Long-Term Trend and Abrupt Change for Major Climate ... - CiteSeerX

204

ACTA METEOROLOGICA SINICA

VOL.21

Long-Term Trend and Abrupt Change for Major Climate Variables in the Upper Yellow River Basin∗ ZHAO Fangfang† (

), XU Zongxue (

), and HUANG Junxiong(

)

Key Laboratory of Water and Sediment Sciences of the Ministry of Education, College of Water Sciences, Beijing Normal University, Beijing 100875 (Received March 14, 2007)

ABSTRACT On the basis of the mean air temperature, precipitation, sunshine duration, and pan evaporation from 23 meteorological stations in the upper Yellow River Basin from 1960 to 2001, the feasibility of using hypothesis test techniques to detect the long-term trend for major climate variables has been investigated. Parametric tests are limited by the assumptions such as the normality and constant variance of the error terms. Nonparametric tests have not these additional assumptions and are better adapted to the trend test for hydro-meteorological time series. The possible trends of annual and monthly climatic time series are detected by using a non-parametric method and the abrupt changes have been examined in terms of 5-yr moving averaged seasonal and annual series by using moving T-test (MTT) method, Yamamoto method, and Mann-Kendall method. The results show that the annual mean temperature has increased by 0.8 ◦ C in the upper Yellow River Basin during the past 42 years. The warmest center was located in the northern part of the basin. The nonlinear tendency for annual precipitation was negative during the same period. The declining center for annual precipitation was located in the eastern part and the center of the basin. The variation of annual precipitation in the upper Yellow River Basin during the past 42 years exhibited an increasing tendency from 1972 to 1989 and a decreasing tendency from 1990 to 2001. The nonlinear tendencies for annual sunshine duration and pan evaporation were also negative. They have decreased by 125.6 h and 161.3 mm during the past 42 years, respectively. The test for abrupt changes by using MTT method shows that an abrupt warming occurred in the late 1980s. An abrupt change of the annual mean precipitation occurred in the middle 1980s and an abrupt change of the mean sunshine duration took place in the early 1980s. For the annual mean pan evaporation, two abrupt changes took place in the 1980s and the early 1990s. The test results of the Yamamoto method show that the abrupt changes mostly occurred in the 1980s, and two acute abrupt changes were tested for the spring pan evaporation in 1981 and for the annual mean temperature in 1985. According to the Mann-Kendall method, the abrupt changes of the temperature mainly occurred in the 1990s, the pan evaporation abrupt changes mostly occurred in the 1960s, and the abrupt changes of the sunshine duration primarily took place in the 1980s. Although the results obtained by using three methods are different, it is undoubted that jumps have indeed occurred in the last four decades. Key words: climate change, trend, abrupt change, the Yellow River

1. Introduction The working meeting, co-organized by the International Geosphere-Biosphere Program (IGBP) and World Climate Research Program (WCRP), was held at Venice City, Italy in November 1994. There were six issues confirmed in the meeting, of which climate abrupt dynamics and climate change evaluation were two important issues (Wang, 1997). In recent years, the regional climate change issues related with the activities of people property have become one of the most important issues (Yan et al., 2001). The cli-

mate system has typical characteristics of multi-scale in space, multi-layer in configuration, nonlinearity in nature, with complex mutual connection and effect (Li, 2001). Many researchers investigated the trend of climate variables and the characteristics of the climate abrupt changes. For example, the summer climate jumps in the Northern Hemisphere in summer of the 1960s (Yan et al., 1990, 1992), the tendencies and climate jumps of four main climate variables in the Sanjiang Plain using accumulated departure, Jy parameter, Yamamoto method, and Mann-Kendall method (Yan et al., 2001, 2003), the

∗ Supported by the “Jingshi scholar” Leading Professor Program, Beijing Normal University and the National Basic Research Program (973) of China under Grant No. 1999043601. † Corresponding author: [email protected].

