1. Introduction
In general, the observed climate variability on decadal and multidecadal time scales can be generated by both internal climate processes and time-varying external forcing. Globally, contributions from externally forced climate change due to anthropogenic forcing may have played a more dominant role on the long-term trend over the internally generated variability (Santer et al. 2011; Meehl et al. 2013), whereas on a regional level internally generated variability may be at least as important as externally forced climate change on decadal or longer time scale (Deser et al. 2012a,b; Hu and Deser 2013; Lu et al. 2014). The interdecadal climate variability has raised concerns from policy makers in order to make prompt formulation of long-term plans for the adaptation and mitigation of potential threat from climate change (Meehl et al. 2009). To better understand the physical bases of the interdecadal climate variability and to predict its behavior in the near future, it becomes essential for us to identify which part of the regional climate variability is associated with externally forced climate change, and which part is associated with internally generated climate variability (Solomon et al. 2011).
Decadal and multidecadal oceanic modes have been shown to be major drivers of the interdecadal climate variability (e.g., Delworth et al. 2007; Meehl et al. 2016a,b,c). For example, the interdecadal Pacific oscillation (IPO), which is defined as the first mode of the empirical orthogonal function (EOF) analysis on the detrended sea surface temperature (SST) in the Pacific between 40°S and 65°N (Power et al. 1999; Meehl and Hu 2006), and the Atlantic multidecadal oscillation (AMO; defined as the area-weighted mean SST in the Atlantic between the equator and 65°N) (Delworth and Mann 2000; Enfield et al. 2001) can strikingly alter the regional and global interdecadal climate variations (Liu and Chiang 2012; Meehl et al. 2013, 2016a,b,c; Dai 2013; Si and Ding 2016).
The underlying mechanisms governing IPO and AMO are still under intensive debate because of the lack of the long-term reliable observations. As IPO, there are three schools of thought currently:
The IPO originated from atmospheric processes, since the long-time SST variation may be induced from the response of the oceanic surface layers to short-time stochastic atmospheric forcing (Hasselmann 1976; Frankignoul and Hasselmann 1977);
The IPO originated from oceanic processes. Many mechanisms have been proposed that can give rise to decadal time scales in SST, such as advection of the upper-ocean circulation (Saravanan and McWilliams 1998) and oceanic gyre dynamics (Dewar 2001; Hogg et al. 2005; Taguchi et al. 2005; Ceballos et al. 2009) and oceanic Rossby wave adjustment (Qiu 2003; Qiu et al. 2007; Schneider and Cornuelle 2005).
The IPO originates from atmosphere–ocean coupled process, as this decadal variability generates from a cycle involving unstable atmosphere–ocean interactions over the North Pacific (Latif and Barnett 1994, 1996).
The most popular thought is that the AMO is associated with oceanic process, especially the Atlantic meridional overturning circulation (AMOC) (Delworth et al. 1993; Delworth and Mann 2000; Latif et al. 2004; Knight et al. 2005). Changes of the AMOC strength can affect the amount of heat transported into the North Atlantic, thus affecting the North Atlantic SST.
The second view is that anthropogenic aerosols are a prime driver of the AMO (Mann and Emanuel 2006; Booth et al. 2012) since its indirect effects can influence long-term variation of SST over the North Atlantic (Evan et al. 2009).
The third explanation is that the AMO may be the response to stochastic forcing from the midlatitude atmospheric circulation (Clement et al. 2015), but this has been disputed in recent studies (Zhang et al. 2016).
Multiple previous studies have indicated the significant role that IPO has played in the slowdown of global mean temperature increase (or so-called global warming hiatus) in the past decade and half (Meehl et al. 2011, 2016a; Kosaka and Xie 2013; England et al. 2014; Dai et al. 2015). A recent study also suggests that the negative phase of the IPO may have contributed to the expansion of the Southern Ocean sea ice (Meehl et al. 2016c). On the other hand, some studies have indicated that the AMO has a significant influence on the Eurasian and North African monsoons (Liu and Chiang 2012; Lu et al. 2006; Zhang and Delworth 2006; Liu et al. 2014) and interdecadal climate variability over Europe (Sutton and Dong 2012) and the North Pacific (Zhang and Delworth 2007). Recently, Si and Ding (2016) found that the AMO acts as a source of interdecadal climate variability over the Northern Hemisphere by exciting a circumglobal stationary baroclinic atmospheric teleconnection from the Atlantic through Eurasia and extending to North America. Further previous observational studies show that East Asian summer rainfall exhibits prominent interdecadal variability in the twentieth century (Hu 1997; Ding et al. 2008, 2009; Si et al. 2016). One example is the wetting trend along the Yangtze–Huaihe River valley (YHRV; 28°–32°N, 109°–120°E) of China and the drying trend in northern China since the late 1970s (Wang 2001; Zhou et al. 2009; Wang et al. 2013). Since the late 1990s, another interdecadal change has occurred, with summer rainfall decreasing along the YHRV but increasing to the north of it (Zhu et al. 2011, 2015; Si and Ding 2013; Huang et al. 2013). Studies also show that the IPO and AMO are two major drivers of the interdecadal variability of summer rainfall over East Asia. The first leading mode of this interdecadal variability is attributed to the IPO, while the second leading mode is attributed to the AMO (Si and Ding 2016). Moreover, it is important to know whether these interdecadal oceanic modes are generated by the internal climate processes or by external forcing.
Our focus in this research is to investigate the role of these climate modes on the East Asian summer monsoon rainfall by analyzing the large-ensemble simulation using the Community Earth System Model, version 1 (CESM1) (hereafter CESM-LE; Kay et al. 2015). Specifically, we want to identify to what extent the observed East Asian summer monsoon rainfall variability is related to the internally generated climate processes and to what extent this rainfall variability is associated with external forcing. The rest of this paper is organized as follows: Section 2 provides details of the CESM-LE simulations, observational datasets, and methodology. Results are given in section 3 and summarized in section 4.
