Abstract

This investigation examined effects of climate change, measured as annual, seasonal, and monthly air temperature and precipitation from 1958 to 2010, on water resources (i.e., runoff) in the Bosten Lake Basin. Additionally, teleconnections of hydrological changes to large-scale circulation indices including El Nino Southern Oscillation (ENSO), Arctic Oscillation (AO), North Atlantic Oscillation (NAO), Tibetan High (XZH), westerly circulation index (WI), and northern hemisphere polar vortex area index (VPA) were analyzed in our study. The results showed the following. (1) Annual and seasonal air temperature increased significantly in the Bosten Lake Basin. Precipitation exhibited an increasing trend, while the significance was less than that of temperature. Abrupt changes were observed in 1996 in mountain temperature and in 1985 in plain temperature. (2) Runoff varied in three stages, decreasing before 1986, increasing from 1987 to 2003, and decreasing after 2003. (3) Precipitation and air temperature have significant impacts on runoff. The hydrological processes in the Bosten Lake Basin were (statistically) significantly affected by the northern hemisphere polar vortex area index (VPA) and the Tibetan High (XZH). The results of this study are good indicators of local climate change, which can enhance human mitigation of climate warming in the Bosten Lake Basin.

INTRODUCTION

The IPCC Fifth Assessment Report (IPCC5) indicated that the global average surface temperature increased by 0.11 ± 0.03°C/10a from 1951 to 2012, which was almost two times the warming rate from 1880 to 1950 (Flato et al. 2013). The projections of IPCC5 also forecast that the global average surface temperature will rise by 0.3–0.7°C from 2016 to 2035 compared with 1986–2005 (Qin & Stocker 2014). Warming climate intensifies the hydrological cycle because of higher rates of evaporation and greater extremes of precipitation, which will alter river systems (Chiew & McMahon 1993; Wang et al. 2013; Menéndez et al. 2016). Mountain glaciers have continued to retreat in recent years (Richman 1986) and are likely to significantly influence the global water supply. Therefore, estimating the possible effects of climatic change on water resources is particularly important for long-term planning of water resources, especially in arid regions.

Bosten Lake Basin is one of the most populous inland river basins in China and has the highest degree of water resource utilization; it also has serious environmental problems. Changes in hydrological processes, important signals of climate transition in the arid region of northwestern China, have stimulated climate research in these regions (Shi et al. 2007; Shen et al. 2010). In the past 60 years, rainfall and runoff in the Bosten Lake Basin generally increased, while after 2002, rainfall and runoff sharply decreased (Zhang et al. 2007; Li et al. 2011). Because runoff variation increased, exploitation of groundwater and surface water has accelerated. Water resources can be managed more sustainably by utilizing data from comprehensive analyses of hydrological characteristics of the region.

In the context of global warming, there was substantial fluctuation of the water cycle in Tianshan Mountains, leading to increased runoff instability (Liu et al. 2006; Dou et al. 2011). Underlying causes for hydrological changes must be understood to realize sustainable water resources development. Teleconnection models may account for the temporal variability and hydrological trends in various areas throughout the world (Piechota et al. 1997, 1998; Tootle et al. 2005, 2009; Ghanbari & Bravo 2008; Ríos-Cornejo et al. 2015). This paper also analyzes the teleconnections to large-scale circulation indices using correlation coefficients and wavelet coherence analysis.

STUDY AREA, DATA, AND METHODS

Study area

The Bosten Lake Basin (Figure 1) is located in the southern slope of Tianshan Mountains, Xinjiang, China. The study area is between 40°25′ N and 43°21′ N latitudes and 82°57′ E and 90°39′ E longitudes. It spans an area of 7.7 × 104 km2, with mountain area of 3.47 × 104 km2. Because of the mountain barrier, the Bosten Lake Basin has a typical continental arid climate that is not affected by the Pacific Ocean. The area has a typical inner-continental climate, featuring dry climate, low precipitation, strong evaporation, and shortage of water resources. Water vapor mainly comes from the westerly circulation and Arctic Ocean. The water vapor is uplifted and blocked by Tianshan Mountains, resulting in an uneven spatial distribution of precipitation. Various parallel mountains affect the upper, midstream, and downstream meteorological characteristics of this basin. Precipitation and altitude both decrease from northwest to southeast. The northwestern part of the basin has the highest precipitation, with more than 700 mm per year, while in some southeastern regions, precipitation decreases to 50–60 mm.

Figure 1

Location of meteorological stations, hydrological stations, and hydrological basins in the Bosten Lake Basin, China.

Figure 1

Location of meteorological stations, hydrological stations, and hydrological basins in the Bosten Lake Basin, China.

The basin is mainly composed of the Kaidu River Basin (also known as the Yanqi basin, which includes Bosten Lake Wetland) and the Kongque River Basin. Bosten Lake is a huge ‘regulating reservoir’, being the home of Kaidu River and source of the Kongque River (Figure 1). The administrative region consists of Yanqi, Hejing, Heshuo, Bohu, Yuli County, and also the Korla city. By the end of 2010, the gross domestic product (GDP) was 302.1 × 108 RMB, the primary industry was 86.4 × 108 RMB (accounting for 28.60%), the second industry was 105.3 × 108 RMB (accounting for 34.86%), and the tertiary industry was 110.4 × 108 RMB (accounting for 36.54%).

More than 10 rivers gather into Bosten Lake Basin, including Kaidu River, Qinshui River, Huangshuigou River, Haharengou River, Wulasitai River, Quhuigou River etc. Kaidu River, Huangshuigou River and Qinshui River are the main sources for the whole basin. Kaidu River is the only perennial river, accounting for 84.7% of the inflowing water. Kongque River is the only outlet for Bosten Lake. Since the west pumping station was installed in 1983, the flow has been completely controlled by human activities.

