Next Article in Journal
The Precarious Survival of an Ancient Cultural Landscape: The Thousand-Year-Old Olive Trees of the Valencian Maestrat (Spain)
Previous Article in Journal
Assessing Trade-Offs between Agricultural Productivity and Ecosystem Functions: A Review of Science-Based Tools?
 
 
Search for Articles:
Title / Keyword
Author / Affiliation / Email
Journal
Article Type
 
 
Advanced Search
 
Section
Special Issue
Volume
Issue
Number
Page
 
Journals
Land
Volume 12
Issue 7
10.3390/land12071330
land-logo
Submit to this Journal Review for this Journal Propose a Special Issue
► Article Menu

Article Menu

share Share announcement Help format_quote Cite question_answer Discuss in SciProfiles thumb_up
...
Endorse textsms
...
Comment
first_page
settings
Order Article Reprints
Font Type:
Arial Georgia Verdana
Font Size:
Aa Aa Aa
Line Spacing:
Column Width:
Background:
Article

Spatiotemporal Transfer of Source-Sink Landscape Ecological Risk in a Karst Lake Watershed Based on Sub-Watersheds

by
Zhongfa Zhou
1,2,*,
Weiquan Zhao
1,2,3,
Sisi Lv
3,
Denghong Huang
1,2,
Zulun Zhao
3 and
Yaopeng Sun
1,2,4
1
School of Karst Science, Guizhou Normal University, Guiyang 550001, China
2
National Engineering Research Center for Karst Rocky Desertification Control, Guiyang 550001, China
3
Institute of Mountain Resources in Guizhou Province, Guiyang 550001, China
4
The Engineering Branch of the Third Institute of Surveying and Mapping of Guizhou Province, Guiyang 550001, China
*
Author to whom correspondence should be addressed.
Land 2023, 12(7), 1330; https://doi.org/10.3390/land12071330
Submission received: 3 June 2023 / Revised: 26 June 2023 / Accepted: 27 June 2023 / Published: 1 July 2023
(This article belongs to the Topic Karst Environment and Global Change)
Browse Figures
Versions Notes

Abstract

:
Non-point source pollution is an important source of ecological risk in karst lakes. The process of source–sink landscapes is the main pathway of pollution migration and plays an important role in water quality. In this study, the ecological risk evolution in the past 30 years was studied in a karst lake watershed with 495 sub-basins as the basic evaluation unit, and the risk assessment model of non-point source pollution was improved by using rainfall and fertilizer application. The results show that (1) the area of cultivated land shrank significantly, with forest land and construction land showing an upward trend in general; (2) the layout of the sink landscape continuously shrank, while the source landscape gradually expanded, and the space of high load values further increased and shifted from a flower-like layout distribution to concentrated contiguity, with some values exceeding 0.5; (3) the 252 sub-watersheds of the sink landscape migrated from very low risk to low risk, while the risk of the source landscape changed from medium risk to high and very high risk in 48 sub-watersheds; and (4) in terms of the overall trend of ecological risk transformation of the source–sink landscape, the transformation of sink landscapes to source landscapes was greater than that of source landscapes to sink landscapes, and the overall ecological risk showed an increasing trend.

1. Introduction

The watershed environment is an important source of water quality problems in lakes, as multiple pollutants are often detected in watershed water sample environments, seriously threatening the ecological safety of lakes and human health [1,2]. These pollutants interact with each other, making the ecological risk of watershed water extremely complex and difficult to assess, while the risk management mechanism for complex pollution lacks mature research [3]. Ecological risk assessment (ERA) refers to the likelihood or probability of adverse impacts on health, productivity, and genetics when a population, ecosystem, or even an entire landscape or part thereof is stressed by one or more factors, including qualitative or quantitative assessment and prediction [4,5,6,7]. The ERA framework was promulgated by the U.S. Environmental Protection Agency in 1992 to clearly define the meaning of ERA and assessing the likelihood of negative ecological impacts that have occurred or are occurring, attributing these impacts to receptor exposure to single or multiple stressors with the intention to be used to support environmental decision making [8]. The EPA developed a new evaluation framework by conducting a series of thematic and case studies, including ERAs in aquatic and watershed ecosystems [9]. The current academic community is more concerned with risk sources caused by anthropogenic activities and risk receptors with strong human interference, such as heavy metal pollution, persistent organic pollution, urban and regional ecosystem degradation, soil formation rate, etc. [10,11]. Kayumba et al. used the CA–Markov model to simulate changes in the Bayanbulak field landscape and developed an ERA assessment model based on sensitivity–hazard index relationships [12]. Remote sensing has also been used as an important data source [13,14,15].
The main source of ecological risk in lakes or reservoirs is pollution in their watersheds, and with the improvement in the level and technology of point source pollution control, non-point sources have been recognized as a serious threat to the aquatic environment and a major factor in the eutrophication of water bodies [16,17,18] and have become an important source of pollution in water bodies worldwide [19,20,21]. The stochastic nature of surface source pollution; its large spatial and temporal variability; and the difficulty of its monitoring, modeling, and management mean that it has become a hot spot for domestic and international research [22,23].
As an important subfield of ERA at the regional scale, landscape ERA emphasizes the spatiotemporal heterogeneity and scale effects of risk and is dedicated to the integrated characteristics of multisource risk and its spatial visual representation [24,25]. The evolution of landscape patterns leads to changes in the spatial structure of the landscape, which ultimately affect ecological security [26]. Land use data and landscape indices are currently the most commonly applied methods for regional ERAs because of their association with many ecological problems, such as land degradation, shrinking ecological lands, increased soil erosion, and elevated ecological vulnerability [27,28]. Land use/cover change is often a driving mechanism or response to ecological risk, and source–sink landscape ecological risk is a dynamic response to such change. The source–sink landscape ecological risk is influenced by many factors, including landscape type, elevation, slope, and distance, as well as rainfall. Different source–sink landscape types and objective conditions have different effects on the generation and hindrance of risk [29,30,31]. Therefore, it is highly relevant to assess the impact of landscape change on ecological risk for different land types.
The ongoing expansion of human society has subjected natural ecosystems in most regions to varying degrees of pressure and stress, either directly or indirectly, resulting in varying degrees of ecological risk [24]. Rapid urbanization and agriculturalization have accelerated the deterioration of surface water quality due to non-point source pollution [32]. For instance, 2693 natural lakes with an area greater than 1 km2 and a total area of 81,414.6 km2 in China play a crucial role in socioeconomic development and biodiversity maintenance [33]. In the mountainous regions of southwest China, lake reservoirs are the main sources of drinking water, but eutrophication is an important source of risk for surface lakes and reservoirs, and it can easily cause cyanobacterial blooms and pose a serious threat to the quality of the water environment [25,34,35]. According to the China Ecological Environment Status Bulletin, among 209 lakes and reservoirs monitored in 2021, 27.3% had a eutrophic status above mild. According to MODIS data monitoring, the trophic statuses of key lakes and reservoirs nationwide in the summer of 2018 showed a high proportion of grades, among which eutrophication in drinking water sources occurred in 83% of lakes and 47% of reservoirs, with considerable potential for improvement [25]. The total phosphorus and total nitrogen discharged from agricultural cultivation activities are the main sources of water quality deterioration [36]. The landscape pattern is also a sensitive predictor of water quality, and surface landscape changes are highly correlated with related water quality parameters [37].
China is a country with severe soil erosion, and the southwest karst region is the area most affected by water erosion. Soil erosion has become the number-one problem limiting regional development. The unique aboveground and belowground binary hydrological structure, humid climatic conditions, and carbonate substrate in karst regions are the intrinsic causes of soil erosion [38], and non-point source pollution loads under different precipitation scenarios have varying effects [39]. Although water quality risk problems occur in water, the source is onshore, and the interrelationship between surface landscape changes in shore areas and ecological risks in waters needs to be studied in depth. In karst areas, the characteristics of the surface–subsurface binary structure hydrological cycle, the complex combination of geological landforms, and the process and probability of pollutant migration to lakes through the surface environment directly affect the changes in water quality. Therefore, the dynamic assessment of ecological risk in the lake reservoir watershed in the karst region is crucial to the health of the lake reservoir ecosystem.