2. Model, experiment, data, and method
The CESM-LE simulations are performed using CESM1 (Hurrell et al. 2013), with the Community Atmosphere Model, version 5.2 (CAM5.2), Parallel Ocean Program, version 2 (POP2), and Community Land Model, version 4, Community Sea Ice Code (CICE4). The horizontal resolution for all components is nominally 1°. The CESM-LE project includes an 1800-yr fully coupled preindustrial control run and a 40-member ensemble simulation from 1920 to 2100 forced by twentieth-century observed time-evolving external forcing from 1920 to 2005 (Lamarque et al. 2010) and by representative concentration pathway 8.5 (RCP8.5) from 2006 to 2100 (Meinshausen et al. 2011; Lamarque et al. 2011). The purpose of the CESM-LE project is to advance our understanding of internal climate variability and climate change. The fully coupled preindustrial control simulation was carried out under a constant 1850 forcing, as the CO2 concentration and the solar forcing are set to 1850 levels and kept constant throughout this control simulation. The initial condition of this control run is from a previous CESM1 control simulation for the atmosphere, land, and sea ice models, but with present-day January mean potential temperature and salinity from the Polar Science Center Hydrographic Climatology, version 2 (PHC2; Steele et al. 2001), for the ocean initial condition. All CESM-LE ensemble members have the same external-forcing condition: historical forcing from 1920 to 2005 and RCP8.5 forcing from 2006 to 2100. The specified ozone forcing in all simulations is calculated using a high-top coupled chemistry–climate model, the Whole Atmosphere Community Climate Model (WACCM) (Marsh et al. 2013). Ensemble member 1 is integrated from 1850 to 2100, using the initial condition from a randomly selected date (1 January of year 402) in the preindustrial control run. Ensemble members 2–40 are all started on 1 January 1920 with round-off levels of perturbation added into the atmospheric temperature field, but keeping the initial conditions in other model components identical with the ensemble member 1, and were run to 2100. Further details on the experiment design are provided in Kay et al. (2015).
The observed precipitation data used in this study are the NOAA Precipitation Reconstruction over Land (PREC/L) dataset from 1948 to 2011 (Chen et al. 2002). The observed monthly mean SST data is the HadISST dataset from 1871 to 2011 with 1° horizontal resolution (Rayner et al. 2003).
The IPO index is defined as the normalized time coefficient of the first leading EOF mode of May–July (MJJ) mean SST in the Pacific Ocean (40°S–65°N, 100°E–80°W). Before performing EOF analysis, the SST linear trend at each grid point is removed. To emphasize the interdecadal signal, a low-pass symmetric filter with 13 weights and a half amplitude point at about 12-yr period is used [see appendix A of chapter 3 of Trenberth et al. (2007)]. The AMO index is defined as the locally linearly detrended MJJ mean SST anomalies averaged over the North Atlantic (0°–60°N, 0°–80°W) as in Enfield et al. (2001).
The significance test for various indices and regional mean rainfall used here is the Monte Carlo method. To apply this method, we need first to randomly scramble the original rainfall or IPO/AMO index time series and obtain a new time series. Then we compute the correlation coefficient between this new time series and the IPO or AMO index. We repeat this calculation 1000 times to get a distribution of correlation values. The quoted confidence levels indicate the percentage of randomized correlation coefficients that exceed the values being tested.
3. Results
a. The nature of IPO and AMO modes
The IPO is characterized as an out-of-phase SST anomaly pattern between the central-western parts of the subtropical Pacific and the central-eastern parts of the tropical Pacific (Fig. 1a). The time-evolving observed IPO index shows remarkable interdecadal variability, such as negative/cold phases from the 1950s to 1970s and during the early twenty-first century and positive/warm phases from the 1930s to 1940s and from the 1980s to 1990s (Fig. 1b). The CESM-LE preindustrial control run reproduces the spatial pattern of the observed IPO well, except in tropical Pacific, where in observation it is a conventional El Niño–like pattern with maximum anomalous SST in the eastern tropical Pacific but in the model it shows a more central Pacific El Niño–like pattern with maximum anomalous SST in the central tropical Pacific (Fig. 1c). Power spectral analysis in Figs. 1b,d shows that the IPO index in the CESM-LE control run displays a similar decadal variability as in the observation with dominant period of 20–30 yr (about 25 yr) (spectral figure is not shown). Figures 1e and 1f are the simulated ensemble mean IPO patterns and the IPO index from each individual member, respectively, in historical all-forcing runs. Although the IPO pattern and period (about 30 yr) in these historical runs are very similar to those in the control run and observations, their temporal evolution is quite different (Fig. 1f). Of the 40 ensemble members, there are only five of them (members 1, 7, 9, 10, and 22; gray curves in Fig. 1b) captured the observed temporal evolution of the IPO. The IPO in other members varies randomly in phase, leading to a small residue of the ensemble mean IPO index, implying that if the ensemble number is large enough the ensemble mean IPO index may approach zero. This suggests that the IPO is primarily an internally generated oceanic mode (Meehl and Hu 2006; Dai et al. 2015), and the external forcing may have minimum influence on its time evolution.
Differing from the IPO, the simulated AMOs in historical runs show very similar time evolutions to each other, in good agreement with observations. The observed AMO spatial pattern is characterized by a horseshoe-like pattern with a lower SST in the central and western midlatitude North Atlantic and a higher SST in the rest of the North Atlantic (Fig. 2a). This horseshoe-like pattern is well captured by both the control run and historical ensemble (Figs. 2c and 2e, respectively). The observed AMO index shows a warm–cold–warm variation from 1920 to 2011, with warm phases in 1930–62 and 1995–2011 and a cold phase in 1963–94 (Fig. 2b). Similar decadal to multidecadal time scale variability of the AMO also appears in the control run (Fig. 2d). In the historical ensembles, the observed temporal change of the AMO is well reproduced by all the 40 members as well as their ensemble mean, with correlation coefficients ranging from 0.4 to 0.8 (statistically significant at 95% level) between the simulated AMO index in each ensemble member and the observed AMO index (Fig. 2f). Moreover, the good agreement of the temporal evolution of the ensemble mean AMO index and observations suggests that external forcing, such as anthropogenic aerosols (Booth et al. 2012), may have played an important role in modulating the observed and simulated AMO variations.