NO.2

ZHAO Fangfang, XU Zongxue and HUANG Junxiong

climatic variation tendencies, interdecadal variations, and climate jumps over the middle reaches of the Yarlung Tsangpo River in the Tibetan Plateau (Zhou et al., 2001), and the climate variations, tendencies, and climate jumps in Xinjiang Autonomous Region (Yang, 2003), etc. However, few studies on the climate tendencies and abrupt changes in the upper Yellow River Basin have been done, although Yang and Li (2004b) analyzed the abrupt and periodic changes of the precipitation and runoff series in this area with EOF method and Mann-Kendall method. Therefore, further study should be conducted. In addition, the scarcity of water resources in the Yellow River Basin has been paid more attention to by domestic and international experts in recent years. Headwater catchment of the Yellow River Basin is the “water tower” of the whole basin, in which the streamflow has reduced, and the water level in the lakes has declined. Whether it resulted from climate change or human activities should be further studied. Therefore, on the basis of monthly mean air temperature, precipitation, sunshine duration, and evaporation from 23 meteorological stations in the upper Yellow River Basin (upward of Lanzhou Station) from 1960 to 2001, the feasibility of using hypothesis test techniques to identify the long-term trend for major climate variables during the past 42 years has been investigated in this study. At the same time, the abrupt changes also have been examined by using different methods to quantitatively describe the climate change in the study area. 2. Study area The Yellow River Basin is located in the semi-arid and semi-humid region with severe water scarcity, in which the annual mean precipitation is about 200-600 mm, and the natural streamflow is about 580×108 m3 (Yang and Li, 2004b). The drainage area at upward of Lanzhou Station is about 222551 km2 . The climate belongs to the Qinghai-Tibetan Plateau climate system. In cold seasons, the basin climate has the characteristics of typical continental climate, which is controlled by the high pressure of the Qinghai-Tibetan Plateau, lasting for about seven months. In warm seasons, the climate is affected by southwest monsoon, producing

205

heat low pressure, with abundant water vapor and more precipitation, and thus forms the plateau subtropical humid monsoon climate. The whole climate characteristics of the study area are as follows: long winter nearly without summer, spring immediatly after autumn, low heat, small annual temperature difference, large daily temperature range, long sunshine duration, intense solar radiation, big windy storm, and short plant growth periods. The annual air temperature is 2.68◦ C, with 2554.7 h for sunshine duration, 1428.9 mm for evaporation, and 446 mm for precipitation. The annual precipitation shows an increasing trend from northwest to southeast. The precipitation from June to September accounts for 75% of the annual value. The water resources of the upper Yellow River Basin account for 57.5% of the whole Yellow River Basin (average of 1951-1998), in which the spatio-temporal variations of the water resources are very important for the whole Yellow River Basin (Li, 2003). The variation of climate variables are the main reasons for the water resources change (Li, 2003). Therefore, it is very important to investigate the spatio-temporal variations of climate variables in the upper Yellow River Basin in order to identify the evolvement of the water resources system in the whole Yellow River Basin. 3. Data and methodology There are 23 meteorological stations selected in the upper Yellow River Basin. These stations are spatially well distributed, which can reflect the characteristics of regional climate. The data of monthly mean air temperature, precipitation, sunshine duration, and evaporation come from the China Meteorological Administration, which have been checked by the primary quality control. Considered the reliability and integrality, the observed data from 1960 to 2001 are selected in this study. At the same time, in order to ensure the integrality of the time series, the absent data are interpolated by using the data from nearby stations. From the statistical meaning, it is credible to get the results by the use of so long time series. The location of the study area and the meteorological stations selected are shown in Fig.1.

206


Fig.1. Location of the study area and the meteorological stations selected.

According to the unique climate characteristics in the upper Yellow River Basin, the seasons can be classified as follows: March-April-May for spring, June-July-August for summer, September-OctoberNovember for autumn, and December-JanuaryFeburary (the following year) for winter (Zeng, 2004). For the time series of each season, the data of mean air temperature is the average of three months’ value, but the precipitation, sunshine duration, and evaporation are the sum of three months’ value. In order to reduce the unilateralism of single station record, the regional series are calculated by the spatial average of all the stations in the whole area. When analyzing the climate abrupt changes, the 5-yr moving average series of the regionalized data for seasonal and annual series are estimated, which represents the long-term trends. In this study, the climate tendencies in the study area are analyzed by using nonparametric MannKendall method, the periodic changes are analyzed by using departure curve method, and the climate abrupt changes are analyzed by use of moving T method, Yamamoto method, and Mann-Kendall method. 4. Climate change analysis During the past 20 years, many researchers have investigated the regional climate characteristics on different time scales in China. The results provide favorable basis and direction to exactly grasp the climate characteristics on large scale and further understand the regional climate change (Yan et al., 2001). Nonparametric Mann-Kendall method is widely used to analyze the trends of the environmental time series,

