- Email: [email protected]
Projected changes of precipitation extremes in river basins over China
Quaternary International 244 (2011) 149e158
Contents lists available at ScienceDirect
Quaternary International journal homepage: www.elsevier.com/locate/quaint
Projected changes of precipitation extremes in river basins over China Chonghai Xu a, b, Yong Luo b, Ying Xu b, * a b
Meteorological Observation Center, China Meteorological Administration, Beijing, China National Climate Center, China Meteorological Administration, Beijing, China
a r t i c l e i n f o
a b s t r a c t
Article history: Available online 13 January 2011
Based on daily precipitation data derived from observations and three Coupled General Circulation Model (CGCM)’s outputs (CSIRO_MK3_5, MPI_ECHAM5 and NCAR_CCSM3), some extreme precipitation indices are calculated. Initially, the models’ skills in simulating extreme precipitation during 1961e2000 were assessed. Projected changes before 2050 under the Special Report on Emissions Scenarios (SRES) emission scenario A1B were also analyzed. Results show that although there are some biases in the model results, the three models capture well the geographic distribution of extreme precipitation observed in the last half of the 20th century. Nevertheless, the models tend to show limited skill in reproducing the observed inter-annual variations of such extreme precipitation events, including the dominant patterns and trends. Under SRES A1B, for the period of 2011e2050, the simple daily precipitation intensity (SDII) and the fraction of extreme precipitation in total annual precipitation (R95T) will increase significantly in all the river basins in China, while changes in other indices are different. The periodic changes in consecutive dry days (CDD), consecutive wet days (CWD) and the number of days with precipitation 10 mm/day (R10) are expected to be more pronounced in the Huang-Huai-Hai River Basin. In the South River Basin, such periodic changes maintain similar or slightly weakened magnitudes. From 2001 to 2050 in the Huang-Huai-Hai River Basin, extreme precipitation in spring shows no significant changes. In the South River Basin, R95T, SDII and the maximum 5-days precipitation amount (R5D) in summer show an increasing tendency, but in winter a weak decreasing trend is projected. The changes in precipitation-based indices indicate a higher probability of heavy rainfall or flood occurrence, particularly in the river basins in East China. Ó 2011 Elsevier Ltd and INQUA. All rights reserved.
1. Introduction In recent decades, significant changes in global mean climate and extreme events have taken place, and most are attributed to the impact of human activity (IPCC, 2007). In the context of global warming, extreme climate events in China have also undergone a significant change, such as a longer duration of heat waves, increasing number of heavy rain events and increased precipitation intensity, as well as more frequent “record-breaking” events of extreme temperature. After the 1990s, there have been more disasters which covered large regions and caused great economic losses, especially associated with the expansion of drought areas in North China and the intensification of heavy floods in South China (Zhai et al., 1999; Yan and Yang, 2000; Qian et al., 2007). Since the 1980s, extreme precipitation events have been increasing in the middle and lower reaches of the Yangtze River basin (Sun and Gao, 1998; Jiang
* Corresponding author. E-mail address: [email protected] (Y. Xu). 1040-6182/$ e see front matter Ó 2011 Elsevier Ltd and INQUA. All rights reserved. doi:10.1016/j.quaint.2011.01.002
et al., 2005; Su et al., 2006; Su and Jiang, 2008), but in the Yellow and Haihe river basins there is a tendency towards drought (Qian and Lin, 2005). Extreme precipitation events in summer show an increasing trend in the northwest river and the Yangtze River basins, but only in the middle and lower reaches of Yangtze River Basin is there a significant increasing trend (Ren et al., 2000, 2004; Zhai and Pan, 2003; Zhai et al., 2005; Zhang et al., 2007a). The continuous improvement of observation techniques and the development of coupled climate models provide a strong basis for analysis and projection research on the extreme events in the future under different SRES emission scenarios (Nakicenovic and Swart, 2000). Most of the research has been focused on the Yangtze River Basin. Xu et al. (2003) estimated the mean change in temperature and precipitation for different scenarios based on five global climate models and a regional climate models (Gao et al., 2001); Zhang et al. (2007b, 2008) and Jiang et al. (2008) have projected changes in daily maximum precipitation using ECHAM5/ MPI-OM. Recently, Xu et al. (2009b) analyzed future changes of extremes indices derived from an ensemble of coupled general circulation models (CGCMs). In addition, Tian et al. (2008) analyzed
150
C. Xu et al. / Quaternary International 244 (2011) 149e158
the future climate change under the SRES A2 scenario in YangtzeHuaihe River Basin by RCM-PRECIS. Nevertheless, there are not enough studies done to comprehensively assess such changes for all of the river basins in China over the next several decades. Such studies are necessary in order to assess the affect of human activities on the spatial and temporal distributions of water resources, and to better estimate the future extreme events including the risk of droughts and flooding at a regional scale. In the World Climate Research Programme (WCRP)’s Coupled Model Intercomparison Project phase 3 (CMIP3) multi-model databases contributing to the Fourth Assessment Report of the Intergovernmental Panel on Climate Change (IPCC AR4), daily precipitation simulated for the climate of the 20th century and projections of the 21st century were made available. That provides an opportunity to analyze the future changes of extreme events in the river basins of China, including temporal and spatial changes. In this study, considering the similarity of climate changes on the temporal and spatial scales, research focused on four main river basins in China (Fig. 1): Song-Liao River Basin (including Songhua Jiang Basin and Liaohe Basin), Huang-Huai-Hai River Basin (including Huanghe Basin, Haihe Basin, and Huaihe Basin), South River Basin (including Yangtze River Basin, Zhu Jiang Basin, and Southeast River Basin), and Northwest River Basin. Previous studies indicated that CGCMs showed relatively reasonable performances in simulating the mean climate change over China, except for the mountain regions in the Tibet Plateau and the surrounding area where large deviations exist between simulations and observation (Zhou and Yu, 2006; Xu et al., 2007; Xu et al., 2009a,b; Jiang et al., 2009). Therefore, no analysis was done for the Southwest River Basin. This manuscript is organized as follows. Section 2 is used to introduce the data used in the study as well as the indices of extreme climate events analyzed for the river basins. Section 3
focuses on assessing the performance for the climate models in simulating the present-day extreme climate. In Section 4 projected annual changes in extreme precipitation are described based on the model ensemble (ME). Section 5 is focused on projected seasonal changes, and conclusion and discussions derived from this study are presented in Section 6.
2. Data and extreme indices The data used to calculate the extreme indices are the observed and simulated daily precipitation data. Observed data are the 1961e2002 gridded daily precipitation on a 0.5 latitudee longitude grid over East Asia, derived from gauge observations at over 2200 stations collected from the Global Telecommunication System (GTS) daily summary files archived by the National Oceanic and Atmospheric Administration (NOAA) Climate Prediction Center (CPC), a collection of daily precipitation observations at over 700 Chinese stations archived by the China Meteorological Administration (CMA), and the daily gauge data at over w1000 hydrological stations from the Chinese Yellow River Conservation Commission (YRCC) (Xie et al., 2007). For model simulations, three models were selected: CSIRO_MK3_5 (Commonwealth Scientific and Industrial Research Organisation, Australia, Gordon et al., 2002; Cai et al., 2003), MPI_ECHAM5 (Max Planck Institute for Meteorology, Germany, Roeckner et al., 2003; Marsland et al., 2003; Jungclaus et al., 2006), and NCAR_CCSM3 (National Center for Atmospheric Research, USA, William et al., 2006; Meehl et al., 2006). Each of these models offers a continuous daily-series data, including the simulation of climate of 20th century and SRES A1B scenario experiments of 21st century (mid-range emissions, CO2 concentration of about 700 ppm by 2100). In order to obtain the model
Fig. 1. Locations of the four major River Basins.
C. Xu et al. / Quaternary International 244 (2011) 149e158 Table 1 six extreme indices used in the study. Index
Definition
Unit
CDD CWD R10 R5D R95T
Maximum number of consecutive dry days (Rday < 1 mm) Maximum number of consecutive wet days (Rday 1 mm) Number of days with precipitation 10 mm d1 Maximum 5 d precipitation total Fraction of annual total precipitation due to events exceeding the 1961e1990 95th percentile Simple daily intensity index: annual total/number of Rday 1 mm d1
days days days mm %
SDII
mm/day
ensemble and do the comparison between simulations and observations, all data are interpolated onto a common 1 1 grid. Six extreme indices described by Frich et al. (2002) are calculated in this research as shown in Table 1. Applying the STARDEX Diagnostic Extremes Indices Software Version 3.3.1 designed to calculate the indices for stations (http://www.cru.uea.ac.uk/cru/projects/ stardex) produces those indices for the gridded area in China. The CDD, CWD and R10 indicate the changes in precipitation frequency, and R5D, R95T and SDII the changes of precipitation intensity. All indices mentioned in this paper were calculated on an annual and seasonal basis for observations during 1961e2002, and for the three models during 1961e2050 respectively. To assess the simulation capabilities of climate models, the spatial distributions of the annual indices during 1961e2000 were explored and the time correlation coefficient (reflecting the similarity of two series on temporal changes) for the time series of regional mean indices between observations and simulations were calculated from 1961 to 2000. For the projections, the ensemble results of three models in 2011e2050 (relative to the average of 1980e1999) under SRES A1B scenario were analyzed, including the linear trends (reflecting the increasing or decreasing direction in a certain period, a simple linear regression) from 2001 to 2050 for the regional mean indices and seasonal time series from 1961 to 2050. Empirical Orthogonal Functions (EOFs) are widely used in the oceanographic and meteorological sciences. EOFs are used for decomposing data sets that have two or more dimensions into pairs of loadings (also called the eigenvectors, or the EOFs) and associated principal components (PCs). Typically, a data set is threedimensional, extending over latitude, longitude, and time. The EOFs of the data produce a set of two-dimensional loadings (latitude, longitude space) and one-dimensional principal components (a function of time). The EOFs are ranked with respect to the amount of variance in the original data set that they explain. So, the leading EOF explains the greatest amount of variance that can be captured by one pattern in this way. With the application of EOF analysis, the spatial distribution patterns of annual mean CDD, R10, and SDII during 1961e2000 and 2011e2050 were analyzed. Many time series in geophysics exhibit non-stationarity in their statistics. Although the series may contain dominant periodic signals, these signals can vary in both amplitude and frequency over long periods of time. Wavelet analysis can decompose a time series into time/frequency space simultaneously. Information is available on both the amplitude of any “periodic” signals within the series, and how this amplitude varies with time. By use of Morlet wavelet transform method, the periodic changes of annual extreme precipitation indices from 1961 to 2050 in the Huang-Huai-Hai River Basin and the South River Basin are also discussed in this study.
