The Climate Changes and the Simulation of the Runoff in the Last 50 years (1961–2010) in the Upper Tarim River Basin of Southern Xinjiang, China

Lu, Tianjian; Xiao, Qingquan; Ding, Leiding; Lu, Hanyu; Huang, Ziyue; Yuan, Yongyi

doi:https://doi.org/10.1155/2022/3724960

Scientific Programming

On this page

Abstract Introduction Study Area and Materials Methods Results Conclusions Data Availability Conflicts of Interest Acknowledgments References Copyright Related Articles

Special Issue

Artificial Intelligence for Evaluation Decision-making in Modern Product Design

View this Special Issue

Research Article | Open Access

Volume 2022 | Article ID 3724960 | https://doi.org/10.1155/2022/3724960

The Climate Changes and the Simulation of the Runoff in the Last 50 years (1961–2010) in the Upper Tarim River Basin of Southern Xinjiang, China

Tianjian Lu,¹Qingquan Xiao,¹Leiding Ding,¹Hanyu Lu,^1,2Ziyue Huang,¹and Yongyi Yuan³

Academic Editor: Lianhui Li

Received05 Jul 2022

Revised25 Jul 2022

Accepted29 Jul 2022

Published01 Sept 2022

Abstract

The upper Tarim River basin is supporting approximately 50 million people by melting the glaciers and snow, which are highly vulnerable and sensitive to climate change. Therefore, assessing the relative effects of climate change on the runoff of this region is essential not only for understanding the mechanism of hydrological response over the mountainous areas in Southern Xinjiang but also for local water resource management. This study quantitatively investigated the climate change in the mountainous area of the upper Tarim River basin, using the up-to-date “ground-truth” precipitation and temperature data, the Asian Precipitation Highly Resolved Observational Data Integration Towards Evaluation of Water Resources (APHRODITE, 1961–2010, 0.25°) data; analyzed the potential connections between runoff data, observed at Alar station, and the key climatological variables; and discussed the regression models on simulating the runoff based on precipitation and temperature data. The main findings of this study were as follows—(1) both annual precipitation and temperature generally increase at rates of 0.85 mm/year and 0.25 °C/10a, respectively, while the runoff data measured at the Alar station shows fluctuating decreasing trends. (2) There are significant spatial differences in the temporal trends of precipitation; for example, the larger increasing rates of precipitation occur in the Karakoram mountains, while the larger decreasing rates happen in the northwestern Kashgar county. (3) The decreasing trends of temperature mainly occur in Kashgar county and its surrounding areas in summer. (4) Seasonal correlations in precipitation and temperature trends are more significant than those on a monthly and annual scale. (5) The regression model in simulating the runoff in the upper Tarim River basin based on radial basis function (RBF) is better than that based on the least-squares method, with the predictive values based on RBF models significantly better (correlation coefficient, CC ∼ 0.85) than those by least-squares models (CC ∼ 0.75). These findings will provide valuable information to inform environmental scientists and planners on the climate change issues in the upper Tarim River basin of Southern Xinjiang, China, under a semiarid-arid climate.

1. Introduction

Global climate is changing considerably, characterized by warming over nearly 100 years, and there is a general consensus that increasing average surface air temperature had intensified global hydrological cycles during the 20th century [1–3]. According to the Intergovernmental Panel on Climate Change (IPCC) fifth assessment report [4], by the end of the twenty-first century, the global average surface temperature will increase by 3.7–4.8°C over the 1750 level, and the global average sea level will rise by 0.52–0.98 m because of CO₂ doubling. Global climate change has accelerated regional water circulation and has caused asymmetry in precipitation distribution and high flood frequencies [5–8].

Spatiotemporal changes in global precipitation and temperature have received increasing attention, and the information on trends of precipitation and temperature is the starting point for the accurate assessment of water resources, flood control, and drought relief, and the understanding of climate change and effective management of water resources. Meanwhile, natural surface runoff is also vital to maintaining surface water balance. For temporal and long-term runoff, the physical process of runoff yield always has a close relationship with climate variables, such as precipitation and temperature [9–11]. In arid and semiarid regions of vulnerable ecology, a small climate fluctuation may cause large environmental variation when human activity overwhelms the natural carrying capacity [12]. For this reason, addressing the impact of regional climate change on runoff volume will support scientific and technological sustainable development of local water resources.