Data

Monthly and yearly temperature and precipitation data (1958–2012) covering the study area were provided by National Climate Center (NCC) of China Meteorological Administration (CMA). For this area, six stations (Bayblk, Balt, Yanq, Kuel, Tieglk, and Kums) passed the internal homogeneity check of the departure accumulating method (Buishand 1982), the standard normal homogeneity test (Alexandersson 1986), and the moving t-test (Peterson et al. 1998). Six runoff time series (Dask, Huangsg, Keergt, Yanq, Bosten Lake, and Tasd) were provided by the Hydrological Bureau of the Xinjiang region. We selected six circulation indices to estimate the influence of large-scale circulation on hydrological processes (Chen et al. 2014; Wang et al. 2015), including the El Niño Southern Oscillation (ENSO), the Arctic Oscillation (AO), the North Atlantic Oscillation (NAO), the Tibetan High (XZH), the westerly circulation index (WI), and the northern hemisphere polar vortex area index (VPA). The definition and more details about the teleconnection indices can be found in Table 1.

Table 1

Teleconnection indices used in this study

Index Full name Definition Data download URL 
Nino3.4 Nino3.4 SST index The average sea surface temperature anomaly in the region bounded by 5°N to 5°S, 170°W to 120°W http://www.cgd.ucar.edu/cas/catalog/climind/Nino_3_3.4_indices.html 
NAO North Atlantic Oscillation The difference in the normalized monthly sea level pressure (SLP) regionally zonal-averaged over the North Atlantic sector from 80°W to 30°E, between 35°N and 65°N http://ljp.lasg.ac.cn/dct/page/1 
AO Arctic Oscillation The difference in the normalized monthly zonal-mean SLP between 35°N and 65°N http://ljp.lasg.ac.cn/dct/page/1 
VPA Northern hemisphere polar vortex area The large area of low pressure and cold air surrounding the north pole, which can be calculated by 500 hPa pressure level using the region of 60°E–150°E,150°E–120°W,120°W–30°W, 30°W–60°E http://ncc.cma.gov.cn/cn/ 
XZH Tibetan High The accumulated value of 500 hPa height minus 500 dagpm within 25°N–35°N, 80°E–100°E http://ncc.cma.gov.cn/cn/ 
WI Westerly circulation The dominance for zonal circulation or meridional circulation over Asian westerly belt circulation (45°N–65°N, 60°E –150°E) http://ncc.cma.gov.cn/cn/ 
Index Full name Definition Data download URL 
Nino3.4 Nino3.4 SST index The average sea surface temperature anomaly in the region bounded by 5°N to 5°S, 170°W to 120°W http://www.cgd.ucar.edu/cas/catalog/climind/Nino_3_3.4_indices.html 
NAO North Atlantic Oscillation The difference in the normalized monthly sea level pressure (SLP) regionally zonal-averaged over the North Atlantic sector from 80°W to 30°E, between 35°N and 65°N http://ljp.lasg.ac.cn/dct/page/1 
AO Arctic Oscillation The difference in the normalized monthly zonal-mean SLP between 35°N and 65°N http://ljp.lasg.ac.cn/dct/page/1 
VPA Northern hemisphere polar vortex area The large area of low pressure and cold air surrounding the north pole, which can be calculated by 500 hPa pressure level using the region of 60°E–150°E,150°E–120°W,120°W–30°W, 30°W–60°E http://ncc.cma.gov.cn/cn/ 
XZH Tibetan High The accumulated value of 500 hPa height minus 500 dagpm within 25°N–35°N, 80°E–100°E http://ncc.cma.gov.cn/cn/ 
WI Westerly circulation The dominance for zonal circulation or meridional circulation over Asian westerly belt circulation (45°N–65°N, 60°E –150°E) http://ncc.cma.gov.cn/cn/ 

In addition, the seasonal series were calculated for each station. Seasons were represented as follows: winter is comprised of December, January, and February (DJF); spring is comprised of March, April, and May (MAM); summer is comprised of June, July, and August (JJA); and autumn is comprised of September, October, and November (SON).

Method

In this study, the Mann-Kendall (MK) test (Chen et al. 2006) was used to detect linear trends for hydrological variables. The rank-based nonparametric MK test is a rank-based procedure, which tests trends without requiring normality or linearity (Birsan et al. 2014; Villafuerte & Matsumoto 2015). The presence of a trend was heavily affected by lag-1 autocorrelation. Yue et al. (2002) modified a pre-whitening procedure, in which the slope of trend is first estimated and the record is detrended. The procedure is as follows: the slope b of a trend in sample data (Xt) is estimated by the Theil-Sen approach (TSA). If b differs from zero, then it is assumed to be linear, and the sample data are detrended by  
formula
(1)
The lag-1 serial correlation coefficient r1 of the detrended series is calculated using Equation (3) and then the first-order autocorrelation coefficient is removed from the by Equation (2):  
formula
(2)
 
formula
(3)
where r1 is the lag-1 serial correlation coefficient of the sample data Xt, and E(Xt) is the mean of the sample data. The identified trend Tt and the residual are blended by  
formula
(4)
The blended series Yt can preserve the true trend and is no longer influenced by the effect of autocorrelation. Then the MK test is applied to the blended series to assess the significance of the trend.

Besides, the distribution-free cumulative sum chart (CUSUM) test (Cunderlik & Simonovic 2005) was used to analyze abrupt change in time-series data. In addition, the relationship between hydrological variables and teleconnection indices was explored using a Pearson correlation and wavelet coherence analysis. Detailed algorithms of wavelet coherence analysis can be found in Grinsted et al. (2004).

