The Attribution Identification of Runoff Changes in the Kriya River Based on the Budyko Hypothesis Provides a Basis for the Sustainable Management of Water Resources in the Basin

Abstract

Identifying the impact of climate change and changes in underlying surface conditions on river runoff changes is critical for sustainable water resource use and watershed management in arid regions. The Kriya River is not only a key support for water resources in the arid environment of the Tarim Basin, but also a solid foundation for the survival and development of agricultural oases. In this study, the Kriya River Basin in Xinjiang, China, was taken as the research object, and the Mann–Kendall, Sen’s Slope, Cumulative Sum, and other methods were used to systematically analyze the temporal evolution law and multi-modal characteristics of runoff in the basin. Based on the Budyko hydrothermal coupling equilibrium equation, the contribution of temperature, evaporation, and the underlying surface to runoff variation was quantitatively interpreted. The study found that the annual runoff depth of the Kriya River Basin has shown a significant positive evolution trend in the past 60 years, with an increase rate of 0.5189 mm/a (p ≤ 0.01). Through the identification of mutation points, the runoff time series of the Kriya River was divided into the base period 1957–1999 and the change period 2000–2015. Without considering the supply of snowmelt runoff, the contribution rate of precipitation to runoff change was 75.23%, followed by the change in underlying surface (23.08%), and the potential evapotranspiration was only 1.69%. The results of this study provide a good scientific reference for water resources management and environmental governance in the Kriya River Basin.

1. Introduction

Both anthropogenic activities and climate change drive alterations in runoff dynamics across the entire hydrological system [1], especially in water-scarce arid regions, and are increasingly becoming a matter of concern for water managers and policymakers [2]. Anthropogenic modifications to watershed surface cover (e.g., impervious surfaces, vegetation removal) constitute the main mechanism of hydrological cycle alteration, which covers a variety of aspects such as changes in land use, implementation of water conservation policies, and construction of water conservancy facilities [3]. Interactions among these elements synergistically promote the formation and accumulation of overland flow. By reshaping hydroclimatic variables—specifically precipitation intensity/frequency and evapotranspiration demand—climate change fundamentally alters runoff generation mechanisms [4]. Over longer time scales, runoff generation depends on a dynamic balance between precipitation inputs and evaporative losses [5]. Specifically, precipitation, as the main source of runoff, provides the necessary recharge, while evapotranspiration constitutes the consumptive part of the runoff water volume [6]. There are significant differences in the extent to which rivers in different geographical areas are affected by climate change and human activities [7]. In view of this, it is particularly urgent and important to carry out special research on different rivers to accurately grasp their ecological evolution trends and formulate scientific and reasonable protection and management strategies, so as to provide a solid scientific basis for the sustainable use of river ecosystems and regional sustainable development.

Statistical methods [8], hydrological models [9,10], and conceptual models based on water balance [11] are frequently employed in research examining the hydrological impacts of anthropogenic activities and climatic variability. Statistical methods, such as trend and correlation analyses, despite their simplicity, have limitations in revealing the detailed physical mechanisms behind runoff changes and the quantification process [12]. Process-oriented hydrological modeling is a powerful tool to explore variables and their interactions based on physical mechanisms [13,14]. While this methodological framework enables a more nuanced dissection of how climatic drivers and anthropogenic interventions independently and interactively shape watershed runoff dynamics, the reliability of the results still faces certain challenges due to the complexity of the model architecture, the high demand for data, and the tedious process of model calibration and validation [15]. The Budyko coupled hydrothermal balance equation, as one of the conceptual approaches, provides an intuitive and efficient way to understand and analyze water and energy limitations in long-term hydrological processes [16], while also facilitating a quantitative attribution analysis of climatic versus anthropogenic influences on watershed runoff variability and long-term trends [17]. In recent years, the Budyko framework has gained widespread adoption in hydroclimatic research due to its theoretically robust physical basis and mathematically tractable formulation of water–energy balance principles [18,19]. The significant advantage of this method is its practicality and convenience, which does not require complex parameter adjustment and massive data support, and is especially suitable for watersheds with scarce or restricted data [20].