2. Materials and Methods

2.1. Study Area

The study area is located in the upper reaches of the second tributary of the Yangtze River, mainly covering Guiyang and Anshan, with a small part of the area located in Qiannan (Figure 1). Within the study area, Hongfeng Lake and Baihua Lake are typical karst deepwater reservoirs. Hongfeng Lake is located in the western suburbs of Guiyang and has a water area of 57.2 km2, an average depth of about 70 m, and a maximum depth of more than 100 m. It is 28 km from Guiyang, the first station on the Guizhou western golden tourism road, with a scenic area of 200 km2. It is a picturesque spot integrating a plateau lake, mountains, karst geomorphology, and minority customs. Baihua Lake has a surface area of about 14.5 km2 and a total reservoir capacity of 1.82 × 108 m3. Its average depth is 10.8 m, and its maximum depth is 45 m. Both lakes are typical karst plateau artificial deepwater lakes and are the main drinking water sources of Guiyang City; hence, their ecological functions and value are very important.
The two lakes belong to the same watershed, which is close to Guiyang City. Part of the watershed has been incorporated into the urbanization development direction of Guiyang. It is located in the economically active area of the province, which is exerting great pressure on the water environment. Through joint management by the government and society in recent years, point source pollution around the reservoirs has been effectively controlled, and water quality has improved significantly. Nevertheless, risks still exist, such as garbage dumping around lakes, residential domestic water discharge, and non-point source pollution. In the management and treatment of non-point source pollution of drinking water sources, the difficulty lies in finding the key zone of treatment, and the spatial location of the risk zone needs to be identified. The water quality risk of a lake reservoir occurs in the water, but the source risk is mainly in the shore watershed. Therefore, in this study, we used the natural watershed boundary to determine the study area of the watershed.

2.2. Data Sources and Processing

This study included such data materials as time series satellite remote sensing images, DEM (30 m), rainfall, nitrogen and phosphorus emissions, and statistical information. Satellite remote sensing was used to obtain time series land use type data, which were interpreted and corrected using ENVI5.3 and ArcGIS 10.2software and divided into six categories: arable land, construction land, forest, grassland, water, and unused land. The vegetation index (NDVI) was calculated using satellite remote sensing bands and was used to approximate the time series vegetation cover (FVC). Based on DEM, the area was divided into 495 sub-watersheds according to the principle of merging homogeneous units and ignoring small islands in lakes. Slope extraction was also based on DEM. Time series regional fertilizer application data (nitrogen and phosphorus) were obtained from the Statistical Yearbook of Guizhou Province (1990–2022). Rainfall data were obtained from meteorological stations in Guiyang City and Qingzhen City and collated from the references (Table 1).

2.3. Source-Sink Landscape Type Determination

In this study, the lake and reservoir were the risk receptors, and the process and the relative chance of non-point source pollution entering the lake reservoir were the objects of evaluation. The influencing factors included the total amount of pollution, rainfall, distance, topography, soil conditions, and geological background, among others. According to the source–sink landscape theory, different landscape types have facilitating/hindering effects on pollutants, and there are differences in the pollution capacity of different landscape types. The sub-watershed was the basic unit of pollutant pooling and output in this research, as pollutants can be considered to be released and absorbed first, then pooled into the output of a sub-watershed. Since the pollution weights of various land types are different, both source and sink effects may be present in a sub-watershed, and the respective source–sink load comparisons need to be analyzed. Topography and distance have a deterring effect on pollution migration, whereas rainfall is the main carrier and driver for pollution transportation and is an unfavorable factor for ecological risk [40]. Ground cover mostly plays a hindering role in the pollution migration process, as it helps to slow down the rate and probability of pollution entering a lake reservoir. Regional fertilization (nitrogen and phosphorus) is also a critical background source of pollution. Finally, researchers identified and evaluated the non-point source pollution risk in the lake reservoir basin with the sub-watersheds as the basic units to determine the inversion of spatial and temporal evolution scenarios of ecological risk in the basin. This study integrated the ecological risk migration process and spatial pattern under the action of multiple factors.
A source landscape facilitates the development of a process, while a sink landscape delays or prevents the development of this process [41]. The nature of a source or sink landscape is determined by the specific process; the source landscape of one process may be the sink landscape of another process. The distinction between source and sink landscapes is established by determining whether the landscape type plays a positive role in the evolution of ecological processes or a negative role, i.e., slowing them down. According to the source–sink theory, in the process of non-point source pollution formation, some landscape types play the role of source by generating pollutants, while others play the role of sink by absorbing pollutants. If the spatial distribution of source and sink landscapes in a region reaches a balanced state and forms a reasonable spatial distribution pattern, fewer non-point source pollutants will be generated. In contrast, if the spatial distribution of the landscape in a region is not reasonable (for example, the source landscape is distributed nearer to the water body, while the sink landscape is distributed further away from the water body), more pollutants may be generated [42].
Land use composition and its spatial configuration play an important role in the formation and migration of pollutants, especially in the migration pathways of different land use types and the release, absorption, and retention of pollutants, which will greatly affect the water quality status [43,44]. Combining the data conditions and characteristics of the study area, the watershed landscape was divided into six major categories, namely cultivated land, woodland, grassland, construction land, water, and unused land. Woodland retains and adsorbs nitrogen and phosphorus through infiltration and biological absorption by plant roots. It has the ability to retain, filter, and adsorb particles entering the river, improving water quality, so it can be used as a sink landscape to retain pollutants [45,46]. Grassland effectively reduces nutrient transport by decreasing surface runoff velocity and by other means and has the overall role of a sink, as reduced nutrient migration and water quality are negatively correlated [47]. Unused land and water quality indicators do not have a significant correlation, as both positive and negative correlations exist, and the direction of influence is not clear. Considering the small portion of unused land in the study area, the intensity of nitrogen and phosphorus pollution sources and the retention capacity were very weak, and the overall assessment results had little impact, meaning that unused land was treated as a transport patch without a source–sink effect [48]. Depending on periods of abundant water and dryness, water has varying effects on pollutants. In particular, floods have a considerable impact on the ecological risk of lakes [49]. However, for the waters of the two lakes in the study area, the area of onshore water bodies was small, so these water bodies were also treated as transmission patches. Construction land and water quality indicators are positively correlated, and industrial and domestic pollution sources contribute to the introduction of nitrogen and phosphorus elements into water; thus, construction land is a source landscape of pollution discharge [16]. Agricultural non-point source pollution is primarily generated by rainfall scouring the surface of agricultural land, triggering the loss of nitrogen and phosphorus nutrients in the soil. Furthermore, garden land, paddy fields, dry land, and watered land are influenced by the tillage method and the amount of fertilizer applied, and they are source landscapes of pollution emissions [50,51,52,53]. The source–sink landscape type of the two lakes’ watershed was determined based on the above considerations.

