Runoff is a pivotal ecohydrological cycle feature. Investigating watershed runoff change and its influencing factors from multi-temporal and spatial angles is crucial for water ecology control. The study analyzes hydrological and meteorological changes using the indicators of hydrological alteration and the range of variation approach (IHA–RVA) and RClimDex to explore their dynamic relationship. Finally, using the soil and water assessment tool to quantify climate and human contributions to runoff temporally and spatially, with validation using Budyko-based elasticity coefficients. Results showed that (1) most of the meteorological indices show an upward trend, a change attributable mainly to strong human activity and global warming. The overall hydrological indicators show a moderate degree of change (50.70%). (2) PRCPTOT (the annual total precipitation) and 30-day minimum demonstrate a negative correlation coefficient of 0.91 in the meteorological–hydrological response. (3) On annual/seasonal scales, human activities such as hydraulic projects and land use/cover changes (LUCC) dominate runoff changes. On a monthly scale, climate change prevails in March and November due to temperature/rainfall fluctuations, while human activities dominate other months. Spatially, climate change and LUCC mainly impact runoff in the southeast. The study offers references to improve water management in the Jialing River Basin, effectively addressing the negative impacts of human activities on runoff.

  • The Budyko hypothesis and the soil and water assessment tool model were used for the attribution analysis of runoff changes.

  • A correlation analysis was conducted between hydrological and meteorological indices.

  • Conducting a spatiotemporal attribution analysis of runoff in the Jialing River Basin.

  • Quantitatively analyzing the impacts of climate change and human activities on runoff.

River runoff is not only an integral component of hydrological cycles but also a key factor influencing changes in ecological and socio-economic development. However, the impacts of global warming on precipitation patterns and evaporation, along with human activities related to water resource development, will significantly alter the watershed's hydrological cycle processes and the spatial–temporal distribution of water resources, potentially even exacerbating the complexity of river hydrological characteristics. These changes will not only increase the frequency of drought and flood events but also impact the sustainable utilization of water resources and the healthy development of the watershed's ecological environment. In view of this, the attribution analysis of hydrological regimes has become a focal point of research in the field of water science (Zhang et al. 2020). Research on the impact of climate change and human activities on runoff variation not only facilitates a deeper understanding of the characteristics and dynamics of water resources but also provides a scientific basis for the rational development and utilization of water resources (Zhang et al. 2008).

The variation in runoff is an important condition for the economic development of the watershed. Both hydrological and meteorological indicators have an impact on runoff variation. According to the intergovernmental panel on climate change fifth assessment report, global surface temperatures have increased by approximately 0.85 °C from 1880 to 2012. Rising temperatures alter the evapotranspiration of basin vegetation, indirectly affecting water balance, increasing basin water-holding capacity, and consequently impacting the basin's hydro-environment (Sampson et al. 2021). Temperature rises may also trigger snow and ice melting, altering the seasonal distribution of runoff (Mimikou et al. 1999). Changes in precipitation directly affect the basin's water resources (Li et al. 2020). The increase in rainfall and its intensity may lead to increased river runoff and flooding probability during the flood season, while droughts may prolong during the non-flood season, further contributing to changes in hydrological conditions (Oeurng et al. 2019; Papalexiou & Montanari 2019; Krisnayanti et al. 2021). Previous scholars have conducted research on hydrology and meteorology. However, most current research tends to focus on independent analyses of watershed hydrological or meteorological conditions, establishing overall connections for response assessments such as runoff and precipitation, water temperature, and air temperature (Yang et al. 2024). There is a lack of exploration into the dynamic response relationships of potential detailed features of hydrological and meteorological changes. To address this gap, this study conducted a correlation analysis between hydrological conditions and meteorological indicators to provide more accurate references for water resource management.

Currently, methods for analyzing runoff changes primarily include neural network models (Govindaraju & Rao 2000), the elasticity coefficient method based on the Budyko hypothesis (Xiong et al. 2020), and hydrological modeling approaches (Xu & Jiang 2022). Among these, neural network models can utilize limited data to forecast runoff and demonstrate a certain level of feasibility. Anderson & Radić (2022) employed long short-term memory network models for predicting runoff in southwestern Canada. However, this model lacks a clear physical mechanism and is unable to distinguish the impacts of factors such as precipitation and evapotranspiration on runoff. The elasticity coefficient method based on the Budyko hypothesis offers clear physical significance, simplicity in application, and high feasibility, making it widely used in major river basins (Liu et al. 2019; Yan et al. 2020). Guo et al. (2022) quantitatively analyzed the contributions of climate change and human activities to runoff variations based on six Budyko hypothesis formulas. The results revealed that human activities were the dominant factor influencing runoff changes in the Min River Basin. However, this method can only separate and quantify the impacts of climate change and human activities on runoff changes at an annual scale, with limited applicability at smaller temporal and spatial scales. Hydrological modeling approaches not only possess robust physical mechanisms but also enable accurate simulation of runoff variability across multiple temporal scales, offering high analytical precision. Among these, the soil and water assessment tool (SWAT) stands out for its multifunctionality in simulating both climate change and anthropogenic impacts. Ji & Duan (2019) utilized the SWAT model to develop six simulation scenarios, providing an in-depth analysis of how land use/cover changes (LUCC) and climatic factors influence the hydrological cycle. Their study highlighted that the direct effects of meteorological conditions and human activities are the dominant forces behind the observed reductions in runoff. Therefore, considering the respective strengths and limitations of the Budyko and SWAT models, this study employs both the SWAT model and the elasticity coefficient method based on the Budyko hypothesis for comparative analysis of runoff change attribution, aiming to enhance the precision of the research. In addition, previous research predominantly evaluated the effects of climate change and human activities on runoff from a qualitative perspective or through single-scale quantitative analyses, often neglecting the spatial complexity exhibited by runoff (Xu et al. 2023). This also highlights the importance of quantifying the impacts of various driving factors on runoff across multiple temporal scales (monthly, seasonal, and annual) and spatial scales. Therefore, this study conducts attribution analysis at the subbasin scale based on hydrological simulations under different scenarios. It quantitatively identifies the impacts of climate change and human activities on runoff variations in the Jialing River (JLR) Basin from multiple temporal and spatial perspectives, providing scientific evidence for ecological protection and management planning in the JLR Basin.

As one of the major tributaries of the Yangtze River upstream, the JLR has abundant runoff and precipitation, making it among the tributaries with significant potential for the Yangtze River water resource development (Zhang et al. 2021). Therefore, studying hydrological changes in the JLR Basin and qualitatively and quantitatively assessing various influencing factors on runoff is of paramount importance for the ecological restoration and sustainable development of the Yangtze River Basin. The runoff variations in the JLR Basin are primarily influenced by climate change and human activities. Climate change manifests mainly in alterations in precipitation and temperature. Human intervention has resulted in a series of societal issues, such as extensive degradation of high-quality arable land, accelerated urbanization, severe damage to forests and grasslands, and further intensification of land development and utilization (Qi et al. 2013). These changes affect the water resource balance in the JLR Basin.

What is the correlation between hydrological and meteorological indicators in the JLR Basin? What are the primary driving forces behind the temporal and spatial variations in runoff? To address the aforementioned questions, this study comprehensively employs various methods to delve into the changes and dynamic responses of hydrological and meteorological indicators from different perspectives and makes a quantitative analysis of the hydrological regime, which mainly includes the following three steps: (1) temporal changes and identification of shift points in annual runoff from 1965 to 2019; (2) analysis of the characteristics and correlation of changes in hydrological and extreme meteorological indicators; (3) the Budyko and SWAT models are utilized to analyze the temporal and spatial variations of runoff as well as the primary driving factors causing these changes.

Study area

The JLR is one of the major tributaries of the Yangtze River in China, spanning latitudes 29°18′ to 34°30′ north and longitudes 102°33′ to 109°00′ east (Figure 1). Originating from the Daimiao Mountain in Feng County, Shaanxi Province, at the northern foot of the Qinling Mountains, it flows through Shaanxi, Gansu, Sichuan, and Chongqing before converging with the Yangtze River at Chaotianmen in Chongqing. The JLR spans a total length of 1,345 km, with a mainstream basin area of 39,200 km² and a total basin area of 160,000 km². It stands as the largest basin among the Yangtze River tributaries, while its length is second only to the Yalong River, and its discharge is second only to the Min River. The basin falls within a subtropical climate characterized by moist monsoons. The average annual maximum temperature is 19.4 °C, the average annual minimum temperature is 4.3 °C, the average annual wind speed is 1.1 m/s, and the fluctuating average annual precipitation ranges from 433.6 to 1,320.2 mm. The average annual sunshine duration is 1,450 h (mostly 1991–2018). These climatic features significantly influence the distribution of water resources and the ecological environment in the JLR Basin.
Figure 1

Overview of the JLR Basin.