As a key inland river flowing into the hinterland of the Taklamakan Desert [21], the Kriya River plays a pivotal role in maintaining the ecological environment of the region and promoting local development. Currently, some scholars are committed to analyzing the long-term runoff data in the Kriya River Basin, aiming to explore the seasonal fluctuations, interannual trends, and long-term development of runoff, aiming to achieve a comprehensive characterization of the spatiotemporal heterogeneity in runoff dynamics and its long-term evolutionary trajectories [22,23]. This research systematically investigates the hydroclimatic responses of the Kriya River Basin to global warming, with particular emphasis on disentangling the relative contributions of precipitation regime shifts and temperature-induced evapotranspiration changes to runoff variability [21]. Mathematical and statistical tools were used with the aim of analyzing the combined effect of human activities and climate change on the runoff volume of the Kriya River [24]. However, current studies on the Kriya River runoff still face challenges of data uncertainty, model oversimplification, and methodological limitations, which affect our quantitative understanding of how climate change and human activities specifically affect Kriya River runoff.

This study focuses on the Kriya River Basin, using Mann–Kendall, Sen’s Slope, and Cumulative Sum for trend diagnosis and mutation point identification, and it deeply discusses the trend evolution and stage characteristics of the river runoff. At the same time, with the help of the Budyko hydrothermal coupling equilibrium equation, the precipitation, evaporation, and underlying surface conditions affecting runoff changes are quantitatively evaluated, and the contribution ratio of each factor is clarified. This study focuses on the runoff change process in the upper reaches of the Kriya River, aiming to provide a scientific basis for accurately grasping the water resources dynamics in the region through systematic research. At the same time, it is expected that the results of this study can contribute a new reference perspective to the study of the Budyko framework of runoff change in arid areas from the perspective of sustainable water resources management, promote the innovative application of relevant theories in the practice of sustainable utilization of water resources in arid areas, and help arid areas achieve coordinated and sustainable development of water resources, ecology, and social economy.

2. Materials and Methods

2.1. Study Area

The Kriya River originates from the north side of the Ustenge Mountain in the Kunlun Mountains and meanders from south to north through the southern fringe of the Tarim Basin, flowing through Yutian County [25]. The total length of the river is about 438 km, and its longitudinal specific fall reaches 3.4‰. Langan Hydrological Station is the controlling hydrological station of the Kriya River (81°28′10.26″ E, 36°27′58.45″ N), and the catchment area above the Langan Hydrological Station at the outgoing pass is 7358 km², with a site elevation of 1880 m. The Jiyin Reservoir, constructed upstream of the hydrological station, was lowered and impounded at the end of 2016; therefore, only natural monthly runoff data from 1957 to 2015 at the Langan Hydrological Station were used for the analysis in this paper, which was free from anthropogenic disturbances above the station during that period. The average multi-year runoff of the Kriya River from 1957 to 2015 was 755 million m³, the maximum annual runoff was 1139 million m³, and the minimum annual runoff was 489 million m³; the spring flow was 11.04 m³/s, the summer flow was 63.72 m³/s, the fall flow was 14.28 m³/s, and the winter flow was 6.75 m³/s, which was unevenly distributed within the year. The distribution of hydrological and meteorological stations in the study area is shown in Figure 1.

2.2. Data Sources and Processing

Monthly runoff data of the Kriya River for the period of 1957–2015 were provided by the Langan Hydrological Station under the Hotan Regional Hydrological Bureau. Meanwhile, the meteorological data of the basin were obtained from the Yutian meteorological station, and the detailed data can be obtained by visiting the official website of the National Center for Meteorological Information (http://data.cma.cn/, accessed on 10 August 2024), which covers the daily records of the Yutian meteorological station from 1957 through 2015, including information on air temperature (maximum, minimum, and average), wind speed, sunshine duration, daily precipitation, average relative humidity, and average water vapor pressure. Based on the daily meteorological data from the Yutian meteorological station, the potential evapotranspiration in the study area was calculated using the FAO-modified Penman–Monteith formula as shown below [26]:

$$E{T}_0={\displaystyle \frac{0.408\Delta (R_n-G)+\gamma \frac{900}{T+273}{U}_2(e_s-e_a)}{\Delta +\gamma (1+0.34U_2)}}$$

(1)

where ET₀ is the calculated potential evapotranspiration in mm/d; Δ is the slope of the saturated water vapor pressure curve in kPa/°C; R_n is the net solar radiation in MJ m⁻² d⁻¹; G is the soil heat flux in MJ m⁻² d⁻¹; γ is the dryness and wetness constant in kPa/°C; T is the average air temperature in °C; U₂ is the wind speed at a height of 2 m in m/s; e_s and e_a are the average saturated water vapor pressure and actual water vapor pressure, respectively, both in kPa.

