Impact assessment of climate change and human activities on annual highest water level of Taihu Lake

Transcription

1 Water Science and Engineering, 2009, 2(1): 1-15 doi: /j.issn e-ail: Ipact assessent of cliate change and huan activities on annual highest water level of Taihu Lake Qing-fang HU* 1, 2, Yin-tang WANG 2 1. Departent of Hydraulic Engineering, Tsinghua University, Beijing , P. R. China 2. Nanjing Hydraulic Research Institute, Nanjing , P. R. China Abstract: The annual highest water level of Taihu Lake (Z ) is very significant for flood anageent in the Taihu Basin. This paper first describes the inter-annual and intra-annual traits of Z fro 1956 to Then, using the Mann-Kenall (MK) and Spearan (SP) nonparaetric tests, the long-ter change trends of area precipitation and pan evaporation in the Taihu Basin are deterined. Meanwhile, using the Morlet wavelet transforation, the fluctuation patterns and change points of precipitation and pan evaporation are analyzed. Also, huan activities in the Taihu Basin are described, including land use change and hydraulic project construction. Finally, the relationship between Z, the water level of Taihu Lake 30 days prior to the day of Z (Z 0 ), and the 30-day total precipitation and pan evaporation prior to the day of Z (P and E 0, respectively) is described based on ulti-linear regression equations. The relative influence of cliate change and huan activities on the change of Z is quantitatively ascertained. The results deonstrate that: (1) Z was distinctly higher during the period than during the period, and the 30 days prior to the day of Z are the key phase influencing Z every year; (2) P increased significantly at a confidence level of 95% during the period, while the reverse was true for E 0 ; (3) The relationship between Z, P and E 0 distinctly changed after 1980; (4) Cliate change and huan activities together caused frequent occurrences of high Z after 1980; (5) Cliate change caused a substantially greater Z difference between the and periods than huan activities. Cliate change, as represented by P and E 0, was the doinant factor raising Z, with a relative influence ratio of 83.6%, while huan activities had a saller influence ratio of 16.4%. Key words: cliate change; huan activities; annual highest water level; Taihu Lake 1 Introduction Cliate and huan activities are two eleentary factors in the hydrologic process. Cliate deterines the ain inputs to the hydrologic circle, such as rainfall and radiation. Huan activities have persistent influences on the land surface or directly interfere with the transforation and distribution of runoff, sedient and so on, so the hydrologic process becoes ore coplicated. In the past fifty years, both global cliate change and intensive huan activities have This study was supported by the National Key Technology R & D Progra of the Ministry of Science and Technology of China (Grant No. 2006BAB14B01) and the Innovation Progra of Science and Technology of the Ministry of Water Resources of China (Grant No. XDS ). *Corresponding author (e-ail: hqf_work@163.co) Received Nov. 2, 2008; accepted Jan. 12, 2009

