contaminants of emerging concern listed in recently launched EU legislation. Env Pollut. 2017;227:428–43.10.1016/j.envpol.2017.04.060Search in Google Scholar PubMed
[84] Wang Y, Bian J, Zhao Y, Tang J, Jia Z. Assessment of future climate change impacts on nonpoint source pollution in snowmelt period for a cold area using SWAT. Sci Rep-Uk. 2018;8:2402.10.1038/s41598-018-20818-ySearch in Google Scholar PubMed PubMed Central
[85] Li T, Kim G. Impacts of climate change scenarios on non-point source pollution in the Saemangeum watershed, South Korea. Water-Sui. 2019;11:1982.10.3390/w11101982Search in Google Scholar
[86] Xiang S, Wu Y, Lv X, Gao S, Chu Z, Pang Y. Characteristics and spatial distribution of agricultural non-point source pollution in Erhai Lake basin and its classified control strategy. Res Environ Sci. 2020;33:2474–83.Search in Google Scholar
[87] Xie K, Wu F, Luo Y, Zhang K, Shen S, Wang S, et al. Alternating distribution of farmland and inhabited landscapes in Erhai Lake and the influence of circadian rhythm on the water quality of a ditch during flood. J Agro-Environment Sci. 2018;37:2427–33.Search in Google Scholar
[88] Tong X, Wang Y, Xu W, Liu J, Zhang P. The change law of runoffs and sediments, non-point source nitrogen and phosphorus discharge in Erhai Lake irrigated watershed under different land-use types. China Rural water Hydropower. 2018;93–7.Search in Google Scholar
[89] Chen A, Lei B, Liu H, Wang H, Zhai L, Mao Y, et al. Characteristics of N and P losses from ditch runoff in framlands different in planting patterns offshore of Lake Erhai. J Ecol Rural Environmen. 2017;33:697–705.Search in Google Scholar
[90] Li Q, Wei C, Wang X, Zhao Z, Xie D. Mechanism and condition of agricultural non-point source pollution. Chin J Soil Sci. 2008;39:169–76.Search in Google Scholar
© 2023 Wu Cheng et al., published by De Gruyter
This work is licensed under the Creative Commons Attribution 4.0 International License.
