The impact of land use on water quality at different spatial scales in the area around Chaohu Lake region Open Access

Water quality serves as a critical index of the health of riverine and lacustrine ecosystems and plays a pivotal role in the ecological and environmental safety as well as in the sustainable development of economic and social systems (Wu 2010; Sakke et al. 2023). However, rapid population growth and continuous urban expansion have led to significant changes in land use, exerting substantial pressure on local aquatic environments and exacerbating water quality degradation, thereby presenting a prominent environmental challenge (Bwapwa 2018). Research has identified non-point source pollution as a key factor affecting water quality, with land use playing a decisive role in the generation, migration, and transformation of such pollution (Chen et al. 2016; Oliveira et al. 2017; Mainali & Chang 2018; Wu & Lu 2019; Weiwei et al. 2020). By strategically managing land use, it is possible to control the input, output, and transformation of non-point source pollution, thereby achieving improvements in water quality (Gang 2016). Therefore, exploring the relationship between land use and water quality not only provides crucial guidance for the management and planning of land use but also offers a significant scientific basis for the enhancement of water quality.

In recent years, studies, both domestic and international, have primarily explored the relationship between land use and water quality from two perspectives: the natural attributes (types of land use) and spatial scales (proportional area) (Mengjun et al. 2024). For instance, research by Li et al. (2008) examined the relationship between land use and water quality in China's Han River Basin, while Wei et al. (2020) investigated the impact of land degradation in catchment areas on the water quality of lakes and reservoirs. Additionally, Sliva & Williams (2001) employed buffer zone and whole-watershed approaches to study the impact of land use on river water quality. These studies have found that an increase in built-up and agricultural lands intensifies the risk of river water quality degradation, whereas a higher proportion of forest and grassland areas positively contributes to water quality improvement. Owing to spatial scale effects varying by region, there is debate over the spatial scale at which land use has the strongest impact on water quality. Pratt & Chang (2012) observed that, compared to buffer zone scales, the impact of land use changes on river water quality is more pronounced at the watershed scale. Shi et al. (2017) demonstrated that the correlation between land use and river water quality is highest within a 1,000 m riparian buffer zone, and that different land use indices have varying scale effects on water quality; Zhang et al. (2019) similarly noted that the strongest spatial scale impacts of land use on water quality occur within 1,000 m buffer zones and sub-watershed scales. Moreover, the influence of land use on surface water quality also exhibits seasonal variations, with seasonal changes affecting the relationship between land use and water quality indices (Wang et al. 2009; Kang et al. 2010; Huang et al. 2013).

Chao Lake, one of China's five major freshwater lakes, is situated in an economically advanced and densely populated watershed with a high degree of intensive land use, which has led to notable water quality deterioration (Jianshu 2022; Xiao 2022; Yanzhao 2023). However, with the Chinese government's increasing emphasis on ecological and environmental protection, the management and preservation of Chao Lake's ecological environment have become focal points of interest. Current scholarly research in the Chao Lake Basin has explored changes in land use and ecological condition assessments (Jia 2018; Chunqiu 2023), as well as estimations of agricultural non-point source pollution loads (Linlin 2018). Nevertheless, the impacts of land use practices at various spatial scales on surface water within the surrounding Chao Lake region remain unclear. The watershed contains numerous rivers, with 39 major and minor tributaries flowing into the lake, making the water quality of these inflows critical to the overall quality of Chao Lake (Rui 2009; Zhen 2017). Therefore, this study focuses on the area around Chaohu Lake region, utilizing land use patterns and river monitoring section water quality data to apply mathematical statistics and spatial analysis methods (Aiping et al. 2023; Mingzhu et al. 2024; Xuda et al. 2024). This research aims to examine the relationships between land use structures at different spatial scales and surface water quality in the area around Chaohu Lake region. It seeks to identify the primary land use types and spatial patterns that influence changes in river water quality in this area, providing a scientific basis for the rational allocation of soil and water resources and the improvement of water quality in the basin.

This study is based on 10 national surface water assessment sections located in the main rivers within the area around Chaohu Lake region, all of which are distributed within the study area. The characteristics of the water quality assessment sections are presented in Table 1. Based on the monthly average precipitation in the study area, the rainy season is defined as May to October, and the dry seasons are January to April and November to December. Data for all water quality assessment sections cover January to December 2021, sourced from the China National Environmental Monitoring Station's integrated surface water data (https://www.cnemc.cn/). The water quality monitoring data include nine parameters: water temperature, pH, dissolved oxygen, conductivity, turbidity, permanganate index (COD_Mn), ammonia nitrogen (NH₃-N), TP, and TN.

