Changes in water quality in Dianchi Lake over a long temporal scale have been significantly influenced by climate change and policy issuance. This study utilized the Mann -Kendall test and Theil -Sen estimation to examine water quality trends at 10 monitoring sites within Dianchi Lake from 1989 to 2018. The corresponding impacts of policies, including point source (PS) and non-point source (NPS) pollution control, ecological restoration (ER), and enforcement and supervision (ES), along with climatic conditions, were evaluated. Analysis of regression curves of water quality indexes indicates a slight increase in ammonia (NH4+-N) and total nitrogen (TN), while total phosphorus (TP), biochemical oxygen demand (BOD5), permanganate index (CODMn), and chlorophyll-a (Chl-a) decreased from the baseline time (year 1989). During the mid-phase (2001 -2006) of the study period, the impacts of policies on water quality were inconspicuous, with water temperature affecting Chl-a concentration variations. During the 11th Five-Year Plan, pollutants like NH4+-N, TN, TP, and CODMn peaked and then significantly declined between 2009 and 2015, likely due to comprehensive issuance of NPS and ER policies. This study can provide a reference for water quality management in eutrophic lakes.

  • The water quality of Dianchi Lake improved during the 13th Five-Year Plan (FYP) compared with the 7th FYP.

  • Water quality responses to policy issuance can be categorized into rapid or negligible/delayed patterns.

  • Water temperature-driven changes in water quality were only found in the middle of the study period.

Lakes play a vital role in supporting the function of the earth's ecosystems and social environment. They serve a variety of purposes such as maintaining biodiversity, regulating local climate, flood management, water storage, generating electricity, providing a recreational landscape, and fisheries production and breeding (Hu et al. 2023). In the long term, human activities and climate change are the primary factors that influence the dynamics of lake water quality (Hou et al. 2022; Tang et al. 2022). The introduction of enriched pollution loads from human activities into lakes leads to eutrophic aquatic systems. Excess nutrients, particularly nitrogen and phosphorus, can result in frequent outbreaks of algal blooms and lead to deterioration in water quality. These phenomena occur in lake water bodies globally, especially in lakes that are located in undeveloped regions where the economy is dependent on agricultural fertilizers (Jenny et al. 2020; Hou et al. 2022). Warming may exacerbate algal bloom outbreaks under conditions of high nutrient concentrations (Paerl & Paul 2012; Gobler 2020; Liu et al. 2024). When the concentration of pollutants in lake waters reaches a certain threshold, the normal functioning of the lake ecosystem is disrupted. Understanding the influencing factors and long-term trends of lake water quality can lay the foundation for future water resource management.

During the early 1980s, the water quality of Dianchi Lake rapidly deteriorated owing to the discharge of large amounts of wastewater. As a result, it became one of the most polluted lakes in China (Liu 2017). Since the 9th Five-Year Plan, Dianchi Lake has consistently maintained the highest eutrophic status among three lakes in China (i.e., Taihu Lake, Chaohu Lake, and Dianchi Lake). Since 1986, the Chinese government has invested billions of dollars to combat water pollution in Dianchi Lake. The implementation of projects such as the interception of wastewater around the lake and the upgrading of wastewater plants has made it possible to reuse part of the water in Dianchi Lake. Water reuse has alleviated the water shortage and ameliorated the level of pollution in Dianchi Lake in some ways. Benefiting from these management measures, the water quality has improved from class V to class III (Kong et al. 2021). Over the past 30 years, the government has implemented a series of policies to improve water quality based on more scientific and effective methods. These include the Comprehensive Management Program for Pollution in Dianchi Lake and the Yunnan Provincial Regulations for Dianchi Lake Protection, which were formulated in 1989. Subsequently, projects such as the interception of wastewater around the lake, dredging of substrate, and recharging of the Niulanjiang–Dianchi Lake have been carried out (Liu & Wang 2016). However, the relationship between changes in water quality and algal growth with these policies has not been thoroughly studied at this time. Previous studies have evaluated the response of water quality to nonpoint-source (NPS) pollution controls and found such controls can be effective in the initial stages. However, this efficiency diminishes as inputs increase (Yan et al. 2024). Furthermore, a previous study assessed that optimizing dynamic water transfers not only decreased total nitrogen (TN) and total phosphorus (TP) in Dianchi Lake by 7 and 6%, respectively, but also reduced the quantity of water diversion by 75% (Jiang et al. 2024). However, the mechanism of long-term water quality change and algal bloom dynamics in Dianchi Lake from the perspective of anthropogenic intervention, specifically referring to policy issuance, still remains.

The Mann–Kendall (M–K) test is commonly used to analyze trends in water quality variations and to detect abrupt-changing points of water quality indexes at the long-term timescales (Halil Ibrahim 2022). It has been applied to various water body types, including lakes, rivers, and groundwater systems (Kisi & Ay 2014; Lopez et al. 2015). This test is typically used in combination with Sen's slope method to determine water quality trends synergistically. For instance, a previous study used this method to examine 10-year trends in organic indicators (biochemical oxygen demand (BOD5) and permanganate index (CODMn)), nutrients, suspended solids, dissolved oxygen (DO), and precipitation in rivers (Mustapha 2013) and found that changes in water quality were significantly influenced by human activities. In addition, the trends variations of organic pollution indicators (BOD5 and CODMn), nitrogen and phosphorus nutrient salts, and pH indicators in Dongting Lake over a long period of time (1991–2018) were analyzed (Geng et al. 2021), and found that most of the change in water quality occurred after the construction of Three Gorges Dam. The concentrations of TN and BOD5 showed a significant increase after the construction of the dam. The aforementioned approach was also used to identify changes in lake water storage, which significantly influenced the DO, total suspended solids, and Secchi depth (SD) in a lake (Yenilmez et al. 2011). The temporal trends of nutrient and DO variations were analyzed in a highly modified river in Europe from 1970 to 2014. They discovered a significant decrease in nitrogen and phosphorus concentrations over the past 20 years, particularly in ammonia nitrogen. In addition, an increase in DO was observed following the implementation of the European Water Framework Directive (WFD). The operation of WFD led to the disappearance of hypoxia caused by summer algal blooms. However, NPS pollution control had less effect in reduction of nitrate loads; instead, nitrate concentrations had increased by almost 150% since 1980 (Romero et al. 2016).