Xinjiang Province, located in Northwest China, is under an inland, arid, or semiarid desert climate, which is not directly affected by the monsoon system [13, 14]. The Tarim basin in Southern Xinjiang Province is a typical inland watershed in an arid area, and the hydrologic processes of the upper Tarim River basin are typical among those in other regions at mid- and high latitudes of the Northern Hemisphere [11]. Various studies have focused on the impacts of climate change on water resources in the Tarim River basin [15, 16]. They concluded that the temperature and precipitation show an upward tendency during the past several decades and a significant jump has been detected for both the two variables around 1986 [10, 17]. Although the streamflow from the headwaters of the Tarim River shows a significant increase and is sensitive to precipitation [18], the streamflow along the mainstream of the river has decreased. This implies that anthropogenic activities such as irrigation and increased population instead of climate change dominated the streamflow change of the river [19]. Recently, various reports have shown a widespread climatic and hydrologic change in the Tian Shan Mountains during the past few decades [20]. For example, temperature demonstrated a significant rising trend (significant level is smaller than 0.001) at a rate of 0.33∼0.34°/10a during 1960∼2010, which is higher than those of China (0.25°/decade) and the entire globe (0.13°/10a) [21]; precipitation increased substantially in most regions especially for the middle and high latitudes, at a rate of 0.61 mm/a [22]; glacier area decreased by 11.5%, and the thickness of snowpack has also decreased [23]. The climate in Northwest China changed dramatically from a warm-dry mode to a warm-wet mode around 1987, and the Chinese Tian Shan Mountains experienced the most dramatic changes during this transition [20].

However, the upper Tarim River basin is relatively unknown in terms of recent climate changes simply due to a lack of meteorological observations in high-altitude areas (Wang et al., [24]. Fortunately, the state-of-the-art dataset (updated in September 2018), Asian Precipitation Highly Resolved Observational Data Integration Towards Evaluation of Water Resources (APHRODITE, 0.25°/daily [25]), provided great valuable precipitation and air temperature data in the Asian over the last half century, from 1951 to 2015. And APHRODITE has been demonstrated to replicate “ground-truth” observations very well [26] and represents the best tool for analyzing historical precipitation variability and change. Therefore, this study aims at revealing the in-depth climate change principles in the upper Tarim River basin and determining the response of runoff to climate change in the region through the analyses of precipitation and temperature.

2. Study Area and Materials

2.1. Study Area

Xinjiang Province of Northwest China accounts for one-sixth of China’s land area and has an inland, arid, or semiarid desert climate that is not directly affected by the monsoon system. Local precipitation of a large spatiotemporal variation is concentrated in mountainous areas [15]. The Tarim River basin in Southern Xinjiang is one of the world’s largest closed drainage hydrographical systems without outflow. The basin is composed of 114 streams belonging to nine river systems: the Aksu River, Kashgar River, Yarkand River, Hotan River, Kaidu River, Dina River, Weigan River, Kuqa River, and Keriya River. The landforms of the Tarim River basin include mountains (47%), plains (22%), and deserts (31%) [27]. The Tarim River lies entirely within a landlocked area and has a mainstream length of 1,321 km. Water has been imported to the mainstream from nine systems throughout history, and the main causes of this import were climate change and human economic activities. The Qarqan, Keriya, and Dina rivers successively lost surface connections to the mainstream prior to the 1940s, as did the Kashgar, Kongqi, and Weigan thereafter. As a result, only the Yarkant River, Akesu River, Hotan River, and Kaidu River have links with the mainstream of the Tarim River. Three tributary river systems (Aksu, Hotan, and Yarkant River) contributing to the Tarim River converge just above the Alar gauging station, while the Kaidu River flows into the Tarim River at the lower reaches.

The study area (73° E–82° E, 35° N–43° N) is the upper mountain areas of the Tarim River basin (Figure 1(a)). The region is surrounded by high mountains like Tianshan, Eastern Pamir, and Karakoram mountains, which leads to orographic precipitation. Based on the Asian Precipitation Highly Resolved Observational Data Integration Towards Evaluation of Water Resources, the precipitation in the mountainous regions can exceed 300 mm/year in some areas and is mostly in the form of snowfall (Figure 1(b)). However, the average annual precipitation in the region is below 100 mm. The average annual temperatures can range from −12°C to 16°C (Figure 1(c)). The mainstream of the Tarim River is a typical pure dissipation inland river that does not yield water resources by itself and is supplied only by runoff from its upper basin [11, 14, 28]. This runoff is primarily from glacial meltwater and precipitation in mountainous areas. Therefore, hydrologic processes of the mountainous areas are typical of those in other regions at mid- and high latitudes of the Northern Hemisphere [11]. With the backdrop of global climate change, a detailed investigation on changing spatial and temporal features of precipitation will not only help to understand the relationship between climate change and the hydrological cycle, but also be helpful in formulating a regional strategy for water resource management in Southern Xinjiang.

2.2. Data

This study employed the new state-of-the-art daily dataset with a resolution of 0.25° (updated in September 2018), Asian Precipitation Highly Resolved Observational Data Integration Towards Evaluation of Water Resources [25, 29, 30], which provided great valuable precipitation and temperature data in the Asian over the last half century, from 1961 to 2010, to reveal the in-depth climate change principles in Southern Xinjiang. The APHRODITE project develops state-of-the-art daily precipitation datasets and daily mean temperature datasets with local meteorological/hydrological agencies for Asia. In terms of precipitation, the APHRODITE's Water Resources project has been executed by the Research Institute for Humanity and Nature (RIHN) and the Meteorological Research Institute of Japan Meteorological Agency (MRI/JMA) since 2006. The datasets are created primarily with data obtained from in situ rain-gauge-observation networks [31].