2 caused a rearkable influence on hydrology and water resources all over the world (IPCC WGI 2007; Ding 2008; Wang et al. 2004). Detecting and evaluating these changes, and especially scientifically distinguishing the ipacts of cliate change and huan activities, is a challenge for hydrologists and eteorologists. Many approaches and odels have been established for the purpose of investigating the operating echaniss aong hydrologic processes, cliate change and huan activities and predicting the future hydrologic scenario under changing conditions. These range fro coplicated atospheric-hydrologic coupled odels and distributed hydrologic odels (Milly et al. 2005; Mori et al. 2008) to relatively siple luped hydrologic odels (Guo et al. 2002; Wang et al. 2006) and statistical ethods (Burn and Hag Elnur 2002; Xu 2005; Zhang et al. 2001; Liu 2007). This study aied to investigate cliate change and huan activities in the Taihu Basin fro 1956 to 2000 and assess their ipacts on Z. The value of Z is of great iportance for flood control and anageent of this basin, so its variability in response to cliate change and huan activities is worth our attention. This paper begins with a brief description of the Taihu Basin and the data used in this study. Then the ain ethods used to analyze variation of precipitation and pan evaporation in the Taihu Basin and the quantitative assessent of their influence on Z are presented. Finally, the results are discussed and soe conclusions are suarized. 2 Study area and data availability 2.1 Study area The Taihu Basin, which has an area of about k 2, is located in the Yangtze River Delta in southeast China (Fig. 1). The cliate is a typical sei-tropical onsoon cliate with an annual pan evaporation fluctuating around 820 and annual rainfall varying fro 1050 to About 60 percent of the annual precipitation is concentrated in the rainy season fro May to Septeber. As one of the ost developed regions in China, its population density exceeds persons per k 2 and its urbanization ratio reached 66.5% in Located in the center of the basin, Taihu Lake is the third largest fresh water lake in China with a surface water area of about k 2. The storage capacity of Taihu Lake akes up about 80 percent of that of all the lakes in this basin, so it is the ost iportant flood storage reservoir in the basin. 2.2 Data availability There is a an extensive rainfall onitoring network coposed of 115 rain gauges in the Taihu Basin (Fig. 1). Fro these gauges, area rainfall data over the basin fro 1956 to 2000 at a daily teporal resolution can be obtained. There are nine total pan evaporation stations providing daily pan evaporation data. The daily series of Taihu Lake s water level is coputed by the five gauges shown in Fig. 1. However, daily water level records are issing for soe of 2 Qing-fang HU et al. Water Science and Engineering, Mar. 2009, Vol. 2, No. 1, 1-15

3 the years fro 1956 to 2000, naely , 1974, 1981 and , so the available water level data actually covers thirty-seven years. This is true for daily precipitation and pan evaporation, but the annual and rainy season total values of the for these years are available. Fig. 1 River network and hydroetric stations in Taihu Basin Land use data and socioeconoic indexes fro the 1950s were also collected. The land use area data of three different periods are shown in Table 1, while population and the gross doestic product (GDP) can be seen in Table 2. Generally, changes of land use and socioeconoic indexes anifest the strength of huan activities which have becoe ore and ore intense since the 1950s. Paddy field Table 1 Land use type in Taihu Basin over different periods Dry land and non-plowland Area (k 2 ) Ratio (%) Construction land Surface water area Paddy field Dry land and non-plowland Construction land Surface water area 1950s s Table 2 Socioeconoic indexes in Taihu Basin Population (10 4 ) Urbanization ratio (%) GDP (10 8 RMB) 1950s < s s Methodology 3.1 MK and SP nonparaetric tests The MK and SP nonparaetric tests are two effective approaches for detecting possible trends of hydrologic and cliatologic variables (Ding and Deng 1988; Yue et al. 2002; Liu 2007; Burn and Hag Elnur 2002). These two ethods were used to inspect the trends of Qing-fang HU et al. Water Science and Engineering, Mar. 2009, Vol. 2, No. 1,