Figure 1

Overview of the JLR Basin.

Close modal

Data preparation

The hydrological data collected from the Beibei Hydrological Station for the period 1965–2019 were sourced from the Yangtze River Commission. The Beibei hydrological station serves as a control station for the confluence of the JLR, Fu River, and Qu River, and it is also a downstream control station for the mainstream of the JLR. It can effectively reflect the hydrological situation of the JLR and is crucial for flood control in both the main urban area of Chongqing and the Three Gorges Reservoir area (Wang et al. 2019; Wu & Pu 2019).

Meteorological data were collected from ten meteorological stations in the JLR Basin for the period 1965–2019. These data include daily averages of temperature, daily high and low temperatures, rainfall, sunshine hours, and other variables. The data were sourced from the China Meteorological Information Center (http://data.cma.cn/).

Spatial data encompass soil, digital elevation model (DEM), and LUCC data. The DEM data were obtained from the China Academy of Sciences Resource and Environmental Sciences Data Center (http://www.resdc.cn/) at a resolution of 500 m × 500 m. Land use data and the normalized difference vegetation indicator (NDVI) data covering three periods (1980, 2000, and 2020) were also provided by the China Academy of Sciences Resource and Environmental Sciences Data Center, with a spatial resolution of 1 km × 1 km. Land use types were classified into six primary categories: arable land, forest land, grassland, water bodies, urban and rural construction land, and unused land. Soil data were derived from the Harmonized World Soil Database (http://www.fao.org/) at a scale of 1:1,000,000. Meteorological and hydrological data cover the period from 1965 to 2019.

Detection of change points

This study employs the Mann–Kendall (MK) non-parametric test (Mann 1945; Kendall 1990) and the cumulative anomaly method (Chen et al. 2018) to analyze the change characteristics and identify change points in the annual mean runoff at the Beibei station. The MK is a non-parametric test method applied in the fields of hydrology and meteorology. This method has the advantages that the sample sequence does not have to follow a specific distribution, a small amount of missing data and abnormal data have little impact on the results, and can objectively show the overall change trend of the sample sequence. Therefore, it is widely used in trend analysis and change point detection in time series. Details of the calculation process can be found in the relevant literature (Chen et al. 2018; Zhuo et al. 2020).

Climate indicator analysis

The study selected 27 extreme climate indices jointly developed by the Expert Team on Climate Change Detection and Indices of the World Meteorological Organization. Based on daily data from 10 meteorological stations in the JLR Basin, calculations were performed using the RClimDex indicators developed by the Canadian Meteorological Center. Detailed information about these meteorological indices can be found on the website (http://etccdi.pacificclimate.org/list_27_indices.shtml). In this study, the icing days (ID) in the study are zero; therefore, ID is not considered in the meteorological analysis.

Hydrological situation analysis

In 1997, the indicators of the hydrologic alteration (IHA) method proposed by Richter et al. (1997) was widely utilized to assess the degree of hydrological changes in ecosystems under the influence of human activities. It encompasses five aspects: magnitude, timing, frequency, duration, and rate of change, totaling 33 hydrological indicators for the quantitative evaluation of hydrological alterations. In the Range of Variability Approach (RVA) analysis, calculations can be performed using either parametric (mean) or non-parametric (median) methods. Since parametric statistics require datasets to follow a normal distribution, whereas many hydrological datasets are skewed in reality, non-parametric statistics represent a better choice. Details of the IHA hydrological indicators are presented in Table 1.

Table 1

Hydrological regime indicators and ecological impacts

Parameter groupIHA ParametersEcological influences
Magnitude of monthly water level conditions Monthly average streamflow (precipitation) Provides habitat for aquatic organisms; maintains the soil moisture required for plant growth 
Magnitude and duration of annual extreme water level conditions Annual average 1, 3, 7, 30, 90 days minimum and maximum streamflow (precipitation), base indicator Maintains competitive balance between organisms; provides tectonic channels, geomorphology and physical habitat conditions 
Timing of annual extreme water level conditions Date of occurrence of the maximum and minimum 1 day of the year (Roman day) Access to specific habitats for breeding or to avoid predation; spawning opportunities for migrating fish 
Frequency and duration of high and low pulses Average number of high and low pulses per year and the duration of the pulses Exchange of nutrients and organic matter between the river and the floodplain; good for waterfowl feeding, roosting, nesting, and breeding 
Rate and frequency of water level condition changes Annual average rates of increase and decrease and the number of reversals Drought stress on plant production; drought stress on organic matter at the margins of low-mobility riverbeds 
Parameter groupIHA ParametersEcological influences
Magnitude of monthly water level conditions Monthly average streamflow (precipitation) Provides habitat for aquatic organisms; maintains the soil moisture required for plant growth 
Magnitude and duration of annual extreme water level conditions Annual average 1, 3, 7, 30, 90 days minimum and maximum streamflow (precipitation), base indicator Maintains competitive balance between organisms; provides tectonic channels, geomorphology and physical habitat conditions 
Timing of annual extreme water level conditions Date of occurrence of the maximum and minimum 1 day of the year (Roman day) Access to specific habitats for breeding or to avoid predation; spawning opportunities for migrating fish 
Frequency and duration of high and low pulses Average number of high and low pulses per year and the duration of the pulses Exchange of nutrients and organic matter between the river and the floodplain; good for waterfowl feeding, roosting, nesting, and breeding 
Rate and frequency of water level condition changes Annual average rates of increase and decrease and the number of reversals Drought stress on plant production; drought stress on organic matter at the margins of low-mobility riverbeds 

On the basis of the 33 hydrological indicators in the IHA, the RVA is utilized to quantitatively assess hydrological changes before and after hydrological regime shifts at the Beibei station. The formula is as follows:
(1)
Di represents the hydrological alteration magnitude of the ith hydrological indicator in the IHA; N2,i denotes the number of years post-hydrological shift where the ith hydrological indicator falls within the RVA target threshold range. N represents the length of the hydrological time series expected to fall within the RVA range post-shift, where N = 50% × NA (NA is the total number of years in the post-shift runoff series).

To quantitatively analyze the degree of change in IHA indicators, criteria for hydrological alteration are established. Di values in the range of 0–33% are classified as low alteration; 33–67% as moderate alteration; and 67–100% as high alteration (Wang et al. 2023b).

Let n be the number of indicators, the overall degree of hydrological alteration Dall is defined by the following equation:
(2)

Correlation analysis

Grey relational analysis

The grey theory proposed by Deng (1989) has been widely applied in various fields (Ip et al. 2009; Suh 2004) and is a simple yet accurate method for multi-attribute decision-making problems. The grey correlation analysis method, which involves non-dimensionalizing two sets of data through grey processing, as proposed by Lu & Wevers (2007), is a method for measuring the correlation between multiple factors by assessing the similarity or dissimilarity of the development trends between two sequences. Its objective is to identify the correlation between factors and determine the significant influence of certain factors. Therefore, by employing grey correlation analysis, it is possible to identify meteorological–hydrological factors with a high degree of correlation. For specific steps, refer to the literature by Liu et al. (2005).

Cross-wavelet analysis

Cross-wavelet transformation is constructed based on wavelet transformation and cross-spectrum, enabling the analysis of the distribution patterns of two sets of time series in different frequency domains. Additionally, it reveals the consistency and correlation of different time series frequencies at different time scales and analyzes the phase relationship of different signals in the time–frequency domain through wavelet phase angle analysis (Huang et al. 2017). In this study, cross-wavelet analysis is employed to analyze several sets of meteorological and hydrological indicators with high correlation at the annual scale.
(3)
(4)

In the equation, represents the complex conjugate of , denotes the cross product of the fluctuation amplitudes of two time series at a certain frequency, and and represent the amplitudes of the oscillating waves of the two time series.

Attribution analysis of runoff changes

Elasticity coefficient method based on the Budyko hypothesis

The Choudhury–Yang formula is used to calculate the influencing factors of annual-scale runoff changes. The principle of calculation involves computing the elasticity coefficients of the respective influencing factors, which are then used to calculate the runoff changes caused by precipitation, potential evapotranspiration, and the underlying surface (Roderick & Farquhar 2011).
(5)
where represents the multi-year average actual evapotranspiration (mm); denotes the multi-year average annual precipitation (mm); n is the underlying surface parameter, reflecting the overall conditions of vegetation, soil, terrain, and land use in the watershed; ET0 is the multi-year average annual potential evapotranspiration (mm), as detailed in Guo et al. (2022).
Let the watershed aridity indicator φ = ET0/P. According to the definition of elasticity coefficient, the elasticity coefficients corresponding to land surface, precipitation, and potential evapotranspiration can be calculated using the total differential form of the water-heat coupling balance equation:
(6)
(7)
(8)
Using ωp, ωET0, and ωn, the runoff change caused by the corresponding factors can be calculated:
(9)

In the equation, ΔP, ΔET0, and Δn represent the changes in precipitation, potential evapotranspiration, and land surface characteristics before and after the shift, respectively. In this context, the sum of precipitation and potential evapotranspiration changes represents the variation in runoff induced by climate change, while the runoff change induced by land surface alterations signifies the impact of human activities on runoff.