3. Model simulation for the present extreme precipitation For global extreme climate, the studies by Kharin et al. (2007) and Kiktev et al. (2007) indicate CGCMs have shown reasonably
151
good agreement with extreme temperature trends but poor agreement with observed extreme precipitation patterns or trends. In regional scales, studies have suggested that such model skills depend on the extreme indices under consideration (Sillmann and Roekner, 2008; Meehl and Tebaldi, 2004; Meehl et al., 2004, 2007). Regional biases in the simulated mean climate state contribute to biases in precipitation extremes. For example, the R5D, SDII or R10 would be underestimated or overestimated where the simulated mean precipitation amount was more or less than observed. For China, three CGCMs show similar characteristics. For extreme precipitation, they can capture the main features of geographic distributions, while they show limited skill in reproducing observed patterns and trends. 3.1. Simulated spatial and temporal changes during 1961e2000 The spatial distributions of the observed and simulated indices for extreme precipitation (CDD, CWD, R10, R5D, R95T and SDII) during 1961e2000 are shown in Fig. 2. It is clear that the dominant features of the geographic distribution for extreme precipitation indices are well captured by models. MPI_ECHAM5 and CSIRO_MK3_5 have better performance than NCAR_CCSM3. The R10, R5D, R95T and SDII results in NCAR_CCSM3 in river basins across eastern China are generally lower than the observations, especially in South River Basin across the coastal monsoon area. There are large model discrepancies between model simulated and observed CDD spatial patterns. Significant CDD is observed in the Northwest River Basin, but models only show limited CDD in parts of the region. In the northern region, CDD is less (about 10 days) for CSIRO_MK3_5 and MPI_ECHAM5, but more (about 15 days) for NCAR_CCSM3. In the southern region, it is less (about 80e100 days) for the three models’ simulation than observations, even about 120 days in some parts. Good agreement is seen between simulations and observations in the South River Basin for MPI_ECHAM5 and NCAR_CCSM3 where CDD are more or less by about 10 days. In the Song-Liao River Basin, CDD is underestimated, about 20 days for CSIRO_MK3_5 and MPI_ECHAM5, and about 30 days for NCAR_CCSM3. Overall, the models show a better performance for CWD except for the Southwest River Basin, and the bias is about 5 w þ5 days in most areas. NCAR_CCSM3 overestimates the amount of CWD by about 10 days in the Huang-Huai-Hai and South River basins. Observed spatial distributions of R10, R5D and SDII exhibit a common feature, with a notable decline from southeastern to northwestern China. There are more extreme precipitation events in the eastern part of the South River Basin. To some extent, the three models show the same characteristics. The spatial correlation coefficient for R10 between the modeled and the observed by CSIRO_MK3_5 is up to 0.8, but it is about 20 days lower than observations in the South River Basin. MPI_ECHAM5 and NCAR_CCSM3 can reproduce the large numbers in the lower reaches of the Yangtze River Basin, while R10 in the upper and middle reaches of the Yangtze River Basin is overestimated by more than 20 days. In Zhu Jiang River Basin, R10 is lower by about 10, 20, and 30 days for CSIRO_MK3_5, MPI_ECHAM5, and NCAR_CCSM3 respectively. For R5D and SDII, the spatial correlation coefficient of CSIRO_MK3_5 and MPI_ECHAM5 is about 0.7. NCAR_CCSM3 has limited skill in reproducing the features of R5D (less about 100 mm) and SDII (lower about 5 mm/day) in the coastal area of South River Basin. With regard to the regional inter-annual variations of extreme precipitation indices, models tend to show different results for different indices (Table 2). In the Song-Liao River Basin and HuangHuai-Hai River Basin, the three models show good skill for CDD. The temporal correlation coefficient of CSIRO_MK3_5 is generally negative in every region, which means the model can reasonably
152
C. Xu et al. / Quaternary International 244 (2011) 149e158
Fig. 2. Spatial distributions of the observed and simulated indices (CDD, CWD, R10, R5D, R95T and SDII) for extreme precipitation during 1961e2000.
C. Xu et al. / Quaternary International 244 (2011) 149e158 Table 2 Time correlation coefficient of annual extreme indices from 1961 to 2000 between observation and simulations.
I
II
III
IV
ME CSIRO_MK3_5 MPI_ECHAM5 NCAR_CCSM3 ME CSIRO_MK3_5 MPI_ECHAM5 NCAR_CCSM3 ME CSIRO_MK3_5 MPI_ECHAM5 NCAR_CCSM3 ME CSIRO_MK3_5 MPI_ECHAM5 NCAR_CCSM3
CDD
CWD
R10
R5D
R95T
SDII
0.329 0.074 0.257 0.356 0.359 0.216 0.303 0.137 0.167 0.052 0.143 0.173 0.011 0.117 0.173 0.138
0.203 0.083 0.006 0.234 0.015 0.060 0.082 0.011 0.117 0.041 0.071 0.133 0.229 0.274 0.020 0.086
0.003 0.079 0.198 0.322 0.081 0.200 0.006 0.044 0.051 0.131 0.078 0.003 0.099 0.125 0.130 0.149
0.089 0.037 0.136 0.076 0.020 0.061 0.036 0.004 0.094 0.049 0.023 0.156 0.148 0.084 0.094 0.249
0.069 0.225 0.237 0.106 0.108 0.167 0.018 0.078 0.131 0.061 0.197 0.116 0.067 0.054 0.254 0.262
0.058 0.011 0.241 0.143 0.068 0.150 0.044 0.204 0.087 0.104 0.193 0.128 0.071 0.151 0.214 0.207
(I: Song-Liao River Basin, II: Huang-Huai-Hai River Basin, III: South River Basin, IV: Northwest River Basin. For the 80% confidence level, r ¼ 0.207. For the 90% confidence level, r ¼ 0.264. For the 95% confidence level, r ¼ 0.312. Italics values represent negative correlation.)
153
reproduce the temporal change features of extreme precipitation. MPI_ECHAM5 appears to have the ability to simulate the time changes of extreme precipitation (with a positive correlation with observation) and shows good skill in the Song-Liao River Basin and the Northwest River Basin. The annual variability of extreme precipitation indices in Huang-Huai-Hai River Basin and South River Basin can be reasonably reproduced by NCAR_CCSM3. 3.2. EOF analysis of CDD, R10 and SDII during 1961e2000 EOF analysis results of annual mean CDD, R10 and SDII during 1961e2000 are shown in Fig. 3, including the spatial patterns of the principal component and its corresponding time coefficient series which passes the significance test and has the largest variance contribution. The EOF1 of the observed CDD (Fig. 3a) reflects the reverse distributions between Northwest e Song-Liao River Basin and Huang-Huai-Hai e South River Basin. The strongest negative signal appears in Northwest River Basin, and a decreasing trend of CDD is shown in accordance with its time coefficient series (Fig. 3g). Model simulated CDD (Fig. 3b) also shows this distribution pattern, despite the fact that in the middle reaches of the Yangtze River the signal characteristic is contrary to observations. For the time coefficient series (Fig. 3h), good agreement is seen between simulated and
Fig. 3. Spatial pattern and the corresponding time coefficient series for the principal component of EOF analysis of annual mean CDD, R10 and SDII between observation and simulation from 1961 to 2000 (a, c, e: EOF1 of the observed CDD, R10, SDII; b, d, f: EOF1, EOF3, EOF3 of the simulation; e, f, g, h: corresponding time coefficient).
154
C. Xu et al. / Quaternary International 244 (2011) 149e158
Fig. 4. Spatial distributions of extreme precipitation indices in 2011e2050 under SRES A1B scenario (CDD, CWD, R10, R5D, R95T and SDII, relative to 1980e1999).
observed, except for the period 1975e1985. For the observed R10 (Fig. 3c), the main feature is the reversed pattern between the HuangHuai-Hai River Basin and the southeastern South River Basin, where the strongest positive signal appears. The simulated R10 (Fig. 3d) offers similar characteristics, but the negative area extends to the Yangtze River basin and the signal is strengthened. With regard to SDII (Fig. 3e, f), the model is neither well skilled at reproducing the observed pattern on the strength of positive or negative signal, nor for the changes of time coefficient series. 4. Projected annual changes under SRES A1B To assess the future changes in annual mean extreme precipitation, the spatial patterns of changes for the period 2011e2050, and the linear trend from 2001 to 2050 are discussed. Secondly, the results for CDD, R10 and SDII are discussed based on the EOF analysis for the period of 2011e2050. Finally, periodic changes in the Huang-Huai-Hai and South River basins are presented. 4.1. Changes in spatial and temporal patterns The changes of annual mean indices in 2011e2050 are shown in Fig. 4, which shows result from the ensemble of three models (ME). The linear trends from 2001 to 2050 are presented in Table 3. In the Song-Liao River Basin, the linear trends (Table 3) in extreme precipitation show an increasing tendency from 2001 to 2050, with the exception of CDD (0.35 days/10a). For 2011e2050, CDD is reduced, especially in the western region by about 4 days (Fig. 4a). CWD does not show significant changes. R10, R5D, R95T and SDII show positive changes, increasing by 2 days, 5 mm, 2% and 0.3 mm/day respectively. SDII shows a remarkable increase in the lower reaches of the Songhua River and Liaohe River basin (Fig. 4f). In the Huang-Huai-Hai River Basin, all the indices exhibit increasing tendencies from 2001 to 2050 (Table 3). In the period
of 2011e2050, there will be more consecutive dry days (CDD, about 4 days) in Huanghe River Basin and Haihe River Basin (Fig. 4a), but 2 days fewer in Huaihe River Basin. The change of consecutive wet days (CWD) is less significant, only 1e2 days more (Fig. 4b). R5D, R95T and SDII show a notable increasing trend, especially for SDII in the eastern part (Fig. 4d, e, f). The changes of precipitation-based indices in Haihe River Basin are more remarkable than in other areas, particularly showing significant increases in R95T and SDII. Those trends indicate that consecutive dry day periods will be longer, but the intensity of precipitation will be stronger in future. In South River Basin, the linear trends (Table 3) show an increasing tendency except for CWD from 2001 to 2050. In the
Table 3 Linear trends of annual extreme indices from 2001 to 2050 for CGCM’s simulations under SRES A1B scenario.
I
II
III
IV
ME CSIRO_MK3_5 MPI_ECHAM5 NCAR_CCSM3 ME CSIRO_MK3_5 MPI_ECHAM5 NCAR_CCSM3 ME CSIRO_MK3_5 MPI_ECHAM5 NCAR_CCSM3 ME CSIRO_MK3_5 MPI_ECHAM5 NCAR_CCSM3
CDD
CWD
R10
R5D
R95T
SDII
0.35 0.38 0.28 0.41 0.32 0.22 0.43 0.32 0.98 1.28 0.54 1.10 0.02 0.62 0.27 0.30
0.20 0.40 0.12 0.07 0.12 0.06 0.11 0.43 0.09 0.16 0.06 0.05 0.06 0.01 0.01 0.19
0.47 0.68 0.22 0.96 0.38 0.12 0.03 1.06 0.15 0.34 0.14 0.03 0.15 0.10 0.06 0.29
2.07 1.87 0.59 3.76 1.45 1.91 1.69 0.76 0.92 2.11 1.63 0.97 0.61 0.33 0.35 1.14
0.70 0.58 0.32 1.19 0.70 0.40 1.08 0.60 0.49 0.70 0.73 0.05 0.53 0.42 0.25 0.90
0.09 0.09 0.04 0.13 0.11 0.02 0.17 0.11 0.08 0.10 0.12 0.02 0.03 0.02 0.03 0.05
(I: Song-Liao River Basin, II: Huang-Huai-Hai River Basin, III: South River Basin, IV: Northwest River Basin. Units: CDD: day10yr1, CWD: day10yr1, R10: day10yr1, R5D: mm10yr1, R95T: %10yr1, SDII: mmd110yr1).