4 precipitation and pan evaporation in the Taihu Basin. 3.2 Wavelet transforation Wavelet transforation is a ulti-tie-scale analysis ethod for tie series (Wang et al. 2005), which can be used to analyze evolutionary oscillations at various tie scales. Kuar and Foufoula-Georgiou (1993) introduced this ethod into spatial rainfall field analysis. It has been widely used to analyze the variation traits of different hydrologic factors, such as runoff, precipitation, sedient flux, and water level. In our study, we applied the Morlet continuous wavelet (Kuar and Foufoula-Georgiou 1993; Wang et al. 2005) to identify the breakpoints and periodicity of precipitation and pan evaporation series. 3.3 Ordinal clustering Regarding the intervention of cliate change and huan activities, the relationship between Z, basin precipitation, and potential evapotranspiration (represented by pan evaporation) changed during the period fro 1956 to Therefore, identifying the change point is iportant for reasonable evaluation of the ipact on Z of cliate change and huan activities. This could be considered an ordinal clustering process for ulti-diensional tie series. There are any kinds of ordinal clustering ethods for ulti-diensional tie series clustering, including the fuzzy set analysis (Chen 1998) and Fisher s ethod (Liu et al. 2007). Here, a syste clustering ethod is proposed to search for the change point for the relationship between Z, precipitation and pan evaporation. For the 4D tie sequence consisting of the four variables below, the splitting point is deterined by Eq. (1): 2 2 t n St () = in i t + i n t 1 t n 1 Y Y Y Y (1) i= 1 i= t+ 1 where Yi Yt and Yi Y n t represent the Euclidian distance, n is the nuber of total saples aounting to 37, St ( ) is the variance suation of the two classes partitioned by point t, and Y i is a 4D vector defined as follows: ( i i i 0i 0i T Y = Z, P, E, Z ) (2) where Z i is the highest water level in year i; P i is the total precipitation for the 30 days prior to the day of Z i ; E 0i is the total pan evaporation spanning the sae period as P i ; and Z 0i is the water level of Taihu Lake 30 days prior to the day of Z i. The ean values of these two classes are calculated as follows: 1 t t i= 1 Yt = Y i (3) 1 n Yn-t = n Y (4) t =+ i i t 1 The reason that Z 0, P and E 0 are chosen as the ain factors influencing Z will be 4 Qing-fang HU et al. Water Science and Engineering, Mar. 2009, Vol. 2, No. 1, 1-15

5 illustrated in the following section. According to Eq. (1), when S(t) reaches the iniu value, the splitting point t is chosen and the relationship between Z and the three influencing factors is distinctly different before and after t. 3.4 Multi-linear regression Multi-linear regression is a classical statistical approach principally based on the least squares approach and the Gauss-Markov assuption (Wang 1999). A prerequisite for eploying ulti-linear regression is that the correlation between the selected variables is weak. After the splitting point t described above is selected, using this ethod, the quantitative relationship between Z and the three influencing factors before and after t is established through the following equation: Z = β + α P + α E + α Z (5) i 1 i 2 0i 3 where α 1, α 2, and α 3 are the regression coefficients of P i, E 0i and Z 0i, respectively, and β is a constant. Eq. (5) can be further written in vector for: (,,, ) Z = I P E Z 0 A 0 = (,,, ) T, Z ( ) T Z ( Z Z L Z ) T where E0 E01 E 02 L E0n 0 = Z01, Z 02, L, Z0n, =1, ( 1,,1) T I L, P ( P P L P ) T,and = ( β, α, α, α ) T = 1, 2,, n A. 0i =,,, n 1 2 Based on Eq. (5) and Eq. (6), the individual effects of P, E 0 and Z 0 on Z can be assessed and copared. The regression coefficients and constants are derived with the following equation: 1 T T I I T T P * T (,, 0, 0 ) P A= I P E Z Z T (7) E0 E0 T T Z0 Z0 For the purpose of illustration, c is uniforly adopted as the unit of P, E 0, Z 0 and Z. The constant β also has a unit of c while the other three coefficients are diensionless. Physical eanings of the four regression coefficients can be interpreted by deconstructing Eq. (5) into two functions as follows: f = β + α Z (8) ϕ = α1pi + α2e 0 i (9) The function f describes the relation between Z and Z 0 when P and E 0 approxiate zero. The paraeters β and α 3 together reflect the integrative relation between Z and Z 0 under the influence of other iplicit factors, especially water concentration and drainage of Taihu Lake due to huan regulation. The function ϕ stands for the contribution of P and E 0 to Z under certain land surface and huan activity conditions. The eaning of α 1 can be explained as the increent of Z with just 1 c change of P while other factors reain constant. The paraeter α 2 has a eaning siilar to that of α i (6), Qing-fang HU et al. Water Science and Engineering, Mar. 2009, Vol. 2, No. 1,