Consequently, the contributions of precipitation, potential evapotranspiration, and land surface alterations to runoff changes are determined as follows:
(10)
(11)
(12)
(13)

The SWAT model

The SWAT (Arnold et al. 1998; Srinivasan et al. 1998) mainly simulates the runoff generation process based on water balance, using land use data, soil data, DEM, and meteorological data as input data. It divides the watershed into several subbasins, further subdividing them into hydrological response units (HRUs) with similar land use, soil, and slope characteristics. With a daily/minute time step, it simulates hydrological processes at annual, monthly, and daily scales at the HRU level. This allows for the simulation of changes in runoff within the watershed. The formula for water balance calculation is:
(14)
where SSWt represents the final soil moisture content at the end of day t; SSW0 represents the initial soil moisture content; i represents the time series; Rd, Qs, Ea, and Qg represent the precipitation, surface runoff, evapotranspiration, and groundwater return flow for time period i, respectively; and Ws represents the surface water infiltrated during time period i.
Using the SWAT allows for quantitatively determining the contributions of both climate change and human activities to runoff variation. This study divides the study period into two periods: base period and variation period. The specific steps are as follows: The hydro-meteorological data, land use types, and soil types of the base period, where human activities have a relatively minor impact on the runoff sequence, are input into the SWAT model. The SWAT-CUP is used to calibrate hydrological parameters, ensuring that the simulated runoff meets the relevant evaluation criteria, thereby establishing a hydrological model for the natural period. Using this model, the hydro-meteorological data from the post-change period is input into the hydrological model, which continues to drive the hydrological model to simulate the natural runoff after the change. Analyzing the impact of climate change on runoff using natural runoff before the change and natural runoff after the change; analyzing the impact of human activities on runoff using measured runoff after the change and natural runoff after the change. The specific formula is as follows:
(15)
(16)
(17)
(18)
where RPost represents the observed runoff after the change, RPre denotes the observed runoff before the change, ΔRC signifies the impact of climate change on runoff, ΔRH indicates the impact of human activities on runoff, Rsim,Post stands for the simulated runoff after the change, Rsim,Pre represents the simulated runoff before the change, ΔRLUCC denotes the impact of land use change on runoff, ΔR2020 signifies the annual average runoff under land use conditions in 2020, ΔR1980 indicates the annual average runoff under land use conditions in 1980, and ΔRother represents the impact of other human activities on runoff. The contributions of climate change and human activities are as follows:
(19)
(20)

where ηC represents the contribution rate of climate change to runoff, ηH denotes the contribution rate of human activities to runoff, ηLUCC signifies the contribution rate of land use changes to runoff, and ηother indicates the contribution rate of other human activities to runoff.

Temporal characteristics of runoff and meteorological data

Based on the MK test combined with the cumulative anomaly method validation, the results of the change point experiment for the JLR runoff sequence are shown in Figure 2. From Figure 2, it can be observed that the UF curve and UB curve intersect multiple times, and 1994 is the change point detected by both methods. Therefore, it is determined that the annual average runoff at the Beibei hydrological station experienced a change point in 1994. In the same year, construction of the Three Gorges Dam officially commenced, and it was fully completed by 2009. The dam controls a watershed area of 1 million km², accounting for 56% of the Yangtze River Basin. Previous studies have demonstrated that the construction of the Three Gorges Dam results in a reduction of the JLR's runoff (Feng et al. 2020). The dam, as a form of human activity, has profoundly altered the hydrological characteristics and water environment of the upstream regions through means such as water storage regulation. This indicates that the construction of large-scale hydraulic projects can significantly impact the ecosystem of the basin, and it also validates the accuracy of the selected change point year. Therefore, in this study, the long-term runoff series of the JLR Basin is divided into two periods: a baseline period (1965–1993) and a variation period (1994–2019). Changes in precipitation, runoff depth, and temperature in the JLR Basin base period and variation period are shown in Figure 3. During the reference period and the variation period, both precipitation and runoff depth exhibited an increasing trend, while temperature remained relatively stable. The overall mean of precipitation, runoff depth, and temperature during the variation period was lower than that during the reference period. The average runoff depth before and after the mutation was 423.93 and 370.79 mm, respectively. Considering the selection of change points and the trends in precipitation, it is evident that, under the coupled effects of human activities and climate change, the runoff depth decreased by 53.14 mm and experienced significant alterations.
Figure 2

Detection of change points in the runoff of the JLR Basin.

Figure 2

Detection of change points in the runoff of the JLR Basin.

Close modal
Figure 3

Interannual variations of precipitation, runoff depth, and air temperature from 1965 to 2019.

Figure 3

Interannual variations of precipitation, runoff depth, and air temperature from 1965 to 2019.

Close modal
Precipitation is the main source of runoff, and the JLR, influenced by local climate, exhibits significant differences in annual precipitation distribution. As shown in Figure 4, precipitation is mainly concentrated in spring and summer, while it is relatively scarce in autumn and winter. Comparing the base period and variation period, the precipitation changes in the JLR Basin are mainly concentrated from May to September. Months with significant changes in runoff depth mainly occur from July to September. While the peak of runoff depth does not show significant temporal changes compared to the base period, overall, there is a decrease in runoff depth at the peak. It can also be observed that during the flood season in August and September when water is abundant, the asymmetry of the distribution curve of runoff depth gradually increases before and after the change point. Based on the observed patterns of precipitation and runoff depth in the JLR Basin, it can be inferred that rainfall is the primary factor influencing runoff variations.
Figure 4

Seasonal variations in runoff depth and precipitation in the JLR Basin occurred during the base period and variation period.

Figure 4

Seasonal variations in runoff depth and precipitation in the JLR Basin occurred during the base period and variation period.

Close modal

Temporal characteristics of meteorological and hydrological indicators during extreme events

Variations in meteorological indices characteristics

This study utilizes daily data from 10 meteorological stations along the JLR and employs the RClimDex method to compute various extreme climate indices.

Upon analyzing the interannual trends of extreme temperature indices (Figure 5), it is observed that, over time, indicators representing cold events in the watershed, such as frost days, cold nights (TN10P), and cold spell duration indicator, exhibit varying degrees of decline. Conversely, indices representing warm events in the watershed, such as the number of summer days (SU), warm nights (TX90P), and warm spell duration indicator, demonstrate varying degrees of increase. Furthermore, the magnitude of the increase in indices representing warm events is greater than the magnitude of the decrease in indices representing cold events. The increased frequency of high-temperature events and the decreased frequency of low-temperature events in the JLR Basin indicate a warming trend in the climate. Additionally, the rate of warming during the night is higher than during the day. The smaller slope value of the cold nights indicator (TN10P) compared to the cold days indicator (TX10P) indicates warming in the Basin, with daytime contributing more than nighttime. Conversely, the higher slope value of the warm nights indicator (TN90P) compared to the warm days indicator (TX90P) suggests a higher rate of warming during the night than during the day. The growing season length shows a pronounced upward trend, indicating an increase in the duration of heat accumulation required for crop growth in the JLR Basin from 1965 to 2019. This has significant implications for the growth of crops and vegetation in the region.
Figure 5

Interannual variation trends of extreme temperature indicators.

Figure 5

Interannual variation trends of extreme temperature indicators.

Close modal
After analyzing the interannual variation trends of extreme precipitation indices (Figure 6), it is evident that, except for R10 mm (the annual count of days when precipitation ≥ 10 mm), other precipitation indices show an increasing trend. This indicates that the frequency of excessive precipitation and prolonged dry events has increased over the past 55 years. Among them, PRCPTOT (the annual total precipitation) and R95P (the annual total precipitation when the daily precipitation > 95p) show significant increases, with rates of 0.48 and 0.63 mm/year, respectively, while SDII (the simple precipitation intensity index) shows the smallest variation amplitude, exhibiting a relatively stable trend. The SDII, which reflects the average conditions of the watershed, remains stable with a variation amplitude of only 0.002 mm/year, indicating that the annual average precipitation intensity in the watershed has not changed significantly. The trends in R95P and R99P (the annual total precipitation when the daily precipitation > 99p) are consistent with those of SDII, indicating a gradual increase in heavy precipitation events in the watershed, leading to a tendency toward precipitation extremes. The CDD (maximum length of dry spell) indicator represents the number of consecutive days without significant precipitation in the region, which has exhibited complex variations over the past 55 years. There are significant interannual differences, with extreme highs and lows adjacent in distribution. This has significant implications for water use in agriculture and industry in the watershed, particularly negatively affecting crop growth.
Figure 6

Interannual variations in extreme precipitation indices.