VOL.21

which is recommended by World Meteorological Organization (WMO) (Liu and Zheng, 2003, 2004; Yu et al., 2002). It is also an efficient tool to examine the monotonic trend of hydro-meteorological series (Xu et al., 2002, 2003). In this study, the climate trends of climate variables from 23 gauging stations in the upper Yellow River Basin for 42 yr are detected at the 95% level of significance in this study. At the same time, the magnitude of long-term trend for climate variables (Kendall slope) from different gauging stations are spatially interpolated in this study in the whole basin by using Kriging method. 4.1 Temperature Figure 2a shows the spatial distribution of nonlinear tendency for the mean air temperature in the upper Yellow River Basin. It shows an increasing trend in most parts of the study area. The Kendall slopes at 21 gauging stations are positive, and only 2 stations (Zhongxin and Henan Stations) are negative. Two warm centers are shown in the whole basin: one is near Qiabuqia Station in the north, and the other is near Lanzhou Station in the east, in which the Kendall slopes are up to 0.48◦C/(10 yr) and 0.44◦ C/(10 yr), respectively. The average Kendall slope for the whole basin is 0.18◦C/(10 yr), i.e., the mean air temperature has increased by 0.76◦C in the upper Yellow River Basin during the past 42 years. Figure 2b shows the departure curves of the mean air temperature in the upper Yellow River Basin. Departure is the difference of climate variables for 42 yr. It is shown in Fig.2b that there are two obvious periods in the study area for the past 42 years. One is the cold period of 1960-1986, in which the negative departures account for more than 80%, and the abnormal cold years are 1967, 1977, and 1983, respectively. The other is the short warm period of 1987-2001, in which the mean air temperature is 3.1◦ C, 0.46◦C higher than the average of the whole basin. In warm period, the negative departures account for more than 87%, in which the highest temperature for 42 yr, and the temperature in 1998 is 1.7◦ C higher than that of the whole basin. The temperature change has an obvious seasonal difference. Comparing with the departure curves, winter temperature has major contribution to the annual

NO.2


207

Fig.2. Distribution of nonlinear tendency for temperature: (a) spatial distribution ( ◦ C/10 yr) and (b) departure curves (◦ C).

mean temperature. This conclusion is consistent with the result obtained by Ding and Dai (1994). 4.2 Precipitation Figure 3a shows the distribution of nonlinear tendency for precipitation in the upper Yellow River Basin. It shows a decreasing trend in most parts of the basin. The Kendall slopes at 17 gauging stations are negative. There are two decreasing centers located near Lintao and Henan Stations along the main stem, in which the Kendall slopes are −28.63 mm/(10 yr) and −28.83 mm/(10 yr), respectively. The average Kendall slope for the whole basin is −4.26 mm/(10 yr). Therefore, there is a slight dry trend in the upper Yellow River Basin since 1960. Figure 3b shows the departure curves of the precipitation in the upper Yellow River Basin. It is shown that the departure curve fluctuates significantly, with the characteristics of three increasing and three de-

creasing abrupt changes since the 1960s. The period of more precipitation i. e., 1970-1989, lasts for long time, in which the mean precipitation is 11.7 mm more than that of the whole basin. The period of less precipitation is from 1990 to 2001, in which the mean precipitation is 18.6 mm less than that of the whole basin. The seasonal departure curves show that autumn precipitation has the greatest contribution to the annual total. 4.3 Sunshine duration Figure 4a shows the distribution of nonlinear tendency for annual sunshine duration in the upper Yellow River Basin. It shows a decreasing trend in most parts of the study area. The decreasing center is located at Minhe and Lanzhou Stations, in which the greatest Kendall slope is up to −104.38 h/(10 yr). Meanwhile, the increasing area centered at Tongde

Fig.3. Distribution of nonlinear tendency for precipitation: (a) spatial distribution (mm/10 yr) and (b) departure curves (mm).