One may also notice that the exact timing of the simulated AMO phase changes differs slightly from that observed, and there are also more shorter-term variations in the simulated AMO. These differences may imply that the internal climate processes, such as changes in AMOC (Zhang et al. 2016) or the atmospheric stochastic forcing (Clement et al. 2015), may also play a role on determining the time evolution of the AMO (Zhang et al. 2016), which will be addressed in more detail in next section.
b. The influence of the internal climate process on AMO
In this section, we examine further the potential influence of the internal climate processes on the AMO, and our focus is primarily the role of the AMOC. Figure 3 shows the correlation pattern between the MJJ AMOC index and the Atlantic SST field in the control run. This correlation pattern is characterized as a horseshoe-like distribution, which agrees well with the AMO spatial pattern in the observations, the CESM1 control run, and the twentieth-century ensemble (Figs. 2a,c,e). The simultaneous correlation coefficient between the AMOC and AMO indices in the 1800-yr control run is 0.4, significant at a 99% level. To quantify the relationship between the AMOC and AMO, the lead–lag correlation of these two indices is calculated. We found that the most significant correlation happens when the AMOC leads AMO by 2 yr with a leading correlation coefficient of 0.44. However, the 2-yr lag correlation between the AMO and AMOC indices in the twentieth-century ensemble runs varies randomly among different ensemble members (Fig. 4). Of the 40 ensemble members, 19 of them (47.5%) produce the positive correlation between AMOC and AMO with significance level of 95%, and 12 out of these 19 members (30%) has the correlation significant at 99% level. On the other hand, a negative correlation between AMOC and AMO appears in about four members. Therefore, although AMOC has played a significant role in modulating AMO with the absence of external forcing, such as in the CESM1 control run, the presence of external forcing adds influences on Atlantic SST in a way that could make the close relationship between AMOC and AMO fail in some occasions (e.g., some members of the twentieth-century all-forcing simulations). In other words, the influence of external forcing on SST could overwrite the role of AMOC.
Further power spectral analysis of AMO and AMOC indices for the long control run shows that both AMO and AMOC have similar frequency spectra with two dominant periods: 30–50 and 13–17 yr (Fig. 5). A further wavelet power spectral analysis shows consistent features between these two indices, such as significant peaks in the 8–16- and 32–64-yr bands (Fig. 6). Both indices also have medium high power in the 64–128-yr band; however, it is only significant at 95% level between model years 500 and 1000 for AMOC. To explore whether the correlation between AMO and AMOC is tighter for selected frequencies, we applied a bandpass filter for the AMO and AMOC indices for the frequencies of 32–64 and 64–128 yr (figure not shown). The correlations for the filter time series between these two indices increase significantly (up to 0.85).
To quantify the common power and relative phase in time–frequency space between AMO and AMOC indices, the cross-wavelet transform and wavelet coherence analyses are applied to the control run data (Grinsted et al. 2004). The assessment of the statistical significance is based on Monte Carlo methods against the red noise backgrounds. Details of these methods are described in Grinsted et al. (2004).
Figure 7a shows the cross-wavelet transform of the AMO and AMOC in the preindustrial control run. As noted before, significant common power in the 8–16- and 32–64-yr bands appears almost in the entire control period, and the common power in the 64–128-yr band appears for model years 500–1400 and 1800–2000. The cross-wavelet transform phase angle within the 95% confidence regions has a mean phase of 0°. This basically confirms that the AMO and AMOC are nearly in phase. The wavelet coherence analysis of the AMO and AMOC shows that the AMO and AMOC are highly coherent in a 8–16-yr band during nearly the entire period; a 32–64-yr band in model years 500–700, 1200–1400, and 1600–2100; and a 64–128-yr band in model years 500–1200 and 1800–2000 (Fig. 7b). The vectors in Fig. 7b indicate the phase difference between AMO and AMOC at each time point and period. In most cases where the coherence is significant, the phase of the AMO and AMOC is also coherent. Therefore, our analysis suggests that to a large extent, the AMO is primarily controlled by the AMOC with the absence of external forcing, which agrees well with previous studies (e.g., Delworth and Mann 2000; Zhang and Delworth 2006; Delworth et al. 2007).
c. Evaluation of East Asian summer rainfall simulated by CESM-LE
Before examining the influence of IPO and AMO on the summer East Asian monsoon rainfall, we first verify the capability of CESM-LE in simulating the East Asian summer rainfall in the historical run against the observations. Figure 8 shows the simulated and observed climatological mean and standard deviation for the summer rainfall over East Asia. The model simulates the observed heavy rainfall over the southern extent of the Tibetan Plateau and over southern China well. In general, the simulated summer rainfall belt has a spatial pattern similar to the observation. However, the differences between the simulated and observed rainfall belt in eastern China can also be found, such as the simulated rainfall belt extending too far northward compared to the observation. In particular, the model simulates excessive rainfall over the Huanghe River valley (HRV; 33°–38°N, 108°–118°E) and northern China. This bias also shows up in the previous versions of this model series (such as CCSM3 and CCSM4) (Meehl et al. 2012). Although significant improvements on the simulated locations and magnitudes of monsoon rainfall over the southern extent of the Tibetan Plateau in CCSM4 compared to CCSM3 have been made, multiple biases still exist, such as a false rainfall center over the eastern extent of the Tibetan Plateau (27°–36°N, 100°–106°E) (Meehl et al. 2012). In CESM-LE simulations, this false rainfall center is still over this region.
The simulated and observed eastern China summer rainfall variability is shown in Figs. 8c and 8d, respectively. In general, the simulated rainfall variance is comparable to the observation in both magnitudes and regional distributions. The total rainfall is larger in CESM simulation in many regions than the observation, except southern China where the simulated rainfall actually is less than observations. However, this larger climatological mean rainfall over China in CESM does not translate into a larger variance. This smaller simulated variance suggests that the simulated rainfall in China is too steady (raining too frequently and intensity too weak) as shown in previous studies (e.g., Neale et al. 2013).