Figure 6

Interannual variations in extreme precipitation indices.

Close modal

Characteristics of hydrological indicator variation

This study utilizes daily runoff data from Beibei station for the period of 1965–2019 to analyze the runoff variations across the entire JLR Basin during the base period. The annual count of the 32 parameters of daily runoff falling within the target range during the change period is calculated to assess the degree of variation in runoff change parameters. Based on this, the IHA–RVA method is applied to calculate the degree of change in runoff during the base period and variation period at the Beibei hydrological station. The number of low pulses at Beibei station increased from 3 to 13, representing a change of 89.86%, which is a high probability. The duration of low pulses decreased from 11 to 2 days, representing a change of 55.38%. Meanwhile, the number of high pulses increased from 9 to 10, with a change of 22.78%, indicating a low change. The duration of high pulses decreased from 6 to 4 days, representing a change of 40.51%. The minimum flow for 1, 3, and 7 days all decreased to varying degrees, while the minimum flow for 30 and 90 days increased. The annual maximum flow after the change was lower than before the change, consistent with the results of the change-year test, indicating a decreasing trend in flow depth in the JLR. The reversal frequency significantly increased compared to before the change, with a change degree of 100%. Due to the limited resilience of ecosystems to external influences, the frequency of changes in JLR runoff will have a significant impact on the ecosystem of this basin.

By analyzing the degree of hydrological change in the JLR Basin during the base period and the variation period, among the 32 hydrological indicators, 8 indicators exhibited a high degree of change, accounting for 25% of all hydrological indicators. Additionally, 12 indicators showed a moderate degree of change, representing 37.5% of the overall hydrological indicators, while 12 indicators demonstrated a low degree of change, also constituting 37.5% of the total hydrological indicators. The specific results are illustrated in Figure 7. Among these, the number of reversals exhibited the highest degree of change, reaching 100%, and it showed a significant correlation with the fall rate, which also had a high change degree of 81.41%, within the same category, likely influenced by both human activities and climate change.
Figure 7

Degree of hydrological change at Beibei station (Group 1 is count and duration time of extreme pulse; Group 2 is time of extreme flow occurring; Group 3 is extreme flow; Group 4 is monthly mid-value flow; Group 5 is reversals and rate of change).

Figure 7

Degree of hydrological change at Beibei station (Group 1 is count and duration time of extreme pulse; Group 2 is time of extreme flow occurring; Group 3 is extreme flow; Group 4 is monthly mid-value flow; Group 5 is reversals and rate of change).

Close modal

Based on the overall hydrological indicator changes and the changes in various hydrological indices (Table 2), it can be concluded that the hydrological condition of the JLR Basin is most affected by the rate and frequency of water flow changes, with a degree of change of 77.34%. Following this are the frequency and duration of high and low pulses, as well as the occurrence time of annual extreme flows, all reaching a moderate level of change with degrees of change of 57.67 and 52.58%, respectively. The degree of change for the other two groups of indices is lower than that of the above three groups, and except for the fifth group of indices, which represents a high level of change, the rest belong to a moderate level of change. The overall degree of change in hydrological indices is 50.70%, indicating a moderate level of change.

Table 2

Overall hydrological alteration of discharge sequence unit (%)

HydrologicalHydrological alteration degree of each group
Overall hydrological alteration degree
Group 1Group 2Group 3Group 4Group 5
Beibei 57.67(M) 52.58(M) 36.13(M) 41.77(M) 77.34(H) 50.70(M) 
HydrologicalHydrological alteration degree of each group
Overall hydrological alteration degree
Group 1Group 2Group 3Group 4Group 5
Beibei 57.67(M) 52.58(M) 36.13(M) 41.77(M) 77.34(H) 50.70(M) 

Note: H, high change; M, moderate change; L, low change.

Meteorological–hydrological correlation analysis

Significant impact characteristics of meteorological–hydrological events

There is a close relationship between meteorological factors and hydrological factors. For example, the variation in runoff is strongly influenced by the coupled effects of precipitation and temperature. Therefore, this study conducts a grey correlation analysis on the hydrological indicator data obtained from the IHA and the meteorological indicator data obtained from the RClimDex method to understand the correlation between various meteorological and hydrological indicators. By comparing the results, the degree of influence between meteorological and hydrological indices can be determined. The correlation coefficient test results, as shown in Figure 8, indicate that the datasets PRCPTOT and 30-day minimum, PRCPTOT and high pulse count, and SDII and 30-day minimum have a relatively high correlation, with correlation coefficients of 0.91, 0.90, and 0.90, respectively. Additionally, the correlation coefficients of the datasets PRCPTOT and June, SDII and high pulse count, and SDII and June are slightly lower than those of the aforementioned three datasets, with correlation coefficients of 0.87, 0.88, and 0.96, respectively.
Figure 8

Correlation between meteorological and hydrological changes during the study period.

Figure 8

Correlation between meteorological and hydrological changes during the study period.

Close modal

The dynamic response of meteorological–hydrological interactions

To elucidate the dynamic response relationship between meteorological and hydrological factors and considering the heterogeneity in the propagation between meteorological and hydrological indicators, we conducted further analysis using cross-wavelet analysis on six pairs of meteorological–hydrological indicators showing relatively high correlation. These pairs include PRCPTOT, SDII, and 30-day minimum, high pulse count, and June.

From Figure 9, the study observes that in the cross-wavelet transform (XWT), at a confidence level of 95%, four pairs of indicators, PRCPTOT and high pulse count, PRCPTOT and 30-day minimum, SDII and high pulse count, and SDII and 30-day minimum, exhibit significant resonance periods during 1965–2019. These pairs share similar distribution patterns, mainly concentrated in the 4–6 month signals from 2000 to 2006. Additionally, within the 4–6-month period, the rough contour lines show higher average energy values but shorter durations, reflecting intermittent cyclical oscillation characteristics. The phase angles in the high wavelet energy areas also indicate that in the dynamic response between meteorology and hydrology, the 30-day minimum leads PRCPTOT, the 30-day minimum leads SDII, and June leads SDII.
Figure 9

At the annual scale, the WTC spectrum and cross-wavelet transform (XWT) of PRCPTOT, SDII and 30-day minimum, high pulse count, and June are depicted.

Figure 9

At the annual scale, the WTC spectrum and cross-wavelet transform (XWT) of PRCPTOT, SDII and 30-day minimum, high pulse count, and June are depicted.

Close modal

Further analysis using the wavelet transform coherence (WTC) spectrum reveals that at a 95% confidence level, PRCPTOT and high pulse count, SDII and High pulse count, PRCPTOT and June, and SDII and June exhibit significant positive correlations primarily during the 4–6-month period from 2003 to 2006. Conversely, PRCPTOT and 30-day minimum and SDII and 30-day minimum show negative correlations at a 95% confidence level. Additionally, the distribution of short-duration and frequent occurrences in the low-frequency energy areas of the six indicators reflects their non-stationary relationship. Relatively, there is a statistically significant relationship within the wavelet cone of influence for all six indicators. The WTC spectrum indicates that throughout the entire period, meteorology and hydrology are indispensable driving forces for each other, demonstrating the existence of a stable correlation between the two.

Hydrological attribution analysis

Budyko hydrological regime change attribution

According to Table 3, during the change period in the JLR Basin, the p-value decreased by 2.14% compared to pre-change; ET0 increased by 3.32% compared to the baseline period; and the aridity indicator (ET0/P) also increased compared to the baseline period. The elasticity coefficients εp, εET0, and εn for the influencing factors on runoff after the change are 1.80, −0.80, and −0.77, respectively. This indicates that for a 1% increase in P, there is a 1.80% increase in runoff depth, a 0.80% decrease for ET0, and a 0.77% decrease for n. It is evident that the variation in JLR runoff is positively correlated with precipitation and negatively correlated with potential evapotranspiration. The change in runoff in the JLR Basin is most sensitive to precipitation and least sensitive to underlying surface conditions. Table 3 also presents the contribution rates of each driving factor to runoff. Compared to the baseline period, the multi-year average runoff during the variation period decreased by 53.14 mm. Precipitation decreased by 21.02 mm, resulting in a runoff decrease of 15.27 mm and a contribution rate of 27.49%. Potential evaporation increased by 30.75 mm, causing a runoff decrease of 9.67 mm with a contribution rate of 17.40%. The underlying surface parameters decreased by 0.13, leading to a runoff decrease of 30.62 mm and a contribution rate of 55.11%. Therefore, climate change-induced runoff changes amount to 24.94 mm, contributing 44.89%, and the underlying surface is the primary factor influencing runoff changes, contributing 55.114%. The sensitivity of runoff changes to underlying surfaces is minimal, while the contribution of underlying surfaces to runoff is predominant. Therefore, a detailed understanding of the impact of underlying surfaces on runoff is necessary.