6 3.5 Evaluation ethod for driving factors of Z The driving factors of Z can be generally classed into cliate change and huan activities as stated above. In this paper, the influence of cliate change is ainly considered P and E 0 while other influences are all considered huan activities. As a statistical odel, the regression equation is still useful for evaluating the ipacts of cliate change and huan activities and can estiate future changes of Z under certain P and E 0 scenarios. Actually, regression coefficients of Eq. (5) iplicitly and synthetically describe the interaction echaniss between the land surface, huan activities and Z under certain P and E 0 conditions. If the linear regression functions of the two different periods before and after t are respectively denoted by F 1 and F 2, then the total difference in the ean Z ( Z ) between the two periods is calculated as: Δ = F X F X (10) where The variables P 1, 01 ( ) ( ) ( P, E, Z ) T ( P, E, Z ) T X = (11) X = (12) E and Z 01 respectively indicate the ean values of P i, E 0i and Z 0i ( i = 1, 2, L, t ) during the first period, and P 2, E 02 and Z 02 respectively indicate the ean values of P,,and Z ( i = t + 1, t+ 2, L, n) during the second period. The partial i E0i 0i difference of Z caused by cliate change and by huan activities can be respectively calculated as follows: ( P, E ) ( P, ) Δc = ϕ ϕ2 1 E 01 (13) Δh = Δ Δc (14) where Δ c is the increent exerted only by cliate change, Δ h is the increent exerted only by huan activities, and ϕ 2 is the function of ϕ in the second period. Thus, the relative influence of cliate change and huan activities on Z can be coputed with the following forulas: Δc ρc = Δc + Δh (15) Δh ρh = Δc + Δh (16) where ρc and ρ h are the relative influence of cliate change and huan activities, respectively. Fro Eq. (15) and Eq. (16), we can postulate that if ρ c is positive, cliate change raises Z ; and if ρc is negative, cliate change reduces Z. This situation is true for huan activities and for ρ h. 4 Results and discussion 4.1 Variability of water level of Taihu Lake Based on the water level records at a daily resolution fro 1956 to 2000, intra-annual and inter-annual variability of Taihu Lake was investigated and suarized. Water level variation 6 Qing-fang HU et al. Water Science and Engineering, Mar. 2009, Vol. 2, No. 1, 1-15

7 of Taihu Lake is rather coplicated. The duration of the water level rise and the date of Z were different every year. However, through analysis of the water level curves of 37 years, it was found that the doinant period of water level rise started about 30 days before Z. This is especially true for soe typical flood years (Fig. 2) and it is the reason why we conclude that P, E 0 and Z 0 are the ain factors influencing Z. Fig. 2 Rising process of water level of Taihu Lake in typical years The Z value of the period fro 1956 to 2000 was 3.74, calculated with the 37-year available records. Anoalous values of Z are shown in Fig. 3(a). Although the records of Z for several years were issing, we can still see that Z had an evident ascending trend over these years. High Z values ostly appeared after The Z value of the latter period was 0.42 higher than that of the previous period. Z was particularly high in the 1990s, indicating that the Taihu Basin suffered severe flood hazards. Additionally, the initial water level Z 0 is shown in Fig. 3(b). With alost no trend, the annual variation of Z 0 is distinctly different fro that of Z. Fig. 3 Anoalous values of Z and Z 0 of Taihu Lake fro 1956 to Dynaics of precipitation and pan evaporation Tie series of precipitation and pan evaporation in the Taihu Basin are presented in Fig. 4. Annual precipitation and pan evaporation are denoted by PY and EY while rainy season precipitation and pan evaporation are denoted by PX and EX, respectively. It can be seen in Fig. 5 that precipitation at different tie steps varied with an upward incline fro 1956 to 2000, while the reverse was true for pan evaporation. Also, standardized curves of precipitation and pan evaporation are drawn in Fig. 5. The standardization forula used is Qing-fang HU et al. Water Science and Engineering, Mar. 2009, Vol. 2, No. 1,

8 P si P P i = (17) σ P where P si is the standardized P i, and P and σ P are the ean and standard deviation of P, respectively. Eq. (17) is also adapted for standardizing PY, EY, PX, EX and E 0. After the effects of ean value and standard derivation are eliinated, we can see that although there are soe differences, the ain changes of P are consistent with PY and PX. This is siilarly true for E 0, EX and EY (Fig. 5). Fig. 4 Tie series of precipitation and pan evaporation and their linear trends fro 1956 to 2000 Fig. 5 Standardized curves of precipitation and pan evaporation fro 1956 to Qing-fang HU et al. Water Science and Engineering, Mar. 2009, Vol. 2, No. 1, 1-15