This study aimed to identify the characteristics of water quality variations and delineating the relevant policy impacts over a span of 30 years. Water quality data from 10 evenly distributed monitoring sites within Dianchi Lake were collected for the duration of 1989–2018. The M–K test and Theil–Sen estimation were used to (1) characterize the trend of water quality variations in Dianchi Lake over the 30-year period; (2) elucidate the underlying reasons for the occurrence of abrupt-changing points and overall trend in water quality variations; and (3) illustrate the response of water quality to shifts in policy.

Study area

Dianchi Lake, located in Kunming, Yunnan Province (102°29′–103°01′ E, 24°28′–25°28′ N), is the sixth largest freshwater lake in China. Dianchi Lake has a water area of approximately 330 km2, with the deepest water depth of approximately 10.1 m and an average water depth of approximately 4.1 m. It is a typical plateau shallow lake (Gong et al. 2009). The Dianchi basin experiences a prevailing southwest wind with an annual average speed of approximately 3 m/s. The average temperature in the basin is approximately 15.8 °C. The basin experiences annual precipitation of approximately 1,064 mm, with 90% of it occurring between May and October. The recharge of Dianchi Lake primarily relies on the rivers and natural precipitation. River recharge constitutes approximately 70% of the total recharge, making it the most significant source of recharge for Dianchi Lake.

Data sources

All water quality data used in this study were derived from the Center for Environmental Monitoring in Kunming, Yunan Province of China. The information on the issuance records of policies and regulations were provided by the Tsinghua University–Kunming Plateau Lakes Joint Research Center. Water quality in 10 national surface water monitoring sites (Figure 1), Duanqiao (DQ), Caohaizhongxin (CHZX), Huiwanzhong (HWZ), Luojiaying (LJY), Guanyinshangdong (GD), Guanyingshanzhong (GZ), Guanyingshangxi (GX), Baiyukou (BYK), Haikouxi (HKX), and Dianchinan (DCN), from 1989 to 2018 (spanning three decades), were obtained in this study. The measurement of the data comply with the requirements of the ‘National Water Quality Sampling Technical Guidelines’ (HJ494-2009) and the long-term monitoring specifications for aquatic ecosystems. Water quality indexes, including water temperature (WT), DO, pH, SD, TN, TP, ammonia nitrogen (), chlorophyll-a (Chl-a), CODMn, and BOD5, were analyzed in this study because they can effectively describe the major environmental issues that have occurred in Dianchi over the past 30 years. The Five-Year Plans are an important part of China's national economic program, with 5-year intervals. The study period spanned seven Five-Year Plans, during which we averaged water quality data to analyze long-term trends in water quality variations. The seven periods are 1985 to 1990, 1991 to 1995, 1996 to 2000, 2001 to 2005, 2006 to 2010, 2011 to 2015, and 2016 to 2020, corresponding to the 7th, 8th, 9th, 10th, 11th, 12th, and 13th, plans respectively.
Figure 1

Location map showing (a) the study area in China, (b) the study area in Yunnan province, and (c) enlargement of the monitoring sites in the study area.

Figure 1

Location map showing (a) the study area in China, (b) the study area in Yunnan province, and (c) enlargement of the monitoring sites in the study area.

Close modal

Water quality evaluation methods

To evaluate water quality conditions and provide convincing results, this study used the water pollution index (WPI) to assess the physical, chemical, and ecological conditions of Dianchi Lake's water body (Hossain & Patra 2020). The specific method for calculating the WPI is shown in Equation (1) (Li et al. 2015):
(1)

The evaluation criteria for WPI are presented in Supplementary material, Table S1.

To evaluate the eutrophication status of the lake, the trophic level index (TLI) was used in this study. The calculation formula and evaluation criteria for TLI are shown in Supplementary material, Tables S2 and S3, respectively.

Data analysis methods

The Mann–Kendall trend test

The M–K trend test is a nonparametric test commonly used to analyze trends in hydrological factors such as changes in atmospheric water. It is also known as a distribution-free test (Sharma & Saha 2017). This study used the M–K trend test to analyze the trend of local long-term water quality indexes and precipitation amount in Dianchi Lake. The aim was to elucidate changes in water quality over the past 30 years, including the abrupt-changing point of the index. The specific calculation principle is outlined in Equations (2) and (3). The order series for the long-term data x with n sample sizes is constructed as follows:
(2)
(3)
The order series represents the cumulative number of times that the value at time i exceeds the value at time j. These statistics are defined assuming stochastic independence within the time series:
(4)
where UF1 = 0, E(Sk) and Var(Sk) are the mean and variance of the cumulative number Sk, respectively, and x1, x2, …, xn are independent of each other and have the same continuous distribution. This is calculated using Equations (5) and (6):
(5)
(6)

The UF series represents a standard normal distribution of statistics calculated from time series data x in the order of x1, x2, …, xn, after being given a significance level α, combined with the normal distribution table. If , it indicates a significant trend change in the series. The processes described earlier is repeated in a reverse order for the time series x, starting from xn and going backward to x1. At the same time, the time series data are also processed in the reverse order , k = n, n − 1, …, 1, UB = 0.