It is well known that it is a rather difficult task to realistically simulate the East Asian summer rainfall (Wang et al. 2005). Here we will further test how well CESM is capable to simulate this summer rainfall in term of the correlation between the observations and the twentieth-century simulations. As shown in Fig. 9, there are high positive correlations over parts of southwestern China, the HRV, northern China, and Mongolia, representing a higher simulated scale, while there is low or no skill over other parts of East Asia, including the southeast coast of China, the YHRV, and the Korea Peninsula. These results suggest that it is still a challenge for CESM to simulate the regional distribution of the summer rainfall realistically in all regions. Although we will discuss the regional patterns of the summer rainfall and their relationship with IPO and AMO, we will more focus on the regional means, such as the mean rainfall in YHRV and HRV.
d. The influence of the IPO and AMO on the summer rainfall over East Asia
The observed summer (June–August) rainfall over East Asia exhibits significant interdecadal variability in the twentieth century. The YHRV of China experienced wetter conditions in the early 1950s and in the 1980s and 1990s, but dryer conditions from the late 1950s to the 1970s and after year 2000 (Fig. 10). Si and Ding (2016) found that the decadal rainfall variation of the YHRV has been positively correlated with the IPO since 1960 (Fig. 11a). This relationship is well simulated in CESM-LE control run and each member of the twentieth-century ensemble simulations (Figs. 11b,c), but not in the ensemble mean (Fig. 11d), suggesting that the external forcing may not have inserted significant influence on the relationship between summer YHRV rainfall and the IPO.
To further investigate the influence of the IPO on YHRV rainfall, we examined how well the temporal evolution of the IPO and YHRV rainfall can be reproduced by the CESM-LE long control run and the twentieth-century ensemble simulations. For the twentieth-century simulations, the time series of IPO and YHRV rainfall is extended from 1920 to 2011 using RCP8.5 simulations from 2006 to 2011 (92 yr). For the 1800-yr control run, a 92-yr running time series is constructed with total of 1708 cases. Then the correlation of these 92-yr-long IPO index and YHRV rainfall with observed IPO and YHRV rainfall is calculated. The significance of these correlations is tested using the Monte Carlo method. Results show that observed IPO temporal evolution can be reproduced by 30% of the samples in the control run at a 90% confidence level and 17% at a 99% confidence level, with similar percentages for the YHRV rainfall. Moreover, we also examine whether the cases with successful reproduction of observed IPO temporal evolution can reproduce the observed YHRV rainfall temporal evolution. It is found that about 35% of these cases can reproduce the YHRV rainfall temporal evolution at a 90% confidence level, and about 19% at a 99% level. This implies that although CESM-LE is capable of reproducing the observed relationship between IPO and YHRV rainfall successfully (the chance is 1 out of 3), to reproduce the observed time evolution of both IPO and YHRV is still a challenge (the chance is about 1 out of 10).
For the twentieth-century historical ensemble simulation, none of the ensemble members can successfully reproduce time evolution of both observed IPO and YHRV rainfall although the observed correlation between IPO and YHRV is well simulated. This suggests that the simulated IPO frequency and magnitude may differ from the observations. On the other hand, the smaller sample size (1708 in the control vs 40 in the twentieth-century simulations) may also reduce the chance to reproduce the observed IPO and YHRV rainfall due to the random nature of the internally generated IPO variability. For IPO, about 15% of the historical ensemble members can reproduce the observed time evolution at the 90% confidence level and 10% at the 99% confidence level, whereas YHRV rainfall can be reproduced by 7.5% of the samples in historical run at both the 90% and 99% significant levels. This suggests that the changes of the external forcing do not insert strong influence on either IPO or YHRV rainfall.
The observed AMO has a significant positive correlation with the summer rainfall over the HRV of China (Fig. 12a; see Si and Ding 2016). In the long preindustrial control run, CESM-LE failed to reproduce this observed correlation pattern (Fig. 12b), suggesting that to realistically reproduce this observed relationship, time-evolving external forcing is needed. As shown in Figs. 12c,d, the observed correlation pattern is well reproduced by both the historical ensemble mean and the 40 individual members, especially over the HRV region (red box). Note that the strong positive correlation between East Asian rainfall and the AMO north of 40°N in CESM-LE historical ensemble mean is associated with the model bias, which leads to the location of the rain belt too far north in comparison with observations (Fig. 8). Additionally, compared to the average pattern, we find that the correlation coefficient increases remarkably in the ensemble mean, and the positive correlation over the HRV exceeds the significant test at a 95% level. This suggests that by eliminating the internal climate variability that is uncorrelated with the external forcing, the positive correlation between the AMO and the summer rainfall over the HRV is enhanced, implying that the HRV rainfall is highly affected by the external forcing–induced AMO variability on interdecadal time scales.
Next we examined the correlation pattern between the AMO and HRV rainfall among the cases that reproduced the observed time evolution of AMO and HRV rainfall in the control run. In general, both the individual cases (Fig. 12e) and their ensemble mean (Fig. 12f) could reproduce the positive correlation between the AMO and the rainfall over HRV to a certain degree. This positive correlation is higher than that in the 1800-yr control run (Fig. 12b) but much lower than that in the historical run (Figs. 12c,d). This result indicates that without external forcing, the internal climate variability could still partially capture the positive relationship between the AMO and HRV rainfall by chance.
Now we ask how well the temporal evolution of the observed AMO and HRV rainfall can be reproduced by CESM-LE long control run and the twentieth-century ensemble simulations. It is found that observed AMO temporal evolution can be reproduced by 32% of the samples in control run at 90% confidence level and 20% at 99% confidence level, with similar percentages for the HRV rainfall. Moreover, for the cases in which AMO temporal evolution is well reproduced in control run, about 30% of these cases can reproduce the HRV rainfall temporal evolution at a 90% significance level and about 22% at a 99% level. Different from IPO, although the basin mean HRV rainfall and AMO temporal evolution could be reproduced by the control run, the spatial pattern of correlation between AMO and HRV rainfall compares poorly with the observations.