RESULTS AND DISCUSSION

Temperature change

Correlation coefficients of air temperature were computed for each weather station as shown in Figure 2. At the annual and seasonal scale, temperatures showed a significant correlation (p < 0.05) between stations. While the temperature at the Bayblk station had a relatively weak correlation with other stations, it did show a higher correlation with the Balt station. The Balt station and Bayblk station have higher elevations, being at 2,458 m and 1,739 m, respectively, so the arithmetic mean temperature of these two stations was considered to be representative of the mountain temperature. The average temperature of the other four stations was considered to be representative of the plain temperature. Annual mountain temperature had an observed step change in 1996 (Figure 3, the lowest point of the curve). Spring, autumn, and winter mountain temperatures also exhibited an abrupt change in 1996. The plain temperature had an observed abrupt change in 1985, while the seasonal abrupt change had some differences, occurring in 1996 in summer, 1987 in fall, and 1984 in winter.

Figure 2

Correlation coefficients of annual and seasonal temperature between meteorological stations (the circle size and color denote the correlation coefficients between meteorological stations; the size of the circle reflects magnitudes of the correlation coefficients, the larger circles indicate larger correlation coefficients). ANN, MAM, JJA, SON and DJF are the annual, spring, summer, autumn and winter CUSUM charts, respectively. Please refer to the online version of this paper to see this figure in color: http://dx.doi.org/10.2166/wcc.2017.140.

Figure 2

Correlation coefficients of annual and seasonal temperature between meteorological stations (the circle size and color denote the correlation coefficients between meteorological stations; the size of the circle reflects magnitudes of the correlation coefficients, the larger circles indicate larger correlation coefficients). ANN, MAM, JJA, SON and DJF are the annual, spring, summer, autumn and winter CUSUM charts, respectively. Please refer to the online version of this paper to see this figure in color: http://dx.doi.org/10.2166/wcc.2017.140.

Figure 3

Cumulative sum (CUSUM) charts for annual and seasonal temperature (ANN, MAM, JJA, SON and DJF are the annual, spring, summer, autumn and winter CUSUM charts, respectively).

Figure 3

Cumulative sum (CUSUM) charts for annual and seasonal temperature (ANN, MAM, JJA, SON and DJF are the annual, spring, summer, autumn and winter CUSUM charts, respectively).

For mountain time series (Figure 4), a significant jump was observed in 1996. Mean annual temperature was 0.74°C before 1996 and 2.12°C after 1996. Mean seasonal temperature in spring, summer, autumn and winter, respectively, was 3.14°C, 13.91°C, 1.67°C and −15.76°C prior to 1996, while after 1996, mean temperature increased to 4.35°C, 14.85°C, 3.28°C and −14.01°C. The temperature in the two periods was relatively stable, with no significant increasing or decreasing trend (p > 0.05). For plain temperature (not shown), abrupt changes in the mean annual, autumn, and winter temperatures were observed around 1985, with mean temperatures increasing from 9.57°C, 9.51°C and −7.90°C to 10.60°C, 10.22°C and −6.57°C, respectively. In spring and summer, the abrupt change occurred in 1996, from 12.98°C and 24.58°C to 13.89°C and 25.58°C, respectively. There was a relatively stable temperature in the previous period, while in the later period, annual temperature had a significantly increasing trend (p = 0.01). Seasonal temperature did not change significantly, indicating a more stable situation at a seasonal scale. Trend analysis indicated that the highest rate of change, 0.039°C/a, in mountain temperature was observed in autumn. The winter exhibited the most significant increasing trend in plain temperature.

Figure 4

Variations of annual and seasonal temperature in mountainous area (ANN, MAM, JJA, SON and DJF are annual, spring, summer, autumn and winter temperature, respectively; the black line is the regional temperature time series; the dashed blue line is the mean value before and after abrupt change year of regional temperature time series; the dot-dash red line is the linear fit before and after abrupt change year of regional temperature time series). Please refer to the online version of this paper to see this figure in color: http://dx.doi.org/10.2166/wcc.2017.140.

Figure 4

Variations of annual and seasonal temperature in mountainous area (ANN, MAM, JJA, SON and DJF are annual, spring, summer, autumn and winter temperature, respectively; the black line is the regional temperature time series; the dashed blue line is the mean value before and after abrupt change year of regional temperature time series; the dot-dash red line is the linear fit before and after abrupt change year of regional temperature time series). Please refer to the online version of this paper to see this figure in color: http://dx.doi.org/10.2166/wcc.2017.140.

Precipitation change

The precipitation correlation diagram (not shown) showed that the correlation coefficients between stations were much smaller than the air temperature correlation coefficients. The correlation coefficients between Bayblk and Balt (and also between the other four stations) are relatively large, similar to temperature. Combined with the site elevation, we considered Bayblk and Balt as mountain precipitation, and the arithmetic average precipitation of the other four stations as plain precipitation. The amount of precipitation in the mountain area increased at 0.8 mm/year (p < 0.05) (Table 2). The precipitation in the plain area also increased, but with no significant change. A step jump year was not significant when evaluating potential abrupt changes.

Table 2

Trend magnitudes for precipitation in plain and mountainous area (mm/year)

  Annual Spring Summer Autumn Winter 
Mountain area 0.8* −0.1146 0.75* 0.1143* 0.073 
Plain area 0.1386 0.0353 0.0229 0.0635 0.024 
  Annual Spring Summer Autumn Winter 
Mountain area 0.8* −0.1146 0.75* 0.1143* 0.073 
Plain area 0.1386 0.0353 0.0229 0.0635 0.024 

*Indicates significant at the 0.05 level.