C. Xu et al. / Quaternary International 244 (2011) 149e158
155
Fig. 5. Spatial pattern and the corresponding time coefficient series for the principal component of EOF analysis of annual mean CDD, R10 and SDII from 2011 to 2050 under SRES A1B scenario (a: EOF2 of CDD; b: EOF1 of R10; c: EOF3 of SDII; d, e, f: corresponding time coefficient).
period of 2011e2050, CDD increases in southeastern part, and CWD does not show significant changes in most areas (Fig. 4a, b). In Southeast River Basin, R10 is reduced by about 2 days (Fig. 4c). SDII and R5D increase significantly in most areas, especially in Zhu Jiang River Basin (about 0.4 mm/day for SDII, 5e10 mm for R5D). The temporal variations of annual averaged CDD in South River Basin show a significantly increasing tendency, and its linear trend is greater than in other regions, up to 1 day/10a, while CWD shows a negative trend. R5D, SDII and R95T show increases to some extent (Table 3). This means it is possible to have shortened periods of consecutive rainy days, but the intensity of precipitation will increase. There may be some more heavy rainfall events, which results in the high possibility of the occurrence of floods.
In Northwest River Basin, the changes in CWD, R10, R5D and SDII are similar, not showing obvious increase or decrease. In the western part, CDD will increase by about 4 days in some regions, and R95T will increase in most areas (about 2%). 4.2. Projected distribution pattern of CDD, R10 and SDII Fig. 5 shows the EOF analysis results of the annual mean CDD (EOF2), R10 (EOF1) and SDII (EOF3) during 2011e2050, which passes the significance test and has the largest variance contribution. CDD shows the patterns with the negative signal (less CDD when time coefficient is positive in Fig. 5d) in most areas of the river basins. The stronger negative signals are located in the Huaihe
Fig. 6. Morlet wavelet transform (real parts) of simulated annual extreme precipitation indices in Huang-Huai-Hai River Basin from 1961 to 2050.
156
C. Xu et al. / Quaternary International 244 (2011) 149e158
River Basin and Yangtze River Basin (Fig. 5a). The time coefficient series seems to show cyclical changes (Fig. 5d), which means there are longer or shorter consecutive dry days in a certain period. R10 shows the opposite change in the Huang-Huai-Hai River Basin (negative signal) compared to other basins (Fig. 5b, e). For example, there are more R10 in South River Basin, and less R10 in HuangHuai-Hai River Basin from 2037 to 2047. SDII (Fig. 5c, f) suggests that intensity of precipitation in Song-Liao River Basin is weakened, but it is remarkably stronger in the Huaihe River Basin and the coastal area from 2025 to 2040.
appears in the present climate. For R5T and SDII (Fig. 6e, f), there is more noise, and the periodic signal is not remarkable. In the South River Basin, for CDD in present climate there were no significant periodic signals, but a remarkable 10-year cycle develops after 2000. The periodic features of CWD and R10 shows no obvious changes from 1961 to 2050, a 5-year cycle for CWD and a 6-year cycle for R10; the strength of periodic signals also seems to be maintained. For R5D, R95T and SDII, there are no representative periodic features, except for an 8-year cycle from 2005 to 2030. 5. Projected seasonal changes under SRES A1B
4.3. Cyclical changes in Huang-Huai-Hai River Basin and South River Basin In order to consider the temporal change features of extreme precipitation, wavelet analysis is a good tool. The real part of Morlet wavelet transform is a set of two-dimensional matrix (periodic, time sires space), and the value are standardized without units. When there are the areas with alternating-successive positive and negative values, the corresponding ordinate means the periodic signals in the period are reflected by the abscissa. The periodic change derived from Morlet wavelet transform analysis can interpret inter-annual change. In Huang-Huai-Hai River Basin, the principal feature is that the periodic variations for CDD, CWD and R10 (Fig. 6a, b, c) under the SRES A1B scenario will become more pronounced, while for R5D, R95T and SDII (Fig. 6d, e, f) there are no significant periodic signals. For CDD (Fig. 6a), in the latter half of the 20th century there were no obvious periodic signals, only a weak 8-year cycle from 1961 to 1990. Under the SRES A1B scenario, a stronger 10-year cycle appears after 2015. For CWD (Fig. 6b), there was a 12-year cycle with weak signals from 1961 to 1990, but from 2005 to 2030 under SRES A1B scenario an 8-year cycle become more obvious. For R10 (Fig. 6c), a w15-year cycle under SRES A1B scenario will occur after 2010. A 25-year periodic signal exists until 2050, but it becomes more and more weak after 2000. For R5D (Fig. 6d), there are no significant periodic signals, though an obvious 18-year cycle
The eastern part of China is the economic and political center, and the densely populated area. The frequent occurrences of extreme precipitation seriously affect human life, and cause huge economic losses. In spring (March-April-May) in North China (Huang-Huai-Hai River Basin) and Northeast China (Song-Liao River Basin), drought was evident since the 1980s (Zhang et al., 2003; Qian and Lin, 2005; Ma and Ren, 2007; Zou and Zhang, 2008). In summer (June-July-August), the frequency of extreme precipitation events is higher in the regions south of 35 N than in other parts of China, and the persistence of extreme precipitation in the south of the Yangtze River is evidently larger, which can induce regional floods (Min and Qian, 2008; Chen et al., 2009; Wang and Qian, 2009). So it is necessary to analyze the future changes in the seasonality of extreme precipitation. In the Song-Liao and Huang-Huai-Hai River Basins during 2011e2050 under the SRES A1B scenario, CDD, R10 and CWD have little change in summer, but the increases in R5D, R95T and SDII are greater than in their annual averages. This means that in the SongLiao River Basin there are heavier precipitation events in summer, and the chance of flood occurrence increases. Fig. 7 shows the time series of regional averaged indices in spring from 1961 to 2050 in the Huang-Huai-Hai River Basin. The indices show no significant changes, but the inter-annual variations are significant. In some years, heavy precipitation occurs based on the model projections. R10, R5D, R95T and SDII display significant
Fig. 7. Time series of regional averaged extreme precipitation indices in spring from 1961 to 2050 in Huang-Huai-Hai River Basin (model ensemble, relative to 1980e1999).
C. Xu et al. / Quaternary International 244 (2011) 149e158
157
Fig. 8. Time series of regional averaged extreme precipitation indices in summer from 1961 to 2050 in South River Basin (model ensemble, relative to 1980e1999).
increases around 2020, but decreases around 2030. Afterwards, they began to increase again (Fig. 7d, e, and f). Time series of precipitation-based indices in South River Basin in summer under SRES A1B scenario are given in Fig. 8. CWD and R10 (Fig. 8b, c) show no obvious trend in future, but CDD from 2020 to 2040 shows an upward trend (Fig. 8a). R95T, R5D and SDII tend to show increasing trends (Fig. 8d, e, and f). In winter, R10, R95T, R5d and SDII show slight decreasing trends. After 2040, such a reduction becomes even more remarkable. Therefore, the present condition of heavy precipitation and floods in summer in the South River Basin will not be significantly changed in future the under SRES A1B scenario, while in winter it is possible that there will be less extreme precipitation. 6. Conclusions and discussion Analyzing six indices for extreme precipitation events (CDD, CWD, R10, R5D, SDII and R95T), a preliminary assessment of three climate model simulations of current and future precipitation extremes was conducted. Projected changes before 2050 under the SRES A1B scenario in large river basins in China were analyzed, including a detailed analysis of annual and seasonal changes, the spatial patterns and periodic change in extreme precipitation. Compared with observation, there models can capture the dominant features of the geographic distributions of the extreme precipitation indices during 1961e2000, but have a poor performance in reproducing strong extreme precipitation in the southeastern parts of South River Basin. For the regional inter-annual change signals, the models show different performance in different regions. Based on the results of EOF analysis, the models can reproduce the main distribution pattern of CDD, R10 and SDII during 1961e2000, but exhibit differences with observations in detailed regional features. Under the SRES A1B scenario, for the annual changes during 2011e2050, R95T and SDII increase significantly in all of the river basins, especially in the Huang-Huai-Hai River Basin, but CWD do not have obvious changes. CDD increases in the Huang-Huai-Hai
River Basin and South River Basin, but decreases in the Song-Liao River Basin. In future, the periodic signals of CDD, CWD and R10 seem to be more pronounced, while the R5D, R95T and SDII do not change significantly, with a weakened trend in the Huang-Huai-Hai River Basin. During 2011e2050 under the SRES A1B scenario, in the SongLiao and Huang-Huai-Hai River basins, the increases of R5D, R95T and SDII in summer are greater than the annual averages. From 2001 to 2050, in the Huang-Huai-Hai River Basin extreme precipitation in spring shows no significant changes, but the inter-annual change is significant. In the South River Basin, R95T, R5D and SDII in summer show an increasing tendency, but in winter weak decreasing trends are projected by the models. This research is a preliminary assessment of future extreme events in large river basins in China. This is essential for water resources research and to develop strategies to adapt to climate change. Compared with Zhang et al. (2007b, 2008) and Jiang et al. (2008), this research also found that extreme precipitation will be dominated by increasing trends in the Yangtze River Basin, although based on different extreme precipitation indexes and different CGCMs. However, there are still some inconsistent results. The study of Zhang et al. (2007b) showed that there might be more extreme precipitation in the south part of the middle and lower reaches of the Yangtze River, but extreme precipitation appears to be probably decreasing, and there may be more droughts in the north part under the SRES A1B scenario before 2050. Furthermore, uncertainties in the model projections unavoidably exist. It is necessary to do further analysis, including ensemble projections by more CGCMs and analysis of uncertainty in model results. Acknowledgments This research was jointly supported by the National Basic Research Program of China (2010CB950102, 2009CB421407, 2010CB428401) and the R&D Special Fund for Public Welfare Industry (meteorology) (GYHY200806010), and the National Natural Science Foundation of China (No. 40921160379, 40875083). We