APHRODITE’s daily gridded precipitation was presently the only long-term continental-scale high-resolution daily product and has been demonstrated to replicate “ground-truth” observations very well [26], which was the best tool for studies such as the diagnosis of climate changes, evaluation of Asian water resources, and satellite precipitation estimates. And the APHRODITE data are available online at http://aphrodite.st.hirosaki-u.ac.jp/download/with high-resolution (0.5° and 0.25°) grids for Asia. Furthermore, measured annual runoff data were furnished by the administration of the Tarim River basin and were collected from Alar stations during 1961–2010. And Alar can be seen as the entrance to the Tarim River. Therefore, the runoff measured at the Alar station was considered to be the runoff of the Tarim River in this article.

3. Methods

3.1. Regression Analysis

Rational allocation of water resources is the foundation for the rational allocation of water rights, while the control of runoff prediction in the hydrological site is a prerequisite for the rational allocation of water resources. Therefore, the mathematical model developed to predict runoff timely and accurately for the right to water management is an extremely important basic job. Runoff prediction is one of the problems in the natural sciences and technology field, and its difficulty is that hydrological changes are subject to various uncertainties and not entirely clear to operation rules. And in mountainous catchments, the quality of runoff modeling depends strongly on the amount and intensity of precipitation and the snow melting.

At present, there are many methods for predicting river runoff such as neural networks, wavelet analysis, and support vector machines. One of the commonly used methods is to calculate the runoff data using statistical methods, and the other is to set up a prediction model according to the evolution of the predicted subject. We used statistical methods including linear regression and nonlinear regression, based on the observed data, to predict and analyze the runoff situations [32–38]. Speaking specifically, based on the monthly runoff data of the upper mountain areas of the Tarim River basin, from the Alar station, during the period of 1961–2010, and precipitation and temperature data in mountainous areas, the influence of the temporal-spatial component of regional precipitation and temperature on monthly runoff was analyzed by using least-squares methods and radial basis function (RBF) neural network.

3.2. Quantitative Error Indicator

In this study, the regression models were established based on the precipitation and temperature data, according to least-squares methods and RBF neural network, and the runoff data were predicted and analyzed by those models. To assess the accuracy of the predictive value of the regression models, four indicators were used in this study: correlation coefficient , bias, root mean square error (RMSE), and mean absolute error (MAE) [39]. Among them, is used to measure the correlation between satellite precipitation data and rain gauge data with the value ranging from 0 to 1. Bias evaluates the degree of bias of satellite precipitation data against the rain gauge data. RMSE and MAE are used to evaluate the overall level of satellite precipitation data error within the range of [0, +∞]. The best possible values are CC = 1, bias = 0, RMSE = 0, and MAE = 0. The equations for the above-mentioned statistics are shown as follows:where represents the amount of runoff observed at the Alar station, represents the average observed runoff, and are the estimated values of the regression models and their average values, respectively, and n represents the number of pairs of the two runoff factors in the analysis.

4. Results

4.1. Analysis of Climate Change from 1961 to 2010 in the Upper Tarim River Basin

Generally, a linear regression analysis of precipitation and temperature data from 1961 to 2010 found that precipitation and temperature demonstrated significant increasing trends in the upper Tarim River basin at a rate of 0.85 mm/year and 0.25°C/10a, respectively (significant level is smaller than 0.001; Figure 2(a) and Figure 2(b)). Figure 2(c) shows the temporal patterns of the runoff observed at the Alar station, demonstrating the streamflow along the mainstream of the upper Tarim River basin fluctuating in decreasing trends. In terms of precipitation, it showed a continuously increasing trend, while there were still several years (e.g., 1975, 1985, and 1997) with the smallest precipitation volume around 70.00 mm/year, and the largest precipitation occurred in 2010 with the volume around 210.00 mm/year. In terms of temperature, there was a clear horizontal trend at the beginning from 1961 to 1995, while in the afterward period from 1996 to 2010, the temperature demonstrated a more significant increasing trend at a rate of 0.51°C/10a, with the lowest and highest annual temperature happened in 1974 around 4.00°C and in 2006 around 6.46°C, respectively. In the period from 1961 to 1995, the annual variation in precipitation and temperature demonstrated negative correlations (CC ∼ -0.24 and P = 0.168). And the negative correlations (CC ∼ -0.40 and ) were more significant in the period from 1996 to 2010. In addition, the Tarim River runoff showed a continuous decreasing trend, while there were still several years (e.g., 1978, 1995, and 2010) with the largest runoff volume around 220.00 m³/s. In the period from 1996 to 2010, the annual variation in runoff and temperature demonstrated significant correlations (CC ∼ 0.59 and ) and the CC value of runoff and precipitation is around −0.27.

(a)

(b)

(c)

The Aksu River, Hotan River, and Yarkand River are all dominated by snowmelt and precipitation replenishment in the mountain areas. Therefore, changes in precipitation and temperature in the river basin have a direct impact on changes in river runoff. However, the increasing trends of precipitation and temperature in the upper Tarim River basin over the last 50 years conflict with the decreasing trend of runoff, indicating that the Tarim River runoff is affected not only by climate change, but also by human activities. Since the 1960s, land reclamation in the upper Tarim River basin has never stopped. The area of an artificial oasis has been continuously expanded, and the area of irrigation has more than doubled. It can be said that human activities are the dominant factor that causes the Tarim River runoff to decrease.