2.4. Source–Sink Pollution Risk Load Index

Construction land and cultivated land are source landscapes that generate non-point source pollution, while woodland and grassland are sink landscapes that trap and absorb pollutants. To compare the effects of different landscape types on non-point source pollution, the emission or absorption capacity of different landscape types needs to be quantitatively assessed and standardized. Thus, a landscape was selected as the standard, and the emission or absorption coefficient of this landscape type was compared with that of other corresponding landscape types; then, a certain weight was assigned to each. Accordingly, considering the proportions and weight relationships of various landscape types in the evaluation unit and the effect of the source–sink landscape pattern on pollution risk, we were able to obtain the risk of pollution source output in the evaluation unit. If the role of the source landscape in the evaluation unit was greater than the role of the sink landscape and the role of pollution was greater than the role of retention, then the sub-watershed was an output risk area. The source–sink pollution load comparison index was calculated using the following formula [54]:
L C I N = ∑ i = 1 n W i N × S i − ∑ j = 1 n W j N × S j L C I P = ∑ i = 1 n W i P × S i − ∑ j = 1 n W j P × S j L C I N P = L C I N + L C I P
where LCIN and LCIp are the pollution loads of total nitrogen and total phosphorus, respectively; i and j are the numbers of source and sink landscape types, respectively; wiN and wip are the weights of total nitrogen and total phosphorus discharged from source i, respectively; wjN and wjp are the weights of total nitrogen and total phosphorus absorbed by sink landscape j, respectively; and Si and Sj are the area shares of source i and sink j landscapes in the sub-watershed, respectively. The nitrogen and phosphorus emission (absorption) weights of the source and sink landscapes were converted directly or indirectly from the nitrogen and phosphorus emission coefficients in the First National Pollution Source Census—Fertilizer Loss Coefficient Manual for Agricultural Pollution Sources, statistical yearbooks, and other similar research results [55] and combined with the main landscape characteristics of the study area. Finally, the emission (absorption) weights of nitrogen and phosphorus of the different landscape types for major pollutants were obtained (Table 2).

2.5. Spatiotemporal Risk Evolution Scenario Simulation

With the lake and reservoir as the main receptor and object of risk, respectively, pollution risk assessment had to consider not only the source load but also the migration pathways of the pollutants, as well as the ease of access to the lake reservoir. Previous studies have shown that the loss of surface nutrients is influenced by a combination of factors, such as the soil’s physical and chemical properties, topography, rainfall intensity, vegetation cover, and tillage management practices [30]. The steeper the surface slope, the faster the migration rate of pollutants. Coupled with the stronger scouring of the surface by rainfall during an abundant water period, nitrogen and phosphorus nutrients are more easily transported into rivers, accelerating the eutrophication of lakes and reservoirs. Conversely, the shallower the surface slope and the flatter the terrain, the weaker the scouring of surface runoff, and the nitrogen and phosphorus pollutants are easily adsorbed by soil and vegetation during the long runoff movement [53,56]. The lower the surface vegetation cover level, the lower the nutrient loss from water sources and water management practices. However, the lower the surface vegetation cover level, the poorer the water containment and soil conservation capacity. In addition, soil is more easily washed away during the rainy season, accelerating the loss of nitrogen and phosphorus elements [57,58]. The distance pollutants have to travel to reach a lake reservoir is also an important factor in water pollution risk, as the further the sub-watershed outlet is from the lake reservoir, the greater the biological absorption and physical retention during migration, and the concentration is continuously reduced [41]. The shorter the distance from the lake reservoir, the more easily the pollutants can enter the water body. Due to the small sizes of the sub-basins, the lowest point of each sub-watershed was taken as the centralized exclusion point of the pollutant sources, and the surface distance of pollutant migration from the sub-watersheds to the lake or reservoir was calculated using a GIS distance analysis tool. Since the soil properties of the study area were relatively similar, the influence of soil property differences on pollutant migration was not considered here. However, the annual fertilization and rainfall of the region directly affect the final effect of pollutant discharge into the lake reservoir; therefore, these factors were considered. Based on the above understanding, the following equation was obtained by combining the non-point source pollution risk index equation [54,59] and using the annual rainfall (Rainfallm) and annual fertilizer application (TPJZm) to correct the equation, reflecting the annual risk parameter differences.
N P P R I m = L C I m N P × ( 1 + S l o p e m S l o p e max ) × ( 1 − D i s t a n c e m D i s t a n c e max ) × ( 1 − F v c m F v c max ) × ( 1 + Rainfall m Rainfall max ) × ( 1 + TPJZ m TPJZ max )
where NPPRIm is the non-point source pollution risk index value; LCImNP is the source–sink pollution load comparison index of the sub-watersheds; Slopem and Slopemax are the average slope value of the sub-watersheds and the maximum slope value of the two lakes’ watersheds, respectively; Distancem and Distancemax are the surface distances from the lowest point of the sub-watersheds to the reservoir and the furthest sub-watersheds, respectively; FVCm and FVCmax are the average and maximum vegetation cover of the sub-watersheds, respectively; Rainfallm and Rainfallmax are the average and maximum annual rainfall, respectively; and TPJZm and TPJZmax are the average and maximum annual fertilizer application of nitrogen and phosphorus, respectively, serving as correction coefficients for annual risk.
Based on the evaluation results and taking into account the actual situation, the entire watershed was divided into five classes: I (extremely low risk: NPPRIm < −0.5), II (low risk: −0.5 ≤ NPPRIm < 0), III (medium risk: 0 ≤ NPPRIm < 0.5), IV (high risk: 0.5 ≤ NPPRIm < 1), and V (extremely high risk: NPPRIm ≥ 1).
The DEM data were used to obtain the sub-watershed boundary and slope information, and the flow distance analysis tool (FlowDirection) in the hydrology analysis module (Hydrology) of ARCGIS was applied to calculate the flow distance from each image element to the reservoir, with the entire reservoir water surface as the lowest point (outlet). Then, the surface distance from each image element was calculated by combining the slope values, based on which the average surface distance (Distancem) from each sub-watershed to the reservoir was obtained (Figure 2).
The FVC was calculated based on Landsat5 data (1990, 1994, 2002, 2007, and 2011) and Landsat8-OLI data (2017 and 2022), with geometric correction and atmospheric correction using an image element dichotomous model [60]. The calculated results were sequentially classified as low cover, low–medium cover, medium cover, medium–high cover, or high cover, and their average cover indicators were calculated in the sub-watersheds as units to participate in the comprehensive evaluation.
The annual average rainfall was obtained by collating historical meteorological data. The heaviest average rainfall of the specific year was used as the reference standard and compared with data from the other years to obtain the standardized index.
The total amount of fertilizer (nitrogen and phosphorous) in the study area’s main regions (Guiyang and Qingzhen) was checked and collated from the statistical yearbook to obtain the annual fertilizer application standardized comparison index.

3. Results

3.1. Spatial and Temporal Evolution of Land use Types in the Two Lakes’ Watershed

The landscape types of the two lakes’ watershed were mostly forest, cultivated land, and construction land, which accounted for 57.84%, 35.13%, and 2.27% of the total watershed area in 1990, respectively, and 31.60%, 44.54%, and 14.91% in 2022, respectively (Table 3).
From 1990 to 2017, cultivated land showed a downward trend. Since 2017, however, it has rebounded. Forest land showed an upward trend in general, but there was a period of downward fluctuation around 2010. Construction land showed a general upward trend, especially from 2007 to 2012, when the rate of increase was the fastest, after which it rose more slowly. Grassland also increased at a generally stable rate, although there was a period of decline from 2007 to 2011. The area of water was at a stable level, with a clear decline in 2007. Unused land fluctuated widely from year to year, but the total size was small and relatively stable overall (Table 3 and Figure 3).
The advancement of urbanization was a strong contributor to landscape change. Rural construction and economic development also affected surface landscape change. Urban expansion and activities characterize the study area, as part of it is located in the suburbs of the city, while another part is an urban development space. The watershed not only serves as an important drinking water source protection site for the city but also as an important urban ecological protection space and agricultural development space, and it is increasingly under the dual pressure of ecological protection and economic development (Figure 3).

3.2. Spatial and Temporal Evolution of Vegetation Cover

From 1990 to 2011, the vegetation cover of the study area increased significantly, and most of the areas were dominated by high levels of vegetation cover. By 2007, the vegetation cover showed a decreasing trend, similar to the spatial layout of vegetation cover in 1994. From 2011 to 2022, the vegetation cover increased, the spatial layout of high vegetation cover expanded significantly, and the ecological environmental quality greatly improved. The vegetation cover in the northeast of the study area showed a decreasing trend, which was consistent with the direction of urbanization expansion (Figure 4).