However, all of the twentieth-century historical ensemble simulations can reproduce the observed AMO time evolution at a 99% confidence level. The chance of reproducing the observed HRV rainfall in these historical runs is 27.5% at a 90% confidence level and 20% at a 99% confidence level. Thus, the number of successful cases reproducing both observed AMO and HRV rainfall is much higher than the cases reproducing both IPO and YHRV rainfall, further demonstrating the importance of the external forcing in generating the AMO interdecadal variability and its contribution to the HRV rainfall.
e. The combined impact of the IPO and AMO
Here we will further examine the combined impact of the IPO and AMO on East Asian rainfall anomalies. Composite analysis is used. From the IPO and AMO time series, we first identify the warm and cold periods of IPO and AMO in both the control run and 40-member historical runs. They are classified into four categories: 1) warm IPO (IPO+) and warm AMO (AMO+), 2) cold IPO (IPO−) and warm AMO, 3) warm IPO and cold AMO (AMO−), and 4) cold IPO and cold AMO. Based on this, composites of the rainfall anomalies are also constructed.
Figure 13 shows the rainfall anomalies under the four different categories of the IPO and AMO in control run. With the warm phase of the IPO, similar positive rainfall anomalies appear over YHRV under both warm (Fig. 13a) and cold (Fig. 13c) AMO phases (the opposite is true for the cold IPO phase; Figs. 13b,d). This suggests that the AMO plays only a secondary role in modulating the YHRV rainfall anomalies. Similar results are also true for AMO and HRV rainfall although the regional pattern differences between AMO+/IPO+ and AMO+/IPO− (or between AMO−/IPO+ and AMO−/IPO−) are a bit larger than those between IPO+/AMO+ and IPO+/AMO− (or between IPO−/AMO+ and IPO−/AMO−). Thus, for HRV rainfall, either the IPO also plays a significant role (but less than AMO) or the lack of time-evolving external forcing makes the model incapable of simulating more realistically the observed relationship between AMO and HRV rainfall, as we notice earlier.
For the historical run, the same composite analysis of the 40-member ensemble shows a much more consistent relationship between the IPO/AMO and the rainfall in YHRV/HRV (Fig. 14). Regardless of the phase of the AMO, the IPO plays the dominant role in controlling YHRV rainfall, especially during the negative IPO phase. The latter may suggest that the response of the YHRV rainfall to the IPO is not symmetric. The negative IPO has more influence on YHRV rainfall than the positive IPO. For HRV rainfall, the AMO is the dominant factor with some modulation from the IPO, and agrees better with the observations. This implies that the time-evolving external forcing is very important to produce the observed relationship between AMO and HRV rainfall.
4. Summary
These unique 40-member large-ensemble experiments using CESM1 give the potential to enable a separation of the internally generated and externally forced variability from the oceanic system. Our study here shows that the CESM-LE is an effective tool to understand and diagnose the possible mechanisms for observed climate variability and trend.
Here, we found that the IPO appears to be a mostly internally generated decadal oceanic mode. The observed IPO pattern is well simulated in both the control run and the twentieth-century all-forcing ensemble simulations. The chance of simulating the observed temporal evolution of the IPO (1920–2011) by the control run is about 30% when the control run is sampled in a running 92-yr chunk, and about only 13% for the twentieth-century ensemble. Moreover, the observed positive correlation between IPO and summer rainfall over the YHRV is well simulated in the control run, as well as in the individual twentieth-century simulations, but not in the ensemble mean. This suggests that external forcing may not have played any significant role in modulating the summer rainfall over the YHRV and the IPO temporal evolution. On the other hand, although there is about a 10% chance of CESM simulating successfully the observed temporal evolution of both IPO and YHRV rainfall, all of the twentieth-century simulations failed to reproduce this temporal evolution. This further indicates that adding external forcing will not increase the chance of reproducing the observed temporal evolution of both IPO and YHRV rainfall.
Meanwhile, the AMO may mostly be an oceanic mode representing the oceanic response to external forcing on multidecadal time scales associated with the Atlantic meridional overturning circulation. With time-evolving external forcing, all 40 historical simulations as well as their ensemble mean are able to capture well the observed temporal evolution of the AMO. Moreover, the positive correlation between the AMO and the summer rainfall over the HRV is reproduced only in the historical ensemble, but not in the control run because of the exclusion of the time-evolving external forcing. Further, to certain degree the model can partially capture the observed temporal evolution of both AMO and HRV rainfall in some periods in the control run, but it failed to capture the observed spatial correlation pattern between the AMO and HRV. This suggests that although the internal variability, such as AMOC, has significant influence on the AMO, the time-evolving external forcing in the twentieth century has modulated the AMO (possibly also the AMOC) in a way that enhances the teleconnection between the AMO and HRV rainfall.
Acknowledgments
Most of this work was finished during a visit to the Climate Change Research Section, CGD/NCAR. D. Si was supported by the National Natural Science Foundation of China (Grants 41405071 and 41575082), the National Basic Research Program of China (Grant 2013CB430202), and the National Key Research and Development Program of China (2016YFA0602204). A. Hu was supported by the Regional and Global Climate Modeling Program (RGCM) of the U.S. Department of Energy Office of Science (BER), Cooperative Agreement DE-FC02-97ER62402. Computing resources were provided by the Climate Simulation Laboratory at NCAR’s Computational and Information Systems Laboratory, sponsored by the National Science Foundation and other agencies. The authors thank the editor and the three anonymous reviewers for their constructive comments and suggestions that led to an improved manuscript.
REFERENCES
Booth, B. B., N. J. Dunstone, P. R. Halloran, T. Andrews, and N. Bellouin, 2012: Aerosols implicated as a prime driver of twentieth-century North Atlantic climate variability. Nature, 484, 228–232, doi:10.1038/nature10946.
Ceballos, L. I., E. Di Lorenzo, C. D. Hoyos, N. Schneider, and B. Taguchi, 2009: North Pacific Gyre Oscillation synchronizes climate fluctuations in the eastern and western boundary systems. J. Climate, 22, 5163–5174, doi:10.1175/2009JCLI2848.1.