158
C. Xu et al. / Quaternary International 244 (2011) 149e158
acknowledge the modeling groups for making their simulations available for analysis, the Program for Climate Model Diagnosis and Intercomparison (PCMDI) for collecting and archiving the CMIP3 model output, and the WCRP’s Working Group on Coupled Modeling (WGCM) in making the WCRP CMIP3 multi-model data set available. The WCRP CMIP3 multi-model data set is supported by the Office of Science, US Department of Energy. We also acknowledge the work undertaken on the diagnostics extremes indices software at King’s College London and the Climatic Research Unit, which was funded by the Commission of the European Union under the STARDEX (Statistical and Regional dynamic Downscaling of Extremes for European regionsesee http://www. cru.uea.ac.uk/projects/stardex) project (contract EUK2-CT-200100115). The European Climate Assessment project (http://www. knmi.nl/samenw/eca/) and US National Climatic Data Centre are thanked for the code, which forms the basis of this software. In addition, EOF software (Empirical Orthogonal Function) was provided by David W.P., and is available at URL: http://meteora. ucsd.edu/wpierce/eof/eofs.html. Wavelet analysis software was provided by Torrence and Compo (1998), and is available at URL: http://atoc.colorado.edu/research/wavelets/. References Cai, W., Collier, M.A., Durack, P.D., Gordon, H.B., Hirst, A.C., O’Farrell, S.P., Whetton, P.H., 2003. The response of climate variability and mean state to climate change: preliminary results from the CSIRO Mark 3 coupled model. CLIVAR Exchanges 28, 8e11. Chen, H.S., Fan, S.D., Zhang, X.H., 2009. Seasonal differences of variation characteristics of extreme precipitation events over china in the last 50 years. Transactions of Atmospheric Sciences 32 (6), 744e751 (in Chinese). Frich, P., Alexander, L.V., Della-Marta, P., Gleason, B., Haylock, M., Klein Tank, A.M.G., Peterson, T., 2002. Observed coherent changes in climatic extremes during the second half of the twentieth century. Climate Research 19 (3), 193e212. Gao, X.J., Zhao, Z.C., Ding, Y.H., Huang, R.H., Giorgi, F., 2001. Climate change due to greenhouse effects in China as simulated by a regional climate model. Advances in Atmospheric Sciences 18 (6), 1224e1230. Gordon, H.B., Rotstayn, L.D., McGregor, J.L., Dix, M.R., Kowalczyk, E.A., O’Farrell, S.P., Waterman, L.J., Hirst, A.C., Wilson, S.G., Collier, M.A., Watterson, I.G., Elliott, T.I., 2002. The CSIRO Mk3 Climate System Model (Electronic publication). CSIRO Atmospheric Research, Aspendale. http://www.dar.csiro.au/publications/gordon_ 2002a.pdf (CSIRO Atmospheric Research technical paper, no. 60) 130 pp. IPCC, 2007. Contribution of Working Group I to the Fourth Assessment Report of the Intergovernmental Panel on Climate Change. In: Solomon, S., Qin, D., Manning, M., Chen, Z., Marquis, M., Averyt, K.B., Tignor, M., Miller, H.L. (Eds.), Climate Change 2007: The Physical Science Basis. Cambridge University Press, Cambridge, United Kingdom and New York, NY, USA. Jiang, D., Zhang, Y., Sun, J., 2009. Ensemble projection of 1e3 warming in China. Chinese Science Bulletin 54 (18), 3326e3334. Jiang, T., Su, B.D., Marco, G., 2008. Trends in precipitation extremes over the Yangtze River basin. Advance in Water Science 19 (5), 650e655 (in Chinese). Jiang, T., Su, B.D., Wang, Y.J., 2005. Trends of temperature, precipitation and runoff in the Yangtze River basin from 1961 to 2000. Advance in Climate Change Research 1 (2), 65e68 (in Chinese). Jungclaus, J.H., Botzet, M., Haak, H., Keenlyside, N., Luo, J.-J., Latif, M., Marotzke, J., Mikolajewicz, U., Roeckner, E., 2006. Ocean circulation and tropical variability in the coupled model ECHAM5/MPI-OM. Journal of Climate 19 (16), 3952e3972. Kharin, V.V., Zwiers, F., Zhang, X., Hegerl, G.C., 2007. Changes in temperature and precipitation extremes in the IPCC Ensemble of Global Coupled Model Simulations. Journal of Climate 20 (8), 1419e1444. Kiktev, D., Caesar, J., Alexander, L.V., Shiogama, H., Collier, M., 2007. Comparison of observed and multimodeled trends in annual extremes of temperature and precipitation. Geophysical Research Letters 34 (10), L10702. doi:10.1029/ 2007GL029539. Ma, Z.G., Ren, X.B., 2007. Drying trend over China from 1951 to 2006. Advances in Climate Change Research 3 (4), 195e201 (in Chinese). Marsland, S.J., Haak, H., Jungclaus, J.H., Latif, M., Röske, F., 2003. The Max Planck Institute global ocean/sea-ice model with orthogonal curvilinear coordinates. Ocean Modelling 5 (2), 91e127. Meehl, G.A., Warren, M.W., Benjamin, D.S., William, D.C., Julie, M.A., 2006. Climate change projections for the twenty-first century and climate change commitment in the CCSM3. Journal of Climate 19 (11), 2597e2616. Meehl, G.A., Arblaster, J.M., Tebaldi, C., 2007. Contributions of natural and anthropogenic forcing to changes in temperature extremes over the United States. Geophysical Research Letters 34 (19), L19709. doi:10.1029/2007GL030948.
Meehl, G.A., Tebaldi, C., 2004. More intense, more frequent and longer lasting heat waves in the 21st century. Science 305 (5686), 994e997. Meehl, G.A., Tebaldi, C., Nychka, D., 2004. Changes in frost days in simulations of twentyfirst century climate. Climate Dynamics 23 (5), 495e511. Min, S., Qian, Y.F., 2008. Regionality and persistence of extreme precipitation events in China. Advances in Water Science 19 (6), 763e771 (in Chinese). Nakicenovic, N., Swart, R., 2000. Special Report on Emissions Scenarios. Cambridge University Press, Cambridge, United Kingdom, pp. 1e612. Qian, W.H., Fu, J.L., Zhang, W.W., Lin, X., 2007. Change in mean climate and extreme climate in China during the last 40 years. Advance in Earth Science 22 (7), 673e684 (in Chinese). Qian, W.H., Lin, X., 2005. Regional trends in recent precipitation indices in China. Meteorology and Atmospheric Physics 90 (3e4), 193e207. Ren, G.Y., Wu, H., Chen, Z.H., 2004. Variation of precipitation and its influence on flood of the Yangtze Basin over the past 50 years. In: Jiang, T., King, H. (Eds.), Climate Change and Yangtze Floods. Science Press, Beijing, pp. 224e234. Ren, G.Y., Wu, H., Chen, Z.H., 2000. Spatial patterns of change trend in rainfall of China. Quarterly Journal of Applied Meteorology 11 (3), 322e330. Roeckner, E., Bäuml, G., Bonaventura, L., Brokopf, R., Esch, M., Giorgetta, M., Hagemann, S., Kirchner, I., Kornblueh, L., Manzini, E., Rhodin, A., Schlese, U., Schulzweida, U., Tompkins, A., 2003. The atmospheric general circulation model ECHAM5. Part I: model description. Max Planck Institute for Meteorology Rep 349, 127. Sillmann, J., Roekner, E., 2008. Indices for extreme events in projections of anthropogenic climate change. Climatic Change 86 (1e2), 83e104. Su, B.D., Jiang, T., 2008. Distribution feature of time series of extreme precipitation over the YangtzeRiver Basin. Journal of Lake Science 20 (1), 123e128 (in Chinese). Su, B.D., Jiang, T., Ren, G.Y., Chen, Z.H., 2006. Observed trends of precipitation extremes in the YangtzeRiver Basin during 1960 to 2004. Advance in Climate Change Research 2 (1), 9e14 (in Chinese). Sun, A.J., Gao, B., 1998. Change trends of extreme climate events in China. ACTA Meteorological Sinica 12 (2), 129e141. Tian, H., Xu, Y.L., Lin, E.D., 2008. Climate change due to greenhouse effects in Changjiang-HuaiheRiver valley projected by a regional climate model. Advances in Climate Change Research 4 (6), 357e362 (in Chinese). Torrence, C., Compo, G.P., 1998. A Practical Guide to wavelet analysis. Bulletin of the American Meteorological Society 79 (1), 61e78. William, D.C., Cecilia, M.B., Maurice, L.B., Gordon, B.B., Christopher, S.B., James, A.C., Chang, P., Scott, C.D., 2006. The community climate system model version 3 (CCSM3). Journal of Climate 19 (11), 2122e2143. Wang, Z.F., Qian, Y.F., 2009. Frequency and intensity of extreme precipitation events in China. Advances in Water Science 20 (1), 1e9 (in Chinese). Xie, P., Yatagai, A., Chen, M.Y., Hayasaka, T., 2007. A gauge-based analysis of daily precipitation over East Asia. Journal of Hydrometeorology 8 (3), 607e626. Xu, C.H., Shen, X.Y., Xu, Y., 2007. An analysis of climate change in East Asia by using the IPCC AR4 simulations. Advances in Climate Change Research 3 (5), 287e292 (in Chinese). Xu, Y., Gao, X.J., Giorgi, F., 2009a. Regional variability of climate change hot-spots in East Asia. Advances in Atmospheric Sciences 26 (4), 783e792. Xu, Y., Gao, X.J., Guo, Z.H., Zhao, Z.C., Ding, Y.H., 2003. Impacts of greenhouse effects upon middle and lower reaches of the Yangtze River. Journal of LakeSciences 15 (S), 30e37 (in Chinese). Xu, Y., Xu, C.H., Gao, X., Lu, Y., 2009b. Projected changes in temperature and precipitation extremes over the Yangtze River Basin of China in the 21st century. Quaternary International 208, 44e52. Yan, Z.W., Yang, C., 2000. Geographic patterns of extreme climate change in China during 1951e1997. Climatic and Environmental Research 5 (3), 267e272 (in Chinese). Zhai, P., Pan, X.H., 2003. Trends in temperature extremes during 1951e1999 in China. Geophysical Research Letters 30 (17), 1e4. CLM9. Zhai, P., Zhang, X.B., Wan, H., 2005. Trends in total precipitation and frequency of daily precipitation extremes over China. Journal of Climate 18 (7), 1096e1108. Zhai, P.M., Ren, F.M., Zhang, Q., 1999. Detection of trend in China’s precipitation extremes. ACTA Meteorological Sinica 57 (2), 208e216. Zhang, Q.Y., Wei, J., Tao, S.Y., 2003. The decadal and inter-annual variations of drought in the northern china and association with the circulations. Climatic and Environmental Research 8 (3), 307e318 (in Chinese). Zhang, Z.X., Fraedrich, L., Jiang, T., Zhang, J.C., 2007b. Projection of future precipitation extremes in the YangtzeRiver Basin for 2001e2050. Advances in Climate Change Research 3 (6), 340e344 (in Chinese). Zhang, Z.X., Luan, Y.L., Jiang, T., 2008. Change of extreme in the Yangtze River delta and its pre-evaluation for future. Journal of NanjingUniversity (Natural Sciences Edition) 32 (3), 5e8 (in Chinese). Zhang, Z.X., Zhang, Q., Jiang, T., 2007a. Changing features of extreme precipitation in the YangtzeRiver basin during 1961e2002. Journal of Geographical Science 17 (1), 33e42. doi:10.1007/s11442-007-0033-x. Zhou, T.J., Yu, R.C., 2006. Twentieth-century surface air temperature over China and the globe simulated by coupled climate models. Journal of Climate 19 (22), 5843e5858. Zou, X.K., Zhang, Q., 2008. Preliminary studies on variations in droughts over China during past 50 years. Journal of Applied Meteorological Science 19 (6), 679e687 (in Chinese).