9 By eans of MK and SP nonparaetric tests and Morlet wavelet transforation, the dynaics of precipitation and pan evaporation were further analyzed. Table 3 shows the test results of these series. Change trends of precipitation and pan evaporation were rather different. For precipitation series, only P had a significant trend at a confidence level of 95%, though all of the tended to ascend. However, the pan evaporation series all had significantly descending trends at the confidence level of 95%, which was alost the opposite of what was observed of the precipitation sequences. At annual and rainy season resolutions, the absolute change ratio of pan evaporation even exceeded that of precipitation. This deonstrates the fact that the decrease of pan evaporation was a crucial aspect of cliate change in the Taihu Basin fro 1956 to Table 3 Nonparaetric tests for change trends of precipitation and pan evaporation Category Precipitation Pan evaporation Series Series length Change ratio MK SP Mean () (year) (/year) trend trend PY PX P EY EX E Note: The confidence level of nonparaetric tests is 95%; 1 eans the trend is statistically significant and 0 eans the trend is statistically insignificant. Fig. 6 presents the Morlet wavelet transforation coefficients of PY, PX, EY and EX. Fig. 7 is the wavelet variance curve of the four sequences. Wavelet coefficient aps clearly describe the fluctuation process of precipitation and pan evaporation at different tie scales fro 1 to 45 years, and also detect the possible change points with zero isolines. Wavelet variance curves of the four series anifest the principal periods that correspond to local peaks. Together with these, a basic evolutionary pattern is deduced, that PY and PX approxiately traversed three different phases at tie scales of 33 years and 28 years, respectively. The two change points eerged in the initial years of the 1960s and the 1980s. Siilarly, it can be inferred that both EY and EX varied through two different phases at a tie scale of 42 years. The years 1980 and 1979 are the corresponding change points of EY and EX. Considering these findings, PY and PX will possibly go through a dry phase after 2003, while EY and EX will presuably go through a phase of abundant rainfall after Detailed suaries of the wavelet analysis are provided in Table 4. Wavelet analysis of P and E 0 was not conducted for issing data. However, the evolutionary rules and future scenario are likely siilar to the annual and rainy season series due to their consistency (Fig. 5). The evolution of P and E 0 over the 45 years is also shown in Table 4. To su up, the abundance of precipitation and diinishent of pan evaporation fro the 1980s to 2000 likely together brought on higher Z. Qing-fang HU et al. Water Science and Engineering, Mar. 2009, Vol. 2, No. 1,

10 Fig. 6 Wavelet coefficient isolines of precipitation and pan evaporation fro 1956 to 2000 Fig. 7 Wavelet variance curves of precipitation and pan evaporation 4.3 Effects of huan activities The effects of huan activities on Z are rather coplicated, but can be analyzed fro two perspectives. On the one hand, with the persistent industrialization and urbanization process of the Taihu Basin, land use types were considerably transfored and the flood storage capacity was reduced. Apparently, these changes of land surface conditions in the Taihu Basin were iportant reasons for the frequent high value of Z fro the 1980s to On the other hand, with extensive construction of hydraulic projects and alteration of the river syste, the 10 Qing-fang HU et al. Water Science and Engineering, Mar. 2009, Vol. 2, No. 1, 1-15