Chen, M., P. Xie, J. E. Janowiak, and P. A. Arkin, 2002: Global land precipitation: A 50-yr monthly analysis based on gauge observations. J. Hydrometeor., 3, 249–266, doi:10.1175/1525-7541(2002)003<0249:GLPAYM>2.0.CO;2.
Clement, A., K. Bellomo, L. N. Murphy, M. A. Cane, T. Mauritsen, G. Radal, and B. Stevens, 2015: The Atlantic multidecadal oscillation without a role for ocean circulation. Science, 350, 320–324, doi:10.1126/science.aab3980.
Dai, A., 2013: The influence of the inter-decadal Pacific oscillation on US precipitation during 1923–2010. Climate Dyn., 41, 633–646, doi:10.1007/s00382-012-1446-5.
Dai, A., J. C. Fyfe, S.-P. Xie, and X. Dai, 2015: Decadal modulation of global surface temperature by internal climate variability. Nat. Climate Change, 5, 555–559, doi:10.1038/nclimate2605.
Delworth, T. L., and M. E. Mann, 2000: Observed and simulated multidecadal variability in the Northern Hemisphere. Climate Dyn., 16, 661–676, doi:10.1007/s003820000075.
Delworth, T. L., S. Manabe, and R. J. Stouffer, 1993: Interdecadal variations of the thermohaline circulation in a coupled ocean–atmosphere model. J. Climate, 6, 1993–2011, doi:10.1175/1520-0442(1993)006<1993:IVOTTC>2.0.CO;2.
Delworth, T. L., R. Zhang, and M. E. Mann, 2007: Decadal to centennial variability of the Atlantic from observations and models. Ocean Circulation: Mechanisms and Impacts, Geophys. Monogr., Vol. 173, Amer. Geophys. Union, 131–148.
Deser, C., R. Knutti, S. Solomon, and A. S. Phillips, 2012a: Communication of the role of natural variability in future North American climate. Nat. Climate Change, 2, 775–780, doi:10.1038/nclimate1562.
Deser, C., A. S. Phillips, V. Bourdette, and H. Y. Teng, 2012b: Uncertainty in climate change projections: The role of internal variability. Climate Dyn., 38, 527–546, doi:10.1007/s00382-010-0977-x.
Dewar, W. K., 2001: On ocean dynamics in midlatitude climate. J. Climate, 14, 4380–4397, doi:10.1175/1520-0442(2001)014<4380:OODIMC>2.0.CO;2.
Ding, Y., Z. Wang, and Y. Sun, 2008: Inter-decadal variation of the summer precipitation in East China and its association with decreasing Asian summer monsoon. Part I: Observed evidences. Int. J. Climatol., 28, 1139–1161, doi:10.1002/joc.1615.
Ding, Y., Y. Sun, Z. Wang, Y. Zhu, and Y. Song, 2009: Inter-decadal variation of the summer precipitation in East China and its association with decreasing Asian summer monsoon. Part II: Possible causes. Int. J. Climatol., 29, 1926–1944, doi:10.1002/joc.1759.
Enfield, D. B., A. M. Mestas-Nuñez, and P. J. Trimble, 2001: The Atlantic multidecadal oscillation and its relation to rainfall and river flows in the continental U.S. Geophys. Res. Lett., 28, 2077–2080, doi:10.1029/2000GL012745.
England, M., and Coauthors, 2014: Recent intensification of wind-driven circulation in the Pacific and the ongoing warming hiatus. Nat. Climate Change, 4, 222–227, doi:10.1038/nclimate2106.
Evan, A. T., D. J. Vimont, A. K. Heidinger, J. P. Kossin, and R. Bennartz, 2009: The role of aerosols in the evolution of tropical North Atlantic Ocean temperature anomalies. Science, 324, 778–781, doi:10.1126/science.1167404.
Frankignoul, C., and K. Hasselmann, 1977: Stochastic climate models, Part II: Application to sea-surface temperature variability and thermocline variability. Tellus, 29, 289–305, doi:10.3402/tellusa.v29i4.11362.
Grinsted, A., J. C. Moore, and S. Jevrejeva, 2004: Application of the cross wavelet transform and wavelet coherence to geophysical time series. Nonlinear Processes Geophys., 11, 561–566, doi:10.5194/npg-11-561-2004.
Hasselmann, K., 1976: Stochastic climate models. I: Theory. Tellus, 28, 473–485, doi:10.3402/tellusa.v28i6.11316.
Hogg, A. M., P. D. Killworth, J. R. Blundell, and W. K. Dewar, 2005: Mechanisms of decadal variability of the wind-driven ocean circulation. J. Phys. Oceanogr., 35, 512–531, doi:10.1175/JPO2687.1.
Hu, A., and C. Deser, 2013: Uncertainty in future regional sea level rise due to internal climate variability. Geophys. Res. Lett., 40, 2768–2772, doi:10.1002/grl.50531.
Hu, Z.-Z., 1997: Interdecadal variability of summer climate over East Asia and its association with 500 hPa height and global sea surface temperature. J. Geophys. Res., 102, 19 403–19 412, doi:10.1029/97JD01052.
Huang, R., Y. Liu, and T. Feng, 2013: Interdecadal change of summer precipitation over Eastern China around the late-1990s and associated circulation anomalies, internal dynamical causes. Chin. Sci. Bull., 58, 1339–1349, doi:10.1007/s11434-012-5545-9.
Hurrell, J. W., and Coauthors, 2013: The Community Earth System Model: A framework for collaborative research. Bull. Amer. Meteor. Soc., 94, 1339–1360, doi:10.1175/BAMS-D-12-00121.1.
Kay, J. E., and Coauthors, 2015: The Community Earth System Model (CESM) large ensemble project: A community resource for studying climate change in the presence of internal climate variability. Bull. Amer. Meteor. Soc., 96, 1333–1349, doi:10.1175/BAMS-D-13-00255.1.