The curves for the and of the 11 water quality indicators were plotted. If the line was greater than 0, then that indicated an upward trend in the water quality indicator. Conversely, if it was less than 0, then that indicated a downward trend. If the line exceeded the confidence interval, then that indicated a significant upward or downward trend. The range beyond the critical line is considered to be the area where the time of mutation occurs. The intersection of two curves within the confidence interval is called the abrupt-changing point.

The statistical detection value of Z is calculated by Equation (7):
(7)

The magnitude of the statistical variable Z indicates the change in the trend of the data. A significance level (α) of 0.05 was set, where Z1 − α/2 = 1.96. The two-sided test does not accept the hypothesis when |Z| ≥ Z1 − α/2.

This study combined the M–K mutation test and the coupled analysis of the statistical test of Z-value to elucidate the water quality status of Dianchi Lake over the past 30 years, detect the water quality trends and their patterns, and analyze and identify the key pollutants affecting the water quality of Dianchi Lake.

Theil–Sen estimation

Theil–Sen estimation was used to obtain the water quality trends in each year relative to that in the first year of the 30-year study period. The Theil–Sen median method, also known as the Sen Slope estimation method, is a robust nonparametric statistical method for calculating trends. The method is computationally efficient, insensitive to measurement error and outlier data, and is widely used in trend analysis of long time series data. The calculation formula of statistics variables is as follows:
(8)
where x and y are time series data, and greater than 0 indicates that the time series has an upward trend, and otherwise, it has a downward trend.

Tendency of 30-year water quality variations

The overall trends in water quality variations over the 30-year period as well as average trend in each 5-year intervals (i.e., corresponding to seven of China's Five-Year Plan) of water quality variations were analyzed (Figure 2).
Figure 2

Average value and corresponding linear regression lines of indexes including (a) precipitation amount, (b) WT, (c) DO, (d) pH, (e) SD, (f) WPI, (g) , (h) TN, (i) TP, (j) CODMn, (k) BOD5, and (l) Chl-a for seven periods (i.e., 7th, 8th, 9th, 10th, 11th, 12th, and 13th Five-Year Plan period) spanning 30 years. The black line and dots in the figure represent the error bars and median value for corresponding periods, respectively.

Figure 2

Average value and corresponding linear regression lines of indexes including (a) precipitation amount, (b) WT, (c) DO, (d) pH, (e) SD, (f) WPI, (g) , (h) TN, (i) TP, (j) CODMn, (k) BOD5, and (l) Chl-a for seven periods (i.e., 7th, 8th, 9th, 10th, 11th, 12th, and 13th Five-Year Plan period) spanning 30 years. The black line and dots in the figure represent the error bars and median value for corresponding periods, respectively.

Close modal

The slopes of the linear regression curves for variations of 5-year average value in WT and pH indexes were 0.05 (R2 = 0.58) and 0.007 (R2 = 0.54), respectively, showing an increasing trend over the 30-year period. Conversely, BOD5 and Chl-a exhibit a decreasing trend with slopes of −0.07 and −0.001, respectively (Figure 2). The R2 for the linear regression curves of all other indexes were low, ranging from 0.00 to 0.22. In terms of pollutants, the slopes of the regression curves for the temporal variations of 5-year average of and TN were positive and approach zero, indicating stable and almost unchanged trends, whereas those for TP, BOD5, and CODMn were negative (Figure 2), denoting gradually decreasing.

The temporal variation of average value in water quality over the seven periods can be categorized into three groups: (i) continual reduction (Chl-a), (ii) the appearance of a single stationary point (, TN, and WPI), and (iii) the appearance of two to more stationary points (TP, BOD5, WT, DO, pH, CODMn, and SD). For the first category, there was a decrease in Chl-a concentration from 0.1 mg/L during the 9th Five-Year Plan period to 0.068 mg/L in the 13th Five-Year Plan period (Figure 2). Notably, there were two significant declines in both average and median concentration of Chl-a from the 9th to 10th and from 12th to 13th Five-Year Plan period (Figure 2). In the first decline period, the median value for WT, TP, CODMn, and BOD5 decreased, whereas that of nitrogen nutrients remained unchanged. In the subsequent decline period, median values for all indexes, including TN, TP, , CODMn, and BOD5, showed a decreased exception of WT, which increased. For the second category, stationary points of all indexes, specifically maximum concentration of , TN, and WPI, were observed during the 11th Five-Year Plan period. In the third category, appearances of stationary points were predominantly from the 8th to the 12th Five-Year Plan (Supplementary material, Table S4). Seventy-five percent of the indexes detected stationing points during the 9th and 11th Five-Year Plan periods (Supplementary material, Table S4).

The abrupt-changing point of water quality indexes over 30 years

Abrupt-changing points were identified in a subset of the metrics, including Pre, WT, DO, pH, SD, WPI, TP, CODMn, and BOD5 (Figure 3 and Table 1). Abrupt-changing points occurring within the first three years of the 30-year period were disregarded due to the great uncertainty (i.e., boundary effect) associated with such early-stage fluctuations in the study period.
Table 1

Time of occurrence of abrupt-changing point of water quality indexes during 30-year period