Runoff change

The runoff can basically be divided into three stages for six typical hydrological stations in the Bosten Lake Basin (Figure 5). Discharge (i.e., runoff) decreased from 1955 to 1986, increased from 1987 to 2003, and then decreased from 2004 to 2010 (Figure 5, Table 3). Dask, Huangsg and Keergt station showed the same hydrological curve, indicating similar conditions of runoff yield. Runoff at those three stations was governed by the combined effects of melt water, precipitation, and groundwater. With the distance increase from stations to origins, the runoff curves of Yanq, Bosten Lake, and Tasd were poorly correlated with the Dask station (the origin hydrological station). At Tasd station, the runoff series was relatively smooth (Figure 5), indicating the flow was entirely controlled by human activities.

Table 3

Runoff trends over different periods in Bosten Lake Basin

Station Time Period
 
1955–1986 1987–2003 2004–2010 
Dask − 0.26** 1.36** −0.83 
Huangsg − 0.03** 0.19** −0.31 
Keergt − 0.03* 0.24** −0.63 
Yanq −0.49 0.96 −0.58 
Bosten Lake − 0.08** 0.25** − 0.34** 
Tasd −0.094 0.86** − 0.95** 
Station Time Period
 
1955–1986 1987–2003 2004–2010 
Dask − 0.26** 1.36** −0.83 
Huangsg − 0.03** 0.19** −0.31 
Keergt − 0.03* 0.24** −0.63 
Yanq −0.49 0.96 −0.58 
Bosten Lake − 0.08** 0.25** − 0.34** 
Tasd −0.094 0.86** − 0.95** 

Note: Units: 108 m3/a in Dask, Huangsg, Keergt, Yanq and Tasd, Units: m/a in Bosten Lake.

*Indicates significant at the 0.05 level.

**Indicates significant at the 0.01 level.

Figure 5

Changes in water quantity at the key stations, Bosten Lake Basin (the red line is the linear trend in the corresponding time period; Dask: Dask hydrological station in the Kaidu River; Huangsg: Huangsg hydrological station in the Huangsg River; Keergt: Keergt hydrological station in the Qingshui River; Yanq: Yanq hydrological station in the lower reaches of Kaidu River; Bosten Lake: water level of Bosten Lake; Tasd: Tasd hydrological station in the Kongque River).

Figure 5

Changes in water quantity at the key stations, Bosten Lake Basin (the red line is the linear trend in the corresponding time period; Dask: Dask hydrological station in the Kaidu River; Huangsg: Huangsg hydrological station in the Huangsg River; Keergt: Keergt hydrological station in the Qingshui River; Yanq: Yanq hydrological station in the lower reaches of Kaidu River; Bosten Lake: water level of Bosten Lake; Tasd: Tasd hydrological station in the Kongque River).

Effects of air temperature and precipitation on runoff

Factors affecting the runoff of Dask and Huangsg stations were analyzed because these two stations are located in the source areas of rivers; therefore, human activities are likely to have a lower effect on the amount of water. These two rivers are the main resources of Bosten Lake, accounting for 84.6% and 6.3% of the total inflow, respectively. Rainfall and temperature significantly affect runoff in the headwater of the Bosten Lake Basin (Table 4). The runoff has a linear relationship with rainfall, and an exponential relationship with temperature (Figure 6), which indicates that rainfall and the glaciers are important water sources in the Bosten Lake Basin. In Kaidu River (Dask station), the correlation coefficients between temperature, precipitation and delayed three season runoff showed that the autumn temperature and summer runoff, and summer temperature and autumn runoff have significant relationships (Table 5). The summer precipitation and winter precipitation also can influence the fall and spring runoff, respectively. Runoff relationships in the Huangsg basin are inconsistent; for example, spring, summer, and autumn temperature have an impact on the winter runoff (Table 5).

Table 4

Synchronous correlation coefficients between temperature, precipitation and runoff in Bosten Lake Basin

Station Temperature
 
Precipitation
 
ANN MAM JJA SON DJF ANN MAM JJA SON DJF 
Dask 0.40** −0.2 0.22 0.49** 0.22 0.69** 0.31* 0.65** 0.27* 0.16 
Huangsg 0.45** 0.10 −0.1 0.43** 0.66** 0.77** 0.10 0.80** 0.17 0.05 
Station Temperature
 
Precipitation
 
ANN MAM JJA SON DJF ANN MAM JJA SON DJF 
Dask 0.40** −0.2 0.22 0.49** 0.22 0.69** 0.31* 0.65** 0.27* 0.16 
Huangsg 0.45** 0.10 −0.1 0.43** 0.66** 0.77** 0.10 0.80** 0.17 0.05 

*Indicates significant at the 0.05 level.

**Indicates significant at the 0.01 level.

Table 5

Time lag effects of the previous three seasons (temperature and precipitation) on runoff in Bosten Lake Basin

Runoff DJF
 
SON
 
JJA
 
MAM
 
T/P MAM JJA SON MAM JJA DJF MAM SON DJF JJA SON DJF 
Dask (T0.25 0.23 0.37* 0.21 0.32* 0.19 0.11 0.34* 0.16 0.16 −0.1 −0.0 
Dask (P0.09 0.07 0.01 0.06 0.56** 0.14 0.20 0.03 0.12 0.19 0.11 0.44* 
Huangsg (T0.42** 0.46** 0.45** 0.28 0.11 0.55 0.08 0.44* 0.40* 0.02 0.16 0.23 
Huangsg (P0.01 0.07 0.02 0.06 0.66 −0.0 0.36* −0.0 −0.2 0.36* 0.49** 0.00 
Runoff DJF
 
SON
 
JJA
 
MAM
 
T/P MAM JJA SON MAM JJA DJF MAM SON DJF JJA SON DJF 
Dask (T0.25 0.23 0.37* 0.21 0.32* 0.19 0.11 0.34* 0.16 0.16 −0.1 −0.0 
Dask (P0.09 0.07 0.01 0.06 0.56** 0.14 0.20 0.03 0.12 0.19 0.11 0.44* 
Huangsg (T0.42** 0.46** 0.45** 0.28 0.11 0.55 0.08 0.44* 0.40* 0.02 0.16 0.23 
Huangsg (P0.01 0.07 0.02 0.06 0.66 −0.0 0.36* −0.0 −0.2 0.36* 0.49** 0.00 

Note:T represents the temperature, P represents the precipitation.