The area around Chaohu Lake region, predominantly a plain with a dense network of rivers, often experiences water quality variations at the same section due to influences from multiple directions. Therefore, a circular buffer zone centered on the water quality monitoring sections is employed as the hydrological unit (Jie et al. 2017). The radius of the buffer zone is primarily determined based on studies of regional water environment and land use scale effects. Scholars both domestically and internationally commonly use a minimum spatial scale of 100–500 m and a maximum spatial scale of 2,000–3,000 m (Jialin et al. 2022; Mengjun et al. 2024). Based on the scope of the study area and the characteristics of riverbank land use types, buffer zones of 300, 500, 1,000, 1,500, 2,000, and 3,000 m were delineated around the water quality monitoring sections, as shown in Figure 2. These zones were then intersected with land use types maps to obtain data on the proportion of different land use types at various spatial scales.

Using the Statistics 26.0 software platform, the water quality data for both the rainy and dry seasons at national surface water assessment cross-sections were subjected to a non-parametric Kolmogorov–Smirnov test (KS). It was found that the water quality data did not conform to a normal distribution, prompting the use of Spearman's rank correlation analysis to elucidate the relationship between the proportional areas of different land use types and water quality indices (Lu et al. 2021). Temporal differences in water quality were assessed using the non-parametric Mann–Whitney U test (Li et al. 2018).

Redundancy analysis (RDA) is now widely employed to determine the relationships between environmental factors and landscape metrics. In this study, after performing detrended correspondence analysis (DCA) on various water quality indices at each monitoring section using Canoco 5 software, it was found that the gradient values of the water quality data were less than 3. Therefore, RDA was selected (Lin & Siyue 2021). The mean values of each water quality index from May to October were used as the indices for the rainy season, and the mean values from January to April and November to December were used as indices for the dry season. Both dry and rainy season water quality indices were treated as response variables, with the proportional areas of land use types serving as explanatory variables to account for the influence of land use patterns on the variability of multiple water quality indices across seasons.

Analyzing the temporal variations of water quality indices, it is observed that the pH levels fluctuate between 7 and 8, with an average pH of 7.4 during the dry season and 7.37 during the rainy season, indicating slightly alkaline water conditions. The DO index exhibits significant seasonal variation, ranging from 3.57 to 10.52 mg/L, with an average of 8.79 mg/L in the dry season compared to 6.03 mg/L in the rainy season (p < 0.01), suggesting increased DO consumption during the rainy season. The fluctuation range for COD_Mn is between 3.15 and 5.45 mg/L and for TP between 0.04 and 0.24 mg/L. Seasonal differences in COD_Mn concentrations are not statistically significant (p > 0.05), while those in TP show significant variation (p < 0.05), with both indices generally lower in the dry season except at monitoring point S5 during the rainy season, where values meet the Class III standard for surface water. The NH₃-N ranges from 0.07 to 1.6 mg/L, and TN from 1.26 to 6.91 mg/L, with both showing significant seasonal differences (p < 0.05) and generally higher concentrations in the dry season. Notably, TN values at all monitoring points did not meet the Class III surface water standard throughout the seasons, indicating severe TN pollution in the area around Chaohu Lake region. In addition, due to poor hydrodynamic conditions in the Chaohu Lake area and other reasons, combined with the increasing and accumulated pollution load, the eutrophication of the water body is exacerbated, which is prone to trigger the outbreak of blue-green algae blooms. In 2021, the Chaohu Lake area as a whole was in a slightly eutrophic state, among which the western half of the lake was in a moderately eutrophic state. In 2021, a total of 33 algal blooms were detected in Chaohu Lake, with the largest algal bloom covering an area of 282.88 km². Overall, there is a marked distinction in water quality indices between the dry and rainy seasons across the area around Chaohu Lake region, particularly for DO, suggesting moderate overall water quality.

Spatially, while pH, DO, and COD_Mn show minimal variation, significant differences are noted in NH₃-N, TP, and TN concentrations across different regions. High pollution concentrations are observed at monitoring points S5 and S8 for NH₃-N, reaching maximum levels of 1.6 and 1.14 mg/L, respectively. TP and TN concentrations are notably higher at points S5, S7, and S8, with TP peaking at 0.24 mg/L during the rainy season at S5, and TN peaking at 6.91 mg/L during the dry season at the same point. Overall, water quality varies significantly across different spatial locations, with the western half of Chao Lake, connected to the densely populated and economically advanced urban area of Hefei, exhibiting inferior water quality compared to the eastern half. This disparity largely results from significant differences in pollution sources between the two areas. Additionally, monitoring point S5 experiences the most severe water quality pollution, located on the Nanfei River, which traverses the city and hosts Hefei's largest port. The developed water transport there contributes to substantial pollution loads, exacerbating non-point source pollution.