2.3. Research Methods

2.3.1. Trend Test and Mutation Test

Mann–Kendall trend test and mutation test analyses are widely used in hydrological time series analysis; they can accurately reveal the trend and mutation characteristics of hydrological series changes, showing high accuracy and strong robustness to outliers, even for non-normally distributed data, and can be effectively applied [27]. The detailed calculation steps can be found in the related literature [28,29].

As a classical nonparametric trend estimation method, Sen’s Slope is not sensitive to outliers and does not depend on data distribution assumptions. It is widely used in long-term trend analysis in meteorology, hydrology, remote sensing, and other fields, especially for data with noise or nonlinear changes [30,31].

The Cumulative Sum method (CUSUM) is a statistical method for detecting change points in data sequences. It monitors the degree of deviation of monitoring data by calculating the cumulative sum of data sequences, and it triggers events when the cumulative sum exceeds the preset threshold [32].

2.3.2. Runoff Change Attribution Recognition

By employing the Budyko coupled hydrothermal equilibrium equations, one can quantify the sensitivity and assess the contribution of climate change and human activities to variations in runoff [33]. This method has been widely adopted and applied in analyzing the extent to which climate change and surface cover changes affect runoff [34,35].

Sensitivity Analysis

Based on the theoretical framework of Budyko’s coupled hydrothermal equilibrium theory, the actual evapotranspiration estimation model applicable to long time scales is constructed with the theoretical expression:

$$ET={\displaystyle \frac{P\times E{T}_0}{{(P^{\omega }+E{T}_0^{\omega })}^{1/\omega }}}$$

(2)

where ET and ET₀ denote the annual actual and potential evapotranspiration of the basin, respectively, and P denotes the annual precipitation in mm; the parameter ω, as a dimensionless characterization of the comprehensive attributes of the basin’s underlying surface, is determined by the coupling of soil physicochemical properties, spatial pattern of topography, and ecological function of vegetation.

On long time scales, closed basins follow the principle of water balance, i.e.,:

$$R=P-ET$$

(3)

where R, P, and ET are the multi-year average runoff depth, multi-year average precipitation, and multi-year average actual evapotranspiration in mm. Substituting Equation (2) into Equation (3) gives the following:

$$R=P-{\displaystyle \frac{P\times E{T}_0}{{(P^{\omega }+E{T}_0^{\omega })}^{1/\omega }}}$$

(4)

Setting the precipitation factor (P), potential evapotranspiration capacity (ET₀), and underlying surface property parameter (ω) to satisfy the statistical independence condition in the theoretical derivation, Equation (3) $R=f$ can be rewritten and the runoff depth can be expressed as follows: $R=f(P,E{T}_0,\omega )$

$$dR={\displaystyle \frac{\partial f}{\partial P}}dP+{\displaystyle \frac{\partial f}{\partial E{T}_0}}dE{T}_0+{\displaystyle \frac{\partial f}{\partial \omega }}d\omega$$

(5)

Defining ${\epsilon }_p={\displaystyle \frac{dR/R}{dP/P}}$, ${\epsilon }_{E{T}_0}={\displaystyle \frac{dR/R}{dE{T}_0/E{T}_0}}$, and ${\epsilon }_{\omega }={\displaystyle \frac{dR/R}{d\omega /\omega }}$ as the elasticity coefficient of precipitation, the elasticity coefficient of potential evapotranspiration, and the elasticity coefficient of underlying surface coefficient of runoff, respectively, Equation (5) can be expressed as follows:

$${\displaystyle \frac{dR}{R}}={\epsilon }_P{\displaystyle \frac{dP}{P}}+{\epsilon }_{E{T}_0}{\displaystyle \frac{dE{T}_0}{E{T}_0}}+{\epsilon }_{\omega }{\displaystyle \frac{d\omega }{\omega }}$$

(6)

By further derivation, the expression for each elasticity coefficient is $(\varphi =E{T}_0/P)$:

$${\epsilon }_p={\displaystyle \frac{{(1+{\varphi }^{\omega })}^{1/\omega +1}-{\varphi }^{\omega +1}}{(1+{\varphi }^{\omega })\left[{(1+{\varphi }^{\omega })}^{1/\omega }-\varphi \right]}}$$