Indexes8th9th10th11th12th13th
Pre 1993 – 2004–2005 ↑ – 2014–2015 ↑ – 
WT 1992–1993 ↑
1994–1995 ↑ 
1996 ↑ – – – – 
DO 1991–1992 ↑ – – – – 2017 
pH 1993–1994 ↑ 2000–2001 ↑ – – – 2017 ↑ 
SD – 1997 – 2007–2008 ↑
2010 ↑ 
2014–2015 ↑ – 
WPI –  – – – 2017–2018 
TP – – – – 2015 – 
CODMn 1993–1994 ↑
1994–1995 ↑ 
– – 2006–2007 – 2017–2018 
BOD5 – – – – 2013–2014 – 
Indexes8th9th10th11th12th13th
Pre 1993 – 2004–2005 ↑ – 2014–2015 ↑ – 
WT 1992–1993 ↑
1994–1995 ↑ 
1996 ↑ – – – – 
DO 1991–1992 ↑ – – – – 2017 
pH 1993–1994 ↑ 2000–2001 ↑ – – – 2017 ↑ 
SD – 1997 – 2007–2008 ↑
2010 ↑ 
2014–2015 ↑ – 
WPI –  – – – 2017–2018 
TP – – – – 2015 – 
CODMn 1993–1994 ↑
1994–1995 ↑ 
– – 2006–2007 – 2017–2018 
BOD5 – – – – 2013–2014 – 

Note: The upward arrow indicates the increasing tendency of water quality at time of the abrupt-changing point (i.e., UF > 0).

Figure 3

Analysis of abrupt changes in the water quality of Dianchi Lake. The orange and blue lines refer to the sequential (UF) and reverse (UB) statistical curves, respectively. The intersection of the two lines indicates the time of the change. The gray dotted lines show the 0.05 significance level. The water quality indexes included (a) precipitation amount, (b) WT, (c) DO, (d) pH, (e) SD, (f) WPI, (g) , (h) TN, (i) TP, (j) CODMn, (k) BOD5, and (l) Chl-a.

Figure 3

Analysis of abrupt changes in the water quality of Dianchi Lake. The orange and blue lines refer to the sequential (UF) and reverse (UB) statistical curves, respectively. The intersection of the two lines indicates the time of the change. The gray dotted lines show the 0.05 significance level. The water quality indexes included (a) precipitation amount, (b) WT, (c) DO, (d) pH, (e) SD, (f) WPI, (g) , (h) TN, (i) TP, (j) CODMn, (k) BOD5, and (l) Chl-a.

Close modal

Most abrupt-changing points occur during the 8th Five-Year Plan period, followed by occurrences in 12th and 13th Five-Year Plan periods (Table 1). For water quality indexes indicating pollutant status, namely, WPI, TP, CODMn, and BOD5, the UF values were observed to be less than zero from the 11th to the 13th Five-Year Plan, indicating a downward trend in pollutants levels relative to the baseline year (1989). Abrupt-changing point of water quality variations occurring after the year 2000 was analyzed in detail in Section 4.2.

The overall trends of water quality variations over the span of 30 years, commencing from 1989, were examined using the Theil–Sen estimation. The trend in all water quality index variations as calculated by the M–K method was consistent with those calculated by Sen's slope method (Supplementary material, Table S5). However, a statistically significant trend was only present for select years.

The trend of water quality variations pre- and post-abrupt-changing points over 30 years

An abrupt-changing point of pH occurred between 2000 and 2001. Throughout this time interval, there was a noticeable decrease in pH, WT, and Chl-a, as well as increases in SD, , TN, and CODMn (Figure 4). From 2005 to 2006, WT, CODMn, BOD5, , TP, TN, WPI, Chl-a, and pH increased, whereas SD decreased (Figure 4). In the following years, between 2006 and 2007, there was an abrupt-changing point of CODMn, characterized by a surge in its concentration (Figure 4). From 2007 to 2008, an abrupt-changing point of SD occurred, marked by a decrease during that period. This period also witnesses increases in BOD5, , and pH, and decreases in TN, TP, and WPI. The SD showed an abrupt-changing point from 2009 to 2010, during which SD increased, while , TN, TP, and CODMn decreased, and Chl-a and BOD5 increased. In 2017, an abrupt-changing point of WPI occurred, which exhibited a downward trend in 2018, accompanied by a decrease in all other pollutant indices over the same period (Figure 4).
Figure 4

Variations of water quality of (a) precipitation amount, (b) WT, (c) DO, (d) pH, (e) SD, (f) WPI, (g) , (h) TN, (i) TP, (j) CODMn, (k) BOD5, and (l) Chl-a from 1989 to 2018. The relative changes of water quality postchange point compared with prechange point are indicated by the red lines.

Figure 4

Variations of water quality of (a) precipitation amount, (b) WT, (c) DO, (d) pH, (e) SD, (f) WPI, (g) , (h) TN, (i) TP, (j) CODMn, (k) BOD5, and (l) Chl-a from 1989 to 2018. The relative changes of water quality postchange point compared with prechange point are indicated by the red lines.

Close modal

Governance policy over 30 years for Dianchi Lake

The water pollution control and management for Dianchi Lake can be divided into three categories: laws/regulations, standards, and policies. Hereafter, these are collectively referred to as policy. Regarding the policy control mechanisms of pollution on water quality, these measures can be classified into four types: point-source (PS), NPS pollution control, ecological restoration (ER), as well as enforcement and supervision (ES) (Supplementary material, Table S6).

The development history of the four policies is significantly different. Policy for PS control developed rapidly before the year 2000 (Figure 5(a)). Since the issuance of the first PS control policy, i.e., ‘Emission control of priority polluting industries’, in 1988, a total of eight policies have been issued by 2000. Post-2010, however, policies targeting PS control entered a refinement phase, characterized by a markedly reduced frequency of policy issuance. ER policies were consistently released from the onset of 2000 until 2020, transitioning into the refinement stage by 2020 (Figure 5(c)). The period from 2008 to 2018 saw a rapid development in the number of enforcement and supervision policies, gradually transitioning into a refinement phase by 2018 (Figure 5(d)). Policies of NPS control were primarily issued between 2010 and 2015, with no subsequent policies released after 2015 (Figure 5(b)). In this study, these policy impacts can be characterized as having either a direct/rapid effect or an indirect/delayed effect on water quality variations.
Figure 5

Time of release of policies and cumulative number of policies released over time. The policies are categorized into (a) point-source pollutant control, (b) nonpoint-source pollutant control, (c) ecological restoration, and (d) enforcement and supervision type.