208


Dari, and Maqu Stations, with the greatest Kendall slope up to 56.39 h/(10 yr). The average Kendall slope for the whole basin is −29.9 h/(10 yr), i.e., the sunshine duration decreased by 125.6 h in the upper Yellow River Basin during the past 42 years. Figure 4b shows the departure curves of the sunshine duration in the upper Yellow River Basin. There are two obvious periods. One is the higher period from 1961 to 1980 lasting for long time, in which the sunshine duration is 35.4 h higher than that of the whole basin. The other is the lower period from 1981 to 1996, in which the sunshine duration is 49.4 h lower than that of the whole basin. In the seasonal trend curves, the spring and winter sunshine durations make the greatest contribution to the annual total. 4.4 Evaporation Figure 5a shows the distribution of nonlinear tendency for annual pan evaporation in the upper Yellow River Basin. It shows a decreasing trend in most parts of the study area, in which the greatest Kendall slope is −115.74 mm/(10 yr). In the eastern, southern, and western marginal area, the evaporation shows an increasing trend. The average Kendall slope for the whole basin is −38.4 mm/(10 yr), i.e., the evaporation decreased by 161.3 mm in the upper Yellow River Basin during the past 42 years. Figure 5b shows the departure curves for the pan evaporation in the upper Yellow River Basin. There are an obvious increasing period from 1960 to 1973 and an obvious decreasing period from 1974 to 1997. The pan evaporation is 74 mm higher in the increasing period and 50 mm lower in the decreasing period than the average. In the seasonal trend curves, the spring and summer pan evaporation has significant contribution to the annual value. 4.5 Relationship among major climate variables The interdecadal variations for the departure time series of the mean air temperature, precipitation, sunshine duration, and pan evaporation are shown in Figs.2b, 3b, 4b, and 5b with dashed lines. It is shown

VOL.21

in Figs.2b and 3b that the temperature in the 1960s and 1980s are increasing, and the precipitation is increasing from the 1960s to 1980s, but deceasing in the 1990s. The relationship between precipitation and temperature is weak, i.e., the precipitation may be high or low when the temperature is high. The result is similar to that obtained by Shi (1996), i.e., the changes of temperature are not directly responsible for the changes of precipitation. Therefore, the variations of precipitation in the future should be further studied. The sunshine duration is one of the important climate factors to evaluate the regional radiation resources. In principle, decreasing of the sunshine duration may result in decreasing of temperature. However, the green house effect leads to the increasing of temperature (Yang et al., 2004). The impact of human activities in the upper Yellow River Basin is relatively small. Therefore, there is inverse relationship among sunshine duration, evaporation, and precipitation. It is shown in Fig.4b that sunshine duration has decreased since the 1960s, especially in the 1980s, and began to increase in the 1990s. The variations of evaporation (Fig.5b) are the same as that of the sunshine duration, which is opposite to the inter-decadal variations of the precipitation as shown in Fig.3b. It is the basis of eco-environmental change in the study area to qualitatively analyze the relationship between temperature, precipitation, sunshine duration, and evaporation. It can help to reasonably predict the future climate changes and establish the corresponding countermeasures. 5. Abrupt changes of climatic variables Climate system is nonlinear and discontinuous. Therefore, it is necessary to analyze and understand the change process of the climate system by using nonlinear theories and methods, such as theory of the abrupt changes and the detection method (Yan et al., 2003). Fu and Wang (1992) discussed the definition and detection methods, which can help to understand and detect the abrupt changes. There are many kinds of methods to detect the

NO.2


209

Fig.4. Distribution of nonlinear tendency for sunshine duration: (a) spatial distribution (h/10 yr) and (b) departure curves (h).