The pH levels exhibit minor variation among different monitoring points and between seasons, indicating stable annual values with minimal influence from land use, thus not prioritizing pH as a key index for exploring response relationships.

Unused land showed no proportional area within the 300, 500, and 1,000 m buffer zones. As the scale increased, the correlation between unused land during the rainy season and various water quality indices was generally weak. Grassland exhibited a negative correlation with COD_Mn, NH₃-N, TP, and TN at smaller scales, while showing a positive correlation with DO. With increasing spatial scale, the strength of these correlations first intensified and then diminished, with the weakest correlations observed in the 3,000 m buffer zone. Notably, grassland showed significant positive correlations with NH₃-N, TP, and TN within the 1,000 and 2,000 m buffer zones (p < 0.05). Water areas were predominantly negatively correlated with COD_Mn, NH₃-N, TP, and TN, and positively correlated with DO. In the 300 m buffer zone, a significant negative correlation with TN was observed (p < 0.05), and in the 500 m buffer zone, significant negative correlations with NH₃-N and TN were noted (p < 0.05). As the spatial scale increased, the correlations with water quality indices generally showed a decreasing trend. Forest land consistently showed negative correlations with COD_Mn, NH₃-N, TP, and TN across all scales, and positive correlations with DO. Particularly, in the 300, 500, 1,000, and 3,000 m buffer zones, forest land exhibited significant negative correlations with COD_Mn and TP (p < 0.05), and at the 2,000 m buffer zone, the negative correlation with COD_Mn was highly significant (p < 0.01). Construction land displayed weaker correlations with water quality indices at smaller scales, but significant and highly significant correlations with COD_Mn, NH₃-N, TP, and TN were observed at and beyond the 1,500 m buffer zone. Cultivated land exhibited fluctuations in correlation strength with water quality indices; at the 500 m and 3,000 m buffer zones, a highly significant negative correlation with TP was observed (p < 0.01), and within the 3,000 m buffer zone, significant positive correlations with COD_Mn and NH₃-N were noted (p < 0.05). A significant negative correlation with DO was evident in the 2,000 m buffer zone (p < 0.05).

This figure illustrates the relationship between various land use types and water quality indices across different spatial scales during the rainy season. In the dry season, the correlation between unused land and water quality indices is stronger than in the rainy season, showing a significant negative correlation with COD_Mn within a 2,000 m buffer zone (p < 0.05). Grassland exhibits a significant positive correlation with TP and a negative correlation with DO within 1,000 and 2,000 m buffer zones, respectively (p < 0.05), maintaining overall consistency with the rainy season, though with a slightly reduced correlation strength. Water areas show a change in correlation from the rainy season primarily within the 300 and 500 m buffer zones, with weakened negative correlations with NH₃-N and TN, and a significant positive correlation with DO (p < 0.05). Forest land maintains consistent correlations with water quality indices as in the rainy season, albeit with notably reduced strength. Construction land demonstrates an increasing correlation with increasing spatial scales, showing significant and highly significant positive correlations with TP and a significant negative correlation with DO at larger spatial scales. Cultivated land, across various spatial scales, shows positive correlations with COD_Mn, NH₃-N, TP, and TN, and a negative correlation with DO, with significant and highly significant positive correlations with TP at the 500 and 3,000 m buffer zones, respectively. Overall, the dry season shows a noticeable reduction in the number and strength of significant and highly significant correlations between land use types and water quality indices compared to the rainy season, with variations in both positive and negative correlations.

The RDA (Table 3) reveals that during the dry season, the buffer zone scales demonstrating the impact of different land use types on water quality are ranked in the following order: 3,000 m > 500 m > 2,000 m > 300 m > 1,500 m > 1,000 m. The interpretation ratio of land use type area to water quality index in the 3,000 m buffer zone is the largest. During the rainy season, the buffer zone scales demonstrating the impact of different land use types on water quality are ranked in the following order: 3,000 m > 500 m > 2,000 m > 1,500 m > 300 m > 1,000 m. Also, the interpretation rate of land use type area ratio in the 3,000 m buffer zone is the largest. Both seasons show that the proportion of land use types within the buffer zones and their explanatory power on water quality indices exhibit a trend of overall fluctuations rising. The maximum explanatory power is reached within the 3,000 m buffer zone, with 95.70% (P = 0.004) during the dry season and 91.00% (P = 0.004) in the rainy season. This indicates that within a 3,000 m buffer zone, land use has the most significant impact on water quality, thereby establishing 3,000 m as the optimal buffer zone scale for assessing the influence of land use on water quality.