Figure 5

Time of release of policies and cumulative number of policies released over time. The policies are categorized into (a) point-source pollutant control, (b) nonpoint-source pollutant control, (c) ecological restoration, and (d) enforcement and supervision type.

Close modal

Significant improvement in water quality over 30-year in Dianchi Lake

Water quality significantly improved in the final 5-year stage (13th, 2016–2020) compared to the initial 5-year stage (7th, 1986–1990) (Figure 2). All pollutants, except for nitrogen nutrients (i.e., TN and ), showed a decreasing trend compared to the initial year of the study period, as indicated by the Theil–Sen results (Supplementary material, Table S5) and the slope of variations in the average values for the Five-Year Plan (Figure 2). In addition, the TLI calculated from the TN, TP, SD, and CODMn indexes indicate that water quality reverted to the initial moderate eutrophication state by the end of the 12th Five-Year Plan period (Supplementary material, Table S7). However, while Chl-a showed a decreasing trend throughout the study period (Figure 2), its TLI indicated a heavy eutrophication level during the initial phase (in 2016 and 2017) of the 13th Five-Year Plan period in Dianchi Lake (Supplementary material, Table S7). The high Chl-a concentration could result from algal blooms and/or endogenous pollution, potentially reflecting a bottleneck in water quality improvement (Liu et al. 2015).

The current reduction in most pollutants currently has been attributed to the completion of PS pollutant control policies (Liu et al. 2021), which have undergone a period of rapid development before entering a phase of refinement and consolidation (Figure 5). Meanwhile, policies for ER are in a rapid development stage and require substantial improvement (Figure 5). The incomplete and fragmented policies of the ER, along with a possible hysteresis response, are likely reasons why water quality remained at a moderate eutrophication status at the end of the study period (i.e., 2018). The most desirable outcome of ER is to stimulate a transition of the ecosystem from a phytoplankton-dominant to a macrophyte-dominant state (Qin et al. 2006). However, due to the significant uncertainty or hysteresis effects in ER (Pelletier et al. 2020; Wang et al. 2021), the government may face extreme challenges in implementing ER.

Chl-a concentration exhibited a notable decline during two periods from the 9th (1996–2000) to the 10th (2001–2005), and from the 12th (2011–2015) to the 13th (2016–2020) Five-Year Plans. The reduction in Chl-a concentration during the decrease from the 9th to the 10th Five-Year Plans can be primarily ascribed to the synergistic effects of WT and constrain of nutrients concentration. Specifically, both the median WT and the concentration of all nutrients, with the exception of nitrogen nutrients, exhibited a decrease during this period (Figure 2). The occurrence of algal bloom, driven by WT, aligns with findings from previous studies (Yang et al. 2018; Duan et al. 2022). Conversely, the period transitioning from the 12th to the 13th Five-Year Plans saw changes in Chl-a predominantly attributed to nutrient regulation. That is, a significant reduction in nutrient concentration was observed despite a significant increase in median WT (Figure 2). These nutrient-induced variations in Chl-a may be the result of the well-developed policy of PS and NPS control.

Nutrient concentration presented a consistent downward trend beginning in the 12th Five-Year Plan, i.e., characterized by significant reductions in TN and TP as well as a decrease in WPI (Figure 4). This trend may be attributed to the initiation of policy aimed at controlling NPS pollution, which entered a phase of rapid development starting in 2010. In addition, the issuance of three ES policies in the same year likely contributed to this observed reduction in nutrient levels.

Rapid, negligible, and delayed effects of policies on water quality

Changes in trends of water quality variations in this study can be partially attributed to the implementation of policy, as proved in previous studies (Harmel et al. 2014; Liu et al. 2021). The water quality response to policy intervention can be classified into categories including rapid, negligible, and delayed response dependent on response characteristics of policies. For pollutants such as , TN, TP, and CODMn, which exhibited peak concentrations during the 11th Five-Year Plan, the pattern of fluctuation followed by sharp decreases from 2009 to 2015 is likely attributable to the centralized issuance of NPS pollution control policies during this interval. This pattern exemplifies a direct and rapid response. Comparatively, peak concentration of BOD5 was recorded during the 8th Five-Year Plan, with a subsequent trend of fluctuation and reduction after 2000 (Figure 4). Until 2015, the decrease in BOD5 may be ascribed to either delayed or negligible effects of those policies. In addition, TP exhibited a rapid decline in 1997–1998, indicative of a rapid immediate reaction to the enactment of the ‘Prohibition and Limitation of Phosphorus’ policy released in 1997 (Supplementary material, Table S6). Concurrently, a surge in SD during this period indicates a significant response to the same policy. This policy likely played a role in reducing other pollutants, including CODMn, BOD5, and TN, alongside a decrease in the WPI throughout this period.