3.3. Spatial and Temporal Patterns of Source–Sink Pollution Risk Loads

In 1990, the landscape loads were mostly sinks, and the load values were concentrated between −0.5 and −0.3. The source landscape load distribution was also relatively concentrated and contiguous. In 1994, the source–sink landscape load value distribution began to fragment, and the sink landscape area shrank, while the source landscape area gradually expanded. The high-load-value area also increased, especially in the upper reaches of the two lakes. The northeast section of the study area is close to the main urban area of Guiyang and is influenced by the expansion of urbanization; thus, the distribution and spatial layout of the load values showed an increasing trend. In the upstream area of Hongfeng Lake, which is influenced by the development of mountain agriculture and rural construction, the landscape load showed a significant and increasing trend, with some load values exceeding 0.5. Moreover, the high load values shifted from an interspersed distribution to a concentrated and continuous spatial pattern (Figure 5).

3.4. Evolution of Spatial and Temporal Patterns of Non-Point Source Pollution Risk

From 1990 to 2022 (Figure 6), the ecological risk in the study watershed increased, followed by a decrease. The ecological risk in the northern part of the watershed changed from extremely low to medium–low, whereas the ecological risk in the southwest was relatively stable, with little change. The ecological risk in the eastern neighboring Baihua Lake watershed gradually increased, transforming from medium–low to high—extremely high. The ecological risk in the southern part also gradually rose from extremely low to medium–low. The ecological risk in the upper part of the Hongfeng Lake watershed showed an increasing then decreasing trend, especially in 2011–2022. The transformation from high and extremely high risk into medium–low risk was evident, and the effect on drinking water source protection was also obvious.
The ecological risk of the watershed was primarily extremely low and low in 1990. Most of the sub-watersheds were dominated by sink landscape risk, with 125 sub-watersheds showing source landscape risk, accounting for 25.89% of the total surface area of the watershed. In 2011, there were 179 sub-watersheds at risk, accounting for 36.19% of the total watershed surface. In 2022, there were 160 sub-watersheds at risk, which accounted for 30.89% of the total watershed surface; that is, most of the study watershed was dominated by risk in the sink landscape. Within the source landscape, the trends of high risk and extremely high risk increased before decreasing, and medium risk decreased before increasing, while the areas with more than medium risk remained stable within the sink landscape (Table 4).
The changes in the number and area of sub-watersheds with different levels of risk were similar (Figure 7 and Figure 8), indicating that the risk levels of most of the sub-watersheds were relatively stable and that no large-scale transformations of ecological risk occurred. The number and area of extremely low-risk areas were relatively stable until 2002, with a trend of slow decline, then a rapid rise, followed by a rapid decline after 2007, a slow rise after 2011, and a declining trend after 2017. The number and area of low-risk areas dominated, and the general trend was initially an increase, then a decrease, followed by a steady rise, similar to the trend of very low-risk areas. Changes in the trend of the medium-risk areas were relatively flat; there was a slow decline in 1990–2002 and a gradual increase after 2002, reaching the highest value in 2012 in terms of the number and area of sub-watersheds. This was followed by a decline and another gradual rise after 2017, with the greatest possibility of transformation from extremely low risk. The high-risk areas were in a stable state from 1990 to 2007, as there were minimal changes. However, an upward trend was noted in the next four years, followed by a continuously decreasing trend from 2011 to 2017. Eventually, it began to rise again. The change trend of the extremely high-risk areas was similar to that of the high-risk areas, but the change magnitude after 2007 was significantly higher than that of the high-risk areas.
From 1990 to 1994, the extremely low-risk and low-risk areas underwent a mutual transformation, and the overall trend was from high risk to low risk. Only one sub-watershed converted from medium risk to high risk, covering only 0.28% of the whole watershed, and 17 sub-watersheds were converted from low risk to extremely low risk, covering 3.69% of the watershed. In 1994–2002, the migrations were mostly between extremely low risk, low risk, and medium risk, and the overall conversion trend was still from high risk to low risk. There were 47 sub-watersheds that migrated from medium risk to low risk and extremely low risk, covering 9.97% of the whole watershed. In contrast, 35 sub-watersheds were converted from extremely low risk to low risk, covering 7.21% of the watershed. In 2002–2007, the transformation mainly occurred between very low risk, low risk, medium risk, and high risk, and the overarching trend was from low risk to high risk. There were 26 sub-watersheds converted from low risk to medium and high risk, accounting for 6.38% of the entire watershed. Fourteen sub-watersheds were converted from medium risk to high and extremely high risk, accounting for 3.03% of the watershed. In 2007–2011, the migration occurred mostly between extremely low risk, low risk, medium risk, and high risk, and the overall transformation trend was from low risk to high risk. Only four sub-watersheds changed from low risk to extremely low risk, covering 0.67% of the whole watershed, while 89 sub-watersheds shifted from extremely low risk to low and medium risk, comprising 20.45%. During the 2011–2017 period, the interconversion mainly occurred between extremely low risk, low risk, medium risk, and high risk, and the overall conversion trend was from high risk to low risk. Accounting for 9.36% of the whole watershed, 36 sub-watersheds converted from low risk to extremely low risk, while 55 sub-watersheds shifted from medium risk to low and extremely low risk, covering 12.91% of the entire watershed (Table 4, Figure 7 and Figure 9). In 2017–2022, each level of risk underwent a mutual transformation, and the overall transformation trend was more obviously from low risk to high risk. Four sub-watersheds were converted from low risk to very low risk, with an area of 0.5% of the whole watershed, and another four sub-watersheds shifted from extremely high risk to high risk, with an area of 0.73%.
A change from extremely low risk to low risk occurred in 82 sub-watersheds, with an area of 17.38% of the whole watershed, and a change from medium risk to high and extremely high risk occurred in two sub-watersheds, with an area of 0.48% of the entire watershed (Table 4 and Figure 9).
Considering the risk-level classifications of the study, extremely low risk and low risk focus on sink landscape risk, which has a blocking effect on the ecological risk of reservoirs in a watershed, while medium risk, high risk, and extremely high risk focus on source landscape risk. Therefore, according to the migration of ecological risks in different periods, the changes in the sink landscape risk indicate a growth trend, as shown by the transformation of 252 sub-watersheds from extremely low risk into low risk and of 153 sub-watersheds from low risk into extremely low risk. This illustrates that the sink landscape risk had an upward trend. Meanwhile, the source landscape risk changed from medium risk to high and extremely high risk in 48 sub-watersheds and from high and extremely high risk to medium risk in 34 sub-watersheds. This confirms that the risk level of the sink landscape also had a gradual upward trend. A migration from sink landscape risks (I and II) to source landscape risks (III, IV, and V) was observed in 154 sub-watersheds, while 119 sub-watersheds were involved in the migration from source landscape risks (III, IV, and V) to sink landscape risks (I and II), indicating that the migration of sink to source was greater than that of source to sink and that the overall ecological risk of the source–sink landscape was increasing (Figure 9).