*Indicates significant at the 0.05 level.

**Indicates significant at the 0.01 level.

Figure 6

Relationships between runoff and climate variables (Dask station: Dask hydrological station in the Kaidu River; Huangsg station: Huangsg hydrological station in the Huangsg River).

Figure 6

Relationships between runoff and climate variables (Dask station: Dask hydrological station in the Kaidu River; Huangsg station: Huangsg hydrological station in the Huangsg River).

Teleconnections with climate indices

According to the correlation analysis, hydrological variables showed significant relationships with VPA, XZH, and WI (Table 6), especially for the VPA and XZH. Seasonal analysis also indicates the same relationships (Table 7), with significant correlation coefficients mostly observed for VPA and XZH, especially for the runoff and temperature. Time-lag correlation analysis showed that the current seasonal hydrological variables were impacted by the formal seasonal circulation indices (Table 8); for example, winter runoff will be affected by spring, summer, and autumn XZH and VPA.

Table 6

Correlation between annual temperature, precipitation, runoff and atmospheric circulation indices

Station Climate variable AO Nino3 NAO VPA XZH WI 
Dask R 0.133 −0.01 0.022 − 0.49** 0.531** 0.248 
T 0.29 0.063 0.186 − 0.47** 0.521** 0.450** 
P 0.086 −0.08 0.134 − 0.37** 0.235 0.027 
Huangsg R 0.109 −0.01 0.053 − 0.51** 0.494** 0.237 
T 0.231 0.154 −0.01 − 0.66** 0.727** 0.402** 
P 0.264 0.033 0.237 − 0.44** 0.309* 0.298* 
Station Climate variable AO Nino3 NAO VPA XZH WI 
Dask R 0.133 −0.01 0.022 − 0.49** 0.531** 0.248 
T 0.29 0.063 0.186 − 0.47** 0.521** 0.450** 
P 0.086 −0.08 0.134 − 0.37** 0.235 0.027 
Huangsg R 0.109 −0.01 0.053 − 0.51** 0.494** 0.237 
T 0.231 0.154 −0.01 − 0.66** 0.727** 0.402** 
P 0.264 0.033 0.237 − 0.44** 0.309* 0.298* 

Note:R: runoff, T: temperature, P: precipitation.

*Indicates significant at the 0.05 level.

**Indicates significant at the 0.01 level.

Table 7

Correlations between seasonal temperature, precipitation, runoff and atmospheric circulation indices in Dask station

Circulation indices Runoff
 
Temperature
 
Precipitation
 
MAM JJA SON DJF MAM JJA SON DJF MAM JJA SON DJF 
AO 0.018 0.109 0.088 0.239 0.175 0.237 0.258 0.258 −0.153 0.025 0.227 −0.073 
Nino3.4 −0.055 0.031 −0.051 0.008 0.013 −0.056 0.098 −0.011 −0.182 0.207 −0.050 −0.107 
NAO −0.116 0.022 −0.116 0.202 0.148 −0.153 −0.047 0.180 −0.250 0.052 0.040 −0.046 
VPA −0.076 − 0.340* − 0.558** − 0.448** −0.205 − 0.541** − 0.30* −0.274 0.063 −0.23 −0.187 −0.050 
XZH 0.32* 0.231 0.394** 0.476** 0.40** 0.186 0.54** 0.39** 0.064 0.084 0.036 0.045 
WI 0.140 0.197 0.136 0.220 0.164 0.205 0.41** 0.254 −0.058 0.199 0.136 −0.034 
Circulation indices Runoff
 
Temperature
 
Precipitation
 
MAM JJA SON DJF MAM JJA SON DJF MAM JJA SON DJF 
AO 0.018 0.109 0.088 0.239 0.175 0.237 0.258 0.258 −0.153 0.025 0.227 −0.073 
Nino3.4 −0.055 0.031 −0.051 0.008 0.013 −0.056 0.098 −0.011 −0.182 0.207 −0.050 −0.107 
NAO −0.116 0.022 −0.116 0.202 0.148 −0.153 −0.047 0.180 −0.250 0.052 0.040 −0.046 
VPA −0.076 − 0.340* − 0.558** − 0.448** −0.205 − 0.541** − 0.30* −0.274 0.063 −0.23 −0.187 −0.050 
XZH 0.32* 0.231 0.394** 0.476** 0.40** 0.186 0.54** 0.39** 0.064 0.084 0.036 0.045 
WI 0.140 0.197 0.136 0.220 0.164 0.205 0.41** 0.254 −0.058 0.199 0.136 −0.034 

*Indicates significant at the 0.05 level.