Fig.5. Distribution of nonlinear tendency for pan evaporation: (a) spatial distribution (mm/10 yr) and (b) departure curves (mm).

abrupt changes, such as low pass filtering method, moving T test method (MTT method), Crammer method, Yamamoto method, Mann-Kendall method, Spearman method, etc. The low pass filtering method is not applicable. MTT method, Crammer method, and Yamamoto method are famous for intuitionistic, simple, and convenient uses. But the results may be different because of the artificial reasons. Therefore, it should depend on the Mann-Kendall method and Spearman method to accurately examine the occurrence of abrupt changes. These methods have merits of broad detecting range, small artificial impact, and high quantitative degree (Wei, 1999). Therefore, the abrupt changes of the climate variables in the study area are detected by using MTT method, Yamamoto

method, and Mann-Kendall method. The detailed theories are referred to Fu and Wang (1992) and Wei (1999). Wei and Cao (1995) analyzed the abrupt changes of mean air temperature in China, Northern Hemisphere, and global area, and got the results that it is creditable when 10-yr mean period is taken for the abrupt index. Therefore, the mean period of time n1 =n2 =10 is adopted in this paper, and n1 =n2 =14 as the comparing period. However, only abrupt changes occurring in 1967-1991 were detected with these two mean periods. Based on these ideas, the abrupt changes of climate variables are detected in terms of 5-yr moving seasonal and annual time series of temperature, precipitation, sunshine duration, and pan

210


evaporation in the upper Yellow River Basin. The t statistics of mean temperature, precipitation, and sunshine duration in the upper Yellow River Basin are at the 1% level of significance in 1985, 1987, and 1982, respectively (see Figs.6a, b, and c). It shows that an abrupt warming occurred in the late 1980s in the study area, which is somewhat consistent with that obtained by You (1998). An abrupt change of the annual precipitation occurred in the mid-1980s and an abrupt change of the sunshine duration took place in the early 1980s. Although the abrupt changes of the mean temperature happened in spring of the early 1970s and winter of the late 1960s and early 1970s, there is no statistical significance. In the same way, the abrupt changes of precipitation occurred in summer of the 1970s and in winter of the late 1960s and early 1970s, but it is not up to the corresponding level of significance. Figure 6d shows that there is an obvious increasing trend for annual pan evaporation in the 1960s-1980s, and there is an abrupt change from high value to low values in the early 1980s. In addition, the t statistic is over 1% level of significance (negative) in the early 1990s, i.e., there is also a significant abrupt change from low to high values for annual evaporation in the same period. The abrupt changes detected by using Yamamoto method are listed in Table 1. Two mean periods of time n=10 and n=14 are used in this study. It is shown

VOL.21

that there are abrupt changes for sunshine duration in all seasons except autumn. For other three climate variables, the abrupt changes occurred in all four seasons, especially acute jumps for annual mean temperature, winter precipitation, and spring pan evaporation. In the 1960s, the abrupt changes occurred for the mean air temperature, precipitation, and evaporation in the upper Yellow River Basin, of which the abrupt changes of the evaporation occurred in spring, winter, and the whole year, but only in winter for the mean air temperature and precipitation. It is mainly because the time series are too short in this study. However, it can also be understood that the abrupt changes of the climate variables occurred in the upper Yellow River Basin in the 1960s. In different periods of the 1970s, the abrupt changes were detected for the mean air temperature, precipitation, sunshine duration, and evaporation in the upper Yellow River Basin, in which the abrupt changes were detected in four seasons and the annual series for evaporation, and in the summer and winter for precipitation. One acute abrupt change was detected in the winter of 1971, with signal-noise ratio (S/N ) of 2.1. In addition, some abrupt changes were detected in the same period for spring temperature and summer sunshine duration. Compared with the abrupt changes in the 1960s and 1970s, the climate jumps in the 1980s are most significant. These results are similar to the results detected in the

Fig.6. The moving t-statistic curve of the climatic factors in the upper Yellow River Basin. (a) Temperature, (b) precipitation, (c) sunshine duration, and (d) evaporation; dashed lines: α=0.01.

NO.2

211


Qinghai-Tibetan Plateau (Tang et al., 1998). The abrupt changes were examined for major climate variables in different seasons and annual series of the 1980s, except for the spring mean temperature, summer precipitation, summer sunshine duration, and summer and winter evaporations. About 47% of the maximum values of S/N happened in this period. In addition, the acute abrupt changes happened for spring evaporation in 1981 and for the annual air temperature in 1985, and the values of S/N are 2.19 and 2.02, respectively. Figure 7 shows the S/N values of annual time series for four climatic factors. The phases of S/N values for the pan evaporation abrupt changes are obviously ahead of other climate factors. The abrupt changes of the evaporation mainly occurred in the 1970s. However, the phases of S/N values for other factors mainly occurred in the 1980s. Different climate factors were detected with different S/N values of the abrupt changes. The abrupt changes can be detected in 1971-1987 simultaneously using the two mean periods of time, with the similar results. However, the abrupt changes of climate factors can not be detected before 1971 and after 1987 when n=14 because of the