(7)

$${\epsilon }_{E{T}_0}={\displaystyle \frac{1}{(1+{\varphi }^{\omega })\left[1-{(1+{\varphi }^{-\omega })}^{1/\omega }\right]}}$$

(8)

$${\epsilon }_{\omega }={\displaystyle \frac{\ln (1+{\varphi }^{\omega })+{\varphi }^{\omega }\ln (1+{\varphi }^{-\omega })}{\omega (1+{\varphi }^{\omega })\left[1-{(1+{\varphi }^{-\omega })}^{1/\omega }\right]}}$$

(9)

Based on Equations (7)–(9), the elasticity coefficients of precipitation, potential evapotranspiration, and underlying surface elements to runoff changes can be determined separately. The symbolic properties of the elasticity coefficient characterize the directionality of the response between the runoff evolution and the driving elements, and its absolute magnitude quantitatively characterizes the intensity of the contribution of different environmental elements to the runoff dynamics [36].

Given the dominance of snowmelt runoff in the runoff composition of the Kriya River [37], we make simplified assumptions based on the snowpack evolution process: (1) the sublimation of the snowpack layer is negligible; (2) snowmelt does not reach the threshold of runoff formation during the snow season; and (3) the snow water during the ablation period is completely transformed into surface runoff, during which the evaporation and infiltration losses can be considered as zero [38,39].

This implies that snow water is not involved in any form of evapotranspiration during the full cycle of snow accumulation–ablation. Within this theoretical framework, the multi-year water balance Equation (3) can be reconstructed as follows:

$$ET=P-R\left(1-r_s\right)$$

(10)

where R and r_s are the total amount of runoff and the proportion of glacial meltwater runoff recharge, respectively, and r_s can be determined by referring to the results of previous research [40].

Contribution Analysis of Runoff Changes

When the runoff series has significant variation, the study period can be divided into the base period and the change period based on the mutation detection point, and the difference in runoff depth between the two periods can be defined as ΔR. The Budyko equation suggests that changes in runoff depth (ΔR) can be decomposed into three types of driving effects: precipitation, potential evapotranspiration, and evolution of the underlying surface environment. The quantitative resolution of the single-factor contribution can be achieved through Equation (11), which has the following mathematical expression:

$$\Delta R_i={\epsilon }_i{\displaystyle \frac{R}{i}}\Delta i$$

(11)

where i is the study variable identifier, corresponding to precipitation (P), potential evapotranspiration (ET₀), and underlying surface characterization parameter (ω), respectively.

After calculation, the amount of change in runoff depth due to rainfall, potential evapotranspiration, and changes in the underlying surface coefficient during the change period compared to the base period can be derived, denoted as ΔRP, ΔRET₀, and ΔR_ω, respectively. Based on this, the quantitative contribution relationship between each driving element and the change in runoff can be further established, and the specific expression is as follows:

$${\eta }_i=\Delta R_i/\Delta R\times 100\%$$

(12)

It represents the contribution weights of precipitation, potential evapotranspiration, and adjustments to underlying surface parameters in influencing runoff variations.

3. Results and Analysis

3.1. Characterization of Hydrometeorological Elements

Table 1. Characteristics of hydrological climate change in the basin.

Basin	Station	Catchment Area/km2	Runoff Depth/mm				Annual Precipitation/mm		Annual Potential Evapotranspiration/mm
			Mean Value	Extreme Ratio	MK Test	Sen’s Slope	MK Test	Sen’s Slope	MK Test	Sen’s Slope
Kriya river	Langan	7358	102.61	2.33	2.9558 ***	0.47358	0.26158	0.04167	−3.5771	−0.00719

Note: *** The significance level was 0.01.

presents the characteristic parameters of runoff in the Kriya River basin, where the average multi-year runoff depth is 102.61 mm and the interannual runoff extreme ratio reaches 2.33. The nonparametric Mann–Kendall (MK) trend test method and Sen’s Slope method were used to analyze the runoff depth time series of the basin. The results showed that the runoff depth exhibited a significant increasing trend (p < 0.01). Figure 2 illustrates the linearly fitted trend of annual runoff depth in the watershed over the period of 1957–2015, which increased at a rate of 0.5189 mm/a. Through MK mutation detection, it was further determined that 1999 was the mutation year of the runoff series, and the relative change range of runoff was more than 30% after the mutation year. Figure 3 shows the runoff CUSUM statistics, further indicating that from 1957, after 43 years, to 1999, the runoff began to show a significant mutation.