4. Discussion

4.1. Spatial and Temporal Migration of Ecological Risks

As seen in the results of the NPPRI index classification, with the passage of time, the trend of the source ecological risk is upward, and that of the sink ecological risk is downward. The ecological risk of the whole watershed is shifting from sink to source, and the scale is gradually increasing. The migration of source–sink ecological risk is closely related to the risk load of source–sink pollution. Due to human activities, the land use types are mutually transformed, resulting in the migration of source–sink pollution risk. This study considered the lakes as risk receptors and investigated the role of non-point source pollution in the watershed in water quality and magnitude. It was found that the risk source requires a complex migration process from its original location to the lake, which involves not only the load of the risk source itself but also various factors, such as distance, topography, surface cover, and rainfall.
Based on the evolution and migration of ecological risks in the watershed time series, on the one hand, the transformation of the source–sink landscape type leads to qualitative changes in the type of ecological risk caused by the manner and intensity of human activities, such as rapid urbanization, land use changes, soil formation rate [15,60], the delineation and management of drinking water reserves, the implementation of a policy of returning cultivated land to forest, the transformation of planting structures, and the improvement of the level of agricultural modernization. On the other hand, the transformation is due to changes in the influencing factors of ecological risk that cause changes in the internal intensity of ecological risk in source–sink landscapes. For example, a change in atmospheric precipitation (Figure 10) is an important factor, as it can cause surface runoff. Changes in fertilizer application in the watershed (Figure 11) and the evolution of surface cover degree are also direct factors. In addition, topographic conditions (Figure 12) and differences in the surface distances from the sub-watersheds to the lakes (Figure 13) are also important factors controlling the occurrence of ecological risks in the landscape [61]. Pollution load is the main component of landscape ecological risk, and it changes over time due to human actions and social activities.
The study area is located in the same watershed, and the surface runoff flows upstream from Hongfeng Lake to Baihua Lake, converges on Wujiang River, and finally flows into the Yangtze River. Locally, the risk source first forms a convergence in the small watershed where it is located, then converges on the lake through rivers and dry canals, causing pollution risk. Influenced by urban expansion, the generated risk increases year by year and is concentrated in the southeastern area of Baihua Lake, where low hills dominate and the terrain is relatively flat. The surface runoff converges on Baihua Lake. The ecological risk of the sink landscape upstream of Hongfeng Lake is transformed into the ecological risk of a source landscape.

4.2. Influencing Factors of Ecological Risk in Karst Watersheds

As shown by the analysis of the formation mechanism of ecological risk of the karst lakes in the watershed source–sink landscape, (1) there are differences in the amount of non-point source pollution in different periods due to changes in fertilizer application. (2) There are differences in the source and sink of the landscape itself due to the different loading or absorption of pollutants in the surface landscape. (3) The geomorphic environment and surface cover have a considerable influence on the process of non-point source pollution migration. Slope, slope direction, mountains, hills, and different ground covers are all factors that control the migration of surface pollutants with soil and water elements. (4) Rainfall amount and intensity are also important factors affecting the migration of pollutants through surface runoff to the lakes, as well as soil and water leakage to underground rivers then into the lakes. (5) The path length of the risk source from the lake is another significant factor in judging the risk role, as some risk sources are too far away from the lake. Considering the surface environment of the pollution sources of dissipation, the final role of the lake can be negligible. However, the risk sources closer to the lake reservoir have a higher probability of entering the lake, and the risk caused to the lake also increases.

4.3. Methodological Improvements

This study built on previous scholarly research, with further in-depth consideration. (1) Because the karst landscape environment was fragmented and complex, we chose sub-watersheds as the ecological risk evaluation units, as they were closer to the migration process of the pollutants and reduced the influence of the landscape environment on the evaluation results. (2) Considering the process of ecological risk occurrence and migration in the source–sink landscape and its influencing factors, the non-point source pollution risk index was refined and corrected based on the surface vegetation cover, rainfall, and fertilizer application. (3) The spatiotemporal sequence was explored, and scenario simulation of the external ecological risk change of the lake was made more scientific. The evolution trend and development direction of the ecological risk were also portrayed.

5. Conclusions

Several conclusions can be drawn as follows. (1) Although landscape type is a determining factor, factors such as distance, landform, ground cover, fertilizer application, and rainfall are also controlling factors for the migration of ecological risk. The distribution of very high-risk areas is closely related to urbanization, but the effective promotion of environmental protection significantly reduces the risk level; even as the landscape type of construction land is expanding, the ecological risk can be reduced by other factors. (2) In the past 30 years, medium–low ecological risk areas have migrated to the upstream of the lake reservoir, and medium–high risk areas have mainly migrated to the downstream, mainly due to the advancement of urbanization in the downstream and the increased protection measures and efforts in the upstream water extraction sites. (3) The investigation of the surface migration of ecological risk was insufficient to accurately portray the actual migration process of ecological risk, and there is room for further research, especially in watersheds with complex underground hydrological environments in karst areas.

Author Contributions

Conceptualization, Z.Z. (Zhongfa Zhou); Methodology, W.Z. and Z.Z. (Zulun Zhao); Investigation, W.Z., S.L., Z.Z. (Zulun Zhao) and Y.S.; Resources, D.H.; Data curation, W.Z., S.L. and D.H.; Writing—review & editing, W.Z.; Project administration, Z.Z. (Zhongfa Zhou). All authors have read and agreed to the published version of the manuscript.

Funding

This research was funded by the Guizhou Provincial Basic Research Program (Natural Science) (No. QKHJC [2020]1Y410), the Guizhou Provincial Science and Technology Projects (No. QKHZC [2020]4Y132), the Program in Guizhou Planning of Philosophy and Social Science (Grant NO. 21GZZD39), the High-level Innovative Talents Training Program in Guizhou Province (Grant NO. 2016-5674), the Guizhou Academy of Sciences Innovation Plate Scientific Research Start-up Fund (QKYC [2021]01), and the Provincial special fund for scientific research of the Guizhou Academy of Sciences (Grant No. QKY-KZH-2021-03).

Data Availability Statement

The data in this study are available from the corresponding authors upon request. Due to the sensitivity of the study area, some data cannot be made public.

Acknowledgments

We appreciate the constructive comments and suggestions from the reviewers that helped to improve the quality of this manuscript. We also would like to offer our sincere thanks to those who participated in data processing and manuscript revisions.

Conflicts of Interest

The authors declare no conflict of interest.