Knight, J. R., R. J. Allan, C. K. Folland, M. Vellinga, and M. E. Mann, 2005: A signature of persistent natural thermohaline circulation cycles in observed climate. Geophys. Res. Lett., 32, L20708, doi:10.1029/2005GL024233.
Kosaka, Y., and S.-P. Xie, 2013: Recent global-warming hiatus tied to equatorial Pacific surface cooling. Nature, 501, 403–407, doi:10.1038/nature12534.
Lamarque, J.-F., and Coauthors, 2010: Historical (1850–2000) gridded anthropogenic and biomass burning emissions of reactive gases and aerosols: Methodology and application. Atmos. Chem. Phys., 10, 7017–7039, doi:10.5194/acp-10-7017-2010.
Lamarque, J.-F., G. P. Kyle, M. Meinshausen, K. Riahi, S. J. Smith, D. P. van Vuuren, A. J. Conley, and F. Vitt, 2011: Global and regional evolution of short-lived radiatively-active gases and aerosols in the representative concentration pathways. Climatic Change, 109, 191–212, doi:10.1007/s10584-011-0155-0.
Latif, M., and T. P. Barnett, 1994: Causes of decadal climate variability over the North Pacific and North America. Science, 266, 634–637, doi:10.1126/science.266.5185.634.
Latif, M., and T. P. Barnett, 1996: Decadal climate variability over the North Pacific and North America: Dynamics and predictability. J. Climate, 9, 2407–2423, doi:10.1175/1520-0442(1996)009<2407:DCVOTN>2.0.CO;2.
Latif, M., and Coauthors, 2004: Reconstructing, monitoring, and predicting multidecadal-scale changes in the North Atlantic thermohaline circulation with sea surface temperature. J. Climate, 17, 1605–1614, doi:10.1175/1520-0442(2004)017<1605:RMAPMC>2.0.CO;2.
Liu, Y., and J. C. H. Chiang, 2012: Coordinated abrupt weakening of the Eurasian and North African monsoons in the 1960s and links to extratropical North Atlantic cooling. J. Climate, 25, 3532–3548, doi:10.1175/JCLI-D-11-00219.1.
Liu, Y., J. C. H. Chiang, C. Chou, and C. M. Patricola, 2014: Atmospheric teleconnection mechanisms of extratropical North Atlantic SST influence on Sahel rainfall. Climate Dyn., 43, 2797–2811, doi:10.1007/s00382-014-2094-8.
Lu, J., A. Hu, and Z. Zeng, 2014: On the possible interaction between internal climate variability and forced climate change. Geophys. Res. Lett., 41, 2962–2970, doi:10.1002/2014GL059908.
Lu, R., B. Dong, and H. Ding, 2006: Impact of the Atlantic multidecadal oscillation on the Asian summer monsoon. Geophys. Res. Lett., 33, L24701, doi:10.1029/2006GL027655.
Mann, M. E., and K. A. Emanuel, 2006: Atlantic hurricane trends linked to climate change. Eos, Trans. Amer. Geophys. Union, 87, 233–241, doi:10.1029/2006EO240001.
Marsh, D. R., M. J. Mills, D. E. Kinnison, J.-F. Lamarque, N. Calvo, and L. M. Polvani, 2013: Climate change from 1850 to 2005 simulated in CESM1(WACCM). J. Climate, 26, 7372–7391, doi:10.1175/JCLI-D-12-00558.1.
Meehl, G. A., and A. Hu, 2006: Megadroughts in the Indian monsoon region and southwest North America and a mechanism for associated multi-decadal Pacific sea surface temperature anomalies. J. Climate, 19, 1605–1623, doi:10.1175/JCLI3675.1.
Meehl, G. A., and Coauthors, 2009: Decadal prediction: Can it be skillful? Bull. Amer. Meteor. Soc., 90, 1467–1485, doi:10.1175/2009BAMS2778.1.
Meehl, G. A., J. M. Arblaster, J. T. Fasullo, A. Hu, and K. E. Trenberth, 2011: Model-based evidence of deep-ocean heat uptake during surface-temperature hiatus periods. Nat. Climate Change, 1, 360–364, doi:10.1038/nclimate1229.
Meehl, G. A., J. M. Arblaster, J. M. Caron, H. Annamalai, M. Jochum, A. Chakraborty, and R. Murtugudde, 2012: Monsoon regimes and processes in CCSM4. Part I: The Asian–Australian monsoon. J. Climate, 25, 2583–2608, doi:10.1175/JCLI-D-11-00184.1.
Meehl, G. A., A. Hu, J. M. Arblaster, J. T. Fasullo, and K. E. Trenberth, 2013: Externally forced and internally generated decadal climate variability associated with the interdecadal Pacific oscillation. J. Climate, 26, 7298–7310, doi:10.1175/JCLI-D-12-00548.1.
Meehl, G. A., A. Hu, B. D. Santer, and S.-P. Xie, 2016a: Contribution of the interdecadal Pacific oscillation to twentieth-century global surface temperature trends. Nat. Climate Change, 6, 1005–1008, doi:10.1038/nclimate3107.
Meehl, G. A., A. Hu, and H. Teng, 2016b: Initialized decadal prediction for transition to positive phase of the interdecadal Pacific oscillation. Nat. Commun., 7, 11718, doi:10.1038/ncomms11718.
Meehl, G. A., J. M. Arblaster, C. M. Bitz, C. T. Y. Chung, and H. Teng, 2016c: Antarctic sea-ice expansion between 2000 and 2014 driven by tropical Pacific decadal climate variability. Nat. Geosci., 9, 590–595, doi:10.1038/ngeo2751.
Meinshausen, M., and Coauthors, 2011: The RCP greenhouse gas concentrations and their extension from 1765 to 2300. Climatic Change, 109, 213–241, doi:10.1007/s10584-011-0156-z.
Neale, R. B., J. Richter, S. Park, P. H. Lauritzen, S. J. Vavrus, P. J. Rasch, and M. Zhang, 2013: The mean climate of the Community Atmosphere Model (CAM4) in forced SST and fully coupled experiments. J. Climate, 26, 5150–5168, doi:10.1175/JCLI-D-12-00236.1.