Table 3

Hydro-meteorological parameters in the JLR Basin

Parameters1965–19931994–2019Contribution rate (%)
R (mm) 423.93 370.79 — 
P (mm) 984.10 963.08 27.49 
ET0 (mm) 926.07 956.82 17.40 
n 1.30 1.44 55.11 
εp 1.64 1.80 — 
εET0 −0.75 −0.80 — 
εn −0.64 −0.77 — 
Parameters1965–19931994–2019Contribution rate (%)
R (mm) 423.93 370.79 — 
P (mm) 984.10 963.08 27.49 
ET0 (mm) 926.07 956.82 17.40 
n 1.30 1.44 55.11 
εp 1.64 1.80 — 
εET0 −0.75 −0.80 — 
εn −0.64 −0.77 — 

Calibration and validation of the SWAT model

In this study, the entire watershed was divided into 60 subbasins based on DEM data, and subbasin 59 outlet simulations were performed for measured flows at the Beibei station. Further subdivision into 470 HRUs was performed based on slope, LUCC types, and soil types within the 60 subbasins.

To calibrate and validate the model, the SWAT model was utilized to simulate the monthly runoff from 1965 to 2019 at the watershed scale. The model was initialized in 1964, with the calibration period spanning from 1965 to 1993 and the validation period from 1994 to 2019. SWAT-CUP software was employed for model calibration and validation, and parameter adjustments were made. To expedite computation and streamline parameter adjustments, sensitivity analysis of model parameters was conducted to identify parameters with the greatest impact on the study results (Bhattacharya et al. 2020). The sensitivity of parameters was determined using two indicators: t-status and p-value. A higher absolute value of t indicates stronger parameter sensitivity, while a smaller absolute value of P indicates greater significance. Ultimately, nine parameters with high sensitivity were identified, as shown in Table 4. The accuracy of the computed results was assessed using the coefficient of determination (R2) and the Nash–Sutcliffe efficiency (NSE) coefficient to comprehensively evaluate the simulation performance of the SWAT model. Generally, a model simulation is considered acceptable when both R2 and NSE exceed 0.5 (Moriasi et al. 2007). As shown in Figure 10, both R2 and NSE during the calibration and validation periods exceeded 0.7, indicating that the SWAT model effectively simulated the runoff processes in the JLR Basin. From Figure 10, it is evident that the monthly simulated runoff process at the Beibei hydrological station exhibits a favourable fit with the observed runoff process, closely mirroring the monthly precipitation variability.
Table 4

Sensitivity analysis of the SWAT model parameters and calibration results

ParametersSensitivity rankingFinal parameter values
R_CN2.mgt 0.1 
V_ALPHA_BF.gw 0.65 
R_SOL_K.sol 2.7 
V_CANMX.hru 65 
V_CH_N2.rte 0.135 
V_ESCO.hru 0.75 
R_SOL_AWC.sol 0.3 
V_SFTMP.bsn − 1 
V_REVAPMN.gw 209.450 
ParametersSensitivity rankingFinal parameter values
R_CN2.mgt 0.1 
V_ALPHA_BF.gw 0.65 
R_SOL_K.sol 2.7 
V_CANMX.hru 65 
V_CH_N2.rte 0.135 
V_ESCO.hru 0.75 
R_SOL_AWC.sol 0.3 
V_SFTMP.bsn − 1 
V_REVAPMN.gw 209.450 
Figure 10

Comparison of simulated and observed runoff at Beibei hydrological station before and after the change in 1965–2019.

Figure 10

Comparison of simulated and observed runoff at Beibei hydrological station before and after the change in 1965–2019.

Close modal

SWAT hydrological regime change attribution

Maintaining meteorological data unchanged, SWAT was employed to simulate runoff using land use data from 1980 and 2020, respectively. This approach facilitated the isolation of changes in runoff attributed to land use and other human activities, quantitatively characterizing the impacts of climate, land use, and other human activities on runoff at different temporal scales (Figure 11). We found that changes in runoff driven by climate, land use, and other human activities differed significantly from pre-disturbance levels at the multi-year scale, with variations of −140.41, 16.04, and −198.53 m3/s, respectively. At the seasonal scale, compared to base period conditions, climate change led to a reduction in winter runoff by 137.72 m3/s, while land use change and other human activities increased runoff by 33.92 and 356.73 m3/s, respectively, indicating that human activities primarily enhanced changes in winter runoff while inhibiting runoff changes in spring, summer, and autumn. At the monthly scale, climate change had a supplementary effect on runoff in March, dominating with a contribution rate of 77.03%. Human activities exhibited adverse effects on runoff during the flood season, especially from July to October, with contributions exceeding 60%. In terms of magnitude changes in November runoff, climate change inhibited the formation of runoff (51.60 m3/s), and human activities also showed an inhibitory effect on runoff (23.71 m3/s). In addition, LUCC contributed to the formation of 21.86 m3/s runoff, and other human activities inhibited the formation of 45.57 m3/s runoff. The contribution of climate change to runoff generally showed a greater character than that of human activities. Hence, the contribution of climate change to runoff exceeded that of human activities. Attribution analysis results indicated that human activities were the primary driver of reduced runoff in the JLR Basin, accounting for 60.45%, while climate contributed 39.55%.
Figure 11

Quantifying the impacts and contributions of climate, land use, and other human activities to runoff changes at monthly, annual, and seasonal scales.

Figure 11

Quantifying the impacts and contributions of climate, land use, and other human activities to runoff changes at monthly, annual, and seasonal scales.

Close modal
Utilizing the established SWAT model enables the simulation of runoff at various subbasin outlets. Figure 12(a) illustrates the relative changes in runoff at subbasin outlets after the shift compared to before the shift due to climate variations. Among the 60 outlets, only 5 detected positive changes in runoff, with the most significant changes observed near outlet 59 (Beibei hydrological station). Figure 12(b) demonstrates the changes in runoff at subbasin outlets after the shift relative to before the shift due to land use alterations. Two-thirds of the subbasins detect positive changes in runoff, albeit significantly lower than the negative changes attributed to climate. Consequently, the variation period annual average runoff is lower than the base period average. Analyzing the contribution rates reveals that the decline in runoff is primarily influenced by human activities. Thus, in conjunction with the analysis presented in Figure 11, it is evident that changes in runoff are significantly influenced by other human activities, aligning with the outcomes depicted in Figure 11(c) and 11(d). Overall, climate variations primarily suppress runoff changes, while land use predominantly induces runoff alterations. Additionally, climate variations and land use exhibit similar spatial distribution characteristics in their impacts on runoff, primarily concentrated in the southeast region.
Figure 12

The spatiotemporal impacts of climate change and land use on runoff (where ‘a’ denotes runoff changes induced by climate variations and ‘b’ represents runoff changes induced by land use alterations).

Figure 12

The spatiotemporal impacts of climate change and land use on runoff (where ‘a’ denotes runoff changes induced by climate variations and ‘b’ represents runoff changes induced by land use alterations).

Close modal

The impact of climate change and anthropogenic activities on runoff

This study quantitatively analyzes the impact of driving factors behind runoff variations on a spatiotemporal scale based on hydrological and meteorological indicators using the SWAT model. The results are compared with the annual-scale results obtained from the Budyko method. The results reveal that the two are broadly similar, with human activities consistently identified as the primary drivers of runoff variations, thereby affirming the reliability of the SWAT model. Building upon this foundation, quantitative attribution of runoff variations at seasonal and monthly scales was achieved using the model. Wang et al. (2023a), based on the Budyko hypothesis, concluded that human activities are the primary influencing factor, contributing 66.31% to the runoff in the mainstream of the JLR. Li et al. (2022) found that climate factors and human activities contributed 42.70 and 57.30%, respectively, to the variation in river discharge, while their contributions to vegetation change were 28.89 and 71.11%, respectively. Our findings not only corroborate their results at the annual scale but also extend the analysis to more detailed temporal scales. Furthermore, this study conducted a detailed quantitative analysis of spatial variations in runoff.