limited time series. For the upper Yellow River Basin, the abrupt changes of four climate factors did not exhibit consistent rules. This is mainly due to the different seasonal changes for four climate factors. For example, the abrupt changes of the precipitation, sunshine duration, and evaporation showed better association in the mid-1970s, i.e., the summer precipitation increased and the summer sunshine duration and evaporation decreased. Table 2 lists the abrupt changes for regional annual and seasonal time series detected by using MannKendall method for four climate factors in the upper Yellow River Basin, with the corresponding years shown in Fig.8. It is shown that the abrupt changes of mean air temperature mainly occurred in the 1990s, which is corresponding to the results for global warming period of 1990-1991 (Wei et al., 1995). The abrupt changes of autumn precipitation were detected in 1986, which is similar to the results obtained by Yang and Li (2004a). In addition, the abrupt changes of evaporation mostly occurred in the 1960s and the abrupt changes of the sunshine duration primarily took place in the 1980s.

Table 1. Climate jumps detected by using Yamamoto method in the upper Yellow River Basin during the period of 1960-2001 Time

n values Periods

Spring

Summer

Years of max.S/N Periods Years of max.S/N Periods

Autumn

Winter

Annual

Mean temperature

Precipitation

n=10 n=14 n=10 1973-1974 1973-1975 1981-1983

n=14

Sunshine duration n=10 n=14 1982-1986 1981-1985

Pan evaporation n=10 1967-1968,1971

1989-1991 1991 1979-1983,1991 1.14(1974) 1.39(1974) 1.4(1982) 1.71(1984) 1.62(1983) 1.54(1967), 1.04(1971) 1.89(1991) 1.14(1991) 2.19(1981), 1.53(1991) 1985-1991 1986-1987 1972-1976 1973-1974 1973-1979 1976-1980 1972-1981, 1989-1991 1.37(1987) 1.37(1987) 1.28(1973) 1.06(1974) 1.23(1973) 1.15(1979) 1.7(1973), 1.22(1991) 1984-1988 1979-1987 1983-1988 1983-1986

1991 Years of 1.65(1986) 1.45(1987) 1.75(1984) max.S/N 1.10(1991) 1969-1970 1983-1987 1969-1974 Periods 1983-1986 1985-1988 Years of 1.27(1970) 1.61(1985) 2.1(1971) max.S/N 1.65(1985) 1.85(1986) 1984-1988 1984-1987 1985-1991 Periods 1991 Years of 2.02(1985) 1.43(1987) 1.59(1987) max.S/N 1.03(1991)

1972,1984-1985

1.42(1985) 1972-1974 1982-1986

1982

1986 1.37(1973) 1.48(1985) 1.01(1982) 1.00(1986) 1986-1987 1981-1985 1980-1983

n=14 1971-1974 1978-1983 1.22(1971) 1.66(1981) 1971-1980 1.7(1973) 1.7(1974) 1971-1972

1990-1991 1985-1987 1.03(1972),1.13(1984) 1.09(1971) 1.15(1991) 1.13(1987) 1967-1974 1971-1973 1.33(1971)

1.44(1971)

1967,1971-1975

1971-1981

1978-1982,1990-1991 1.68(1987) 1.32(1982) 1.45(1981) 1.12(1967), 1.33(1973) 1.45(1973) 1.61(1981), 1.61(1991) 1.45(1980)

212


VOL.21

Fig.7. S/N values of annual time series for four climatic factors for (a) n=10 and (b) n=14.