Contents lists available at ScienceDirect
Quaternary International journal homepage: www.elsevier.com/locate/quaint
Projected changes of precipitation extremes in river basins over China Chonghai Xu a, b, Yong Luo b, Ying Xu b, * a b
Meteorological Observation Center, China Meteorological Administration, Beijing, China National Climate Center, China Meteorological Administration, Beijing, China
a r t i c l e i n f o
a b s t r a c t
Article history: Available online 13 January 2011
Based on daily precipitation data derived from observations and three Coupled General Circulation Model (CGCM)’s outputs (CSIRO_MK3_5, MPI_ECHAM5 and NCAR_CCSM3), some extreme precipitation indices are calculated. Initially, the models’ skills in simulating extreme precipitation during 1961e2000 were assessed. Projected changes before 2050 under the Special Report on Emissions Scenarios (SRES) emission scenario A1B were also analyzed. Results show that although there are some biases in the model results, the three models capture well the geographic distribution of extreme precipitation observed in the last half of the 20th century. Nevertheless, the models tend to show limited skill in reproducing the observed inter-annual variations of such extreme precipitation events, including the dominant patterns and trends. Under SRES A1B, for the period of 2011e2050, the simple daily precipitation intensity (SDII) and the fraction of extreme precipitation in total annual precipitation (R95T) will increase significantly in all the river basins in China, while changes in other indices are different. The periodic changes in consecutive dry days (CDD), consecutive wet days (CWD) and the number of days with precipitation 10 mm/day (R10) are expected to be more pronounced in the Huang-Huai-Hai River Basin. In the South River Basin, such periodic changes maintain similar or slightly weakened magnitudes. From 2001 to 2050 in the Huang-Huai-Hai River Basin, extreme precipitation in spring shows no significant changes. In the South River Basin, R95T, SDII and the maximum 5-days precipitation amount (R5D) in summer show an increasing tendency, but in winter a weak decreasing trend is projected. The changes in precipitation-based indices indicate a higher probability of heavy rainfall or flood occurrence, particularly in the river basins in East China. Ó 2011 Elsevier Ltd and INQUA. All rights reserved.
1. Introduction In recent decades, significant changes in global mean climate and extreme events have taken place, and most are attributed to the impact of human activity (IPCC, 2007). In the context of global warming, extreme climate events in China have also undergone a significant change, such as a longer duration of heat waves, increasing number of heavy rain events and increased precipitation intensity, as well as more frequent “record-breaking” events of extreme temperature. After the 1990s, there have been more disasters which covered large regions and caused great economic losses, especially associated with the expansion of drought areas in North China and the intensification of heavy floods in South China (Zhai et al., 1999; Yan and Yang, 2000; Qian et al., 2007). Since the 1980s, extreme precipitation events have been increasing in the middle and lower reaches of the Yangtze River basin (Sun and Gao, 1998; Jiang
* Corresponding author. E-mail address: [email protected] (Y. Xu). 1040-6182/$ e see front matter Ó 2011 Elsevier Ltd and INQUA. All rights reserved. doi:10.1016/j.quaint.2011.01.002
et al., 2005; Su et al., 2006; Su and Jiang, 2008), but in the Yellow and Haihe river basins there is a tendency towards drought (Qian and Lin, 2005). Extreme precipitation events in summer show an increasing trend in the northwest river and the Yangtze River basins, but only in the middle and lower reaches of Yangtze River Basin is there a significant increasing trend (Ren et al., 2000, 2004; Zhai and Pan, 2003; Zhai et al., 2005; Zhang et al., 2007a). The continuous improvement of observation techniques and the development of coupled climate models provide a strong basis for analysis and projection research on the extreme events in the future under different SRES emission scenarios (Nakicenovic and Swart, 2000). Most of the research has been focused on the Yangtze River Basin. Xu et al. (2003) estimated the mean change in temperature and precipitation for different scenarios based on five global climate models and a regional climate models (Gao et al., 2001); Zhang et al. (2007b, 2008) and Jiang et al. (2008) have projected changes in daily maximum precipitation using ECHAM5/ MPI-OM. Recently, Xu et al. (2009b) analyzed future changes of extremes indices derived from an ensemble of coupled general circulation models (CGCMs). In addition, Tian et al. (2008) analyzed
150
C. Xu et al. / Quaternary International 244 (2011) 149e158
the future climate change under the SRES A2 scenario in YangtzeHuaihe River Basin by RCM-PRECIS. Nevertheless, there are not enough studies done to comprehensively assess such changes for all of the river basins in China over the next several decades. Such studies are necessary in order to assess the affect of human activities on the spatial and temporal distributions of water resources, and to better estimate the future extreme events including the risk of droughts and flooding at a regional scale. In the World Climate Research Programme (WCRP)’s Coupled Model Intercomparison Project phase 3 (CMIP3) multi-model databases contributing to the Fourth Assessment Report of the Intergovernmental Panel on Climate Change (IPCC AR4), daily precipitation simulated for the climate of the 20th century and projections of the 21st century were made available. That provides an opportunity to analyze the future changes of extreme events in the river basins of China, including temporal and spatial changes. In this study, considering the similarity of climate changes on the temporal and spatial scales, research focused on four main river basins in China (Fig. 1): Song-Liao River Basin (including Songhua Jiang Basin and Liaohe Basin), Huang-Huai-Hai River Basin (including Huanghe Basin, Haihe Basin, and Huaihe Basin), South River Basin (including Yangtze River Basin, Zhu Jiang Basin, and Southeast River Basin), and Northwest River Basin. Previous studies indicated that CGCMs showed relatively reasonable performances in simulating the mean climate change over China, except for the mountain regions in the Tibet Plateau and the surrounding area where large deviations exist between simulations and observation (Zhou and Yu, 2006; Xu et al., 2007; Xu et al., 2009a,b; Jiang et al., 2009). Therefore, no analysis was done for the Southwest River Basin. This manuscript is organized as follows. Section 2 is used to introduce the data used in the study as well as the indices of extreme climate events analyzed for the river basins. Section 3
focuses on assessing the performance for the climate models in simulating the present-day extreme climate. In Section 4 projected annual changes in extreme precipitation are described based on the model ensemble (ME). Section 5 is focused on projected seasonal changes, and conclusion and discussions derived from this study are presented in Section 6.
2. Data and extreme indices The data used to calculate the extreme indices are the observed and simulated daily precipitation data. Observed data are the 1961e2002 gridded daily precipitation on a 0.5 latitudee longitude grid over East Asia, derived from gauge observations at over 2200 stations collected from the Global Telecommunication System (GTS) daily summary files archived by the National Oceanic and Atmospheric Administration (NOAA) Climate Prediction Center (CPC), a collection of daily precipitation observations at over 700 Chinese stations archived by the China Meteorological Administration (CMA), and the daily gauge data at over w1000 hydrological stations from the Chinese Yellow River Conservation Commission (YRCC) (Xie et al., 2007). For model simulations, three models were selected: CSIRO_MK3_5 (Commonwealth Scientific and Industrial Research Organisation, Australia, Gordon et al., 2002; Cai et al., 2003), MPI_ECHAM5 (Max Planck Institute for Meteorology, Germany, Roeckner et al., 2003; Marsland et al., 2003; Jungclaus et al., 2006), and NCAR_CCSM3 (National Center for Atmospheric Research, USA, William et al., 2006; Meehl et al., 2006). Each of these models offers a continuous daily-series data, including the simulation of climate of 20th century and SRES A1B scenario experiments of 21st century (mid-range emissions, CO2 concentration of about 700 ppm by 2100). In order to obtain the model
Fig. 1. Locations of the four major River Basins.
C. Xu et al. / Quaternary International 244 (2011) 149e158 Table 1 six extreme indices used in the study. Index
Definition
Unit
CDD CWD R10 R5D R95T
Maximum number of consecutive dry days (Rday < 1 mm) Maximum number of consecutive wet days (Rday 1 mm) Number of days with precipitation 10 mm d1 Maximum 5 d precipitation total Fraction of annual total precipitation due to events exceeding the 1961e1990 95th percentile Simple daily intensity index: annual total/number of Rday 1 mm d1
days days days mm %
SDII
mm/day
ensemble and do the comparison between simulations and observations, all data are interpolated onto a common 1 1 grid. Six extreme indices described by Frich et al. (2002) are calculated in this research as shown in Table 1. Applying the STARDEX Diagnostic Extremes Indices Software Version 3.3.1 designed to calculate the indices for stations (http://www.cru.uea.ac.uk/cru/projects/ stardex) produces those indices for the gridded area in China. The CDD, CWD and R10 indicate the changes in precipitation frequency, and R5D, R95T and SDII the changes of precipitation intensity. All indices mentioned in this paper were calculated on an annual and seasonal basis for observations during 1961e2002, and for the three models during 1961e2050 respectively. To assess the simulation capabilities of climate models, the spatial distributions of the annual indices during 1961e2000 were explored and the time correlation coefficient (reflecting the similarity of two series on temporal changes) for the time series of regional mean indices between observations and simulations were calculated from 1961 to 2000. For the projections, the ensemble results of three models in 2011e2050 (relative to the average of 1980e1999) under SRES A1B scenario were analyzed, including the linear trends (reflecting the increasing or decreasing direction in a certain period, a simple linear regression) from 2001 to 2050 for the regional mean indices and seasonal time series from 1961 to 2050. Empirical Orthogonal Functions (EOFs) are widely used in the oceanographic and meteorological sciences. EOFs are used for decomposing data sets that have two or more dimensions into pairs of loadings (also called the eigenvectors, or the EOFs) and associated principal components (PCs). Typically, a data set is threedimensional, extending over latitude, longitude, and time. The EOFs of the data produce a set of two-dimensional loadings (latitude, longitude space) and one-dimensional principal components (a function of time). The EOFs are ranked with respect to the amount of variance in the original data set that they explain. So, the leading EOF explains the greatest amount of variance that can be captured by one pattern in this way. With the application of EOF analysis, the spatial distribution patterns of annual mean CDD, R10, and SDII during 1961e2000 and 2011e2050 were analyzed. Many time series in geophysics exhibit non-stationarity in their statistics. Although the series may contain dominant periodic signals, these signals can vary in both amplitude and frequency over long periods of time. Wavelet analysis can decompose a time series into time/frequency space simultaneously. Information is available on both the amplitude of any “periodic” signals within the series, and how this amplitude varies with time. By use of Morlet wavelet transform method, the periodic changes of annual extreme precipitation indices from 1961 to 2050 in the Huang-Huai-Hai River Basin and the South River Basin are also discussed in this study.