On a seasonal scale, both precipitation and temperature showed increasing trends in all seasons (Figure 3(a) and Figure 3(b)). As for precipitation, the largest and smallest increasing rates happened in summer (JJA; June, July, and August, 0.32 mm/year) and spring (MAM; March, April, and May, 0.16 mm/year), respectively. As for temperature, adversely, the largest and smallest increasing rates happened in winter (DJF; December, January, and February) and spring, respectively. The climate changed with the temperature in winter at a rate of 0.36°C/10a in the upper Tarim River basin, which may bring a great threat to the glacier’s accumulations. Meanwhile, the runoff showed decreasing trends in all seasons except spring (Figure 2(c)). The possible reasons are the increase in the arable land area and the continuous expansion of agricultural irrigation areas in the upper Tarim River basin. Although the Tarim River flood season mainly occurs in summer, which is consistent with the period of concentrated water use for agriculture in the basin, a large amount of agricultural water not only occupies the river runoff in summer, but also leads to excessive groundwater extraction. Even in the autumn and winter seasons, even if agricultural water is reduced, the infiltration of precipitation in autumn and winter directly replenishes the groundwater that is lacking due to severe summer mining. As a result, the surface runoff in the Tarim River basin is still reduced in the autumn and winter seasons when it is not the peak period of agricultural water use.

(a)

(b)

(c)

Table 1 lists the change rates of monthly mean precipitation, temperature, and runoff. Precipitation increased with the largest trend occurring in June at a rate of approximately 0.16 mm/year and demonstrated large increasing trends in July and September with rates larger than 0.10 mm/year, while the smallest increasing trend happened in April with the rate of 0.02 mm/year. In terms of temperature, the largest increasing trend happened not in the winter months but in November at a rate of 0.47°C/10a, which was larger than those of the rest months, and the smallest increasing trend occurred in July at a rate of 0.06°C/10a. Generally speaking, the increasing rates of temperature in cold seasons were larger than those in warm seasons. In addition, the runoff of the upper Tarim River at a monthly scale decreased with the most significant trend occurring in July at a rate of approximately −1.81 m³/s, and it also showed large decreasing trends in January, August, November, and December with rates smaller than 1.00 m³/s, while the largest increasing trend happened in May with the rate of 1.01 m³/s. Besides, positive trends were observed in April, September, and October.

4.2. Spatial Distributions of the Climate Change and Runoff Trends from 1960 to 2010 in the Upper Tarim River Basin

Figure 4 shows the spatial distributions of precipitation trends, at annual and seasonal scales over the upper Tarim River basin from 1961 to 2010. Overall, the annual precipitation trends generally presented an increasing trend from northwest to south, and the differences in precipitation trends in spatial distributions among the four seasons were significant. The regions with the largest precipitation increasing rates were in the southern region, the Karakoram mountains, in intervals larger than 2.50 mm/year. Although most regions showed increasing trends, the decreasing trends occurred in the northwest edges of the upper basin, with the rates varying from −0.60 mm/year to −0.30 mm/year (Figure 4(a)). On the seasonal scale, the decreasing trends were detected in most regions in spring and winter (Figures 4(b) and 4(d)), and the increasing trends were detected in almost all study areas in summer and autumn (Figures 4(c) and 4(e)). Moreover, the increase in precipitation trends in the Karakoram mountains at all seasons was more significant than that in the other regions.

From 1961 to 2010, the annual average temperature in the upper Tarim River basin showed an overall increasing trend, with the increasing rates varying from 0.00°C/10a to 0.30°C/10a in most areas. And the most significant increasing trends were detected in the southeastern region, with the maximum rate around 0.60°C/10a (Figure 5(a)). On the seasonal scale, the spatial pattern of temperature trends in spring was similar to that at an annual scale (Figure 5(b)). And the temperature in the south, in autumn and winter, showed a clear upward trend (Figures 5(d) and 5(e)). In addition, the temperature in summer dropped slightly in some northern areas (Figure 5(c)).

The correlation coefficients between precipitation and temperature varied both spatially and temporally. The precipitation and temperature demonstrated higher correlations at the seasonal scale in most regions with the values of CC varying from 0.5 to 0.9, while the correlations on the monthly scale were generally between 0.3 and 0.7. Meanwhile, the spatial distribution of correlations generally presented an increasing trend from south to north (Figure 6(b) and Figure 6(c)). However, on an annual scale, the relationship between the precipitation and temperature was negative in most areas especially in northwestern regions varying from −0.5 to −0.3, while in the east Karakoram mountains, the correlations were generally fluctuating between 0.3 and 0.5 (Figure 6(a)).

4.3. Analysis of Simulating Runoff Based on Regression Models and Deep Learning Models from 1961 to 2010 over the Upper Tarim River Basin

The spatial distributions of the 50-year average monthly precipitation and temperature over the upper Tarim River basin and the runoff volume observed at Alar are shown in Figure 7. The results show that the runoff of the Tarim River was mainly concentrated in July and August. Considering the precipitation and temperature, the runoff was dominated by snow accumulation in winter, subsequent snow melting in spring, and high precipitation in summer.