Policy-driven dynamics of abrupt change in water quality

A portion of the abrupt-changing point in water quality observed during the study can be attributed to variations in WT rather than the impacts of policy interventions over a short period of time. However, this temperature effect was predominantly noted in the middle of the study period (2000–2006). During this period, a majority of the pollutants, i.e., CODMn, , and TN, exhibited increases from 2000 to 2001 (Figure 4), a trend that did not respond to policies. Notably, abrupt-changing point of pH was observed (Figure 3), with a subsequent decrease during this interval (Figure 4). This phenomenon was likely triggered by the inhibition of algal growth due to decreased WT, as evidenced by the shrinking Chl-a concentration. Similarly, all pollutants (CODMn, BOD5, , TP, and TN) increased, and thus, WPI increased from 2005 to 2006 (Figure 4). Concurrently, an increase in Chl-a and pH and a decrease in SD during the same period confirmed the occurrence of high WT-induced algal blooms, leading to the subsequent deterioration of water quality. The rise of CODMn concentration from 2006 to 2007 may result from the deterioration of water quality in the previous year (Figure 4). During these intervals, the influence of WT on water quality was evidently more pronounced than that of policy measures.

The detection of abrupt-changing points of water quality after 2006 (during the 11th Five-Year Plan) highlights the predominant influence of policy interventions. Nevertheless, these observed variations in water quality could also be attributed to unobserved factors such as climate conditions and internal pollutants biogeochemical cycling within the lake's internal cross-sections over short temporal scales. A notable instance is the decrease in SD between 2007 and 2008, which may be associated with the biogeochemical processes triggered by a significant rise in pH caused by elevated BOD5 and concentration (Figure 4). Due to limitations in the available data, the precise mechanism behind this phenomenon are beyond the indicative meaning of the data. Contrary to prior instances where algal bloom proliferation was a primary determinant, the mechanisms underlying this phenomenon were not exclusively governed by algal activity. In contrast, the behavior of BOD5 and , and the concentration of other pollutants such as CODMn, TN, and TP, as well as WPI presented a decreasing trend following the abrupt-changing point in SD. This trend could be attributed to alterations in water levels, enhanced denitrification, and organic particle settlement processes, which was not considered in this study (Zhang et al. 2014; Wen et al. 2020).

The observation that SD exhibited an increase subsequent to the abrupt-changing point between 2009 and 2010, along with the decrease in all pollutants (with the exception of BOD5) during this period (Figure 4), can be attributed to the issuance of policies except for points-source pollutant control. The marked refinement in policies targeting NPS pollutant control and ER, as evidenced by the sharp increase in policy initiatives, contributed to the reduction in pollutants during 2013 to 2014 (Figure 5). Combined with the results of reductions in WPI and Chl-a, alongside increases in SD (Figure 4), these trends suggest a decay of algal blooms and an abatement in nutrients. These alterations are likely indicative of a positive response to these two types of policies.

Furthermore, the role of policies on the trend of water quality variations were evident in the period from 2016 to 2018. The improvement of water quality, characterized by a rapid decrease in WPI and a reduction in the concentrations of all pollutants, may be principally related to the issuance of the ‘Eco-compensation Penalties’ in 2017. However, post-policy issuance period witnessed recurrent deteriorations in water quality. The increase in Chl-a, TN, CODMn, and WPI between 2016 and 2017 predominantly reflects the historical legacy of endogenous pollution sources.

This study analyzed the overall trend of variations in water quality indexes and precipitation at 10 monitoring sites over 30 years (1989∼2018) within Dianchi Lake using the M-K test and Theil-Sen method. The impacts triggered by the relevant four types of policies (including (including point source. The impacts triggered by the relevant four types of policies (including (including point source. The study period spans the 7th to 13th Five-Year Plans in China. The and TN in Dianchi Lake slightly increased (indicated by a slope >0), while TP, BOD5, CODMn, and Chl-a decrease (slope <0) when compared to those concentrations at the onset of the study period. Post-2018, the TLI based on Chl-a concentrations, indicating an improvement in the eutrophic status of Dianchi Lake from severe to moderate. Changes in water quality trend in this study can be partially attributed to the issuance of policy. The water quality response to policy intervention can be classified into rapid, negligible, and delayed responses. An example of rapid response is the decrease in all pollutants (with the exception of BOD5) along with elevated SD between 2009 and 2010, which can be attributed to the issuance of policies except for PS pollutant control. The marked refinement in policies targeting NPS pollutant control and ER induced the reduction in pollutants during 2013–2014. It can be evidenced by the results of reductions in WPI and Chl-a, alongside increases in SD. The negligible response type was predominantly found during the middle phase of the study period, spanning from 2000 to 2006. The WT exerted a pronounced effect on water quality in this period. Between 2000 and 2001, an increase in CODMn, , and TN was observed, a pattern that appeared negligible to existing policies. This period also witnessed an abrupt-changing point in pH and its decline therein likely triggered by the suppression of algal proliferation (evidenced by the decrease in Chl-a concentration) due to dropped WT. Similarly, all pollutants (CODMn, BOD5, , TP, and TN) as well as WPI increased from 2005 to 2006. Concurrently, an increase in Chl-a and pH and a decrease in SD during the same period confirmed the occurrence of high WT-induced algal blooms, leading to the subsequent deterioration of water quality evidenced by the elevated CODMn. In light of Kunming City's ongoing development, it is advisable for the local government to integrate water reuse policy, which can reduce the discharge of waste water and improve the utilization rate of water resources. Such measures would yield significant environmental benefits for Kunming, contributing to Kunming city's sustainable development.

This work was supported by the Major Program of National Natural Science Foundation of China (Nos. 52293440, 52293442, and 52270044) and Yunnan Province Science and Technology Project ‘202305AM340008’.

Data cannot be made publicly available; readers should contact the corresponding author for details.

The authors declare there is no conflict.