3. Model simulation for the present extreme precipitation For global extreme climate, the studies by Kharin et al. (2007) and Kiktev et al. (2007) indicate CGCMs have shown reasonably
151
good agreement with extreme temperature trends but poor agreement with observed extreme precipitation patterns or trends. In regional scales, studies have suggested that such model skills depend on the extreme indices under consideration (Sillmann and Roekner, 2008; Meehl and Tebaldi, 2004; Meehl et al., 2004, 2007). Regional biases in the simulated mean climate state contribute to biases in precipitation extremes. For example, the R5D, SDII or R10 would be underestimated or overestimated where the simulated mean precipitation amount was more or less than observed. For China, three CGCMs show similar characteristics. For extreme precipitation, they can capture the main features of geographic distributions, while they show limited skill in reproducing observed patterns and trends. 3.1. Simulated spatial and temporal changes during 1961e2000 The spatial distributions of the observed and simulated indices for extreme precipitation (CDD, CWD, R10, R5D, R95T and SDII) during 1961e2000 are shown in Fig. 2. It is clear that the dominant features of the geographic distribution for extreme precipitation indices are well captured by models. MPI_ECHAM5 and CSIRO_MK3_5 have better performance than NCAR_CCSM3. The R10, R5D, R95T and SDII results in NCAR_CCSM3 in river basins across eastern China are generally lower than the observations, especially in South River Basin across the coastal monsoon area. There are large model discrepancies between model simulated and observed CDD spatial patterns. Significant CDD is observed in the Northwest River Basin, but models only show limited CDD in parts of the region. In the northern region, CDD is less (about 10 days) for CSIRO_MK3_5 and MPI_ECHAM5, but more (about 15 days) for NCAR_CCSM3. In the southern region, it is less (about 80e100 days) for the three models’ simulation than observations, even about 120 days in some parts. Good agreement is seen between simulations and observations in the South River Basin for MPI_ECHAM5 and NCAR_CCSM3 where CDD are more or less by about 10 days. In the Song-Liao River Basin, CDD is underestimated, about 20 days for CSIRO_MK3_5 and MPI_ECHAM5, and about 30 days for NCAR_CCSM3. Overall, the models show a better performance for CWD except for the Southwest River Basin, and the bias is about 5 w þ5 days in most areas. NCAR_CCSM3 overestimates the amount of CWD by about 10 days in the Huang-Huai-Hai and South River basins. Observed spatial distributions of R10, R5D and SDII exhibit a common feature, with a notable decline from southeastern to northwestern China. There are more extreme precipitation events in the eastern part of the South River Basin. To some extent, the three models show the same characteristics. The spatial correlation coefficient for R10 between the modeled and the observed by CSIRO_MK3_5 is up to 0.8, but it is about 20 days lower than observations in the South River Basin. MPI_ECHAM5 and NCAR_CCSM3 can reproduce the large numbers in the lower reaches of the Yangtze River Basin, while R10 in the upper and middle reaches of the Yangtze River Basin is overestimated by more than 20 days. In Zhu Jiang River Basin, R10 is lower by about 10, 20, and 30 days for CSIRO_MK3_5, MPI_ECHAM5, and NCAR_CCSM3 respectively. For R5D and SDII, the spatial correlation coefficient of CSIRO_MK3_5 and MPI_ECHAM5 is about 0.7. NCAR_CCSM3 has limited skill in reproducing the features of R5D (less about 100 mm) and SDII (lower about 5 mm/day) in the coastal area of South River Basin. With regard to the regional inter-annual variations of extreme precipitation indices, models tend to show different results for different indices (Table 2). In the Song-Liao River Basin and HuangHuai-Hai River Basin, the three models show good skill for CDD. The temporal correlation coefficient of CSIRO_MK3_5 is generally negative in every region, which means the model can reasonably
152
C. Xu et al. / Quaternary International 244 (2011) 149e158
Fig. 2. Spatial distributions of the observed and simulated indices (CDD, CWD, R10, R5D, R95T and SDII) for extreme precipitation during 1961e2000.
C. Xu et al. / Quaternary International 244 (2011) 149e158 Table 2 Time correlation coefficient of annual extreme indices from 1961 to 2000 between observation and simulations.
I
II
III
IV
ME CSIRO_MK3_5 MPI_ECHAM5 NCAR_CCSM3 ME CSIRO_MK3_5 MPI_ECHAM5 NCAR_CCSM3 ME CSIRO_MK3_5 MPI_ECHAM5 NCAR_CCSM3 ME CSIRO_MK3_5 MPI_ECHAM5 NCAR_CCSM3
CDD
CWD
R10
R5D
R95T
SDII
0.329 0.074 0.257 0.356 0.359 0.216 0.303 0.137 0.167 0.052 0.143 0.173 0.011 0.117 0.173 0.138
0.203 0.083 0.006 0.234 0.015 0.060 0.082 0.011 0.117 0.041 0.071 0.133 0.229 0.274 0.020 0.086
0.003 0.079 0.198 0.322 0.081 0.200 0.006 0.044 0.051 0.131 0.078 0.003 0.099 0.125 0.130 0.149
0.089 0.037 0.136 0.076 0.020 0.061 0.036 0.004 0.094 0.049 0.023 0.156 0.148 0.084 0.094 0.249
0.069 0.225 0.237 0.106 0.108 0.167 0.018 0.078 0.131 0.061 0.197 0.116 0.067 0.054 0.254 0.262
0.058 0.011 0.241 0.143 0.068 0.150 0.044 0.204 0.087 0.104 0.193 0.128 0.071 0.151 0.214 0.207
(I: Song-Liao River Basin, II: Huang-Huai-Hai River Basin, III: South River Basin, IV: Northwest River Basin. For the 80% confidence level, r ¼ 0.207. For the 90% confidence level, r ¼ 0.264. For the 95% confidence level, r ¼ 0.312. Italics values represent negative correlation.)
153
reproduce the temporal change features of extreme precipitation. MPI_ECHAM5 appears to have the ability to simulate the time changes of extreme precipitation (with a positive correlation with observation) and shows good skill in the Song-Liao River Basin and the Northwest River Basin. The annual variability of extreme precipitation indices in Huang-Huai-Hai River Basin and South River Basin can be reasonably reproduced by NCAR_CCSM3. 3.2. EOF analysis of CDD, R10 and SDII during 1961e2000 EOF analysis results of annual mean CDD, R10 and SDII during 1961e2000 are shown in Fig. 3, including the spatial patterns of the principal component and its corresponding time coefficient series which passes the significance test and has the largest variance contribution. The EOF1 of the observed CDD (Fig. 3a) reflects the reverse distributions between Northwest e Song-Liao River Basin and Huang-Huai-Hai e South River Basin. The strongest negative signal appears in Northwest River Basin, and a decreasing trend of CDD is shown in accordance with its time coefficient series (Fig. 3g). Model simulated CDD (Fig. 3b) also shows this distribution pattern, despite the fact that in the middle reaches of the Yangtze River the signal characteristic is contrary to observations. For the time coefficient series (Fig. 3h), good agreement is seen between simulated and
Fig. 3. Spatial pattern and the corresponding time coefficient series for the principal component of EOF analysis of annual mean CDD, R10 and SDII between observation and simulation from 1961 to 2000 (a, c, e: EOF1 of the observed CDD, R10, SDII; b, d, f: EOF1, EOF3, EOF3 of the simulation; e, f, g, h: corresponding time coefficient).
154
C. Xu et al. / Quaternary International 244 (2011) 149e158
Fig. 4. Spatial distributions of extreme precipitation indices in 2011e2050 under SRES A1B scenario (CDD, CWD, R10, R5D, R95T and SDII, relative to 1980e1999).
observed, except for the period 1975e1985. For the observed R10 (Fig. 3c), the main feature is the reversed pattern between the HuangHuai-Hai River Basin and the southeastern South River Basin, where the strongest positive signal appears. The simulated R10 (Fig. 3d) offers similar characteristics, but the negative area extends to the Yangtze River basin and the signal is strengthened. With regard to SDII (Fig. 3e, f), the model is neither well skilled at reproducing the observed pattern on the strength of positive or negative signal, nor for the changes of time coefficient series. 4. Projected annual changes under SRES A1B To assess the future changes in annual mean extreme precipitation, the spatial patterns of changes for the period 2011e2050, and the linear trend from 2001 to 2050 are discussed. Secondly, the results for CDD, R10 and SDII are discussed based on the EOF analysis for the period of 2011e2050. Finally, periodic changes in the Huang-Huai-Hai and South River basins are presented. 4.1. Changes in spatial and temporal patterns The changes of annual mean indices in 2011e2050 are shown in Fig. 4, which shows result from the ensemble of three models (ME). The linear trends from 2001 to 2050 are presented in Table 3. In the Song-Liao River Basin, the linear trends (Table 3) in extreme precipitation show an increasing tendency from 2001 to 2050, with the exception of CDD (0.35 days/10a). For 2011e2050, CDD is reduced, especially in the western region by about 4 days (Fig. 4a). CWD does not show significant changes. R10, R5D, R95T and SDII show positive changes, increasing by 2 days, 5 mm, 2% and 0.3 mm/day respectively. SDII shows a remarkable increase in the lower reaches of the Songhua River and Liaohe River basin (Fig. 4f). In the Huang-Huai-Hai River Basin, all the indices exhibit increasing tendencies from 2001 to 2050 (Table 3). In the period
of 2011e2050, there will be more consecutive dry days (CDD, about 4 days) in Huanghe River Basin and Haihe River Basin (Fig. 4a), but 2 days fewer in Huaihe River Basin. The change of consecutive wet days (CWD) is less significant, only 1e2 days more (Fig. 4b). R5D, R95T and SDII show a notable increasing trend, especially for SDII in the eastern part (Fig. 4d, e, f). The changes of precipitation-based indices in Haihe River Basin are more remarkable than in other areas, particularly showing significant increases in R95T and SDII. Those trends indicate that consecutive dry day periods will be longer, but the intensity of precipitation will be stronger in future. In South River Basin, the linear trends (Table 3) show an increasing tendency except for CWD from 2001 to 2050. In the
Table 3 Linear trends of annual extreme indices from 2001 to 2050 for CGCM’s simulations under SRES A1B scenario.
I
II
III
IV
ME CSIRO_MK3_5 MPI_ECHAM5 NCAR_CCSM3 ME CSIRO_MK3_5 MPI_ECHAM5 NCAR_CCSM3 ME CSIRO_MK3_5 MPI_ECHAM5 NCAR_CCSM3 ME CSIRO_MK3_5 MPI_ECHAM5 NCAR_CCSM3
CDD
CWD
R10
R5D
R95T
SDII
0.35 0.38 0.28 0.41 0.32 0.22 0.43 0.32 0.98 1.28 0.54 1.10 0.02 0.62 0.27 0.30
0.20 0.40 0.12 0.07 0.12 0.06 0.11 0.43 0.09 0.16 0.06 0.05 0.06 0.01 0.01 0.19
0.47 0.68 0.22 0.96 0.38 0.12 0.03 1.06 0.15 0.34 0.14 0.03 0.15 0.10 0.06 0.29
2.07 1.87 0.59 3.76 1.45 1.91 1.69 0.76 0.92 2.11 1.63 0.97 0.61 0.33 0.35 1.14
0.70 0.58 0.32 1.19 0.70 0.40 1.08 0.60 0.49 0.70 0.73 0.05 0.53 0.42 0.25 0.90
0.09 0.09 0.04 0.13 0.11 0.02 0.17 0.11 0.08 0.10 0.12 0.02 0.03 0.02 0.03 0.05
(I: Song-Liao River Basin, II: Huang-Huai-Hai River Basin, III: South River Basin, IV: Northwest River Basin. Units: CDD: day10yr1, CWD: day10yr1, R10: day10yr1, R5D: mm10yr1, R95T: %10yr1, SDII: mmd110yr1).