11 odes of flood concentration and discharge were altered. Particularly with the establishent of any puping stations and drainage canals, the water quantity relationship between Taihu Lake and the other parts of the basin was adjusted. Generally, this helped to keep Z fluctuating within a proper range. As illustrated in Table 1, the ain feature of land use change is that the area of dry land and non-plowland has been decreasing whereas construction land area has been increasing. Fro the 1950s to 2000, the ratio of dry land and non-plowland to the basin area decreased by about 10%. Conversely, the construction area showed a considerable increase of 11% relative to the basin area. Therefore, the runoff aount increased even if precipitation and potential evapotranspiration did not change. Fro 1980 to 2000, the runoff volue increased by 4.11% because of land use change (Li et al. 2007). At the sae tie, the water area of the Taihu Basin showed a considerable reduction because of the vanishing of lakes. Thus, the flood storage space was cut down. This was also an iportant factor in high Z values. As shown in Table 5 (Li et al. 2006), the total water area of nine ajor lakes decreased by k 2 fro 1971 to Phase P scale (year) Table 4 Results of Morlet wavelet analysis of precipitation and pan evaporation Change point PY PX P Mean () P scale (year) Change point Mean () P scale (year) Change point , 1964, 1963, EY EX E 0 Phase P P P Change Mean Change Mean Change Mean scale scale scale point () point () point () (year) (year) (year) Note: P scale refers to principal scale tie. Table 5 Water area of ajor lakes in Taihu Basin in 1971 and 2002 Water area (k 2 ) Suation Taihu Ge Yangcheng Dingshan Tao Cheng Kuncheng Yuandang Dushu (k 2 ) Lake Lake Lake Lake Lake Lake Lake Lake Lake Siultaneously, the appearance of flood control systes entirely changed fro the 1950s to Thus, the water balance relation between Taihu Lake and the other parts of the basin was distinctly changed. In the western ountainous regions, several large-scale reservoirs were established with total flood control storage of about billion 3, while drainage systes coposed of puping stations, floodgates and canals were gradually developed in the plains region. Particularly after the great flood disaster of 1991, eleven key projects for flood regulation and drainage were established or reconstructed, including soe iportant drainage Mean () Qing-fang HU et al. Water Science and Engineering, Mar. 2009, Vol. 2, No. 1,

12 projects around Taihu Lake and in other sub-regions. The drainage capability of the plains region exceeded /s and floods could ore rapidly discharge into the sea. Therefore, the regulation ability of the entire basin was substantially intensified, helping to keep Z within an appropriate range. 4.4 Relation between Z, P and E 0 during different periods In order to quantitatively analyze the influence of precipitation and pan evaporation on Z by eans of ordinal clustering and least squares regression, two equations were derived fro Eq. (5). First, the 37 available saples of 4D vectors coposed of Z, P, E 0 and Z 0 were divided into two sections according to the ordinal clustering ethod referred to in Section 3.3. The year 1979 was selected as the splitting point (Fig. 8), so the 20 available saples fro the period of were grouped together in one part while the other 17 saples fro the period of were grouped together in another part. The rationality of choosing 1979 as the splitting point can be illustrated by Fig. 9, where dz represents the difference between Z and Z 0. The scatter plot of dz against P for the period is distinct fro that of the period. This situation is likewise true for the scatter plot distribution of dz against E 0. Fig. 8 Splitting points selected by ordinal clustering Fig. 9 Scatter plots of dz against P and dz against E 0 during different periods Table 6 presents regression coefficients of the linear regression equations for the and periods. Also, a coparison of observed and siulated values of Z is shown in Fig. 10, which verifies that the linear regression equations can well express the relation between Z and the three influencing factors. The obvious difference of regression coefficients between the two periods indicates that water quantity relationships in the Taihu 12 Qing-fang HU et al. Water Science and Engineering, Mar. 2009, Vol. 2, No. 1, 1-15