At the temporal scale, precipitation predominates among climatic factors, serving as a key driver of runoff variations (Shenglian et al. 2015). However, with increasing human activities, the relative influence of climatic factors on runoff diminishes, gradually making human activities the dominant factor in runoff variations (Hayashi et al. 2015). In spatial terms, the facilitating effect of land use on runoff is far less significant compared to the inhibitory effect of climate change on runoff. Additionally, there are other factors influencing the runoff variation in the JLR Basin, such as the construction and operation of reservoirs, as well as changes in groundwater levels within the basin. The construction of reservoirs is an important manifestation of human activity, which alters the natural runoff process of a river; when a reservoir is impounded, a large amount of water is stored in the reservoir and does not continue to flow along the river, leading to a reduction in the total amount of runoff from the river. In addition, in order to meet the needs for water for irrigated agriculture, economic development, and residential life, human beings have taken a series of measures to divert water resources away from rivers, which has led to a decrease in the runoff volume of rivers. Therefore, with the construction of cascade reservoirs creating an artificial steady environment devoid of natural extreme phenomena, this may be a contributing factor to the significant changes observed in hydrological indicators (Lu 2005).

In addition to reservoir operations, changes in land use also play a crucial role in influencing runoff. Over the past half-century, there has been a rapid expansion of construction land in the JLR Basin, while the implementation of ecological restoration policies such as ‘returning farmland to forests’ has led to significant changes in land use and vegetation (Yang et al. 2015). To reveal the changes in land use and vegetation cover in the JLR Basin, this study examined the land use changes and the NDVI variations from 1980 to 2020. As illustrated in Figures 13 and 14, over the past 40 years, forest and wetland areas have experienced relatively slow growth rates, with overall increases of 2.59 and 17.34%, respectively. Construction land initially experienced rapid growth, followed by a slower increase with an overall increase of 199.18%. In this case, most of the surfaces of construction land are made of impermeable materials such as cement and asphalt, resulting in the inability of precipitation to infiltrate, at which point it can rapidly form surface runoff. Therefore, the dramatic expansion of construction land in the JLR Basin will significantly increase surface runoff. The area of cropland has shown a continuous decreasing trend, declining by 3.16%. Grassland and unused land both exhibited a trend of initial decrease followed by increase, with overall reductions of 3.25 and 8.12%, respectively. A conversion of 4.97% of grassland to cultivated land, 6.32% of grassland to forest land, and 6.25% of cultivated land to forest land was observed in the watershed. The conversion rates among wetlands, urban construction land, and unused land remained relatively low. Cropland possesses a weak water retention capacity; in contrast, forests and grasslands exhibit excellent water retention capacity, and they inhibit the increase in runoff by trapping rainwater. However, in seasons of high rainfall, the soil water content of cropland, woodland, and grassland is saturated and cannot provide effective water storage and retention, leading to a further increase in runoff. In the southern part of the basin, urban land predominates, characterized by relatively flat terrain, exhibiting a transition trend from southwest to northeast with relatively minor variations. The annual average NDVI shows an increasing trend (Cui et al. 2020), rising from 0.61 in 1980 to 0.77 in 2020, marking a 26.23% increase. It reached its peak in 2020, indicating the significant impact of large-scale ecological restoration projects implemented since the 1980s, which also positively influenced biodiversity development (Wang et al. 2019).
Figure 13

Temporal and spatial changes in land use and vegetation cover from 1980 to 2020.

Figure 13

Temporal and spatial changes in land use and vegetation cover from 1980 to 2020.

Close modal
Figure 14

The rate of change in land use and vegetation cover from 1980 to 2020.

Figure 14

The rate of change in land use and vegetation cover from 1980 to 2020.

Close modal

The impact of runoff changes on ecology

The variation of runoff under natural conditions can serve as one of the crucial indicators for evaluating the level of ecological diversity and the vulnerability of the ecological environment in a watershed (Zhang et al. 2009). Runoff, as the overall manifestation of water flow within a watershed, is influenced by natural factors such as precipitation, evaporation, infiltration, and vegetation cover. Therefore, it can reflect the hydrological processes of the watershed and the state of the ecosystem, thereby assessing the adaptability of the ecosystem to hydrological changes. The majority of the extreme temperature indices in the JLR Basin exhibit an increasing trend, indicating a gradual warming of the basin. This trend may be due to global warming caused by strong anthropogenic emissions of greenhouse gases or changes in LUCC. The rise in temperature may lead to an increase in evapotranspiration, thereby affecting watershed runoff. In this study, it was observed that the runoff during the variability period was lower than that of the baseline period, and the hydrological alteration reached 50.70%, confirming this observation.

As a significant geographical and ecological region in the upper reaches of the Yangtze River, the JLR Basin boasts rich fish diversity, currently representing one-third of the entire Yangtze River Basin (Xu et al. 2012). The ecological environment protection and high-quality development of the basin have been elevated to a national strategic level. Due to the changes in the hydrological environment of the JLR Basin, fish, and vegetation within the basin have exhibited diverse responses. Furthermore, excessive alterations in water flow have modified the original aquatic ecosystem of rivers, affecting habitats suitable for fish growth and reproduction, thereby influencing fish migration patterns and reproductive behaviors (Liu et al. 2021). The reduction in fish habitat has led to alterations in the spawning times of major fish species and a sharp decline in spawning quantity, ultimately resulting in a decrease in the fish diversity indicator in the JLR (Yuan et al. 2019). With the increase in planktonic organisms in reservoirs, there is a likelihood of a corresponding increase in certain stagnant water fish species, which may contribute to stimulating the growth in reservoir fishery yield (Zeng et al. 2014). Taking the middle reaches of the JLR as an example, approximately 28 species of waterfowl have been observed, among which the species including the spot-billed duck, black-headed gull, and cormorant constitute 50.87% of the total bird population in the area (Jiang & He 2008). Recent surveys indicate a gradual decline in the number of waterfowl. This trend is primarily attributed to the ‘channelization’ of the JLR and the construction of cascade reservoirs, resulting in the submergence of large areas of riverine floodplains, central embankments, and other habitats suitable for waterfowl within the basin. The alteration has brought about noticeable changes to the surrounding ecological environment, and in the short term, the process of natural restoration to form new suitable habitats is relatively challenging (Zhou & Fang 1998). Meanwhile, the homogenization of water flow may reduce the water and nutrient availability necessary for the vegetation in the JLR Basin, resulting in a decrease in vegetation coverage (Xu 2019). Therefore, in studying and managing the hydrological environment of the basin, it is necessary to comprehensively consider the integrated impacts of human activities in order to develop effective water resource management strategies and environmental protection measures (Hu et al. 2016).

To investigate the impact of climate change and human activities on the runoff variation in the JLR Basin, this study employs the IHA–RVA method to assess hydrological conditions based on annual variability characteristics and quantifies the extent of runoff process changes in the altered environment. Concurrently, the RClimDex method is utilized to compute various extreme weather indices, providing deeper insights into the impact of climate change on runoff. Subsequently, grey relational analysis and cross-wavelet analysis are employed to analyze the dynamic response between meteorological and hydrological indicators. Finally, the SWAT model is utilized to quantitatively separate the impacts of climate change and human activities on runoff in both temporal and spatial dimensions, with validation conducted using the Budyko model based on the principle of water balance. This study provides scientific guidance for the spatial and temporal regulation of water resources in the JLR Basin through the quantification of meteorological and hydrological indicators and the attribution analysis on the spatial and temporal scales. The research findings indicate:

  • (1) The runoff at the Beibei hydrological station in the JLR Basin exhibited a decreasing trend from 1965 to 2019, and all passed the 95% significance test. Eventually, 1994 was determined to be the year of abrupt change. From 1965 to 2019, the JLR Basin experienced an increasing trend in extreme temperatures, with nighttime warming rates surpassing those of daytime, indicative of a warming climate. Apart from R10 mm, all other extreme precipitation indices showed an upward trend, suggesting a gradual drying trend in the JLR Basin. The hydrological variation degree in the JLR Basin was 50.70%, categorized as a moderate change.

  • (2) The correlation among the six sets of indices, including PRCPTOT and 30-day minimum, PRCPTOT and high pulse count, SDII and 30-day minimum, PRCPTOT and June, SDII – high pulse count, and SDII and June, is notably high. PRCPTOT and High pulse count, PRCPTOT and 30-day minimum, SDII and high pulse count, and SDII and 30-day minimum exhibit significant resonance periods during 1965–2019, mainly concentrated in the 4–6-month signals from 2000 to 2006.

  • (3) Utilizing the Budyko framework for quantitative attribution, it was found that precipitation is the most sensitive driver of runoff variability in the JLR Basin. At the annual and seasonal scales, human activities emerge as significant influencing factors on runoff variations in the JLR. However, at the monthly scale, there is a notable shift in the proportion of different drivers, with climate change exhibiting the highest contribution in March at 77.03%, except for March and November, the remaining months are dominated by human activities. In spatial terms, the impact of climate change on runoff outweighs that of land use and is predominantly concentrated in the southeastern region.