C. Xu et al. / Quaternary International 244 (2011) 149e158
155
Fig. 5. Spatial pattern and the corresponding time coefficient series for the principal component of EOF analysis of annual mean CDD, R10 and SDII from 2011 to 2050 under SRES A1B scenario (a: EOF2 of CDD; b: EOF1 of R10; c: EOF3 of SDII; d, e, f: corresponding time coefficient).
period of 2011e2050, CDD increases in southeastern part, and CWD does not show significant changes in most areas (Fig. 4a, b). In Southeast River Basin, R10 is reduced by about 2 days (Fig. 4c). SDII and R5D increase significantly in most areas, especially in Zhu Jiang River Basin (about 0.4 mm/day for SDII, 5e10 mm for R5D). The temporal variations of annual averaged CDD in South River Basin show a significantly increasing tendency, and its linear trend is greater than in other regions, up to 1 day/10a, while CWD shows a negative trend. R5D, SDII and R95T show increases to some extent (Table 3). This means it is possible to have shortened periods of consecutive rainy days, but the intensity of precipitation will increase. There may be some more heavy rainfall events, which results in the high possibility of the occurrence of floods.
In Northwest River Basin, the changes in CWD, R10, R5D and SDII are similar, not showing obvious increase or decrease. In the western part, CDD will increase by about 4 days in some regions, and R95T will increase in most areas (about 2%). 4.2. Projected distribution pattern of CDD, R10 and SDII Fig. 5 shows the EOF analysis results of the annual mean CDD (EOF2), R10 (EOF1) and SDII (EOF3) during 2011e2050, which passes the significance test and has the largest variance contribution. CDD shows the patterns with the negative signal (less CDD when time coefficient is positive in Fig. 5d) in most areas of the river basins. The stronger negative signals are located in the Huaihe
Fig. 6. Morlet wavelet transform (real parts) of simulated annual extreme precipitation indices in Huang-Huai-Hai River Basin from 1961 to 2050.
156
C. Xu et al. / Quaternary International 244 (2011) 149e158
River Basin and Yangtze River Basin (Fig. 5a). The time coefficient series seems to show cyclical changes (Fig. 5d), which means there are longer or shorter consecutive dry days in a certain period. R10 shows the opposite change in the Huang-Huai-Hai River Basin (negative signal) compared to other basins (Fig. 5b, e). For example, there are more R10 in South River Basin, and less R10 in HuangHuai-Hai River Basin from 2037 to 2047. SDII (Fig. 5c, f) suggests that intensity of precipitation in Song-Liao River Basin is weakened, but it is remarkably stronger in the Huaihe River Basin and the coastal area from 2025 to 2040.
appears in the present climate. For R5T and SDII (Fig. 6e, f), there is more noise, and the periodic signal is not remarkable. In the South River Basin, for CDD in present climate there were no significant periodic signals, but a remarkable 10-year cycle develops after 2000. The periodic features of CWD and R10 shows no obvious changes from 1961 to 2050, a 5-year cycle for CWD and a 6-year cycle for R10; the strength of periodic signals also seems to be maintained. For R5D, R95T and SDII, there are no representative periodic features, except for an 8-year cycle from 2005 to 2030. 5. Projected seasonal changes under SRES A1B
4.3. Cyclical changes in Huang-Huai-Hai River Basin and South River Basin In order to consider the temporal change features of extreme precipitation, wavelet analysis is a good tool. The real part of Morlet wavelet transform is a set of two-dimensional matrix (periodic, time sires space), and the value are standardized without units. When there are the areas with alternating-successive positive and negative values, the corresponding ordinate means the periodic signals in the period are reflected by the abscissa. The periodic change derived from Morlet wavelet transform analysis can interpret inter-annual change. In Huang-Huai-Hai River Basin, the principal feature is that the periodic variations for CDD, CWD and R10 (Fig. 6a, b, c) under the SRES A1B scenario will become more pronounced, while for R5D, R95T and SDII (Fig. 6d, e, f) there are no significant periodic signals. For CDD (Fig. 6a), in the latter half of the 20th century there were no obvious periodic signals, only a weak 8-year cycle from 1961 to 1990. Under the SRES A1B scenario, a stronger 10-year cycle appears after 2015. For CWD (Fig. 6b), there was a 12-year cycle with weak signals from 1961 to 1990, but from 2005 to 2030 under SRES A1B scenario an 8-year cycle become more obvious. For R10 (Fig. 6c), a w15-year cycle under SRES A1B scenario will occur after 2010. A 25-year periodic signal exists until 2050, but it becomes more and more weak after 2000. For R5D (Fig. 6d), there are no significant periodic signals, though an obvious 18-year cycle
The eastern part of China is the economic and political center, and the densely populated area. The frequent occurrences of extreme precipitation seriously affect human life, and cause huge economic losses. In spring (March-April-May) in North China (Huang-Huai-Hai River Basin) and Northeast China (Song-Liao River Basin), drought was evident since the 1980s (Zhang et al., 2003; Qian and Lin, 2005; Ma and Ren, 2007; Zou and Zhang, 2008). In summer (June-July-August), the frequency of extreme precipitation events is higher in the regions south of 35 N than in other parts of China, and the persistence of extreme precipitation in the south of the Yangtze River is evidently larger, which can induce regional floods (Min and Qian, 2008; Chen et al., 2009; Wang and Qian, 2009). So it is necessary to analyze the future changes in the seasonality of extreme precipitation. In the Song-Liao and Huang-Huai-Hai River Basins during 2011e2050 under the SRES A1B scenario, CDD, R10 and CWD have little change in summer, but the increases in R5D, R95T and SDII are greater than in their annual averages. This means that in the SongLiao River Basin there are heavier precipitation events in summer, and the chance of flood occurrence increases. Fig. 7 shows the time series of regional averaged indices in spring from 1961 to 2050 in the Huang-Huai-Hai River Basin. The indices show no significant changes, but the inter-annual variations are significant. In some years, heavy precipitation occurs based on the model projections. R10, R5D, R95T and SDII display significant
Fig. 7. Time series of regional averaged extreme precipitation indices in spring from 1961 to 2050 in Huang-Huai-Hai River Basin (model ensemble, relative to 1980e1999).
C. Xu et al. / Quaternary International 244 (2011) 149e158
157
Fig. 8. Time series of regional averaged extreme precipitation indices in summer from 1961 to 2050 in South River Basin (model ensemble, relative to 1980e1999).
increases around 2020, but decreases around 2030. Afterwards, they began to increase again (Fig. 7d, e, and f). Time series of precipitation-based indices in South River Basin in summer under SRES A1B scenario are given in Fig. 8. CWD and R10 (Fig. 8b, c) show no obvious trend in future, but CDD from 2020 to 2040 shows an upward trend (Fig. 8a). R95T, R5D and SDII tend to show increasing trends (Fig. 8d, e, and f). In winter, R10, R95T, R5d and SDII show slight decreasing trends. After 2040, such a reduction becomes even more remarkable. Therefore, the present condition of heavy precipitation and floods in summer in the South River Basin will not be significantly changed in future the under SRES A1B scenario, while in winter it is possible that there will be less extreme precipitation. 6. Conclusions and discussion Analyzing six indices for extreme precipitation events (CDD, CWD, R10, R5D, SDII and R95T), a preliminary assessment of three climate model simulations of current and future precipitation extremes was conducted. Projected changes before 2050 under the SRES A1B scenario in large river basins in China were analyzed, including a detailed analysis of annual and seasonal changes, the spatial patterns and periodic change in extreme precipitation. Compared with observation, there models can capture the dominant features of the geographic distributions of the extreme precipitation indices during 1961e2000, but have a poor performance in reproducing strong extreme precipitation in the southeastern parts of South River Basin. For the regional inter-annual change signals, the models show different performance in different regions. Based on the results of EOF analysis, the models can reproduce the main distribution pattern of CDD, R10 and SDII during 1961e2000, but exhibit differences with observations in detailed regional features. Under the SRES A1B scenario, for the annual changes during 2011e2050, R95T and SDII increase significantly in all of the river basins, especially in the Huang-Huai-Hai River Basin, but CWD do not have obvious changes. CDD increases in the Huang-Huai-Hai
River Basin and South River Basin, but decreases in the Song-Liao River Basin. In future, the periodic signals of CDD, CWD and R10 seem to be more pronounced, while the R5D, R95T and SDII do not change significantly, with a weakened trend in the Huang-Huai-Hai River Basin. During 2011e2050 under the SRES A1B scenario, in the SongLiao and Huang-Huai-Hai River basins, the increases of R5D, R95T and SDII in summer are greater than the annual averages. From 2001 to 2050, in the Huang-Huai-Hai River Basin extreme precipitation in spring shows no significant changes, but the inter-annual change is significant. In the South River Basin, R95T, R5D and SDII in summer show an increasing tendency, but in winter weak decreasing trends are projected by the models. This research is a preliminary assessment of future extreme events in large river basins in China. This is essential for water resources research and to develop strategies to adapt to climate change. Compared with Zhang et al. (2007b, 2008) and Jiang et al. (2008), this research also found that extreme precipitation will be dominated by increasing trends in the Yangtze River Basin, although based on different extreme precipitation indexes and different CGCMs. However, there are still some inconsistent results. The study of Zhang et al. (2007b) showed that there might be more extreme precipitation in the south part of the middle and lower reaches of the Yangtze River, but extreme precipitation appears to be probably decreasing, and there may be more droughts in the north part under the SRES A1B scenario before 2050. Furthermore, uncertainties in the model projections unavoidably exist. It is necessary to do further analysis, including ensemble projections by more CGCMs and analysis of uncertainty in model results. Acknowledgments This research was jointly supported by the National Basic Research Program of China (2010CB950102, 2009CB421407, 2010CB428401) and the R&D Special Fund for Public Welfare Industry (meteorology) (GYHY200806010), and the National Natural Science Foundation of China (No. 40921160379, 40875083). We