Duan
H.
,
Shang
C.
,
Yang
K.
&
Luo
Y.
2022
Dynamic response of surface water temperature in urban lakes under different climate scenarios – A case study in Dianchi Lake, China
.
International Journal of Environmental Research and Public Health
19, 12142.
doi:10.3390/ijerph191912142
.
Geng
M.
,
Wang
K.
,
Yang
N.
,
Li
F.
,
Zou
Y.
,
Chen
X.
,
Deng
Z.
&
Xie
Y.
2021
Evaluation and variation trends analysis of water quality in response to water regime changes in a typical river-connected lake (Dongting Lake), China
.
Environmental Pollution
268
,
115761
.
https://doi.org/10.1016/j.envpol.2020.115761
.
Gobler
C. J.
2020
Climate change and harmful algal blooms: Insights and perspective
.
Harmful Algae
91
,
101731
.
https://doi.org/10.1016/j.hal.2019.101731
.
Gong
Z.-j.
,
Li
Y.-l.
,
Shen
J.
&
Xie
P.
2009
Diatom community succession in the recent history of a eutrophic Yunnan Plateau lake, Lake Dianchi, in Subtropical China
.
Limnology
10
(
3
),
247
253
.
doi:10.1007/s10201-009-0282-8
.
Halil Ibrahim
B.
2022
The short-term and seasonal trend detection of sediment discharges in Turkish rivers
.
Rocznik Ochrona Środowiska
24
,
214
230
.
doi:10.54740/ros.2022.016
.
Harmel
R. D.
,
Smith
P. K.
,
Migliaccio
K. W.
,
Chaubey
I.
,
Douglas-Mankin
K. R.
,
Benham
B.
,
Shukla
S.
,
Muñoz-Carpena
R.
&
Robson
B. J.
2014
Evaluating, interpreting, and communicating performance of hydrologic/water quality models considering intended use: A review and recommendations
.
Environmental Modelling & Software
57
,
40
51
.
https://doi.org/10.1016/j.envsoft.2014.02.013
.
Hossain
M.
&
Patra
P. K.
2020
Water pollution index – A new integrated approach to rank water quality
.
Ecological Indicators
117
.
doi:10.1016/j.ecolind.2020.106668
.
Hou
X.
,
Feng
L.
,
Dai
Y.
,
Hu
C.
,
Gibson
L.
,
Tang
J.
,
Lee
Z.
,
Wang
Y.
,
Cai
X.
,
Liu
J.
,
Zheng
Y.
&
Zheng
C.
2022
Global mapping reveals increase in lacustrine algal blooms over the past decade
.
Nature Geoscience
15
(
2
),
130
134
.
doi:10.1038/s41561-021-00887-x
.
Hu
C.
,
Hu
T.
&
Liang
L.
2023
Spatial variation and potential ecological risk assessment of trace elements in the sediments of Chaohu Lake in China
.
Journal of Freshwater Ecology
39
(
1
).
doi:10.1080/02705060.2023.2294129
.
Jenny
J.-P.
,
Anneville
O.
,
Arnaud
F.
,
Baulaz
Y.
,
Bouffard
D.
,
Domaizon
I.
,
Bocaniov
S. A.
,
Chèvre
N.
,
Dittrich
M.
,
Dorioz
J.-M.
,
Dunlop
E. S.
,
Dur
G.
,
Guillard
J.
,
Guinaldo
T.
,
Jacquet
S.
,
Jamoneau
A.
,
Jawed
Z.
,
Jeppesen
E.
,
Krantzberg
G.
,
Lenters
J.
,
Leoni
B.
,
Meybeck
M.
,
Nava
V.
,
Nõges
T.
,
Nõges
P.
,
Patelli
M.
,
Pebbles
V.
,
Perga
M.-E.
,
Rasconi
S.
,
Ruetz
C. R.
,
Rudstam
L.
,
Salmaso
N.
,
Sapna
S.
,
Straile
D.
,
Tammeorg
O.
,
Twiss
M. R.
,
Uzarski
D. G.
,
Ventelä
A.-M.
,
Vincent
W. F.
,
Wilhelm
S. W.
,
Wängberg
S.-Å.
&
Weyhenmeyer
G. A.
2020
Scientists’ warning to humanity: Rapid degradation of the world's large lakes
.
Journal of Great Lakes Research
46
(
4
),
686
702
.
https://doi.org/10.1016/j.jglr.2020.05.006
.
Jiang
Q.
,
Li
J.
,
Sun
Y.
,
Huang
J.
,
Zou
R.
,
Ma
W.
,
Guo
H.
,
Wang
Z.
&
Liu
Y.
2024
Deep-reinforcement-learning-based water diversion strategy
.
Environmental Science and Ecotechnology
17
,
100298
.
https://doi.org/10.1016/j.ese.2023.100298
.
Kisi
O.
&
Ay
M.
2014
Comparison of Mann–Kendall and innovative trend method for water quality parameters of the Kizilirmak River, Turkey
.
Journal of Hydrology
513
,
362
375
.
https://doi.org/10.1016/j.jhydrol.2014.03.005
.
Li
J.
,
Meng
X.
,
Zhang
Y.
,
Li
J.
,
Xia
L.
&
Zheng
H.
2015
Analysis of the temporal and spatial distribution of water quality in China's major river basins, and trends between 2005 and 2010
.
Frontiers of Earth Science
9
(
3
),
463
472
.
doi:10.1007/s11707-015-0498-3
.
Liu
X.
2017
The identification of nutrient limitations on eutrophication in Dianchi Lake, China
.
Water and Environment Journal
31
(
4
),
592
597
.
doi:10.1111/wej.12284
.
Liu
X.
&
Wang
H.
2016
Dianchi Lake, China: Geological formation, causes of eutrophication and recent restoration efforts
.
Aquatic Ecosystem Health & Management
19
(
1
),
40
48
.
doi:10.1080/14634988.2016.1145022
.
Liu
W.
,
Wang
S.
,
Zhang
L.
&
Ni
Z.
2015
Water pollution characteristics of Dianchi Lake and the course of protection and pollution management
.
Environmental Earth Sciences
74