The RDA plots reveal the correlations between land use types and water quality indices across various buffer zones. In the dry season (Figure 7), the proportion of forest land and water areas shows a negative correlation with COD_Mn, NH₃-N, TP, and TN, while demonstrating a positive correlation with DO. Notably, in the 1,000 m buffer zone, water areas exhibit a weaker correlation with COD_Mn and TP (angles near 90° indicate a weak correlation). Conversely, the proportion of cultivated land correlates positively with COD_Mn, NH₃-N, TP, and TN, and negatively with DO. Within the 1,500 m buffer zone, cultivated land shows a weak positive correlation with COD_Mn and TN, and strong positive correlations with NH₃-N and TP are evident in the 300, 500, and 3,000 m buffer zones. The proportion of construction land is positively correlated with COD_Mn, NH₃-N, TP, and TN, and negatively with DO. At the smaller spatial scales (300 and 500 m), the correlation with NH₃-N and TP is weaker, but as the spatial scale increases, the correlation initially strengthens before diminishing. Grassland proportions exhibit considerable variability in correlation with water quality indices; at smaller scales (300–500 m), they are negatively correlated with COD_Mn, NH₃-N, TP, and TN, and positively with DO. However, as the scale increases (1,000–3,000 m), these correlations become positive with COD_Mn, NH₃-N, TP, and TN, and negative with DO, with weaker correlations with NH₃-N in the 1,500 and 3,000 m buffer zones. In larger scales (2,000–3,000 m), unused land proportions are negatively correlated with COD_Mn, NH₃-N, TP, and TN, and positively with DO. During the rainy season (Figure 8), the main differences are observed in the 1,000 m buffer zone, where the proportion of water areas is negatively correlated with DO, while construction land and grassland proportions are positively correlated with it. Furthermore, in larger scales (1,500–3,000 m), the correlation between construction land and DO is weaker; however, the correlation between grassland proportions and NH₃-N in the 1,500–3,000 m buffer zones is stronger compared to the dry season. Overall, the correlation between land use types and water quality indices during the rainy season remains largely consistent with the dry season, though the strength of the correlations varies. The results of the RDA are in general agreement with those obtained from Spearman's rank correlation analysis, mutually corroborating the relationships and also revealing the optimal buffer zone scales for studying the influence of land use structure on water quality across different periods and buffer zone scales.

Different land use types influence water quality by affecting surface morphology, hydrological cycles, biodiversity, and the migration and transformation of pollutants in river and lake basins (Mengjun et al. 2024). The findings of this study indicate that an increase in the proportion of cultivated land and construction land heightens the risk of water pollution. Conversely, areas covered by forest land and water bodies are conducive to the improvement of water quality. Grasslands exhibit varying correlations with water quality across different spatial scales. Specifically, grassland areas near shorelines are inversely correlated with water pollution. However, as the spatial scale increases, this correlation changes and the degree of correlation weakens. Construction land, which is predominantly located in urban areas, contributes to non-point source pollution through domestic sewage and waste, and the presence of impervious surfaces accelerates runoff formation. Consequently, during rainfall events, pollutants can enter rivers rapidly with the runoff, deteriorating water quality (Mengjun et al. 2024). Cultivated land, a primary land use type in the area around Chaohu Lake region, contributes to water pollution when unused fertilizers and pesticides are washed into nearby rivers through rainfall runoff, leading to a decline in water quality. Both water areas and forest lands generally exhibit a negative correlation with most water quality indices. An increased proportion of water areas not only dilutes water pollutants but also provides a habitat for aquatic life, which plays a role in the absorption and purification of pollutants. Forest lands contribute to water quality purification through multiple layers, including the canopy, underbrush, surface layer, and soil layer, which intercept and absorb pollutants from precipitation and surface runoff (Wei et al. 2023). Grasslands near shorelines show a negative correlation with water quality indices, suggesting that the root systems actively intercept and absorb pollutants. However, as the spatial scale increases, this purification capability diminishes and exhibits varying degrees of correlation.