Power, S., T. Casey, C. Folland, A. Colman, and V. Mehta, 1999: Interdecadal modulation of the impact of ENSO on Australia. Climate Dyn., 15, 319–324, doi:10.1007/s003820050284.
Qiu, B., 2003: Kuroshio Extension variability and forcing of the Pacific decadal oscillations: Responses and potential feedback. J. Phys. Oceanogr., 33, 2465–2482, doi:10.1175/2459.1.
Qiu, B., N. Schneider, and S. Chen, 2007: Coupled decadal variability in the North Pacific: An observationally constrained idealized model. J. Climate, 20, 3602–3620, doi:10.1175/JCLI4190.1.
Rayner, N. A., D. E. Parker, E. B. Horton, C. K. Folland, L. V. Alexander, D. P. Rowell, E. C. Kent, A. Kaplan, 2003: Global analyses of sea surface temperature, sea ice, and night marine air temperature since the late nineteenth century. J. Geophys. Res., 108, 4407, doi:10.1029/2002JD002670.
Santer, B. D., and Coauthors, 2011: Separating signal and noise in atmospheric temperature changes: The importance of timescale. J. Geophys. Res., 116, D22105, doi:10.1029/2011JD016263.
Saravanan, R., and J. C. McWilliams, 1998: Advective ocean–atmosphere interaction: An analytical stochastic model with implications for decadal variability. J. Climate, 11, 165–188, doi:10.1175/1520-0442(1998)011<0165:AOAIAA>2.0.CO;2.
Schneider, N., and B. D. Cornuelle, 2005: The forcing of the Pacific decadal oscillation. J. Climate, 18, 4355–4373, doi:10.1175/JCLI3527.1.
Si, D., and Y. Ding, 2013: Decadal change in the correlation pattern between the Tibetan Plateau winter snow and the East Asian summer precipitation during 1979–2011. J. Climate, 26, 7622–7634, doi:10.1175/JCLI-D-12-00587.1.
Si, D., and Y. Ding, 2016: Oceanic forcings of the interdecadal variability in East Asian summer rainfall. J. Climate, 29, 7633–7649, doi:10.1175/JCLI-D-15-0792.1.
Si, D., Z. Hu, A. Kumar, B. Jha, P. Peng, W. Wang, and R. Han, 2016: Is the interdecadal variation of the summer rainfall over eastern China associated with SST? Climate Dyn., 46, 135–146, doi:10.1007/s00382-015-2574-5.
Solomon, A., and Coauthors, 2011: Distinguishing the roles of natural and anthropogenically forced decadal climate variability: Implications for prediction. Bull. Amer. Meteor. Soc., 92, 141–156, doi:10.1175/2010BAMS2962.1.
Steele, M., R. Morley, and W. Ermold, 2001: PHC: A global ocean hydrography with a high-quality Arctic Ocean. J. Climate, 14, 2079–2087, doi:10.1175/1520-0442(2001)014<2079:PAGOHW>2.0.CO;2.
Sutton, R. T., and B. Dong, 2012: Atlantic Ocean influence on a shift in European climate in the 1990s. Nat. Geosci., 5, 788–792, doi:10.1038/ngeo1595.
Taguchi, B., S.-P. Xie, H. Mitsudera, and A. Kubokawa, 2005: Response of the Kuroshio Extension to Rossby waves associated with the 1970s climate regime shift in a high-resolution ocean model. J. Climate, 18, 2979–2995, doi:10.1175/JCLI3449.1.
Trenberth, K. E., and Coauthors, 2007: Observations: Surface and atmospheric climate change. Climate Change 2007: The Physical Science Basis, S. Solomon et al., Eds., Cambridge University Press, 237–336.
Wang, B., Q. Ding, X. Fu, I.-S. Kang, K. Jin, J. Shukla, and F. Doblas-Reyes, 2005: Fundamental challenge in simulation and prediction of summer monsoon rainfall. Geophys. Res. Lett., 32, L15711, doi:10.1029/2005GL022734.
Wang, H., 2001: The weakening of the Asian monsoon circulation after the end of 1970s. Adv. Atmos. Sci., 18, 376–386, doi:10.1007/BF02919316.
Wang, T., H. J. Wang, O. H. Otterå, Y. Q. Gao, L. L. Suo, T. Furevik, and L. Yu, 2013: Anthropogenic agent implicated as a prime driver of shift in precipitation in eastern China in the late 1970s. Atmos. Chem. Phys., 13, 12 433–12 450, doi:10.5194/acp-13-12433-2013.
Zhang, R., and T. L. Delworth, 2006: Impact of Atlantic multidecadal oscillations on India/Sahel rainfall and Atlantic hurricanes. Geophys. Res. Lett., 33, L17712, doi:10.1029/2006GL026267.
Zhang, R., and T. L. Delworth, 2007: Impact of the Atlantic multidecadal oscillation on North Pacific climate variability. Geophys. Res. Lett., 34, L23708, doi:10.1029/2007GL031601.
Zhang, R., R. Sutton, G. Danabasoglu, T. L. Delworth, W. M. Kim, J. Robson, and S. G. Yeager, 2016: Comment on “The Atlantic multidecadal oscillation without a role for ocean circulation.” Science, 352, 1527, doi:10.1126/science.aaf1660.
Zhou, T., D. Gong, J. Li, and B. Li, 2009: Detecting and understanding the multi-decadal variability of the East Asian summer monsoon—Recent progress and state of affairs. Meteor. Z., 18, 455–467, doi:10.1127/0941-2948/2009/0396.
Zhu, Y., H. Wang, W. Zhou, and J. Ma, 2011: Recent changes in the summer precipitation pattern in East China and the background circulation. Climate Dyn., 36, 1463–1473, doi:10.1007/s00382-010-0852-9.
Zhu, Y., H. Wang, J. Ma, T. Wang, and J. Sun, 2015: Contribution of the phase transition of Pacific decadal oscillation to the late 1990s’ shift in East China summer rainfall. J. Geophys. Res. Atmos., 120, 8817–8827, doi:10.1002/2015JD023545.