13 Basin have changed due to huan activities. Another point worth our attention is that the absolute partial differential of Z against E 0 exceeds P. Table 6 Regression coefficients of linear regression equations Regression coefficient β (c) α 1 α 2 α 3 Nash coefficient The evaluation results of effects on Z of cliate change and huan activities are shown in Table 7. Between the two periods, the total difference of Z caused by P and E 0 change and huan activities together was 42.4 c. Both cliate change and huan activities during the period generally raised Z as copared with the period. Cliate change partially increased Z by 35.5 c while the corresponding value was only 6.9 c for huan activities. Their relative influence on the change of Z was, respectively, 83.6% and 16.4%, so cliate change was evidently the doinant influencing factor of Z change. The net influence of huan activities on Z was rather sall relative to their high intensity. However, it can be seen through the interaction aong land use change, flood regulation easures and other factors introduced by huan society. Fig. 10 Scatter plots of observed and siulated Z during different periods Table 7 Evaluation results of effects on Z of cliate change and huan activities P (c) E 0 (c) Z 0 (c) Z (c) Δ (c) Δ c (c) Δ (c) h ρ c (%) ρ (%) h Conclusions Cliate change and huan activities have significantly influenced the annual highest water level of Taihu Lake fro 1956 to This study deonstrates that precipitation and potential evaporation in the Taihu Basin substantially changed after Meanwhile, huan activities also had iportant ipact on Z. Soe conclusions can be suarized: (1) The ean annual highest water level of Taihu Lake after the 1980s was higher than the ean value fro 1956 to The converse was true for the ean annual highest water level before the 1980s. Z fro 1980 to 2000 was 0.42, higher than that fro 1956 to The Qing-fang HU et al. Water Science and Engineering, Mar. 2009, Vol. 2, No. 1,

14 30 days prior to the day of Z are the key phase of the water level rise process in the rainy season every year, but the initial water level 30 days prior to the day of Z showed a rather weak incline fro 1956 to (2) Precipitation at different tie resolutions showed an ascending trend fro 1956 to 2000, while pan evaporation showed a descending trend. Based on the MK and SP nonparaetric tests, the change trend of P, E 0, EY, and EX was proven to be significant at a confidence level of 95%. The Morlet wavelet transforation detected the annual change point. The ain change point of rainy season precipitation and pan evaporation ostly occurred in the initial years of the 1980s. After 2000, precipitation in the Taihu Basin will gradually enter a dry period, while pan evaporation will have the reverse evolutionary rule. Also, P 0 and E 0 have siilar evolution traits to PX and EX, respectively. (3) Huan activities in the Taihu Basin have intensified. On the one hand, huan activities have transfored land use fro surface water area and dry plowland into construction land, which is an iportant factor in frequent occurrences of high Z. On the other hand, the extensive construction of flood drainage projects helps to aintain a proper value of Z. These two effects of huan activities counteract each other, so the net influence of huan activities on Z actually weakened over the study period. (4) The influence of cliate change and huan activities on Z was evaluated and assessed by eans of ulti-linear regression. Selecting P, E 0 and Z 0 as independent variables and Z as the deterined variable, regression equations for the two periods of and were established. Based on these, it was deterined that the partial difference of Z between these two periods caused by cliate change is 35.5 c while the corresponding value is just 6.9 c for huan activities. With a relative influence level of about 83.6%, cliate change was evidently the doinant driving factor of Z change fro 1956 to References Burn, D. H., and Hag Elnur, M. A Detection of hydrologic trends and variability. Journal of Hydrology, 255(1-4), Chen, S. Y Engineering Fuzzy Set Theory and Application. Beijing: National Defense Industry Press. (in Chinese) Ding, J., and Deng, Y. R Stochastic Hydrology. Chengdu: University of Science and Technology Press. (in Chinese) Ding, Y. H Huan activity and global cliate change and its ipact on water resources. China Water Resources, (2), (in Chinese) Guo, S. L., Wang, J. X., Xiong, L. H., and Yin, A. W A acro-scale and sei-distributed hydrological odel and cliate change ipact study in China. Journal of Hydrology, 268(1-4), Kuar, P., and Foufoula-Georgiou, E A ulticoponent decoposition of spatial rainfall fields: 1. segregation of large- and sall-scale features using wavelet transfors. Water Resources Research, 29(8), [doi: /93WR00548] Labat, D., Godderis, Y., Probst, J. L., and Guyot, J. L Evidence for global runoff increase related to cliate waring. Advances in Water Resources, 27(6), [doi: /j.advwatres ] Li, D. F., Tian, Y., and Liu, C. M Distributed hydrological siulation of the source regions of the 14 Qing-fang HU et al. Water Science and Engineering, Mar. 2009, Vol. 2, No. 1, 1-15