References

  1. Conley, J.M.; Evans, N.; Cardon, M.C.; Rosenblum, L.; Iwanowicz, L.R.; Hartig, P.C.; Schenck, K.M.; Bradley, P.M.; Wilson, V.S. Occurrence and in vitro bioactivity of estrogen, androgen, and glucocorticoid compounds in a nationwide screen of United States stream waters. Environ. Sci. Technol. 2017, 51, 4781–4791. [Google Scholar] [CrossRef] [PubMed]
  2. Li, Q.; Dai, J.F.; Zhang, H.Y.; Wan, Z.P.; Xu, J.X. Potentially toxic elements in lake sediments in China: Spatial distribution, ecological risks, and influencing factors. Sci. Total Environ. 2023, 868, 161596. [Google Scholar] [CrossRef] [PubMed]
  3. Van den Brink, P.J.; Choung, C.B.; Landis, W.; Mayer-Pinto, M.; Pettigrove, V.; Scanes, P.; Smith, R.; Stauber, J. New approaches to the ecological risk assessment of multiple stressors. Mar. Freshw. Res. 2016, 67, 429–439. [Google Scholar] [CrossRef]
  4. Chen, S.Q.; Chen, B.; Fath, B.D. Ecological risk assessment on the system scale: A review of state-of-the-art models and future perspectives. Ecol. Modell. 2013, 250, 25–33. [Google Scholar] [CrossRef]
  5. Del Signore, A.; Hendriks, A.J.; Lenders, H.J.R.; Leuven, R.; Breure, A.M. Development and application of the SSD approach in scientific case studies for ecological risk assessment. Environ. Toxicol. Chem. 2016, 35, 2149–2161. [Google Scholar] [CrossRef]
  6. Haque, M.R.; Ali, M.M.; Ahmed, W.; Rahman, M.M. Assessment of metal(loid)s pollution in water and sediment from an urban river in Bangladesh: An ecological and health risk appraisals. Case Stud. Chem. Environ. 2022, 6, 100272. [Google Scholar] [CrossRef]
  7. How, C.M.; Kuo, Y.H.; Huang, M.L.; Liao, V.H.C. Assessing the ecological risk and ecotoxicity of the microbially mediated restoration of heavy metal-contaminated river sediment. Sci. Total Environ. 2023, 858, 159732. [Google Scholar] [CrossRef]
  8. Suter, G.W. Applicability of indicator monitoring to ecological risk assessment. Ecol. Indic. 2001, 1, 101–112. [Google Scholar] [CrossRef]
  9. Sergeant, A. Management objectives for ecological risk assessment-developments at US EPA. Environ. Sci. Policy 2000, 3, 295–298. [Google Scholar] [CrossRef]
  10. Jahandari, A.; Abbasnejad, A.; Jamasb, R. Concentration, likely sources, and ecological risk assessment of potentially toxic elements in urban soils of Shiraz City, SW Iran: A preliminary assessment. Arab. J. Geosci. 2020, 13, 936. [Google Scholar] [CrossRef]
  11. Ran, C.; Bai, X.Y.; Tan, Q.; Luo, G.J.; Cao, Y.; Wu, L.H.; Chen, F.; Li, C.J.; Luo, X.L.; Liu, M.; et al. Threat of soil formation rate to health of karst ecosystem. Sci. Total Environ. 2023, 887, 163911. [Google Scholar] [CrossRef]
  12. Kayumba, P.M.; Chen, Y.N.; Mind’je, R.; Mindje, M.; Li, X.Y.; Maniraho, A.P.; Umugwaneza, A.; Uwamahoro, S. Geospatial land surface-based thermal scenarios for wetland ecological risk assessment and its landscape dynamics simulation in Bayanbulak Wetland, Northwestern China. Landscape Ecol. 2021, 36, 1699–1723. [Google Scholar] [CrossRef]
  13. Watson, C.S.; Quincey, D.J.; Carrivick, J.L.; Smith, M.W. The dynamics of supraglacial ponds in the Everest region, central Himalaya. Glob. Planet. Change 2016, 142, 14–27. [Google Scholar] [CrossRef]
  14. Wangchuk, S.; Bolch, T.; Zawadzki, J. Towards automated mapping and monitoring of potentially dangerous glacial lakes in Bhutan Himalaya using Sentinel-1 Synthetic Aperture Radar data. Int. J. Remote Sens. 2019, 40, 4642–4667. [Google Scholar] [CrossRef]
  15. Li, C.J.; Bai, X.Y.; Tan, Q.; Luo, G.J.; Wu, L.H.; Chen, F.; Xi, H.P.; Luo, X.L.; Ran, C.; Chen, H.; et al. High-resolution mapping of the global silicate weathering carbon sink and its long-term changes. Glob. Change Biol. 2022, 28, 4377–4394. [Google Scholar] [CrossRef] [PubMed]
  16. Hobbie, S.E.; Finlay, J.C.; Janke, B.D.; Nidzgorski, D.A.; Millet, D.B.; Baker, L.A. Contrasting nitrogen and phosphorus budgets in urban watersheds and implications for managing urban water pollution. Proc. Natl. Acad. Sci. USA 2017, 114, 4177–4182. [Google Scholar] [CrossRef] [PubMed]
  17. Zou, L.L.; Liu, Y.S.; Wang, Y.S.; Hu, X.D. Assessment and analysis of agricultural non-point source pollution loads in China: 1978–2017. J. Environ. Manag. 2020, 263, 110400. [Google Scholar] [CrossRef] [PubMed]
  18. Bouwman, L. Exploring global changes in nitrogen and phosphorus cycles in agriculture induced by livestock production over the 1900–2050 period. Proc. Natl. Acad. Sci. USA 2013, 110, 20882–20887. [Google Scholar] [CrossRef]
  19. Ahmad, Z.U.; Shou, W.; Gang, D.D. Nonpoint source pollution. Water Environ. Res. 2017, 86, 1875–1894. [Google Scholar] [CrossRef] [Green Version]
  20. Jia, Z.; Tang, S.; Luo, W.; Hai, Y. Water quality improvement through five constructed serial wetland cells and its implications on nonpoint-source pollution control. Hydrol. Sci. J. 2016, 61, 2946–2956. [Google Scholar] [CrossRef] [Green Version]
  21. Liu, R.; Yu, W.; Shi, J.; Wang, J.; Xu, F.; Shen, Z. Development of regional pollution export coefficients based on artificial rainfall experiments and its application in North China. Int. J. Environ. Sci. Technol. 2017, 14, 823–832. [Google Scholar] [CrossRef]
  22. Ongley, E.D.; Zhang, X.; Yu, T. Current status of agricultural and rural non-point source Pollution assessment in China. Environ. Pollut. 2010, 158, 1159–1168. [Google Scholar] [CrossRef] [PubMed]
  23. Collick, A.S.; Fuka, D.R.; Kleinman, P.J.A.; Buda, A.R.; Weld, J.L.; White, M.J.; Veith, T.L.; Bryant, R.B.; Bolster, C.H.; Easton, Z.M. Predicting phosphorus dynamics in complex terrains using a variable source area hydrology model. Hydrol. Process. 2015, 29, 588–601. [Google Scholar] [CrossRef] [Green Version]
  24. Peng, J.; Dang, W.; Liu, Y.; Zong, M.; Hu, X. Review on landscape ecological risk assessment. Acta Geogr. Sin. 2015, 70, 664–677. [Google Scholar]
  25. Cao, C.; Wang, S.; Li, J.; Zhao, H.; Shen, W.; Xie, Y. MODIS-based monitoring of spatial distribution of trophic status in 144 key lakes and reservoirs of China in summer of 2018. J. Lake Sci. 2021, 33, 405–413. [Google Scholar]
  26. Ma, L.; Bo, J.; Li, X.; Fang, F.; Cheng, W. Identifying key landscape pattern indices influencing the ecological security of inland river basin: The middle and lower reaches of Shule River Basin as an example. Sci. Total Environ. 2019, 674, 424–438. [Google Scholar] [CrossRef]
  27. Bhattachan, A.; Emanuel, R.E.; Ardon, M.; Bernhardt, E.S.; Anderson, S.M.; Stillwagon, M.G.; Ury, E.A.; BenDor, T.K.; Wright, J.P. Evaluating the effects of and-use change and future climate change on vulnerability of coastal landscapes to saltwater intrusion. Elementa-Sci. Anthrop. 2018, 6, 62. [Google Scholar] [CrossRef] [Green Version]
  28. Zhang, W.; Chang, W.J.; Zhu, Z.C.; Hui, Z. Landscape ecological risk assessment of Chinese coastal cities based on land use change. Appl. Geogr. 2020, 117, 102174. [Google Scholar] [CrossRef]
  29. Xu, Y.; Tang, H.; Wang, B.; Chen, J. Effects of landscape patterns on soil erosion processes in a mountain-basin system in the North China. Nat. Hazards 2017, 87, 1567–1585. [Google Scholar] [CrossRef]
  30. Chen, L.; Sun, R.; Lu, Y. A conceptual model for a process-oriented landscape pattern analysis. Sci. China Earth Sci. 2019, 62, 2050–2057. [Google Scholar] [CrossRef]
  31. Jiang, Z.; Liu, H.; Wang, H.; Peng, J.; Meersmans, J.; Green, S.M.; Quine, T.A.; Wu, X.; Song, Z. Bedrock geochemistry influences vegetation growth by regulating the regolith water holding capacity. Nat. Commun. 2020, 11, 2392. [Google Scholar] [CrossRef]
  32. Kast, J.B.; Apostel, A.M.; Kalcic, M.M.; Muenich, R.L.; Dagnew, A.; Long, C.M.; Evenson, G.; Martin, J.F. Source contribution to phosphorus loads from the Maumee River watershed to Lake Erie. J. Environ. Manag. 2021, 279, 111803. [Google Scholar] [CrossRef]
  33. Ma, R.; Yang, G.; Duan, H.; Jiang, J.; Wang, S.; Feng, X.; Li, A.; Kong, F.; Xue, B.; Wu, J.; et al. China’s lakes at present: Number, area and spatial distribution. Sci. China Earth Sci. 2011, 54, 283–289. [Google Scholar] [CrossRef]
  34. Shi, K.; Zhang, Y.; Zhou, Y.; Liu, X.; Zhu, G.; Qin, B.; Gao, G. Long-term MODIS observations of cyanobacterial dynamics in Lake Taihu: Responses to nutrient enrichment and meteorological factors. Sci. Rep. 2017, 7, 40326. [Google Scholar] [CrossRef] [PubMed] [Green Version]
  35. Wei, Z.; Yu, Y.; Yi, Y. Spatial distribution of nutrient loads and thresholds in large shallow lakes: The case of Chaohu Lake, China. J. Hydrol. 2022, 613, 128466. [Google Scholar] [CrossRef]
  36. Xue, L.H.; Hou, P.F.; Zhang, Z.Y.; Shen, M.X.; Liu, F.X.; Yang, L.Z. Application of systematic strategy for agricultural non-point source pollution control in Yangtze River basin, China. Agric. Ecosyst. Environ. 2020, 304, 107148. [Google Scholar] [CrossRef]
  37. Xu, S.; Li, S.L.; Zhong, J.; Li, C. Spatial scale effects of the variable relationships between landscape pattern and water quality: Example from an agricultural karst river basin, Southwestern China. Agric. Ecosyst. Environ. 2020, 300, 106999. [Google Scholar] [CrossRef]
  38. Wei, S.; Pan, J.; Liu, X. Landscape ecological safety assessment and landscape pattern optimization in arid inland river basin: Take Ganzhou district as an example. Hum. Ecol. Risk Assess. 2020, 26, 782–806. [Google Scholar] [CrossRef]
  39. Bai, X.Y.; Shen, W.; Wang, P.; Chen, X.H.; He, Y.H. 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. [Google Scholar] [CrossRef]
  40. Hou, M.J.; Ge, J.; Gao, J.L.; Meng, B.P.; Li, Y.C.; Yin, J.P.; Liu, J.; Feng, Q.S.; Liang, T.G. Ecological Risk Assessment and Impact Factor Analysis of Alpine Wetland Ecosystem Based on LUCC and Boosted Regression Tree on the Zoige Plateau, China. Remote Sens. 2020, 12, 368. [Google Scholar] [CrossRef] [Green Version]
  41. Wang, J.L.; Ni, J.P.; Chen, C.L.; Xie, D.T.; Shao, J.A.; Chen, F.X.; Lei, P. Source-sink landscape spatial characteristics and effect on non-point source pollution in a small catchment of the Three Gorge Reservoir Region. J. Mount. Sci. 2018, 15, 327–339. [Google Scholar] [CrossRef]
  42. Wang, R.J.; Wang, Y.; Sun, S.Y.; Cai, C.J.; Zhang, J.F. Discussing on “source-sink” landscape theory and phytoremediation for non-point source pollution control in China. Environ. Sci. Pollut. Res. 2020, 27, 44797–44806. [Google Scholar] [CrossRef]