Precipitation and evapotranspiration processes, as key meteorological elements, have a direct driving effect on river runoff changes. Based on the MK, Sen’s Slope trend test (Table 1), the annual precipitation in the study area did not show a significant trend, with a rate of change of 0.2149 mm/a (Figure 4). In contrast to the change in precipitation, the annual potential evapotranspiration showed a significant decreasing trend (p < 0.01) with a decreasing rate of −3.3199 mm/a.

3.2. Sensitivity Analysis of Runoff to Climate and Underlying Surface Changes

According to the principle of Budyko hydrothermal coupling equilibrium equation, the hydrological characteristics of the Kriya River Basin in different time periods were calculated and analyzed, and the sensitivity coefficients of runoff to precipitation, potential evapotranspiration, and underlying surface changes were systematically analyzed (see

Table 2. Characteristics of meteorological and hydrological variables.

Basin	Period	P/mm	R/mm	ET0/mm	ω	R/P	ET0/P	ε Elastic Coefficient
								εP	εET0	εω
Kriya river	1957–1999	48.23	94.46	1240.47	0.34	1.96	25.72	4.11	−0.65	−4.32
	2000–2015	57.95	124.49	1204.93	0.32	2.15	20.79	4.06	−0.56	−3.73

). The results show that under the influence of the deep intervention of human activities, the precipitation and runoff depth in the change period of the Kriya River Basin exhibit significant positive growth characteristics compared with the base period. At the same time, the annual potential evapotranspiration is characterized by a decrease. This change led to an increase in the runoff coefficient, which indicates a significant shift in precipitation distribution patterns, with more water being exported through surface runoff pathways rather than being consumed in evaporation processes. At the same time, the ratio of precipitation to potential evapotranspiration shows a significant downward trend, and this change directly confirms two important features of the hydrological cycle process: the enhancement of runoff generating capacity and the weakening of the evapotranspiration water consumption process, indicating that a systematic transformation of the basin’s water allocation pattern has occurred.

The change in runoff showed a negative correlation with potential evapotranspiration (ET₀) and underlying surface parameter ω, while maintaining a positive correlation with precipitation (P), suggesting that watershed runoff is primarily driven directly by precipitation inputs, while being inhibited by enhanced evaporative capacity and altered underlying surface conditions. Taking the year of abrupt change as the cut-off point, the precipitation elasticity coefficient undergoes a significant jump: from 4.11 in the base period (1957–1999) to 4.06 in the change period (2000–2015). This change resulted in a shift in the pattern of response of the hydrologic system to precipitation—under a scenario with the same 10% increase in precipitation, the increase in runoff could be as high as 41.1% before 1999, while this increase was suppressed to 40.6% after 1999. This indicates a 0.5% reduction in the efficiency of precipitation driving runoff, possibly reflecting degradation of the basin’s water storage capacity or the role of human activities in regulating hydrologic processes.

The response of runoff shows a similar trend when the potential evapotranspiration (ET₀) or underlying surface parameter ω is increased by 10%. Specifically, before 1999, a rise in potential evapotranspiration (or underlying surface parameter ω) resulted in a reduction in runoff of 6.5% (or 43.2%), respectively; whereas since 1999, the reduction in runoff has been adjusted to 5.6% (or 37.3%).

A comparative analysis of the absolute values of the elasticity coefficients of precipitation, potential evapotranspiration, and underlying surface parameter ω shows that the absolute value of the elasticity coefficient of the underlying surface parameter ω is the largest, the absolute value of the elasticity coefficient of precipitation is the second largest, and the absolute value of the elasticity coefficient of potential evapotranspiration is the smallest. The results of this study reveal that the Kriya River basin has a high sensitivity to precipitation recharge and underlying surface changes. The increase or decrease in precipitation and changes in the underlying surface, covering the process of precipitation interception and interception, filling of depressions and infiltration in the watershed, changes in the water storage capacity of the soil, as well as adjustments to hydrological characteristics such as surface roughness, stream length and direction of flow, can significantly affect the production and catchment mechanisms of the watershed.