(a)

(b)

(c)

Meanwhile, Figure 8 shows the spatial patterns of correlations between monthly precipitation and runoff and those between monthly temperature and runoff for the period 1961–2010. In terms of precipitation, Figure 8(a) shows the correlation coefficients of precipitation and runoff at the corresponding time and the Figure 8(b) shows the correlation coefficients between precipitation and runoff, with one-month delay to the precipitation. The differences in coefficients in spatial distributions were significant, and the coefficients generally presented an increasing trend from south to north. It can be seen that the impact of precipitation, which occurred in the Tian Shan Mountains, had a more significant impact on the runoff than that occurred in the Karakoram mountains. However, the values of correlations shown in Figure 8(b) were larger than those shown in Figure 8(a) demonstrating that the precipitation might have a great potential impact on the runoff with one-month delay at Alar. In terms of temperature, the runoff-temperature relationship exhibited an overall positive correlation. Similar to the relationship between precipitation and runoff, the temperature had larger correlations with runoff with a one-month delay than those of temperature in the corresponding months.

In the present study, another focus was on comparing the least-squares method models with the deep learning models, radial basis function (RBF) neural network models. A training period of 1961–2000 and an evaluation period of 1961–2000 were identified. We regarded the runoff data as the dependent variable, and the precipitation, temperature, and runoff in the current or previous month as the independent variables. And Table 2 lists a total of 12 regression models based on different independent variables and methods. The weighted data were the correlation coefficient of precipitation or temperature and runoff at each grid point as shown in Figure 8.

While comparing the results, it is to be kept in mind that the data used for calibration and verification are the same among different models. The data from 1961 to 2000 were used to build models and validated using data from 2001 to 2010. Figure 9 demonstrated the temporal patterns of monthly runoff at Alar, in the period from 2001 to 2010, based on observations and different prediction models. The least-squares method models significantly overestimated the runoff in winter and spring, and underestimated the runoff in summer especially in August. In terms of accuracy and consistency, the overall performances of the models based on RBF were better than those of the least-squares method models.

(a)

(b)

Based on different models, the four quantitative error indicators (CC, bias, RMSE, and MAE) of predictions against the runoff observations were also analyzed (Table 3). In terms of least-squares method models, the models using weighted independent variables outperformed the models using unweighted independent variables. In addition, an increase in the number of independent variables improved the accuracy of the models. The CC (∼0.75), bias (∼-17.15%), RMSE (∼136.40 m³/s), and MAE (∼89.33 m³/s) of the PreRunW-LS model generally performed better than the other models. In terms of RBF models, although the CC value of PreW-RBF (0.84) is less than the CC of Pre-RBF (0.85), the weighted independent variables still improved the models. Unlike the least-squares method, preceding runoff data as one of the independent variables did not significantly improve the quality of the models in terms of quantitative error indicators. Generally speaking, the PreW-RBF model (CC ∼0.84) outperformed the PreRunW-LS model (CC ∼ 0.75). Although the bias value of the PreW-RBF model (∼30%) was around two times larger than that of the PreRunW-LS model (∼17%), and both the RMSE and MAE values of PreW-RBF were significantly smaller than those of the PreRunW-LS, respectively.

5. Discussions

The Tarim River played a crucial role in the economic and social development of Xinjiang, and the accurate prediction of long-term runoff variation in this basin was important for watershed and flood management. However, frequent floods and droughts in the Tarim River basin were caused by complex forcing factors, so it was challenging to predict when runoff disasters would happen. Many current studies focused on setting up prediction models according to the evolution of the predicted subject. Statistical methods including simple linear or nonlinear methods, in this article, were also considered; it is of great significance for the prediction of runoff in the Tarim River. And the purpose of this study was to be able to quickly predict monthly runoff based on precipitation and temperature data.

Figure 10 demonstrates the temporal patterns of the statistical accuracy indicators evaluating the least-squares method models and RBF models against observed runoff data at the monthly scale. And PreRunW-LS and PreW-RBF models represented the best of the linear and nonlinear models, respectively. The temporal patterns of the CC of least-squares method models and RBF models were different significantly (Figures 10(b) and 10(b)), while the CC values of PreRunW-LS (∼-0.20 − ∼0.0.5) were significantly larger than those of PreW-RBF (∼-0.50 − ∼0.50) in winter and spring, especially in February. In terms of bias, PreRunW-LS overestimated the runoff in January and February (as high as ∼300%), and PreW-RBF overestimated the runoff in April, May, and October (up to ∼350%) against the observed data. However, the bias values of PreW-RBF models were closer to 0.0, in January, February, July, August, and November (Figure 10(d)). As for indicators of RMSE and MAE, the temporal patterns of PreRunW-LS and PreW-RBF were very similar. The RMSE values of PreRunW-LS (∼20 m³/s − ∼350 m³/s) were much larger than those of PreW-RBF (∼10 m³/s − ∼200 m³/s), especially in August (Figure 10(f)); similarly, the MAE values of PreW-RBF (∼10 m³/s − ∼150 m³/s) were also much smaller than those of PreRunW-LS (∼20 m³/s − ∼260 m³/s). In general, neither PreRunW-LS nor PreW-RBF performed better in summer and autumn than in spring and winter. The possible reason might be that in the upper mountainous basin, when the temperature was much lower than the snow melting temperature, the simple linear or nonlinear models were difficult to fit the relationship between temperature and runoff.

