1. Introduction
Improved understanding of decadal variability can help with better prediction of decadal climate variations and adaptation to climate change (Goddard et al. 2009; Hurrell et al. 2009; Meehl et al. 2009a). In addition to the well-known Pacific decadal oscillation (PDO)/interdecadal Pacific oscillation (IPO) (e.g., Mantua et al. 1997; Power et al. 1999) and the Atlantic multidecadal oscillation (AMO) (e.g., Enfield et al. 2001), decadal variations in the Indian Ocean have recently been identified (e.g., Lee and McPhaden 2008; Han et al. 2010; Feng et al. 2010; Han et al. 2014a,b; Li and Han 2015; Dong et al. 2016; Krishnamurthy and Krishnamurthy 2016). In particular, a basinwide pattern dominates the decadal variability in the Indian Ocean sea surface temperature (SST) that Han et al. (2014b) dubbed the “decadal Indian Ocean Basin (IOB) mode” (p. 1692). Attempts have been made to explain the decadal IOB mode as a response to remote forcing from the IPO, with basinwide warm anomalies related to a positive IPO and cool anomalies related to a negative IPO. The teleconnection between the two basins on decadal time scales is analogous to the impact of El Niño–Southern Oscillation (ENSO) on the interannual IOB mode, with IPO-induced atmospheric processes changing surface heat fluxes and wind stresses in the Indian Ocean (Dong et al. 2016).
While the decadal IOB mode was positively correlated with the IPO before 1985, the correlation became negative after 1985, which was first recognized by Han et al. (2014a). That is to say, an anomalously warm Indian Ocean coincided with a negative IPO pattern in the Pacific on decadal time scales after 1985. Such anomalous SST patterns in the Indian and Pacific Oceans together drive intensified easterlies and rapid sea level rise in the western tropical Pacific (Timmermann et al. 2010; Han et al. 2014a; England et al. 2014). The reasons behind the changing relationship between the decadal IOB mode and the IPO after 1985 have not been thoroughly explored in previous studies. Although Han et al. (2014a) suspected that persistent warming in the Indian Ocean driven by anthropogenic forcing may have played some role in affecting the changed relationship, they did not provide evidence in their paper to support this hypothesis. Dong et al. (2014b) and Dong and Zhou (2014), on the other hand, found that external forcing, mainly anthropogenic in nature, can project onto the IOB and induce a basinwide warming trend in the Indian Ocean. Therefore, we hypothesize that external forcing may modulate the evolution of the decadal IOB mode as well as its relationship with the IPO.
We also note that changes in relationships between the IOB and the IPO occur in parallel with the recent global warming hiatus, which featured a slowdown in global-averaged surface air temperature warming despite increasing radiative forcing (e.g., Easterling and Wehner 2009; Meehl et al. 2011; Kosaka and Xie 2013; England et al. 2014). Both the Pacific and Indian Oceans have been implicated in this hiatus. Many studies have attributed this hiatus to a cooling trend in the eastern Pacific (Kosaka and Xie 2013) or enhanced Pacific trade winds (England et al. 2014). Internal variability in tropical Pacific, mainly associated with the IPO, is likely to play an important role in causing these climatic changes in the Pacific (Trenberth et al. 2014). However, the faster warming rate in the Indian Ocean compared to the tropical Pacific in recent decades, due to anthropogenic forcing, also may have contributed to the global warming hiatus (Luo et al. 2012). Furthermore, Lee et al. (2015) and Nieves et al. (2015) highlighted the role of the Indonesian Throughflow (ITF) in regulating the oceanic heat budget during the recent hiatus. The stronger negative IPO in the 2000s and a slowdown in the rate of global warming have also been partly attributed to a series of moderate volcanic eruptions (Santer et al. 2014, 2015) and stratospheric water vapor (Solomon et al. 2010). Since it has been argued that both interbasin warming contrasts and IPO-like natural variability played a role in the current global warming hiatus, it is important to determine the relative importance of these mechanisms in the recent intensification of Pacific trade winds. Whether the changed relationship between the IOB and IPO and the recent global warming hiatus are dynamically related to each other or simply coincidental is likewise an open question.
The purpose of this paper therefore is to explore whether the changed relationship between the decadal IOB mode and the IPO in recent decades is due to increased Indian Ocean warming or internal variability in the climate system. Observations and coupled climate models are analyzed to explain the mechanisms responsible for the changed relationship and clarify the individual effects of external forcing and internal variability. We also address question of the relationship between Indo-Pacific decadal variability and the recent hiatus in global warming.