Figure 5 presents the interannual variability of the aridity index (characterized by ET₀/P), precipitation, potential evapotranspiration, and underlying surface elasticity coefficients for the river outflow pass region of the Kriya River Basin for the period from 1957 to 2015. During this period, the aridity index showed a clear downward trend, which signifies that the basin is undergoing a significant warming and humidification process. In addition, the absolute values of the elasticity coefficients of precipitation and potential evapotranspiration likewise showed a decreasing trend, which indicates that the sensitivity of runoff to meteorological factors (i.e., precipitation and potential evapotranspiration) within the watershed has weakened.

3.3. Runoff Change Attribution Identification

Based on the Budyko coupled hydrothermal equilibrium equations, a quantitative assessment of the degree of contribution of climate change and underlying surface characteristics in the runoff formation mechanism was implemented (see

Table 3. Attribution identification of runoff change.

Basin	Base Period	Change Period	dRP/mm	dRET0/mm	dRω/mm	dR/mm	dR′/mm	CP/%	CET0/%	Cω/%
Kriya river	1957–1999	2000–2015	11.59	0.26	3.55	15.40	2.82	75.23	1.69	23.08

for details). Based on the comparative analysis of the table’s data, compared with the base period, the contribution of precipitation, potential evapotranspiration (ET), and underlying surface conditions to runoff formation showed significant differences. From the perspective of driving factors, the contribution of precipitation change to runoff increase is the most significant, accounting for 75.23%; the change in underlying surface conditions is the second, with a contribution rate of 23.08%. The contribution of potential evapotranspiration (ET) is the smallest, accounting for only 1.69%.

4. Discussion

4.1. Physical Response of Runoff to Precipitation and Watershed Characteristic Parameter ω

In arid zone watersheds, runoff recharge is heavily dependent on snowmelt recharge and strong control of precipitation events, especially the outlet of a river basin at the mountain pass, where snowmelt water accounts for nearly 30% of the recharge [40]. When the snow cover conditions on the glacier surface are favorable, seasonal snow contributes more significantly to runoff formation. Apart from such conditions, runoff is predominantly composed of pure glacial meltwater [41]. The above research conclusions provide strong support for the simplified hypothesis of the snow evolution process mentioned above. Based on this, in the runoff recharge pattern of arid areas, snowmelt water recharge dominates, and its contribution ranks first, while precipitation recharge ranks second. The runoff characteristics of the Kriya River also follow this rule. In the runoff recharge system of the river, snowmelt water occupies the primary position and is the largest source of runoff recharge. After excluding the runoff formed by snowmelt water recharge, the quantitative analysis shows that the contribution of precipitation change to the increase in runoff in the Kriya River is as high as 75.23%. Particularly during periods of variable precipitation, increases in precipitation can directly contribute to enhanced runoff.

The Budyko equation is an empirical formulation used to depict the hydrological cycle processes in a watershed, especially the correlation between evapotranspiration and runoff [42]. The parameter ω in this equation is regarded as a comprehensive indicator of the regional underlying surface characteristics, which is influenced by a variety of factors such as topography, soil type, and vegetation status [43]. The value of ω determines the proportional distribution between evapotranspiration and runoff, thus revealing the uniqueness of the hydrological cycle in a particular region [44]. Notably, vegetation cover is one of the key elements that influence and regulate the value of ω [45]. Vegetation influences the water vapor content of the atmosphere through transpiration activity, contributes to the accumulation of surface moisture, and regulates the water balance of the soil [46]. Yang et al. [47] showed that an increase in the ω parameter is usually associated with an increase in vegetation cover. However, in arid or semi-arid regions, an increase in vegetation cover often triggers a reduction in runoff. Therefore, a decrease in the ω parameter in a watershed usually implies a decrease in vegetation cover from the base period to the change period. Vegetation degradation results in weaker canopy retention and lower intensity of transpiration activities, and these changes in turn contribute to runoff generation.

4.2. Uncertainties and Limitations

Despite the results achieved in this study, it still faces some limitations. First, the framework adopted omits some of the key elements in the water balance in its construction. For example, the study conducted by Xu et al. [48] carefully compared the glacier changes in the basins along the southern margin of the Tarim Basin for the period of 1970–1999. The study noted that significant glacier retreat has occurred in all basins of the region, with the most significant change in glacier area occurring in the Kriya River basin. However, given the scarcity of long-term and high-quality glacier observations, this critical information was not effectively integrated with the Budyko framework when analyzing runoff changes. This constitutes a major constraint in this study, which needs to be explored and supplemented in the future.