  43. Haruna, A.J.; Biswajeet, P.; Azmin, S.W.N.; Rohaizah, J.N. Long-term runoff dynamics assessment measured through land use/cover (LULC) changes in a tropical complex catchment. Environ. Syst. Decis. 2018, 39, 16–33. [Google Scholar]
  44. Ortolani, V. Land use and its effects on water quality using the BASINS model. Environ. Earth Sci. 2014, 71, 2059–2063. [Google Scholar] [CrossRef]
  45. Lin, X.; Wang, Z. Landscape ecological risk assessment and its driving factors of multi-mountainous city. Ecol. Indic. 2023, 146, 109823. [Google Scholar] [CrossRef]
  46. Kaendler, M.; Blechinger, K.; Seidler, C.; Pavlu, V.; Sanda, M.; Dostal, T.; Krasa, J.; Vitvar, T.; Stich, M. Impact of land use on water quality in the upper Nisa catchment in the Czech Republic and in Germany. Sci. Total Environ. 2017, 586, 1316–1325. [Google Scholar] [CrossRef] [PubMed]
  47. Sun, R.; Cheng, X.; Chen, L. A precipitation-weighted landscape structure model to predict potential pollution contributions at watershed scales. Landsc. Ecol. 2018, 33, 1603–1616. [Google Scholar] [CrossRef]
  48. Shen, Z.Y.; Zhong, Y.C.; Huang, Q.; Chen, L. Identifying non-point source priority management areas in watersheds with multiple functional zones. Water Res. 2015, 68, 563–571. [Google Scholar] [CrossRef]
  49. Rounce, D.R.; Watson, C.S.; McKinney, D.C. Identification of hazard and risk for glacial lakes in the Nepal Himalaya using satellite imagery from 2000–2015. Remote Sens. 2017, 9, 654. [Google Scholar] [CrossRef] [Green Version]
  50. Li, Z.; Du, P.; Huang, H.; Ge, Y.; Li, X. Modeling agricultural non-point source pollution in a high-precipitation coastal area of China. Front. Environ. Sci. Eng. 2014, 8, 922–928. [Google Scholar] [CrossRef]
  51. Li, H.; Liu, L.; Ji, X. Modeling the relationship between landscape characteristics and water quality in a typical highly intensive agricultural small watershed, Dongting lake basin, south central China. Environ. Monit. Assess. 2015, 187, 129. [Google Scholar] [CrossRef]
  52. Wang, W.; Chen, L.; Shen, Z. Dynamic export coefficient model for evaluating the effects of environmental changes on non-point source pollution. Sci. Total Environ. 2020, 747, 141164. [Google Scholar] [CrossRef]
  53. Wang, J.L.; Chen, C.L.; Ni, J.P.; Xie, D.T.; Guo, X.; Luo, Z.J.; Zhao, Y. Assessing effects of “source-sink” landscape on non-point source pollution based on cell units of a small agricultural catchment. J. Mount. Sci. 2019, 16, 2048–2062. [Google Scholar] [CrossRef]
  54. Xu, F.; Zhou, X.; Meng, Q.; Zhang, Y. Remote sensing identification and evaluation of non-point source pollution risk of drinking water source based onsource-sinklandscape. Acta Geogr. Sin. 2020, 40, 2609–2620. (In Chinese) [Google Scholar]
  55. Zhang, B.L.; Cui, B.H.; Zhang, S.M.; Wu, Q.Y.; Yao, L. Source apportionment of nitrogen and phosphorus from non-point source pollution in Nansi Lake Basin, China. Environ. Sci. Pollut. Res. 2018, 25, 19101–19113. [Google Scholar] [CrossRef] [PubMed]
  56. Hao, G.R.; Li, J.K.; Li, S.; Li, K.B.; Zhang, Z.H.; Li, H.E. Quantitative assessment of non-point source pollution load of PN/PP based on RUSLE model: A case study in Beiluo River Basin in China. Environ. Sci. Pollut. Res. 2020, 27, 33975–33989. [Google Scholar] [CrossRef] [PubMed]
  57. Wang, J.; Shao, J.a.; Wang, D.; Ni, J.; Xie, D. Identification of the “source” and “sink” patterns influencing non-point source pollution in the Three Gorges Reservoir Area. J. Geogr. Sci. 2016, 26, 1431–1448. [Google Scholar] [CrossRef]
  58. Wang, Z.; Yang, S.; Zhao, C.; Bai, J.; Lou, H.; Chen, K.; Wu, L.; Dong, G.; Zhou, Q. Assessment of non-point source total phosphorus pollution from different land use and soil types in a mid-high latitude region of China. Water 2016, 8, 505. [Google Scholar] [CrossRef] [Green Version]
  59. Zhao, W.; Zhou, Z.; Zhao, Z.; Li, W.; Li, Q. Risk assessment of non-point source pollution in karst reservoirs based on ‘source-sink’ landscape theory. Water Supply 2022, 22, 6094–6110. [Google Scholar] [CrossRef]
  60. Ding, Y.L.; Zheng, X.M.; Zhao, K.; Xin, X.P.; Liu, H.J. Quantifying the impact of NDVIsoil determination methods and NDVIsoil variability on the estimation of fractional vegetation cover in northeast China. Remote Sens. 2016, 8, 29. [Google Scholar] [CrossRef] [Green Version]
  61. Xiao, B.Q.; Bai, X.Y.; Zhao, C.W.; Tan, Q.; Li, Y.B.; Luo, G.J.; Wu, L.H.; Chen, F.; Li, C.J.; Ran, C.; et al. Responses of carbon and water use efficiencies to climate and land use changes in China’s karst areas. J. Hydrol. 2023, 617, 128968. [Google Scholar] [CrossRef]
Land 12 01330 g001 550
Figure 1. Map of the geographical location of the study area.
Figure 1. Map of the geographical location of the study area.
Land 12 01330 g001
Land 12 01330 g002 550
Figure 2. Diagram of terrain index extraction process.
Figure 2. Diagram of terrain index extraction process.
Land 12 01330 g002
Land 12 01330 g003 550
Figure 3. Trend of landscape evolution.
Figure 3. Trend of landscape evolution.
Land 12 01330 g003
Land 12 01330 g004 550
Figure 4. Distribution of vegetation cover in the watershed landscape (1990–2022).
Figure 4. Distribution of vegetation cover in the watershed landscape (1990–2022).
Land 12 01330 g004
Land 12 01330 g005 550
Figure 5. Distribution of pollution loads in the source–sink landscape (1990–2022).
Figure 5. Distribution of pollution loads in the source–sink landscape (1990–2022).
Land 12 01330 g005
Land 12 01330 g006 550
Figure 6. Spatial patterns of ecological risk in the source–sink landscape (1990–2022).
Figure 6. Spatial patterns of ecological risk in the source–sink landscape (1990–2022).
Land 12 01330 g006
Land 12 01330 g007 550
Figure 7. Changes in the number of sub-watersheds with different risk levels. Note: I = extremely low risk; II = low risk; III = medium risk; IV = high risk; V = extremely high risk.
Figure 7. Changes in the number of sub-watersheds with different risk levels. Note: I = extremely low risk; II = low risk; III = medium risk; IV = high risk; V = extremely high risk.
Land 12 01330 g007
Land 12 01330 g008 550
Figure 8. Changes in the area of sub-watersheds with different risk levels. Note: I = extremely low risk; II = low risk; III = medium risk; IV = high risk; V = extremely high risk.
Figure 8. Changes in the area of sub-watersheds with different risk levels. Note: I = extremely low risk; II = low risk; III = medium risk; IV = high risk; V = extremely high risk.
Land 12 01330 g008
Land 12 01330 g009 550
Figure 9. Ecological risk time series transformation. Note: I = extremely low risk; II = low risk; III = medium risk; IV = high risk; V = extremely high risk.
Figure 9. Ecological risk time series transformation. Note: I = extremely low risk; II = low risk; III = medium risk; IV = high risk; V = extremely high risk.
Land 12 01330 g009
Land 12 01330 g010 550
Figure 10. Annual average rainfall trend.
Figure 10. Annual average rainfall trend.
Land 12 01330 g010
Land 12 01330 g011 550
Figure 11. Regional fertilizer application (nitrogen and phosphorous) variations.
Figure 11. Regional fertilizer application (nitrogen and phosphorous) variations.
Land 12 01330 g011
Land 12 01330 g012 550
Figure 12. Watershed slope map.
Figure 12. Watershed slope map.
Land 12 01330 g012
Land 12 01330 g013 550
Figure 13. Surface distance from sub-watersheds to lake.
Figure 13. Surface distance from sub-watersheds to lake.
Land 12 01330 g013
Table
Table 1. Data sources and descriptions.
Table 1. Data sources and descriptions.
Data TypeContextSource
Remote sensing dataLandsat5 (1990/09, 1994/10, 2002/08, 2007/05, 2009/11); Landsat8-OLI (2017/4, 2020/03, 2022/9)http://www.gscloud.cn (accessed on 11 October 2022
Meteorological dataAnnual rainfall (1990–2022)Meteorological Bureau of Guizhou Province and references
Topographic and soil data DEMhttp://www.gscloud.cn (accessed on 11 October 2022)
Amount of fertilizerAnnual nitrogen and phosphorus inputs (1990–2022)Guizhou Statistical Yearbook (1990–2022) and other references
Table
Table 2. Pollution output (absorption) weights for different landscape types.
Table 2. Pollution output (absorption) weights for different landscape types.
Landscape TypeRatio of TN Discharge
(Absorption) Coefficient		Ratio of TP Discharge
(Absorption) Coefficient		Weight of TN Discharge
(Absorption)		Weight of TP Discharge
(Absorption)
Construction land7.131.3211
Cultivated land0.840.250.120.19
Forest5.560.530.780.4
Grassland4.530.450.640.34
Water000.010.03
Unused land000.240.1
Note: TN = total nitrogen (kg/mu·a−1); TP = total phosphorus (kg/mu·a−1). The weight of the TN or TP discharge (absorption) of a landscape type is the ratio of the TN or TP discharge (absorption) coefficient of the landscape type to the maximum TN or TP discharge (absorption) coefficient in all landscapes.
Table
Table 3. Landscape area statistics (1990–2022) in km2.
Table 3. Landscape area statistics (1990–2022) in km2.
Landscape Type1990199420022007201120172022
Cultivated land1107.561016.90838.60835.73746.54583.04605.20
Forest672.80715.11907.06921.86826.90915.65853.04
Grassland35.9045.3635.7713.0111.68101.5695.68
Construction land43.3955.2550.5691.95242.57250.68285.57
Unused land0.100.020.151.748.630.011.49
Water55.2782.3682.8750.7278.6864.0874.04
Table
Table 4. Number and area statistics of risk-rated sub-watersheds (number, km2).
Table 4. Number and area statistics of risk-rated sub-watersheds (number, km2).
YearRisk LevelIIIIIIIVV
1990Number of sub-watersheds103267106127
Area380.341038.82437.5538.5219.78
1994Number of sub-watersheds105272102106
Area403.551073.74387.5032.2417.99
2002Number of sub-watersheds943306335
Area391.031277.09224.089.5313.27
2007Number of sub-watersheds15624274158
Area639.94905.92286.5756.5925.99
2011Number of sub-watersheds712451182734
Area261.1050960.79445.08124.28123.76
2017Number of sub-watersheds98275851027
Area397.991089.58295.0834.1898.17
2022Number of sub-watersheds203151241323
Area74.791248.56457.2350.1884.26
Note: I = extremely low risk; II = low risk; III = medium risk; IV = high risk; V = extremely high risk.
Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content.