15 Yellow River under environental changes. Acta Geographica Sinica, 59(4), (in Chinese) Li, H. P., Yang, G. S., and Jin, Y Siulation of hydrological response of land use change in Taihu Basin. Journal of Lake Sciences, 19(5), (in Chinese) Li, X. G., Jiang, N., Cao, K., and Lü, H Study on surface area change of ajor lakes in Taihu Basin based on topographical and reotely sensed inforation. Water Resources Protection, 22(3), (in Chinese) Liu, C. Z The advances in studying detection of streaflow trend influenced by cliate change. Advances in Earth Science, 22(8), (in Chinese) Liu, K. L., Wang, Y. T., Hu, S. Y., and Gao, B Application of Fisher optial dissection ethod to flood season division. Advances in Science and Technology of Water Resources, 27(3), (in Chinese) Milly, P. C. D., Dunne, K. A., and Vecchia, A. V Global pattern of trends in streaflow and water availability in a changing cliate. Nature, 438, [doi: /nature04312] Mori, E., Kojiri, T., and Tanaka, K Ipact assessent of cliate considering reservoir operation and water resources circustances in the Yodo River. Proceedings of International Session for Annual Conference 2008 of Chinese Hydraulic Engineering Society, Haikou: China WaterPower Press. Morlet, J., Arens, G., Fourgeau, E., and Giard, D Wave propagation and sapling theory, Part II: Sapling theory and coplex waves. Geophysics, 47(2), [doi: / ] Sun, J. H., Yan, Z. J., and Ni, S. H Report of Huan activities and Water Resources Evolutionary Rules in the Taihu Basin. Nanjing: Hydrology and Water Resources Departent, Nanjing Hydraulic Research Institute. (in Chinese) The Intergovernental Panel on Cliate Change, Working Group I. (IPCC WGI) Suary for policyakers. Cliate Change 2007: The Physical Science Basis. Cabridge: Cabridge University Press. Wang, G. Q., Zhang, J. Y., and He, R. M Ipacts of environental change on runoff in Fenhe river basin of the iddle Yellow River. Advances in Water Science, 17(6), (in Chinese) Wang, H., Wang, C. M., and Wang, J. H Theory of annual runoff evolution under natural-artificial dual ode and case study of Wuding River basin on the iddle Yellow River. Science in China (Series E: Technological Sciences), 34(s1), (in Chinese) [doi: S1-004] Wang, H. W Partial Least-Squares Regression Method and Application. Beijing: National Defense Industry Press. (in Chinese) Wang, S. P., Zhang, Z. Q., Sun, G., McNulty, S. G., Zhang, H. Y., Li, J. L., and Zhang, M. L Long-ter streaflow response to cliatic variability in the Loess Plateau, China. Journal of the Aerican Water Resources Association, 44(5), [doi: /j x] Wang, W. S., Ding, J., and Li, Y. Q Hydrology Wavelet Analysis. Beijing: Cheical Industry Press. (in Chinese) Xu, J. X The water fluxes of the Yellow River to the sea in the past 50 years, in response to cliate change and huan activities. Environental Manageent, 35(5), [doi: /s y] Yue, S., Pilon, P., and Cavadias, G Power of the Man-Kendall and Spearan s rho tests for detecting onotonic trends in hydrological series. Journal of Hydrology, 259(1-4), Zhang, J. Y., Zhang, S. L., Wang, J. X., and Li, Y Study on runoff trends of the six larger basins in China over the past 50 years. Advances in Water Science, 18(2), (in Chinese) Zhang, X. B., Harvey, K. D., Hogg, W. D., and Yuzyk, T. R Trends in Canadian streaflow. Water Resources Research, 37(4), [doi: /2000WR900357] Qing-fang HU et al. Water Science and Engineering, Mar. 2009, Vol. 2, No. 1,