**Indicates significant at the 0.01 level.

Table 8

Time-lag correlation coefficients between hydrological and circulation variables in previous three seasons in dask station

Circulation indices JJA time series
 
Circulation indices DJF time series
 
Runoff Temperature Precipitation Runoff Temperature Precipitation 
AO1 0.072 0.096 0.046 AO1 0.058 0.094 −0.259 
AO3 0.001 0.139 0.106 AO2 0.163 −0.145 0.446** 
AO4 0.170 0.266 0.253 AO3 −0.001 0.007 0.077 
Nino1 −0.004 0.045 −0.090 Nino1 0.098 −0.111 0.051 
Nino3 −0.042 0.280 −0.275 Nino2 0.157 0.099 0.042 
Nino4 −0.103 0.091 −0.072 Nino3 0.115 0.206 0.025 
NAO1 0.088 −0.079 0.129 NAO1 −0.005 0.107 −0.216 
NAO3 0.044 −0.173 −0.104 NAO2 −0.071 −0.168 0.181 
NAO4 0.025 0.189 0.300* NAO3 −0.232 −0.035 0.039 
XZH1 0.342* 0.377** 0.258 XZH1 0.452** 0.339* −0.121 
XZH3 0.402** 0.448** 0.415** XZH2 0.331* 0.044 0.130 
XZH4 0.405** 0.330* 0.255 XZH3 0.509* 0.273 0.101 
WI1 0.259 0.012 0.133 WI1 0.143 0.153 0.307* 
WI3 0.013 0.028 0.197 WI2 0.194 −0.076 0.087 
WI4 0.230 0.347* 0.226 WI3 0.129 0.294* −0.016 
VPA1 −0.258 − 0.528** −0.179 VPA1 − 0.473** − 0.292* 0.043 
VPA3 −0.211 − 0.334* − 0.329* VPA2 − 0.643** −0.114 0.303* 
VPA4 −0.270 −0.266 − 0.487** VPA3 − 0.479** −0.167 0.000 
Circulation indices JJA time series
 
Circulation indices DJF time series
 
Runoff Temperature Precipitation Runoff Temperature Precipitation 
AO1 0.072 0.096 0.046 AO1 0.058 0.094 −0.259 
AO3 0.001 0.139 0.106 AO2 0.163 −0.145 0.446** 
AO4 0.170 0.266 0.253 AO3 −0.001 0.007 0.077 
Nino1 −0.004 0.045 −0.090 Nino1 0.098 −0.111 0.051 
Nino3 −0.042 0.280 −0.275 Nino2 0.157 0.099 0.042 
Nino4 −0.103 0.091 −0.072 Nino3 0.115 0.206 0.025 
NAO1 0.088 −0.079 0.129 NAO1 −0.005 0.107 −0.216 
NAO3 0.044 −0.173 −0.104 NAO2 −0.071 −0.168 0.181 
NAO4 0.025 0.189 0.300* NAO3 −0.232 −0.035 0.039 
XZH1 0.342* 0.377** 0.258 XZH1 0.452** 0.339* −0.121 
XZH3 0.402** 0.448** 0.415** XZH2 0.331* 0.044 0.130 
XZH4 0.405** 0.330* 0.255 XZH3 0.509* 0.273 0.101 
WI1 0.259 0.012 0.133 WI1 0.143 0.153 0.307* 
WI3 0.013 0.028 0.197 WI2 0.194 −0.076 0.087 
WI4 0.230 0.347* 0.226 WI3 0.129 0.294* −0.016 
VPA1 −0.258 − 0.528** −0.179 VPA1 − 0.473** − 0.292* 0.043 
VPA3 −0.211 − 0.334* − 0.329* VPA2 − 0.643** −0.114 0.303* 
VPA4 −0.270 −0.266 − 0.487** VPA3 − 0.479** −0.167 0.000 

*Indicates significant at the 0.05 level.

**Indicates significant at the 0.01 level, Numbers 1–4 following the circulation indices represent spring, summer, autumn, winter, respectively; JJA indicates June, July and August; DJF indicates December, January and February.

Correlation analyses often exhibit spurious correlations, so wavelet analysis was used to further analyze the relationships between the circulation indices and hydrological variables. Continuous wavelet analysis of circulation indices is shown in Figure 7. Significant power was observed at the 1-year period for VPA and XZH. Most of the significant powers were observed at the 2–7-year band for Nino3.4, while for other indices, significant power was only distributed at the 0.2–1-year period, and was distributed separately over the years. For the characteristics of wavelet coherence between VPA and runoff (Figure 8), significant common power was observed in the 0.7–1.7-year period for all the observed time. The pattern of wavelet power of XZH was the same as VPA. While VPA exhibited a significant anti-phase relationship, XZH showed a significant in-phase relationship around the 1-year time scale. This indicates that VPA and XZH have an impact on runoff, especially at annual time scales. Significant coherence was also found at the 2–7-year scale between Nino3.4 and runoff, but phase relations were ambiguous. Additionally, high coherence was also found for AO, NAO, and WI at the 1–7-year period, while those relationships were also ambiguous. So for teleconnection indices of AO, NAO, Nino3.4 and WI, there was significant wavelet power but the response was not robust.

Figure 7

The continuous wavelet power spectrum of teleconnection indices. The thick contour designates the 5% significance level against red noise, and the cone of influence (COI) where edge effects might distort the picture is shown as a lighter shade.

Figure 7

The continuous wavelet power spectrum of teleconnection indices. The thick contour designates the 5% significance level against red noise, and the cone of influence (COI) where edge effects might distort the picture is shown as a lighter shade.

Figure 8

Wavelet coherence power spectrum between the regional atmospheric circulation and runoff in Dask station (thick contours enclose the areas with coherence statistically significant at 95% confidence level against red noise. Semi-transparent areas indicate the ‘cone of influence’ where the edge effects of wavelet transforms become important. The relative phase relationship is shown as arrows, with in-phase pointing right, anti-phase pointing left).