(
5
),
3767
3780
.
doi:10.1007/s12665-015-4152-x
.
Liu
Y.
,
Jiang
Q.
,
Sun
Y.
,
Jian
Y.
&
Zhou
F.
2021
Decline in nitrogen concentrations of eutrophic Lake Dianchi associated with policy interventions during 2002–2018
.
Environmental Pollution
288
,
117826
.
https://doi.org/10.1016/j.envpol.2021.117826
.
Liu
Y.
,
Jiang
Q.
,
Sun
Y.
,
Jian
Y.
&
Zhou
F.
2024
Hysteresis analysis reveals how phytoplankton assemblage shifts with the nutrient dynamics during and between precipitation patterns
.
Water Research
121099
.
https://doi.org/10.1016/j.watres.2023.121099
.
Lopez
B.
,
Baran
N.
&
Bourgine
B.
2015
An innovative procedure to assess multi-scale temporal trends in groundwater quality: Example of the nitrate in the Seine–Normandy basin, France
.
Journal of Hydrology
522
,
1
10
.
https://doi.org/10.1016/j.jhydrol.2014.12.002
.
Mustapha
A.
2013
Detecting surface water quality trends using Mann-Kendall tests and Sen's slope estimates
.
International Journal of Advanced and Innovative Research
2
,
108
114
.
Paerl
H. W.
&
Paul
V. J.
2012
Climate change: Links to global expansion of harmful cyanobacteria
.
Water Research
46
(
5
),
1349
1363
.
https://doi.org/10.1016/j.watres.2011.08.002
.
Pelletier
M. C.
,
Ebersole
J.
,
Mulvaney
K.
,
Rashleigh
B.
,
Gutierrez
M. N.
,
Chintala
M.
,
Kuhn
A.
,
Molina
M.
,
Bagley
M.
&
Lane
C.
2020
Resilience of aquatic systems: Review and management implications
.
Aquatic Sciences
82
(
2
),
44
.
doi:10.1007/s00027-020-00717-z
..
Qin
B.
,
Yang
L.
,
Chen
F.
,
Zhu
G.
,
Zhang
L.
&
Chen
Y.
2006
Mechanism and control of lake eutrophication
.
Chinese Science Bulletin
51
(
19
),
2401
2412
.
doi:10.1007/s11434-006-2096-y
..
Romero
E.
,
Le Gendre
R.
,
Garnier
J.
,
Billen
G.
,
Fisson
C.
,
Silvestre
M.
&
Riou
P.
2016
Long-term water quality in the lower Seine: Lessons learned over 4 decades of monitoring
.
Environmental Science & Policy
58
,
141
154
.
https://doi.org/10.1016/j.envsci.2016.01.016
.
Sharma
S.
&
Saha
A. K.
2017
Statistical analysis of rainfall trends over Damodar River basin, India
.
Arabian Journal of Geosciences
10
, 319.
doi:10.1007/s12517-017-3096-8
.
Tang
X.
,
Xie
G.
,
Deng
J.
,
Shao
K.
,
Hu
Y.
,
He
J.
,
Zhang
J.
&
Gao
G.
2022
Effects of climate change and anthropogenic activities on lake environmental dynamics: A case study in Lake Bosten Catchment, NW China
.
Journal of Environmental Management
319
,
115764
.
https://doi.org/10.1016/j.jenvman.2022.115764
.
Wang
Y.
,
Wang
W.
,
Zhou
Z.
,
Xia
W.
&
Zhang
Y.
2021
Effect of fast restoration of aquatic vegetation on phytoplankton community after removal of purse seine culture in Huayanghe Lakes
.
Science of the Total Environment
768
,
144024
.
https://doi.org/10.1016/j.scitotenv.2020.144024
.
Wen
D.
,
Jin
C.
,
Shixiang
G.
,
Gang
C.
,
Jinming
C.
&
Mi
Z.
2020
Hydrological variability of water level of Dianchi lake and Its application
.
IOP Conference Series: Materials Science and Engineering
780
(
6
),
062048
.
doi:10.1088/1757-899X/780/6/062048
..
Yan
W.
,
Wang
J.
,
Zou
H.
,
Min
M.
&
Duan
X.
2024
Modeling economic-environmental-ecological trade-offs for non-point source control strategies: A case study of Dianchi lake watershed, China
.
Ecological Indicators
158
,
111494
.
https://doi.org/10.1016/j.ecolind.2023.111494
.
Yang
K.
,
Yu
Z.
,
Luo
Y.
,
Yang
Y.
,
Zhao
L.
&
Zhou
X.
2018
Spatial and temporal variations in the relationship between lake water surface temperatures and water quality – A case study of Dianchi Lake
.
Science of the Total Environment
624
,
859
871
.
https://doi.org/10.1016/j.scitotenv.2017.12.119
.
Yenilmez
F.
,
Keskin
F.
&
Aksoy
A.
2011
Water quality trend analysis in Eymir Lake, Ankara
.
Physics and Chemistry of the Earth, Parts A/B/C
36
(
5
),
135
140
.
https://doi.org/10.1016/j.pce.2010.05.005
.
Zhang
T.
,
Zeng
W. H.
,
Wang
S. R.
&
Ni
Z. K.
2014
Temporal and spatial changes of water quality and management strategies of Dianchi Lake in southwest China
.
Hydrology and Earth System Sciences
18
(
4
),
1493
1502
.
doi:10.5194/hess-18-1493-2014
.
This is an Open Access article distributed under the terms of the Creative Commons Attribution Licence (CC BY 4.0), which permits copying, adaptation and redistribution, provided the original work is properly cited (http://creativecommons.org/licenses/by/4.0/).

Supplementary data