The authors thank their brothers at North China University of Water Resources and Electric Power for their comments and help with this study.

H. W. rendered support in funding acquisition project administration, and prepared the resources. Y. M. conceptualized the process, rendered support in data curation and formal analysis, investigated the study, developed the methodology, prepared the resources and software, validated and visualized the work, wrote the original draft, wrote the review and edited the article. W. Y. investigated the process, rendered support in formal analysis, developed the methodology, validated the work, prepared the resources. H. Y. rendered support in data curation and formal analysis, investigated and validated the work. H. Y. and Y. L. investigated the work, rendered support in formal analysis, validated the study, prepared the resources. W. G. rendered support in funding acquisition and project administration.

This study was supported by the Basic Research Project of Key Scientific Research Projects of Colleges and Universities of Henan Province (24ZX007).

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

The authors declare there is no conflict.

Anderson
S.
&
Radić
V.
(
2022
)
Evaluation and interpretation of convolutional long short-term memory networks for regional hydrological modelling
,
Hydrology and Earth System Sciences
,
26
(
3
),
795
825
.
doi:10.5194/hess-26-795-2022
.
Arnold
J. G.
,
Srinivasan
R.
,
Muttiah
R. S.
&
Williams
J. R.
(
1998
)
Large area hydrologic modeling and assessment part I: Model development
,
JAWRA Journal of The American Water Resources Association
,
34
(
1
),
1
17
.
doi:10.1111/j.1752-1688.1998.tb05961.x
.
Bhattacharya
R. K.
,
Chatterjee
N. D.
&
Das
K.
(
2020
)
Sub-basin prioritization for assessment of soil erosion susceptibility in Kangsabati, a plateau basin: A comparison between MCDM and SWAT models
,
Science of The Total Environment
,
734
,
139474
.
doi:10.1016/j.scitotenv.2020.139474
.
Chen
L.
,
Wang
Y.
,
Touati
B.
,
Guan
H.
,
Leng
G.
,
Liu
W.
,
Lv
S.
,
Huang
S.
&
Pan
Z.
(
2018
)
Temporal characteristics detection and attribution analysis of hydrological time-series variation in the seagoing river of southern China under environmental change
,
Acta Geophysica
,
66
(
5
),
1151
1170
.
doi:10.1007/s11600-018-0198-y
.
Cui
L.
,
Wang
Z.
&
Deng
L.
(
2020
)
Vegetation dynamics based on NDVI in Yangtze River Basin (China) during 1982–2015
,
IOP Conference Series Materials Science and Engineering
,
780
,
062049
.
doi:10.1088/1757-899X/780/6/062049
.
Deng
J. L.
(
1989
)
Introduction to Grey system theory
,
The Journal of Grey System
,
1
(
1
),
1
24
.
doi:10.1007/978-3-642-16158-2_1
.
Feng
R.
,
Rong
Y.
&
Yin
Y.
(
2020
)
Influence of three Gorges Dam construction on runoff distribution in the upper reaches of the Yangtze River
,
Journal of Beijing Normal University (Natural Science)
,
56
(
3
),
334
342
.
doi:10.12202/j.0476-0301.2020167
.
Govindaraju
R. S.
&
Rao
A. R.
(
2000
)
Artificial neural networks in hydrology
,
Journal of Hydrologic Engineering
,
5
(
2
),
124
137
.
doi:10.1061/(ASCE)1084-0699(2000)5:2(124)
.
Guo
W.
,
Zhou
H.
,
Jiao
X.
,
Huang
L.
&
Wang
H.
(
2022
)
Analysis of alterations of the hydrological situation and causes of river runoff in the Min River, China
,
Water
,
14
(
7
),
1093
.
doi:10.3390/w14071093
.
Hayashi
S.
,
Murakami
S.
,
Xu
K. Q.
&
Watanabe
M.
(
2015
)
Simulation of the reduction of runoff and sediment load resulting from the Gain for Green Program in the Jialingjiang catchment, upper region of the Yangtze River, China
,
Journal of Environmental Management
,
149
(
Feb.1
),
126
137
.
doi:10.1016/j.jenvman.2014.10.004
.
Hu
Y.
,
Feng
J.
,
Wang
M.
,
Tian
F.
&
He
X.
(
2016
)
Influences of climate and land surface change on runoff and sediment in Jialing River Basin
,
Science of Soil and Water Conservation
,
14
(
4
),
75
83
.
doi:10.16843/j.sswc.2016.04.010
.
Huang
S.
,
Li
P.
,
Huang
Q.
,
Leng
G.
,
Hou
B.
&
Ma
L.
(
2017
)
The propagation from meteorological to hydrological drought and its potential influence factors
,
Journal of Hydrology
,
547
,
184
195
.
doi:10.1016/j.jhydrol.2017.01.041
.
Ip
W. C.
,
Hu
B. Q.
,
Wong
H.
&
Xia
J.
(
2009
)
Applications of grey relational method to river environment quality evaluation in China
,
Journal of Hydrology
,
379
,
284
290
.
doi:10.1016/j.jhydrol.2009.10.013
.
Ji
L.
&
Duan
K.
(
2019
)
What is the main driving force of hydrological cycle variations in the semiarid and semi-humid Weihe River Basin, China?
,
Science of The Total Environment
,
684
(
Sep.20
),
254
264
.
doi:10.1016/j.scitotenv.2019.05.333
.
Jiang
G.
&
He
X.
(
2008
)
Status of fish resources in the lower reaches of the Jialing River
,
Freshwater Fisheries
,
S966(2), 3
.
doi:10.3969/j.issn.1000-6907.2008.02.001
.
Kendall
M. G.
(
1990
)
Rank correlation methods
,
British Journal of Psychology
,
25
(
1
),
86
91
.
doi:10.2307/2333282
.
Krisnayanti
D. S.
,
Bunganaen
W.
,
Frans
J. H.
,
Seran
Y. A.
&
Legono
D.
(
2021
)
Curve number estimation for ungauged watershed in semi-arid region
,
Civil Engineering Journal
,
7
(
6
),
1070
1083
.
doi:10.28991/cej-2021-03091711
.
Li
Z.
,
Li
Q.
,
Wang
J.
,
Feng
Y.
&
Shao
Q.
(
2020
)
Impacts of projected climate change on runoff in upper reach of Heihe River Basin using climate elasticity method and GCMs
,
The Science of The Total Environment
,
716
(
May10
),
137072.137071
137072.137013
.
doi:10.1016/j.scitotenv.2020.137072
.
Li
Q.
,
Gao
H.
,
Chai
S.
,
He
Z.
&
Ji
G.
(
2022
)
Attribution analysis of different driving forces on vegetation and streamflow variation in the Jialing River Basin, China
,
Open Geosciences
,
14
(
1
),
1513
1523
.
doi:10.1515/geo-2022-0443
.
Liu
Q.
,
Singh
V.
&
Xiang
H.
(
2005
)
Plot erosion model using gray relational analysis method
,
Journal of Hydrologic Engineering 10 (4), 288–294
.
doi:10.1061/(ASCE)1084-0699(2005)10:4(288)
.
Liu
J.
,
Chen
J.
,
Xu
J.
,
Lin
Y.
&
Zhou
M.
(
2019
)
Attribution of runoff variation in the headwaters of the Yangtze River Based on the Budyko Hypothesis
,
International Journal of Environmental Research and Public Health
,
16
(
14
),
2506
.
doi:10.3390/ijerph16142506
.
Liu
Y.
,
Feng
X.
,
Pu
D.
,
Gu
H.
,
Tian
J.
,
Zhao
Z.
,
Huang
J.
,
Zhu
J.
&
Wang
Z.
(
2021
)
Characteristics and resource status of main commercial fish in the middle reaches of Jialing River
,
Chinese Journal of Applied and Environmental Biology
,
27 (04), 837–847
.
doi:10.19675/j.cnki.1006-687x.2020.10010
.
Lu
M.
&
Wevers
K.
(
2007
)
Grey system theory and applications: A way forward
,
Journal of Grey System
,
10
,
47
53
.
doi:10.30016/JGS.200706.0007
.
Mann
H. B.
(
1945
)
Nonparametric test against trend
,
Econometrica
,
13
(
3
),
245
259
.
doi:10.2307/1907187
.