158
C. Xu et al. / Quaternary International 244 (2011) 149e158
acknowledge the modeling groups for making their simulations available for analysis, the Program for Climate Model Diagnosis and Intercomparison (PCMDI) for collecting and archiving the CMIP3 model output, and the WCRP’s Working Group on Coupled Modeling (WGCM) in making the WCRP CMIP3 multi-model data set available. The WCRP CMIP3 multi-model data set is supported by the Office of Science, US Department of Energy. We also acknowledge the work undertaken on the diagnostics extremes indices software at King’s College London and the Climatic Research Unit, which was funded by the Commission of the European Union under the STARDEX (Statistical and Regional dynamic Downscaling of Extremes for European regionsesee http://www. cru.uea.ac.uk/projects/stardex) project (contract EUK2-CT-200100115). The European Climate Assessment project (http://www. knmi.nl/samenw/eca/) and US National Climatic Data Centre are thanked for the code, which forms the basis of this software. In addition, EOF software (Empirical Orthogonal Function) was provided by David W.P., and is available at URL: http://meteora. ucsd.edu/wpierce/eof/eofs.html. Wavelet analysis software was provided by Torrence and Compo (1998), and is available at URL: http://atoc.colorado.edu/research/wavelets/. References Cai, W., Collier, M.A., Durack, P.D., Gordon, H.B., Hirst, A.C., O’Farrell, S.P., Whetton, P.H., 2003. The response of climate variability and mean state to climate change: preliminary results from the CSIRO Mark 3 coupled model. CLIVAR Exchanges 28, 8e11. Chen, H.S., Fan, S.D., Zhang, X.H., 2009. Seasonal differences of variation characteristics of extreme precipitation events over china in the last 50 years. Transactions of Atmospheric Sciences 32 (6), 744e751 (in Chinese). Frich, P., Alexander, L.V., Della-Marta, P., Gleason, B., Haylock, M., Klein Tank, A.M.G., Peterson, T., 2002. Observed coherent changes in climatic extremes during the second half of the twentieth century. Climate Research 19 (3), 193e212. Gao, X.J., Zhao, Z.C., Ding, Y.H., Huang, R.H., Giorgi, F., 2001. Climate change due to greenhouse effects in China as simulated by a regional climate model. Advances in Atmospheric Sciences 18 (6), 1224e1230. Gordon, H.B., Rotstayn, L.D., McGregor, J.L., Dix, M.R., Kowalczyk, E.A., O’Farrell, S.P., Waterman, L.J., Hirst, A.C., Wilson, S.G., Collier, M.A., Watterson, I.G., Elliott, T.I., 2002. The CSIRO Mk3 Climate System Model (Electronic publication). CSIRO Atmospheric Research, Aspendale. http://www.dar.csiro.au/publications/gordon_ 2002a.pdf (CSIRO Atmospheric Research technical paper, no. 60) 130 pp. IPCC, 2007. Contribution of Working Group I to the Fourth Assessment Report of the Intergovernmental Panel on Climate Change. In: Solomon, S., Qin, D., Manning, M., Chen, Z., Marquis, M., Averyt, K.B., Tignor, M., Miller, H.L. (Eds.), Climate Change 2007: The Physical Science Basis. Cambridge University Press, Cambridge, United Kingdom and New York, NY, USA. Jiang, D., Zhang, Y., Sun, J., 2009. Ensemble projection of 1e3 warming in China. Chinese Science Bulletin 54 (18), 3326e3334. Jiang, T., Su, B.D., Marco, G., 2008. Trends in precipitation extremes over the Yangtze River basin. Advance in Water Science 19 (5), 650e655 (in Chinese). Jiang, T., Su, B.D., Wang, Y.J., 2005. Trends of temperature, precipitation and runoff in the Yangtze River basin from 1961 to 2000. Advance in Climate Change Research 1 (2), 65e68 (in Chinese). Jungclaus, J.H., Botzet, M., Haak, H., Keenlyside, N., Luo, J.-J., Latif, M., Marotzke, J., Mikolajewicz, U., Roeckner, E., 2006. Ocean circulation and tropical variability in the coupled model ECHAM5/MPI-OM. Journal of Climate 19 (16), 3952e3972. Kharin, V.V., Zwiers, F., Zhang, X., Hegerl, G.C., 2007. Changes in temperature and precipitation extremes in the IPCC Ensemble of Global Coupled Model Simulations. Journal of Climate 20 (8), 1419e1444. Kiktev, D., Caesar, J., Alexander, L.V., Shiogama, H., Collier, M., 2007. Comparison of observed and multimodeled trends in annual extremes of temperature and precipitation. Geophysical Research Letters 34 (10), L10702. doi:10.1029/ 2007GL029539. Ma, Z.G., Ren, X.B., 2007. Drying trend over China from 1951 to 2006. Advances in Climate Change Research 3 (4), 195e201 (in Chinese). Marsland, S.J., Haak, H., Jungclaus, J.H., Latif, M., Röske, F., 2003. The Max Planck Institute global ocean/sea-ice model with orthogonal curvilinear coordinates. Ocean Modelling 5 (2), 91e127. Meehl, G.A., Warren, M.W., Benjamin, D.S., William, D.C., Julie, M.A., 2006. Climate change projections for the twenty-first century and climate change commitment in the CCSM3. Journal of Climate 19 (11), 2597e2616. Meehl, G.A., Arblaster, J.M., Tebaldi, C., 2007. Contributions of natural and anthropogenic forcing to changes in temperature extremes over the United States. Geophysical Research Letters 34 (19), L19709. doi:10.1029/2007GL030948.
Meehl, G.A., Tebaldi, C., 2004. More intense, more frequent and longer lasting heat waves in the 21st century. Science 305 (5686), 994e997. Meehl, G.A., Tebaldi, C., Nychka, D., 2004. Changes in frost days in simulations of twentyfirst century climate. Climate Dynamics 23 (5), 495e511. Min, S., Qian, Y.F., 2008. Regionality and persistence of extreme precipitation events in China. Advances in Water Science 19 (6), 763e771 (in Chinese). Nakicenovic, N., Swart, R., 2000. Special Report on Emissions Scenarios. Cambridge University Press, Cambridge, United Kingdom, pp. 1e612. Qian, W.H., Fu, J.L., Zhang, W.W., Lin, X., 2007. Change in mean climate and extreme climate in China during the last 40 years. Advance in Earth Science 22 (7), 673e684 (in Chinese). Qian, W.H., Lin, X., 2005. Regional trends in recent precipitation indices in China. Meteorology and Atmospheric Physics 90 (3e4), 193e207. Ren, G.Y., Wu, H., Chen, Z.H., 2004. Variation of precipitation and its influence on flood of the Yangtze Basin over the past 50 years. In: Jiang, T., King, H. (Eds.), Climate Change and Yangtze Floods. Science Press, Beijing, pp. 224e234. Ren, G.Y., Wu, H., Chen, Z.H., 2000. Spatial patterns of change trend in rainfall of China. Quarterly Journal of Applied Meteorology 11 (3), 322e330. Roeckner, E., Bäuml, G., Bonaventura, L., Brokopf, R., Esch, M., Giorgetta, M., Hagemann, S., Kirchner, I., Kornblueh, L., Manzini, E., Rhodin, A., Schlese, U., Schulzweida, U., Tompkins, A., 2003. The atmospheric general circulation model ECHAM5. Part I: model description. Max Planck Institute for Meteorology Rep 349, 127. Sillmann, J., Roekner, E., 2008. Indices for extreme events in projections of anthropogenic climate change. Climatic Change 86 (1e2), 83e104. Su, B.D., Jiang, T., 2008. Distribution feature of time series of extreme precipitation over the YangtzeRiver Basin. Journal of Lake Science 20 (1), 123e128 (in Chinese). Su, B.D., Jiang, T., Ren, G.Y., Chen, Z.H., 2006. Observed trends of precipitation extremes in the YangtzeRiver Basin during 1960 to 2004. Advance in Climate Change Research 2 (1), 9e14 (in Chinese). Sun, A.J., Gao, B., 1998. Change trends of extreme climate events in China. ACTA Meteorological Sinica 12 (2), 129e141. Tian, H., Xu, Y.L., Lin, E.D., 2008. Climate change due to greenhouse effects in Changjiang-HuaiheRiver valley projected by a regional climate model. Advances in Climate Change Research 4 (6), 357e362 (in Chinese). Torrence, C., Compo, G.P., 1998. A Practical Guide to wavelet analysis. Bulletin of the American Meteorological Society 79 (1), 61e78. William, D.C., Cecilia, M.B., Maurice, L.B., Gordon, B.B., Christopher, S.B., James, A.C., Chang, P., Scott, C.D., 2006. The community climate system model version 3 (CCSM3). Journal of Climate 19 (11), 2122e2143. Wang, Z.F., Qian, Y.F., 2009. Frequency and intensity of extreme precipitation events in China. Advances in Water Science 20 (1), 1e9 (in Chinese). Xie, P., Yatagai, A., Chen, M.Y., Hayasaka, T., 2007. A gauge-based analysis of daily precipitation over East Asia. Journal of Hydrometeorology 8 (3), 607e626. Xu, C.H., Shen, X.Y., Xu, Y., 2007. An analysis of climate change in East Asia by using the IPCC AR4 simulations. Advances in Climate Change Research 3 (5), 287e292 (in Chinese). Xu, Y., Gao, X.J., Giorgi, F., 2009a. Regional variability of climate change hot-spots in East Asia. Advances in Atmospheric Sciences 26 (4), 783e792. Xu, Y., Gao, X.J., Guo, Z.H., Zhao, Z.C., Ding, Y.H., 2003. Impacts of greenhouse effects upon middle and lower reaches of the Yangtze River. Journal of LakeSciences 15 (S), 30e37 (in Chinese). Xu, Y., Xu, C.H., Gao, X., Lu, Y., 2009b. Projected changes in temperature and precipitation extremes over the Yangtze River Basin of China in the 21st century. Quaternary International 208, 44e52. Yan, Z.W., Yang, C., 2000. Geographic patterns of extreme climate change in China during 1951e1997. Climatic and Environmental Research 5 (3), 267e272 (in Chinese). Zhai, P., Pan, X.H., 2003. Trends in temperature extremes during 1951e1999 in China. Geophysical Research Letters 30 (17), 1e4. CLM9. Zhai, P., Zhang, X.B., Wan, H., 2005. Trends in total precipitation and frequency of daily precipitation extremes over China. Journal of Climate 18 (7), 1096e1108. Zhai, P.M., Ren, F.M., Zhang, Q., 1999. Detection of trend in China’s precipitation extremes. ACTA Meteorological Sinica 57 (2), 208e216. Zhang, Q.Y., Wei, J., Tao, S.Y., 2003. The decadal and inter-annual variations of drought in the northern china and association with the circulations. Climatic and Environmental Research 8 (3), 307e318 (in Chinese). Zhang, Z.X., Fraedrich, L., Jiang, T., Zhang, J.C., 2007b. Projection of future precipitation extremes in the YangtzeRiver Basin for 2001e2050. Advances in Climate Change Research 3 (6), 340e344 (in Chinese). Zhang, Z.X., Luan, Y.L., Jiang, T., 2008. Change of extreme in the Yangtze River delta and its pre-evaluation for future. Journal of NanjingUniversity (Natural Sciences Edition) 32 (3), 5e8 (in Chinese). Zhang, Z.X., Zhang, Q., Jiang, T., 2007a. Changing features of extreme precipitation in the YangtzeRiver basin during 1961e2002. Journal of Geographical Science 17 (1), 33e42. doi:10.1007/s11442-007-0033-x. Zhou, T.J., Yu, R.C., 2006. Twentieth-century surface air temperature over China and the globe simulated by coupled climate models. Journal of Climate 19 (22), 5843e5858. Zou, X.K., Zhang, Q., 2008. Preliminary studies on variations in droughts over China during past 50 years. Journal of Applied Meteorological Science 19 (6), 679e687 (in Chinese).