Share and Cite

MDPI and ACS Style

Zhou, Z.; Zhao, W.; Lv, S.; Huang, D.; Zhao, Z.; Sun, Y. Spatiotemporal Transfer of Source-Sink Landscape Ecological Risk in a Karst Lake Watershed Based on Sub-Watersheds. Land 2023, 12, 1330. https://doi.org/10.3390/land12071330

AMA Style

Zhou Z, Zhao W, Lv S, Huang D, Zhao Z, Sun Y. Spatiotemporal Transfer of Source-Sink Landscape Ecological Risk in a Karst Lake Watershed Based on Sub-Watersheds. Land. 2023; 12(7):1330. https://doi.org/10.3390/land12071330

Chicago/Turabian Style

Zhou, Zhongfa, Weiquan Zhao, Sisi Lv, Denghong Huang, Zulun Zhao, and Yaopeng Sun. 2023. "Spatiotemporal Transfer of Source-Sink Landscape Ecological Risk in a Karst Lake Watershed Based on Sub-Watersheds" Land 12, no. 7: 1330. https://doi.org/10.3390/land12071330

APA Style

Zhou, Z., Zhao, W., Lv, S., Huang, D., Zhao, Z., & Sun, Y. (2023). Spatiotemporal Transfer of Source-Sink Landscape Ecological Risk in a Karst Lake Watershed Based on Sub-Watersheds. Land, 12(7), 1330. https://doi.org/10.3390/land12071330

Note that from the first issue of 2016, this journal uses article numbers instead of page numbers. See further details here.

Article Metrics

Back to TopTop