Mimikou
M. A.
,
Kanellopoulou
S. P.
&
Baltas
E. A.
(
1999
)
Human implication of changes in the hydrological regime due to climate change in Northern Greece
,
Global Environmental Change Guildford
,
9 (2), 139–156
.
doi:10.1016/S0959-3780(98)00034-X
.
Moriasi
D. N.
,
Arnold
J. G.
,
Liew
M. W. V.
,
Bingner
R. L.
,
Harmel
R. D.
&
Veith
T. L.
(
2007
)
Model evaluation guidelines for systematic quantification of accuracy in watershed simulations
,
Transactions of The ASABE
,
50
(
3
),
885
900
.
doi:10.13031/2013.23153
.
Oeurng
C.
,
Cochrane
T. A.
,
Chung
S.
,
Kondolf
M.
,
Piman
T.
&
Arias
M. E.
(
2019
)
Assessing climate change impacts on river flows in the Tonle Sap Lake Basin, Cambodia
,
Water
,
11 (3), 618
.
doi:10.3390/W11030618
.
Papalexiou
S. M.
&
Montanari
A.
(
2019
)
Global and regional increase of precipitation extremes under global warming
,
Water Resources Research
,
55
(
6
),
4901
4914
.
doi:10.1029/2018wr024067
.
Qi
S.
,
Zhang
X.
,
Xia
D.
,
Liu
Y.
&
Liu
B.
(
2013
)
Characteristics and driving forces of land use/cover change in Jialingjiang River Basin (Chinese)
,
Journal of Yangtze River Scientific Research Institute
,
30
(
01
),
1
7
.
doi:10.3969/j.issn.1001-5485.2013.01.001
.
Richter
B. D.
,
Baumgartner
J. V.
,
Wigington
R.
&
Braun
D. P.
(
1997
)
How much water does a river need?
,
Freshwater Biology
,
37
,
231
249
.
doi:10.1046/J.1365-2427.1997.00153.X
.
Roderick
M. L.
&
Farquhar
G. D.
(
2011
)
A simple framework for relating variations in runoff to variations in climatic conditions and catchment properties
,
Water Resources Research
,
47
(
12
),
1–11
.
doi:10.1029/2010wr009826
.
Sampson
A. P.
,
Weli
V. E.
,
Nwagbara
M. O.
&
Eludoyin
O. S.
(
2021
)
Sensations of air temperature variability and mitigation strategies in urban environments
,
Journal of Human Earth and Future, 2 (2), 100–113
.
doi:10.28991/HEF-2021-02-02-02
.
Shenglian
G.
,
Jiali
G.
,
Yukun
H.
,
Lihua
X.
&
Xingjun
H.
(
2015
)
Prediction of future runoff change based on Budyko hypothesis in Yangtze River Basin
,
Advances in Water Science
,
26 (2), 151–160
.
doi:10.14042/j.cnki.32.1309.2015.02.001
.
Srinivasan
R.
,
Ramanarayanan
T. S.
,
Arnold
J. G.
&
Bednarz
S. T.
(
1998
)
Large area hydrologic modeling and assessment part II: model application1
,
JAWRA Journal of the American Water Resources Association
,
34
,
91–101
.
doi:10.1111/j.1752-1688.1998.tb05962.x
.
Suh
S. H.
(
2004
)
Fuzzy-bayes fault isolator design for BLDC motor fault diagnosis
,
International Journal of Control Automation & Systems
,
2
(
3
),
354
361
.
doi:10.1002/acs.820
.
Wang
Y.
,
Wang
J.
,
Wu
M.
&
Wang
S.
(
2019
)
Impacts of the land use and climate changes on the hydrological characteristics of Jialing River Basin
,
Research of Soil and Water Conservation
, 26 (01), 135–142.
doi:10.13869/j.cnki.rswc.2019.01.019
.
Wang
H.
,
Ma
Y.
,
Yang
H.
,
Hong
F.
&
Guo
W.
(
2023a
)
Quantitative evaluation of the impact of climate change and human activities on Jialing river runoff changes in the past 60 years, China
,
Journal of Water and Climate Change
,
14
(
2
),
590
609
.
doi:10.2166/wcc.2023.452
.
Wang
H.
,
Yuan
W.
,
Chen
W.
,
Hong
F.
,
Bai
X.
&
Guo
W.
(
2023b
)
Responseofhydrological regimes to land use change: A case study of the Han River Basin
,
Journal of Water and Climate Change
,
14
(
12
),
4708
.
doi:10.2166/wcc.2023.483
.
Wu
Y.
&
Pu
H.-W.
(
2019
)
The influence of hydropower station on sand content detection in Jialing River
,
Technological Development of Enterprise
,
38
(
09
),
55
58
.
doi:10.14165/j.cnki.hunansci.2019.09.014
.
Xu
X.
(
2019
)
Comprehensive assessment of the water ecological security of the Xiangjiang River Basin based on physico-chemistry and organism indices
,
Applied Ecology and Environmental Research
,
17
(
2
),
4547
4574
.
doi:10.15666/aeer/1702_45474574
.
Xu
Z.
&
Jiang
Y.
(
2022
)
Studies on runoff evolution mechanism under changing environment: A state-of-the-art review
,
Hydro-Science and Engineering
,
1
,
10
.
doi:10.12170/20210105008
.
Xu
W.
,
Qiao
Y.
&
Gong
Y.
(
2012
)
Changes of fish resources in upper Yangtze River and its protection
,
Yangtze River
,
43(01), 67–71
.
doi:10.3969/j.issn.1001-4179.2012.01.020
.
Xu
Y.
,
Chen
Y.
,
Ren
Y.
,
Tang
Z.
,
Yang
X.
&
Zhang
Y.
(
2023
)
Attribution of streamflow changes considering spatial contributions and driver interactions based on hydrological modeling
,
Water Resources Management
,
37
(
5
),
1859
1877
.
doi:10.1007/s11269-023-03459-3
.
Yang
D.
,
Zhang
S.
&
Xu
X.
(
2015
)
Attribution analysis for runoff decline in Yellow River Basin during past fifty years based on Budyko hypothesis
,
Scientia Sinica
,
45
(
10
),
1024
.
doi:10.1360/N092015-00013
.
Yang
H.
,
Guo
W.
&
Wang
H.
(
2024
)
Anthropogenic intensification of the eco-hydrothermal regime transition in regulated rivers: The cumulative effect of cascade reservoirs
,
Journal of Environmental Management
,
349
,
119478. doi:10.1016/j.jenvman.2023.119478
.
Yuan
X.
,
Yang
X.
,
Ge
H.
,
Li
H.
&
Deng
D.
(
2019
)
Temporal distribution of fish community structure in Dongting Lake Estuary
,
Agricultural Sciences
,
10 (03), 294–301
.
doi:10.4236/AS.2019.103025
.
Zeng
Y.
,
Chen
Y.
&
Li
Z.
(
2014
)
Utilization and protection status of fish resources in Jialing River
,
Tianjin Agricultural Sciences
,
20 (02), 60–62
.
doi:10.3969/j.issn.1006-6500.2014.02.014
.
Zhang
L.
,
Chen
X.
,
Zhao
Z.
&
Hu
Z.
(
2008
)
Progress in study of climate change impacts on hydrology and water resources
,
Progress in Geography
,
03
,
60
67
.
doi:10.11820/dlkxjz.2008.03.009
.
Zhang
X.
,
Wang
K.
,
Zhang
W.
,
Chen
H.
&
He
X.
(
2009
)
The quantitative assessment of eco-environment vulnerability in karst regions of Northwest Guangxi
,
Acta Ecologica Sinica
,
29 (2), 749–757
.
doi:10.1360/972008-2465
.
Zhang
H.
,
Meng
C.
,
Wang
Y.
,
Wang
Y.
&
Li
M.
(
2020
)
Comprehensive evaluation of the effects of climate change and land use and land cover change variables on runoff and sediment discharge
,
Science of The Total Environment
,
702
,
134401. doi:10.1016/j.scitotenv.2019.134401
.
Zhang
J.
,
Zhang
M.
,
Song
Y.
&
Lai
Y.
(
2021
)
Hydrological simulation of the Jialing River Basin using the MIKE SHE model in changing climate
,
Journal of Water and Climate Change
,
12
(
6
),
2495
2514
.
doi:10.2166/wcc.2021.253
.
Zhou
F.
&
Fang
H.
(
1998
)
Prediction of impacts of Changzhou Water Conservancy Project on waterbird in the reservoir area
,
Chinese Biodiversity
,
6 (1), 42–48
.
doi:10.3321/j.issn:1005-0094.1998.01.008
.
Zhuo
Z.
,
Qian
Z.
,
Jiang
H.
,
Wang
H.
&
Guo
W.
(
2020
)
Evaluation of hydrological regime in Xiangjiang Basin based on IHA-RVA method
,
China Rural Water Hydropower
,
8
,
188
192
.
doi:1007-2284(2020) 08-0188-05
.
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/).