In addition, this study ignores the potential impact of snow sublimation on runoff. Existing studies have shown that, at the watershed scale, sublimation is of non-negligible importance to the composition of the winter water balance and that its interannual variability directly influences the magnitude of spring river runoff [49]. Specific quantitative analyses have shown that the percentage of sublimation losses from winter snowfall can range from 10% to 90% at different locations [50]. At the same time, the influence of groundwater storage on runoff changes on long time scales should not be underestimated [51]. Furthermore, the interactions and coupling effects among various meteorological factors also add complexity to runoff changes [52]. The existing runoff change attribution analysis system based on Budyko’s hypothesis cannot fully cover the previous types of influencing factors, so it is an urgent need to deepen the exploration and improve it in subsequent research.

Although this study was unable to conduct a comprehensive and accurate quantitative analysis of the contribution of various factors to runoff changes, it explored an effective way to use the runoff attribution identification method based on the Budyko hypothesis in arid areas under the condition of a lack of high-quality glacier observation data. In addition, the quantitative analysis results of this study can provide a scientific basis for water resources management and environmental governance in the Kriya River Basin. Specifically, the basin needs to strengthen the monitoring and forecasting of snowmelt water and precipitation to improve the flood control and disaster reduction efficiency of the reservoir; at the same time, with the help of ecological engineering measures such as vegetation restoration and soil and water conservation, the water storage function of the underlying surface is enhanced, thereby maintaining the stability of runoff.

4.3. Limitations and Feasibility of Utilizing Annual Runoff Data for Analysis

In the absence of daily or monthly runoff data, using annual runoff data has certain limitations, which are specifically manifested as follows: (1) it covers the seasonal variation characteristics of runoff and cannot reflect the inhomogeneity of annual distribution (such as flood peak in rainy season and dry season); (2) annual data are usually based on a single site or watershed average, ignoring spatial distribution differences; (3) ignoring the dynamic impact of human activities, annual data may obscure the short-term effects of human activities such as reservoir regulation and inter-basin water transfer.

Based on the above limitations, does it mean that the use of annual data has no research value? The answer is negative. Because of the relevant research and the actual situation of this study, the feasibility of using annual runoff data in this paper is as follows: (1) many studies have used annual runoff data to analyze the overall trend of runoff from decades to centennials, which indicates that annual runoff data can be used to analyze the long-term trend and macro characteristics of river runoff [53]. (2) Annual runoff data can be employed to quantify the impacts of climatic factors (precipitation and temperature) as well as underlying surface conditions on annual runoff volume. In particular, the elasticity coefficient method based on the Budyko water–heat balance equation can be used to evaluate the sensitivity of watershed runoff to climate change [54]. (3) The annual runoff data can be used to separate the cumulative impact of natural changes and human activities on runoff, that is, for the attribution analysis of the long-term impact of human activities on rivers [55].

5. Conclusions

The Kriya River showed a significant upward trend in its runoff depth during the period of 1957–2015, with the interannual rate of change in runoff depth reaching 0.5189 mm/a. During the same period, although the annual precipitation in the basin did not show a significant change, the annual potential evapotranspiration showed a significant decrease with a significance level of p ≥ 0.01.

During the period of 1957–1999, the interference degree of human activities on annual runoff changes in the basin was relatively low, and it was at the stage dominated by natural hydrological processes; from 2000 to 2015, the intervention of human activities in the hydrological cycle of the basin increased significantly, and the influence on the annual runoff changes became more and more prominent as a key factor influencing the runoff dynamics.

Precipitation was the dominant factor driving the increase in runoff in the watershed, with a contribution rate of 75.23%, indicating that spatial and temporal variations of precipitation have a decisive influence on the runoff dynamics in the watershed. Changes in underlying surface conditions were the second driving factor for the increase in runoff, with a contribution rate of 23.08%, reflecting the fact that changes in underlying surface attributes have a significant impact on the hydrological cycle of the watershed under the combined effect of human activities and natural processes. In contrast, changes in potential evapotranspiration had a relatively limited contribution to the increase in runoff, contributing only 1.69%.

This result is of great significance to the management of water resources in arid areas, which can provide scientific support for the optimization of regional water resources strategies, and also provide reference value for the sustainable water supply of similar ecological hydrology basins.