The comparison of the results obtained by Ya-

those detected by using Yamamoto method, in which

mamoto method with those obtained by Mann-

the abrupt changes could not be detected by using

Kendall method exhibited that some of the abrupt

Mann-Kendall method for mean air temperature in the

changes detected by using two methods were quite

1980s, sunshine duration in the 1970s, and pan evap-

similar, especially the sunshine duration in the 1980s.

oration in the 1970s and 1980s. Different detection

However, some abrupt changes detected by using

methods may get different results, and each method

Mann-Kendall method, including pan evaporation in

has its own virtues and shortages (Fu et al., 1992; Yan

the 1960s and mean temperature and precipitation in

et al., 2001). Figure 8 shows the abrupt changes of cli-

the 1990s, could not be detected by Yamamoto method

matic factors detected by using Mann-Kendall method

because of the limited data. The abrupt changes de-

at a significance level of 5% in the study area.

tected by using Mann-Kendall method are only 50% of Table 2. The years with climate jump detected by using Mann-Kendall method in the upper Yellow River Basin Time

Mean temperature

Precipitation

Pan evaporation

Sunshine duration

Spring

1998

-

1968

1984

Summer

1997

-

1965

1977

Autumn

1995

1986

-

-

Winter

1986

1975

1969

1982

Annual

1994

1994

1965

1981

Fig.8. Jump of climatic factors detected by using Mann-Kendall method in the upper Yellow River Basin. (a) Temperature, (b) precipitation, (c) sunshine duration, and (d) evaporation; dashed lines: α=0.05.

NO.2


6. Discussions and conclusions Some conclusions can be drawn by analyzing the climate data in the upper Yellow River Basin from 1960 to 2001. (1) Mean air temperature: There are two increasing centers in the study area: one is near Qiabuqia Station in the north, and the other is near Lanzhou Station in the east. The Kendall slope of the mean air temperature is 0.18◦ C/(10 yr) in the whole area, increasing by 0.76◦ for 42 yr. The winter temperature has the greatest contribution to the annual total. The detection results show that there is an obvious abrupt warming occurring in the late 1980s. An acute abrupt change was detected by using Yamamoto method in 1985. (2) Precipitation: The climate exhibits a dry tendency in the upper Yellow River Basin since the 1960s. The mean Kendall slope of precipitation is −4.26 mm/(10 yr) for the whole basin. The winter precipitation has the greatest contribution to the annual total. The detecting results show that there is an obvious abrupt change for precipitation in the mid1980s, changing from wet to dry. (3) Sunshine duration: The Kendall slope of annual sunshine duration is −29.9 h/(10 yr). There are two obvious periods: one is the high period of 19611980, and the other is the low period of 1981-1996. An obvious abrupt change occurred in the early 1980s, changing from high to low value. (4) Pan evaporation: The annual evaporation decreased by 161.3 mm for the past 42 years, in which the spring and summer evaporation have great contribution to the annual evaporation. In the results obtained by using MTT method, two abrupt changes occurred in the 1980s and the early 1990s. The test results obtained by using the Yamamoto method show that the abrupt changes of evaporation mainly occurred in the 1970s. According to the Mann-Kendall method, the abrupt changes mostly occurred in the 1960s. In conclusion, there is a warm and dry tendency in the upper Yellow River Basin for the past 42 years, i.e., increasing temperature and decreasing precipitation. One interesting phenomenon is that there is

213

a decreasing trend for evaporation in the study area during the past 42 years, although the temperature is increasing. This may result from possible climate change or the impact of human activities. It should be pointed out that it is not easy to distinguish from abrupt changes and monotonic trends, and thus further investigation is required to identify these trends more precisely. However, it is confident that the approaches presented in this paper may be useful tools for further examining the impact of climate change on hydrological processes. REFERENCES Ding Yihui and Dai Xiaosu, 1994: Temperature variation in China during the last 100 years. Meteorological Monthly, 20(12), 19-26. (in Chinese) Fu Congbin and Wang Qiang, 1992: The definition and detection of the abrupt climatic change. Scientia Atmospherica Sinica, 16(4), 482-493. (in Chinese) Li Chunhui, 2003: The reproducible evaluation of the surface water resources in the Yellow River Basin. Dissertation, Beijing Normal University. (in Chinese) Li Daofeng, 2003: Impact of climate and land-cover change on runoff of the source regions of the Yellow River. Dissertation, Beijing Normal University. (in Chinese) Li Yueqing, 2001: A phase space EOF method and its application to climate diagnosis. Plateau Meteorology, 20(1), 88-93. (in Chinese) Liu Changming and Zheng Hongxing, 2004: Changes in components of the hydrological cycle in the Yellow River Basin during the second half of the 20th century. Hydrological Processes, 18, 2337-2345. Liu Changming and Zheng Hongxing, 2003: Trend analysis of hydrological components in the Yellow River Basin. Journal of Natural Resources, 18(2), 129135. (in Chinese) Shi Yafeng, 1996: Chinese Historical Climatic Changes. Shandong Science & Technology Press, 443-467. (in Chinese) Tang Maocang, Bai Chongyuan, Feng Song, and Caiying, 1998: Climate abrupt change in the QinghaiTibetan Plateau in recent century and its relation to astronomical factors. Plateau Meteorology, 17(3), 250-257. (in Chinese) Wang Shaowu, 1997: Scientific intersection of PAGES and CLIVAR. Acta Meteorologica Sinica, 55(6),