Figure 8

Wavelet coherence power spectrum between the regional atmospheric circulation and runoff in Dask station (thick contours enclose the areas with coherence statistically significant at 95% confidence level against red noise. Semi-transparent areas indicate the ‘cone of influence’ where the edge effects of wavelet transforms become important. The relative phase relationship is shown as arrows, with in-phase pointing right, anti-phase pointing left).

The wavelet coherence analysis was also conducted for other climate variables of temperature and precipitation in Dask and in Huangsg station. As with the runoff in Dask station, significant coherence occurred between hydrological indices and teleconnection indices of VPA and XZH. VPA changes result in changes in the mid-latitude westerly trough and ridge system, which will influence the climate in Bosten Lake Basin. When VPA is high, the strengthened northwest winds will decrease the watershed temperature and then runoff will decrease. When VPA is low, the southwest wind from the Arabian Sea to the Bosten Lake will be strengthened, which will increase the water vapor transport, and then the precipitation and runoff will increase (Wang et al. 2015). The Tibetan Plateau is a heat source in summer and a weak heat sink in winter in Asia. The anticyclonic circulation in the northern Tibetan Plateau and cyclone circulation in the southern Tibetan Plateau can increase the moisture from the Caspian Sea into the arid region, which will result in wetter and colder conditions in the region (Villafuerte & Matsumoto 2015). So, among the teleconnection indices, only the XZH and the northern hemisphere VPA are the main teleconnection indices that affect the hydrological circulation in the Bosten Lake Basin (Figures 9 and 10). The curve fitting is better for XZH and VPA, which also verifies the above conclusion.

Figure 9

Relationships between XZH and hydrological variables in the headstream of Bosten Lake Basin (XZH: the Tibetan High; Dask: Dask hydrological station in the Kaidu River; Huangsg: Huangsg hydrological station in the Huangsg River).

Figure 9

Relationships between XZH and hydrological variables in the headstream of Bosten Lake Basin (XZH: the Tibetan High; Dask: Dask hydrological station in the Kaidu River; Huangsg: Huangsg hydrological station in the Huangsg River).

Figure 10

Relationships between VPA and hydrological variables in the headstream of Bosten Lake Basin (VPA: Northern hemisphere polar vortex area; Dask: Dask hydrological station in the Kaidu River; Huangsg: Huangsg hydrological station in the Huangsg River).

Figure 10

Relationships between VPA and hydrological variables in the headstream of Bosten Lake Basin (VPA: Northern hemisphere polar vortex area; Dask: Dask hydrological station in the Kaidu River; Huangsg: Huangsg hydrological station in the Huangsg River).

CONCLUSIONS

In this paper characteristics of linear trend and step jump of climate variables were analyzed by the MK test and CUSUM test. In addition, the effects of climate change and teleconnections on water resources in the Bosten Lake Basin were investigated using correlation coefficients and wavelet coherence analysis. The main results are as follows:

  • (1)

    Bosten Lake Basin can be divided into a mountain area and a plain area from the viewpoint of climatology. The temperature in the past 55 years has significantly increased in the mountain area and the plain area. The precipitation trend is less significant than the temperature trend, with only the annual mountain area precipitation exhibiting a significantly increasing trend. Abrupt change of the mountainous and plain temperatures was observed in 1996 and 1985, respectively.

  • (2)

    Changes in runoff were observed in three stages: decrease from 1955 to 1986, increase from 1987 to 2003, and decrease from 2004 to 2010. Changes in runoff were strongly associated with precipitation and temperature changes, a linear relationship with precipitation, and exponential relationship with temperature.

  • (3)

    Through correlation analysis and wavelet coherence analysis, we prove that the XZH and the northern hemisphere VPA may be the main factors affecting hydrological circulation changes in the Bosten Lake Basin.

ACKNOWLEDGEMENTS

The research is supported by the Natural Science Foundation of the Jiangsu Higher Education Institutions of China (Grant No. 16KJB170001), the Science & Technology Plan Projects of Huaian, China (HAS2015005-2), and the National Natural Science Foundation of China (41471030). The authors thank the National Climate Central, China Meteorological Administration, for providing the meteorological data for this study. We also thank LetPub (www.letpub.com) for its linguistic assistance during the preparation of this manuscript.