2. Observations, model experiments, and methods
a. Observations and model experiments
The SST data we use are monthly analyses from the Hadley Centre Global Sea Ice and Sea Surface Temperature (HadISST; 1° latitude × 1° longitude) dataset (Rayner et al. 2003), the National Oceanic and Atmospheric Administration Extended Reconstructed SST version 3 (ERSST; 2° latitude × 2° longitude; Smith et al. 2008), and the Kaplan Extended SST version 2 (Kaplan; 5° latitude × 5° longitude; Kaplan et al. 1998). These three SST datasets exhibit some different characteristics (Zhang 2016) because of the different instrumental measurements used (e.g., bucket samples and ship engine room intakes, and whether satellite retrievals are used or not) and how they are spatially and temporally smoothed, interpolated, and analyzed using optimal statistical procedures (Deser et al. 2010). Further information on data inputs and processing is given in the references for each dataset. For our purposes, comparisons among different datasets help to confirm the robustness of our results regardless of the different mix of measurements and processing techniques used.
External forcing of the climate system includes anthropogenic forcing [mainly greenhouse gases (GHG), anthropogenic aerosols (AA), ozone, and land use] and natural forcing (solar radiation and volcanic aerosols). To examine the effects of external forcing on Indian Ocean SST, we performed three sets of “all forcing” runs using the NCAR CESM1.2 (Hurrell et al. 2013), a fully coupled state-of-the-art climate system model. The all forcing runs were driven by the historical radiative forcing for 1871–2005 and the representative concentration pathway 4.5 (RCP4.5) scenario for 2006–12 based on phase 5 of the Coupled Model Intercomparison Project (CMIP5; Taylor et al. 2012), which has been often used in previous studies (e.g., Kosaka and Xie 2013; Meehl et al. 2014) to update beyond 2005 with the moderate external forcing effect. We also analyzed a total of 129 realizations from eight CMIP5 models for 1861–2005, comprising 45 all forcing runs (Hist), 28 GHG-only forcing runs (Hist_GHG), 26 AA-only forcing runs (Hist_AA), and 30 natural-only forcing runs (Hist_Nat) to isolate the effects of each external forcing function (Table 1). These historical climate simulations are forced by observed atmospheric composition changes reflecting both natural forcing and anthropogenic forcing (Zhou and Yu 2006; Jha et al. 2014). The Hist_GHG, Hist_AA, and Hist_Nat simulations are only forced by well-mixed GHGs, AAs forcing, or natural agents respectively, with other forcings fixed at the preindustrial level (Taylor et al. 2012). We select these eight models as all the four experiments we need—Hist, Hist_GHG, Hist_AA, and Hist_Nat—are available including both the direct and indirect effects of aerosols. The results from eight models are representative of the full suite of CMIP5 models and our approach has been used in previous studies (e.g., Dong et al. 2014a; Wang et al. 2016a,b).
Model names, the institution from which they originate, horizontal resolution of atmospheric components, and the experiments from CMIP5 used in this study. The number denotes the available members for each experiment. Hist stands for historical, Hist_GHG for greenhouse gas forcing, Hist_AA for anthropogenic aerosol forcing, and Hist_Nat for natural forcing (including solar radiation and volcanic eruption). [Expansions of acronyms are available online at http://www.ametsoc.org/PubsAcronymList.]
Because the temporal evolution of fully coupled ocean and atmosphere model simulations is realization dependent, in order to reproduce the observed evolution in models we conducted another experiment with CESM1.2, called Hist_EP. Note that CESM1.2 was not used for CMIP5 but is an updated version of CESM1.0 that was used in CMIP5. This experiment used historical external forcing plus SSTs that were prescribed as the model climatology plus observed daily anomalies in the tropical eastern Pacific (TEP) domain (15°N–15°S, 80°–180°W) based on HadISST (Rayner et al. 2003). Over other basins, the historical external forcing is also prescribed. The SST anomaly was blended with and relaxed to the modeled anomaly in a buffer zone around the TEP domain within five grid boxes. The oceans in other basins (including the Indian Ocean) were fully coupled with the atmosphere. We obtained the daily SST anomalies from the monthly dataset via linear interpolation by putting with the monthly mean at the middle day of each month, which is commonly used in CMIP5 experiments (http://www-pcmdi.llnl.gov/projects/amip/AMIP2EXPDSN/BCS/).
To compare changes in the relationship between Indian Ocean decadal variability and the IPO under different external forcing levels, we also analyzed the historical simulations, RCP2.6 runs, and RCP8.5 runs of CCSM4 (Meehl et al. 2012) from CMIP5 (Taylor et al. 2012). Monthly outputs for all the above datasets and simulations are then averaged to yearly values for analysis, since the seasonality of decadal time scale variations in the Indian Ocean can be neglected (not shown).
b. Methods
Internal variability is not synchronous across the individual realizations in a given set of simulations, so we use the multimodel ensemble mean (MME) to suppress internal variability and isolate the effects of external forcing. The MME is calculated as the arithmetic mean of the equally weighted eight CMIP5 models. In generating the MME, all the models were regridded at a horizontal resolution of 2.8° × 2.8°, which is the lowest resolution among the eight CMIP5 models. We then define the external forcing signal as the MME of the 45 Hist simulations from the eight CMIP5 models. Similarly, the MME of Hist_GHG runs defines the GHG-only forcing signal, the MME of Hist_AA runs defines the response to AA forcing, and the MME of Hist_Nat runs defines the response to natural forcing (including solar radiation and volcanic aerosols). The same method to obtain external forcing effects has been broadly used in climate change attribution studies (e.g., Xie et al. 2013; Dong and Zhou 2014; Dong et al. 2014a).