214


662-669. (in Chinese) Wei Fengying and Cao Hongxing, 1995: Detection of abrupt changes and trend prediction of the air temperature in China, the Northern Hemisphere and the globe. Scientia Atmospherica Sinica, 19(2), 140-148. (in Chinese) Wei Fengying, 1999: Modern Technology of Statistics, Diagnosis and Forecast for Climate. China Meteorological Press, 62-76. (in Chinese) Xu Z. X., K. Takeuchi, and H. Ishidaira, 2002: Long-term trends of annual temperature and precipitation time series in Japan. Journal of Hydroscience and Hydraulic Engineering, 20(2), 11-26. Xu Z. X., K. Takeuchi, and H. Ishidaira, 2003: Monotonic trend and step changes in Japanese precipitation. Journal of Hydrology, 279, 144-150. Yan Minhua, Deng Wei, and Chen Panqin, 2003: Analysis of climate jumps in the Sanjiang Plain. Scientia Geographica Sinica, 23(6), 661-667. (in Chinese) Yan Minhua, Deng Wei, and Ma Xuehui, 2001: Climate variation in the Sanjiang Plain disturbed by large scale reclamation during the past 45 years. Acta Geographica Sinica, 56(2), 159-170. (in Chinese) Yan Zhongwei, Ji Jinjun, and Ye Duzheng, 1990a: The climatic jump of the Northern Hemisphere in summer during the 1960s. I: Precipitation and temperature change. Chinese Science (B), 1, 97-103. (in Chinese) Yan Zhongwei, Ji Jinjun, and Ye Duzheng, 1990b: The climatic jump of the Northern Hemisphere in summer during the 1960s. II: Sea level pressure and 500 hPa height change. Chinese Science (B), 8, 879-885. (in Chinese) Yan Zhongwei, 1992: A primary analysis of the process of the 1960s Northern Hemispheric summer climatic

VOL.21

jump. Scientia Atmospherica Sinica, 16(1), 111119. (in Chinese) Yang Lianmei, 2003: Climate change of extreme precipitation in Xinjiang. Acta Geographica Sinica, 58(4), 577-583. (in Chinese) Yang Yanwu, Yu Qiang, and Wang Jing, 2004: Spatiotemporal variations of principal climatic factors in North China and part of East China within the past 40 years. Resources Science, 26(4), 45-50. (in Chinese) Yang Zhifeng and Li Chunhui, 2004a: Abrupt and periodic changes of the annual natural runoff in the subregions of the Yellow River. Journal of Mountain Science, 22(2), 140-146. (in Chinese) Yang Zhifeng and Li Chunhui, 2004b: The spatial and temporal structure of precipitation in the Yellow River Basin. Progress in Geography, 23(2), 27-33. (in Chinese) You Weihong, 1998: Multi-scale Diagnosis of Climatic Change and Some Technologies and Methods for Its Forecast. China Meteorological Press, 1-17. (in Chinese) Yu, P. S., T. C. Yang, and C. K. Wu, 2002: Impact of climate change on water resources in southern Taiwan. Journal of Hydrology, 260, 161-175. Zeng Yan, 2004: Distributed modeling of actual evapotranspiration over rugged terrain of the Yellow River Basin. Dissertation, Chinese Academy of Sciences. (in Chinese) Zhou Shunwu, Jia La, and Du Jun, 2001: Analyses of climatic trend and jump over the middle reaches of Yarlung Tsangpo River in the Tibetan Plateau in recent 42 years. Plateau Meteorology, 20(1), 71-75. (in Chinese)