(a)

(b)

(c)

(d)

(e)

(f)

(g)

(h)

Least-squares models and RBF models used a 40-year training period (1961–2000), and forecasts were conducted for the 10-year validation period from 2001 to 2010. The RBF models exhibited promising predictive skills with the correlation coefficient between the observation and the predicted values close to 0.85, which was better than those of the least-squares models. The results have shown that the PreW-RBF model built with the RBF neural networks had some significant advantages for the long-term prediction of runoff and might provide a new and effective tool for the prediction of drought and flood events in the Tarim River basin.

It should be mentioned that the variability of the Tarim catchment hydrology is not only closely linked with the large-scale atmospheric circulation over Asia but also with headstream snow cover because the Tarim River was fed by the glaciers of the mountainous areas. Thus, in addition to the precipitation and temperature data, snow cover or other anomalies may also be an important independent variable for predicting runoff. And we may conduct a further study in the future to analyze the importance of snow cover to the simulation of the runoff.

6. Conclusions

This study quantitatively investigates the climate change in the upper Tarim River basin, using the up-to-date “ground-truth” precipitation and temperature data, the Asian Precipitation Highly Resolved Observational Data Integration Towards Evaluation of Water Resources (APHRODITE, 1961–2010, 0.25°) data; analyses the potential connections between runoff data, observed at Alar station and the key climatological variables; and discusses the models in simulating the runoff based on precipitation and temperature data. The main findings of this study are as follows:(1)Both annual precipitation and temperature generally increased at rates of 0.85 mm/year and 0.25 °C/10a, respectively, while the runoff data measured at Alar station showed fluctuating decreasing trends.(2)There were significant spatial differences in the temporal trends of precipitation, for example, the larger increasing rates of precipitation occurred in the Karakoram mountains, while the larger decreasing rates happened in northwestern Kashgar county.(3)The decreasing trends of temperature mainly occurred in the Kashgar county and its surrounding areas in summer.(4)The seasonal correlations in trends of precipitation and temperature were more significant than those on a monthly and annual scale.(5)The regression model in simulating the runoff in the upper Tarim River basin based on RBF was better than that based on the least-squares method, with the predictive values based on the RBF models significantly better (correlation coefficient, CC, ∼ 0.85) than those by least-squares models (CC ∼ 0.75).

These findings would provide valuable information to environmental scientists and planners on the climate change issues over the Tarim River basin. [40] [41].

Data Availability

All the data used in this study can be obtained from the corresponding author upon request.

Conflicts of Interest

The authors declare that they have no conflicts of interest.

Acknowledgments

This study was financially supported by the National Natural Science Foundation of China (Grant no. 41901343) and Natural Science Foundation of Guizhou Province (Grant no. [2020]4001) The authors appreciate the contribution of the data providers, including the Chinese Meteorological Data Sharing Service System (http://cdc.nmic.cn/home.do) and the APHRODITE data provider (HYPERLINK “http://aphrodite.st.hirosaki-u.ac.jp/download/“ \o “http://aphrodite.st.hirosaki-u.ac.jp/download/“http://aphrodite.st.hirosaki-u.ac.jp/download/).