The impact of land use types on river water quality varies across different spatial scales. As the spatial scale changes, the relationship between the watershed and water quality indices evolves in accordance with the varying structures of land use (Li et al. 2018). This study, employing RDA, reveals that across both the dry and wet seasons, the highest explanatory power for water quality (93.35%) is observed at a 3,000 m buffer zone spatial scale. This finding aligns with research conducted in the Haihe River Basin (Tianjin section) (Mengjun et al. 2024), but differs from results in the Baihe River Basin (Mingzhu et al. 2024). The disparity may be attributed to regional human activities altering land use patterns, which in turn affect hydrological characteristics, material transfer, and energy flows, thereby causing spatial differentiation in watershed ecological processes and water quality (Zhou et al. 2012). The area around Chaohu Lake region, characterized by flat terrain, fertile soil, high population density, and substantial areas of cultivated land and construction land within a 3,000 m buffer zone around water quality monitoring points, is heavily impacted by human activities. This area is critical for future aquatic environmental and ecological protection. Recommendations include increasing the efficiency of fertilizer and pesticide use during agricultural production to reduce quantities and promote green planting to lessen the agricultural impact on surface water pollution. In urban development, enhancing the efficiency of sewage treatment, addressing deficiencies in municipal sewage treatment facilities, advancing waste segregation and collection, and strengthening the construction of urban green spaces are advised.

Research across multiple regional watersheds indicates seasonal variations in the impact of land use types on water quality, as seen in the Wuding and Yanhe River Basins (Xuda et al. 2024). In this study, the explanatory power of land use structure on water quality is higher during the rainy season compared to the dry season under the same spatial scale (see Table 2). This seasonal pattern is consistent with findings from the Haihe River Basin (Tianjin section) and Baihe River Basin (Mingzhu et al. 2024). The primary influence of rainfall on seasonal differences in water quality is its intensification of surface runoff strength. Pollutants on the ground are washed away by water, accumulating in runoff that converges into nearby rivers, thereby polluting the aquatic environment. The study concludes that pollutant concentrations are relatively higher during the rainy season, suggesting that water quality in the area around the Chaohu Lake region is principally affected by surface scouring caused by rainfall. Rainfall during the rainy season in the Chao Lake Basin is generally higher than in the dry season, with precipitation accounting for more than 70% of the annual total. Intense rainfall combined with high soil moisture content facilitates the formation of surface runoff and nutrient output.

• (1) In the area around the Chaohu Lake region, the predominant land use types are cultivated land, construction land, and water areas. Across various buffer zone scales, cultivated land accounts for an average of 31.83% of the area, construction land constitutes 23.57%, and water areas make up 31.9%.

• (2) There are notable differences in water quality indices between dry and rainy seasons, as well as across different spatial locations. Overall, water quality during the dry season is superior to that during the rainy season. Concentrations of COD_Mn and TP are higher during the rainy season, whereas concentrations of DO, NH₃-N, and TN are higher during the dry season. Relative to the Class III surface water standards, TN levels exceed the permissible limits to varying degrees in both seasons. The water quality in the eastern half of Chao Lake is generally better than that in the western half, with particularly notable conditions at the Nanfei River monitoring points.

• (3) Land use types significantly influence river water quality, with more pronounced effects during the rainy season. Both cultivated land and construction land show a positive correlation with various water quality indices, especially with NH₃-N and TP concentrations. Conversely, forest lands and water areas generally exhibit a negative correlation with water quality indices. The correlation between grasslands and water quality indices alternates as the spatial scale increases.

• (4) RDA reveals similar patterns of explanatory power for variations in water quality across different spatial scales in both dry and rainy seasons. During the dry season, land use types have a higher overall explanatory power for variations in water quality compared to the rainy season. Considering both periods together, a buffer zone scale of 3,000 m within the area around Chaohu Lake region provides the maximum explanatory capacity for changes in water quality, identifying it as the optimal buffer zone scale for assessing the impact of land use on water quality in this region.

This research was funded by the self-dependent Research and Development plan of Anhui and Huaihe River Institute of Hydraulic Research (KY202306), Jianghuai Meteorological Joint Project of Anhui Natural Science Foundation (2208085UQ12).

First Author and Corresponding author: Sheng Li, born in October, 1995, postgraduate degree, water resources and hydropower engineer. The research direction for water ecology and remote sensing of water conservancy. E-mail: [email protected].

All relevant data are included in the paper or its Supplementary Information.

The authors declare there is no conflict.