REFERENCES

REFERENCES
Alexandersson
,
H.
1986
A homogeneity test applied to precipitation data
.
Journal of Climatology
6
(
6
),
661
675
.
Birsan
,
M. V.
,
Zaharia
,
L.
,
Chendes
,
V.
&
Branescu
,
E.
2014
Seasonal trends in Romanian streamflow
.
Hydrological Processes
28
(
15
),
4496
4505
.
Buishand
,
T. A.
1982
Some methods for testing the homogeneity of rainfall records
.
Journal of Hydrology
58
(
1–2
),
11
27
.
Chen
,
Y.
,
Takeuchi
,
K.
,
Xu
,
C.
,
Chen
,
Y.
&
Xu
,
Z.
2006
Regional climate change and its effects on river runoff in the Tarim Basin, China
.
Hydrological Processes
20
(
10
),
2207
2016
.
Chen
,
Y.
,
Deng
,
H.
,
Li
,
B.
,
Li
,
Z.
&
Xu
,
C.
2014
Abrupt change of temperature and precipitation extremes in the arid region of Northwest China
.
Quaternary International
336
(
12
),
35
43
.
Chiew
,
F. H. S.
&
McMahon
,
T. A.
1993
Detection of trend or change in annual flow of Australian rivers
.
International Journal of Climatology
13
(
13
),
643
653
.
Dou
,
Y.
,
Chen
,
X.
,
Bao
,
A.
&
Li
,
L.
2011
The simulation of snowmelt runoff in the ungauged kaidu River Basin of tianShan Mountains, China
.
Environmental Earth Sciences
62
(
5
),
1039
1045
.
Flato
,
G.
,
Marotzke
,
J.
,
Abiodun
,
B.
,
Braconnot
,
P.
,
Chou
,
S. C.
,
Collins
,
W. J.
,
Cox
,
P.
,
Driouech
,
F.
,
Emori
,
S.
&
Eyring
,
V.
2013
Evaluation of climate models
. In:
Climate Change 2013: The Physical Science Basis. Contribution of Working Group I to the Fifth Assessment Report of the Intergovernmental Panel on Climate Change, Cambridge University Press
,
Cambridge, UK and New York, NY
, pp.
741
882
.
Ghanbari
,
R. N.
&
Bravo
,
H. R.
2008
Coherence between atmospheric teleconnections, Great Lakes water levels, and regional climate
.
Advances in Water Resources
31
(
10
),
1284
1298
.
Grinsted
,
A.
,
Moore
,
J. C.
&
Jevrejeva
,
S.
2004
Application of the cross wavelet transform and wavelet coherence to geophysical time series
.
Nonlinear Processes in Geophysics
11
(
5/6
),
561
566
.
Li
,
X.
,
Li
,
L.
,
Guo
,
L.
,
Zhang
,
F.
,
Adsavakulchai
,
S.
&
Shang
,
M.
2011
Impact of climate factors on runoff in the Kaidu River watershed: path analysis of 50-year data
.
Journal of Arid Land
03
(
2
),
132
140
.
Liu
,
J. S.
,
Wei
,
W. S.
,
Huang
,
Y. Y.
&
Shang
,
S. C.
2006
Hydrological response of winter streamflow to climate change and permafrost degradation in Manas Watershed, Tianshan Mountains
.
Journal of Glaciology & Geocryology
28
(
5
),
656
662
.
In Chinese with English abstract
.
Peterson
,
T. C.
,
Easterling
,
D. R.
,
Karl
,
T. R.
,
Groisman
,
P.
,
Nicholls
,
N.
,
Plummer
,
N.
,
Torok
,
S.
,
Auer
,
I.
,
Boehm
,
R.
&
Gullett
,
D.
1998
Homogeneity adjustments of in situ atmospheric climate data: a review
.
International Journal of Climatology
18
(
13
),
1493
1517
.
Piechota
,
T. C.
,
Dracup
,
J. A.
&
Fovell
,
R. G.
1997
Western US streamflow and atmospheric circulation patterns during el niño-southern oscillation
.
Journal of Hydrology
201
(
1–4
),
249
271
.
Piechota
,
T.
,
Chiew
,
F.
,
Dracup
,
J.
&
McMahon
,
T.
1998
Streamflow forecasting in eastern Australia and El Nino Southern Oscillation
.
Water Resources Research
34
(
11
),
3035
3044
.
Qin
,
D.
&
Stocker
,
T.
2014
Highlights of the IPCC working group I fifth assessment report
.
Progressus Inquisitiones De Mutatione Climatis
10
(
1
),
1
6
.
Richman
,
M. B.
1986
Rotation of principal components
.
Journal of Climatology
6
(
3
),
293
335
.
Ríos-Cornejo
,
D.
,
Penas
,
Á.
,
Álvarez-Esteban
,
R.
&
Río
,
S. D.
2015
Links between teleconnection patterns and precipitation in Spain
.
Atmospheric Research
156
(
1
),
14
28
.
Shen
,
Y.
,
Liu
,
C.
,
Min
,
L.
,
Yan
,
Z.
&
Tian
,
C.
2010
Change in pan evaporation over the past 50 years in the arid region of China
.
Hydrological Processes
24
(
2
),
225
231
.
Shi
,
Y.
,
Shen
,
Y.
,
Kang
,
E.
,
Li
,
D.
,
Ding
,
Y.
,
Zhang
,
G.
&
Hu
,
R.
2007
Recent and future climate change in northwest China (EI)
.
Climatic Change
80
(
3
),
379
393
.
Tootle
,
G. A.
,
Piechota
,
T. C.
&
Singh
,
A.
2005
Coupled oceanic-atmospheric variability and U.S. streamflow
.
Water Resources Research
41
,
308
324
.
Tootle
,
G. A.
,
Piechota
,
T. C.
,
Aziz
,
O.
,
Miller
,
W. P.
,
Lakshmi
,
V.
,
Dracup
,
J. A.
&
Jerla
,
C.
2009
The 2009–2010 El Niño: Hydrologic Relief to U.S. Regions?
EOS Transactions American Geophysical Union
90
(
50
),
481
482
.
Villafuerte
,
M. Q. I.
&
Matsumoto
,
J.
2015
Significant influences of global mean temperature and ENSO on extreme rainfall in Southeast Asia
.
Journal of Climate
28
(
5
),
1905
1919
.
Wang
,
H.
,
Chen
,
Y.
,
Li
,
W.
&
Deng
,
H.
2013
Runoff responses to climate change in arid region of Northwestern China during 1960–2010
.
Chinese Geographical Science
23
(
3
),
286
300
.
Yue
,
S.
,
Pilon
,
P.
,
Phinney
,
B.
&
Cavadias
,
G.
2002
The influence of autocorrelation on the ability to detect trend in hydrological series
.
Hydrological Processes
16
(
9
),
1807
1829
.
Zhang
,
Y. C.
,
Li
,
B. L.
,
Bao
,
A. M.
,
Zhou
,
C. H.
,
Xi
,
C.
&
Zhang
,
X. R.
2007
Study on snowmelt runoff simulation in the Kaidu River basin
.
Science in China
50
(
S1
),
26
35
.