References

J. D. Milliman, K. L. Farnsworth, P. D. Jones, K. Xu, and L. Smith, “Climatic and anthropogenic factors affecting river discharge to the global ocean, 1951–2000,” Global and Planetary Change, vol. 62, no. 3-4, pp. 187–194, 2008.
View at: Publisher Site | Google Scholar
S. J. Déry, M. A. H. Hernández, J. E. Burford, and E. F. Wood, “Observational evidence of an intensifying hydrological cycle in northern Canada,” Geophysical Research Letters, vol. 36, no. 13, p. L13402, 2009.
View at: Publisher Site | Google Scholar
T. G. Huntington, “Evidence for intensification of the global water cycle: r,” Journal of Hydrology, vol. 319, no. 1-4, pp. 83–95, 2006.
View at: Publisher Site | Google Scholar
Edenhofer and Ottmar, “IPCC, 2014: climate change 2014: mitigation of climate change,” Contribution of Working Group III to the Fifth Assessment Report of the Intergovernmental Panel on Climate Change, Cambridge University Press, Cambridge, New York, NY, USA, p. 72, 2014.
View at: Google Scholar
D. Labat, Y. Godderis, J. L. Probst, and J. L. Guyot, “Evidence for global runoff increase related to climate warming,” Advances in Water Resources, vol. 27, no. 6, pp. 631–642, 2004.
View at: Publisher Site | Google Scholar
L. Menzel and G. Bürger, “Climate change scenarios and runoff response in the Mulde catchment (Southern Elbe, Germany),” Journal of Hydrology, vol. 267, no. 1-2, pp. 53–64, 2002.
View at: Publisher Site | Google Scholar
V. Semenov and L. Bengtsson, “Secular trends in daily precipitation characteristics: greenhouse gas simulation with a coupled AOGCM,” Climate Dynamics, vol. 19, no. 2, pp. 123–140, 2002.
View at: Publisher Site | Google Scholar
Q. Zhang, Y. David Chen, and J. Chen, “Flood/drought variability in the Yangtze Delta and association with the climatic changes from the Guliya ice core: a wavelet approach,” Quaternary International, vol. 189, no. 1, pp. 163–172, 2008.
View at: Publisher Site | Google Scholar
H. Karambiri, S. G. G. Galiano, and J. D. Giraldo, “Assessing the impact of climate variability and climate change on runoff in West Africa: the case of Senegal and Nakambe River basins,” Atmospheric ence Letters, vol. 12, no. 1, pp. 109–115, 2015.
View at: Google Scholar
C. Xu, Y. Chen, W. Li, and Y. Chen, “Climate change and hydrologic process response in the Tarim River Basin over the past 50 years,” Chinese Science Bulletin, vol. 51, no. S1, pp. 25–36, 2006.
View at: Publisher Site | Google Scholar
H. Xu, B. Zhou, and Y. Song, “Impacts of climate change on headstream runoff in the Tarim River Basin,” Hydrology Research, vol. 42, no. 1, pp. 20–29, 2011.
View at: Publisher Site | Google Scholar
T. Y. Gan, M. Ito, S. Huelsmann et al., “Possible climate change/variability and human impacts, vulnerability of drought-prone regions, water resources and capacity building for Africa,” International Association of entific Hydrology Bulletin, vol. 61, no. 5-8, pp. 1209–1226, 2016.
View at: Google Scholar
S. Reinhard, L. Daniel, V. Pier Luigi, and S. Christoph, “The precipitation climate of Central Asia—intercomparison of observational and numerical data sources in a remote semiarid region,” International Journal of Climatology, vol. 28, no. 3, 2010.
View at: Google Scholar
R. Wang and X. Lu, “Quantitative estimation models and their application of ecological water use at a basin scale,” Water Resources Management, vol. 23, no. 7, pp. 1351–1365, 2009.
View at: Publisher Site | Google Scholar
X. Jianhua, L. Weihong, M. Ji, F. Lu, and S. Dong, “A comprehensive approach to characterization of the nonlinearity of runoff in the headwaters of the Tarim River, western China,” Hydrological Processes, vol. 24, 2010.
View at: Google Scholar
Z. X. Xu, Y. N. Chen, and J. Y. Li, “Impact of climate change on water resources in the Tarim river basin,” Water Resources Management, vol. 18, no. 5, pp. 439–458, 2004.
View at: Publisher Site | Google Scholar
Y. N. Chen, L. I. Hong, X. U. Chun, and X. M. Hao, “Effects of climate change on water resources in Tarim river basin, northwest China,” Journal of Environmental Sciences, vol. 19, no. 04, pp. 106–111, 2007.
View at: Google Scholar
L. Chen, C. Liu, and H. Fanghua, “Change of the baseflow and it’s impacting factors in the source regions of Yellow River,” Journal of Glaciology and Geocryology, vol. 28, no. 2, pp. 141–148, 2006.
View at: Google Scholar
X. Hao, Y. Chen, C. Xu, and W. Li, “Impacts of climate change and human activities on the surface runoff in the Tarim river basin over the last fifty years,” Water Resources Management, vol. 22, no. 9, pp. 1159–1171, 2008.
View at: Publisher Site | Google Scholar
Y. Shi, Y. Shen, E. Kang et al., “Recent and future climate change in northwest China,” Climatic Change, vol. 80, no. 3-4, pp. 379–393, 2007.
View at: Publisher Site | Google Scholar
B. Li, Y. Chen, and X. Shi, “Why does the temperature rise faster in the arid region of northwest China,” Journal of Geophysical Research: Atmospheres, vol. 117, 2012.
View at: Google Scholar
B. Li, Y. Chen, Z. Chen, H. Xiong, and L. Lian, “Why does precipitation in northwest China show a signiﬁcant increasing trend from 1960 to 2010,” Atmospheric Research, vol. 167, pp. 275–284, 2016.
View at: Publisher Site | Google Scholar
S. Wang, M. Zhang, Z. Li et al., “Glacier area variation and climate change in the Chinese Tianshan Mountains since 1960,” Journal of Geographical Sciences, vol. 21, no. 2, pp. 263–273, 2011.
View at: Publisher Site | Google Scholar
Y. Zhong, B. Wang, C. B. Zou, B. X. Hu, Y. Liu, and Y. Hao, “On the teleconnection patterns to precipitation in the eastern Tianshan Mountains, China,” Climate Dynamics, vol. 49, no. 9-10, pp. 3123–3139, 2017.
View at: Publisher Site | Google Scholar
A. Yatagai, K. Kamiguchi, O. Arakawa, A. Hamada, N. Yasutomi, and A. Kitoh, “APHRODITE: constructing a long-term daily gridded precipitation dataset for Asia based on a dense network of rain gauges,” Bulletin of the American Meteorological Society, vol. 93, no. 9, pp. 1401–1415, 2012.
View at: Publisher Site | Google Scholar
J. M. A. Duncan and E. M. Biggs, “Assessing the accuracy and applied use of satellite-derived precipitation estimates over Nepal,” Applied Geography, vol. 34, pp. 626–638, 2012.
View at: Publisher Site | Google Scholar
F. Xie, W. Mao, J. Zhang et al., “Analysis of streamflow from four source rivers to mainstream of the Tarim River,” Glaciol. Geocryol, vol. 29, pp. 559–569, 2007.
View at: Google Scholar
Y. Chen, K. Takeuchi, C. Xu, Y. Chen, and Z. Xu, “Regional climate change and its effects on river runoff in the Tarim Basin, China,” Hydrological Processes, vol. 20, no. 10, pp. 2207–2216, 2006.
View at: Publisher Site | Google Scholar
Y. Ma, G. Tang, D. Long et al., “Similarity and error intercomparison of the GPM and its predecessor-TRMM multisatellite precipitation analysis using the best available hourly gauge network over the Tibetan plateau,” Remote Sensing, vol. 8, no. 7, p. 569, 2016.
View at: Publisher Site | Google Scholar
Z. Ma, Y. Xu, J. Peng et al., “Spatial and temporal precipitation patterns characterized by TRMM TMPA over the Qinghai-Tibetan plateau and surroundings,” International Journal of Remote Sensing, vol. 39, no. 12, pp. 3891–3907, 2018.
View at: Publisher Site | Google Scholar
Z. Ma, L. Zhou, W. Yu, Y. Yang, H. Teng, and Z. Shi, “Improving TMPA 3B43 V7 data sets using land-surface characteristics and ground observations on the qinghai-tibet plateau,” IEEE Geoence & Remote Sensing Letters, vol. 15, no. 99, pp. 1–5, 2018.
View at: Google Scholar
L. Li, C. Mao, H. Sun, Y. Yuan, and B. Lei, “Digital twin driven green performance evaluation methodology of intelligent manufacturing: hybrid model based on fuzzy rough-sets AHP, multistage weight synthesis, and PROMETHEE II,” Complexity, vol. 2020, no. 6, Article ID 3853925, pp. 1–24, 2020.
View at: Publisher Site | Google Scholar
H. J. Lin, L. W. Chou, K. M. Chang, J. F. Wang, S. H. Chen, and R. Hendradi, “Visual fatigue estimation by eye tracker with regression analysis,” Journal of Sensors, vol. 2022, Article ID 7642777, pp. 1–7, 2022.
View at: Publisher Site | Google Scholar
L. Li, B. Lei, and C. Mao, “Digital twin in smart manufacturing,” Journal of Industrial Information Integration, vol. 26, no. 9, Article ID 100289, 2022.
View at: Publisher Site | Google Scholar
L. Li, T. Qu, Y. Liu et al., “Sustainability assessment of intelligent manufacturing supported by digital twin,” IEEE Access, vol. 8, pp. 174988–175008, 2020.
View at: Publisher Site | Google Scholar
M. K. Devi, V. P. Vemuri, M. Arumugam et al., “Design and implementation of advanced machine learning management and its impact on better healthcare services: a multiple regression analysis approach (mraa),” Computational and Mathematical Methods in Medicine, vol. 2022, Article ID 2489116, pp. 1–7, 2022.
View at: Publisher Site | Google Scholar
L. Li and C. Mao, “Big data supported PSS evaluation decision in service-oriented manufacturing,” IEEE Access, vol. 8, pp. 154663–154670, 2020.
View at: Publisher Site | Google Scholar
S. Zhou and C. Kang, “Research on the influencing factors of Russian foreign trade based on R language regression analysis,” Mathematical Problems in Engineering, vol. 2021, Article ID 5638831, 11 pages, 2021.
View at: Publisher Site | Google Scholar
Z. Ma, Z. Shi, Y. Zhou, J. Xu, W. Yu, and Y. Yang, “A spatial data mining algorithm for downscaling tmpa 3b43 v7 data over the qinghai-tibet plateau with the effects of systematic anomalies removed,” Remote Sensing of Environment, vol. 200, 2017.
View at: Publisher Site | Google Scholar
A. Rutkowska, P. Willems, C. Onyutha, and W. Mlocek, “Temporal and spatial variability of extreme river flow quantiles in the Upper Vistula River basin, Poland,” Hydrological Processes, vol. 31, no. 7, pp. 1510–1526, 2017.
View at: Publisher Site | Google Scholar
E. Łupikasza, T. Niedźwiedź, and I. Pinskwar, “Observed changes in AirTemperature and precipitation and relationship between them,” The Upper Vistula Basin, Springer International Publishing, Berlin, Germany, 2016.
View at: Google Scholar

Copyright

Copyright © 2022 Tianjian Lu et al. This is an open access article distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

PDF Download Citation

Download other formats

Order printed copies