The limited time duration of CMIP5 historical runs (1861–2005) makes it difficult to quantitatively clarify the contributions of each forcing (such as GHG, AA, and natural external forcing) to the changing relationship between the decadal IOB mode and the IPO after 2000. Thus, longer time span historical runs are needed for analyzing climate change during recent decades. We obtain these runs from the ensemble mean of the three all-forcing runs based on CESM1.2, which largely reduces the magnitude of internal variability, producing a measure of external forcing. Therefore, we can remove this forced signal based on CESM1.2 from the observations and from the CESM1.2 historical simulations with prescribed eastern Pacific SSTs (Hist_EP) to obtain a measure of internal natural variability. We realize that the average over a relatively small number (three) of ensemble runs used for CESM1.2 may not completely remove internal variations. However, we checked the individual runs and found nevertheless that the ensemble means greatly reduce the internal signal [see supplementary Fig. 10 in Dong et al. (2016)]. Zhang et al. (2013) also used a similar three-member ensemble mean to filter out internal variability from CMIP5 simulations.
To identify the decadal time scale signals, we first removed the long-term linear trend and then smoothed with a 13-yr low-pass Lanczos filter (Hamming 1989) to remove interannual variations related to ENSO, the IOB mode, and the Indian Ocean dipole (IOD) mode. We then computed confidence limits using a Monte Carlo technique as described in Dong et al. (2014a), which involves 1) generating two random sequences using a normal distribution and applying the 13-yr low-pass filter to them, 2) computing the correlation coefficient of the two sequences, 3) repeating the first two steps 5000 times, and 4) sorting the correlation coefficients in ascending order. The values for the 95th percentiles are chosen as the significant correlation values at the 5% level.
Following Han et al. (2014a), the dominant mode of Indian Ocean decadal variability was defined as the first empirical orthogonal function (EOF) mode over the Indian Ocean (30°N–30°S, 40°–120°E) on decadal time scales. The leading EOF mode represents basinwide warming and cooling indicative of the decadal IOB mode. The IPO used here is defined as the first EOF mode of decadal SSTs over the Pacific Ocean (60°N–60°S, 120°E–70°W) with the long-term linear trend removed (Power et al. 1999; Meehl et al. 2009b; Bonfils and Benjamin 2011). Note that all the principal components (PCs) of EOF modes in this study are standardized by removing the mean from the total time period and dividing the resulting deviations from the mean by the standard deviation. We then get the EOF patterns by regressing SST anomalies onto the standardized PCs. As a result, the units of PCs are dimensionless and the units of EOF patterns are in °C. Our results are not sensitive to moving the eastern boundary of the Indian Ocean to 100°E to exclude the South China Sea and ITF region east and north of Sumatra and Java.
3. Results
The evolution of the decadal IOB mode and the IPO in three different observational datasets shows results are almost the same as found in Han et al. (2014a) using just HadISST (Fig. 1). All the three SST datasets reveal synchronous evolution of the two indices before 1985 and a much less consistent relationship after 1985. Indeed, a 21-yr running correlation between the decadal IOB mode and IPO (Figs. 1c,f,i,l) shows a transition to negative values around 1985, a changepoint also identified in Han et al. (2014a). In particular, while the IPO enters a negative phase after 2000, the decadal IOB mode stays warm and does not follow the IPO index (Figs. 1c,f,i). The IPO phase transition around 2000 has been discussed in previous studies (Dong et al. 2014a; Dai et al. 2015). In addition, this changed relationship also occurred earlier in 1960s when the IPO was also in its negative phase and CO2 emission increased rapidly (Wang et al. 2016a). Note that the three datasets show a similar evolution after the 1950s, whereas discrepancies are more evident before 1950s, which might be due to the sparse observations in that time period (Deser et al. 2010).
The leading EOF pattern (in °C) and associated standardized principal component (PC) of SST for the Indian Ocean and Pacific Ocean, based on the 13-yr low-pass filtered annual mean SSTs after the long-term trend was removed for 1900–2012: (a)–(c) HadISST, (d)–(f) ERSST, (g)–(i) Kaplan, and (j)–(l) Hist_EP run from CESM1.2. The values given on the top right of the panels represent the explained percentage variance explained by each mode. Red lines show the PC of the decadal IOB mode, black lines show the IPO evolution, and blue lines show the 21-yr running correlation coefficient between the IOB index and the IPO index.
Citation: Journal of Climate 30, 6; 10.1175/JCLI-D-16-0313.1
To evaluate the ability of the global coupled climate model CESM1.2 to simulate this changed decadal relationship, we added observed TEP SST variations to the model in order to prescribe the evolution of the IPO. With the observed IPO included, the model not only captured the patterns of leading modes in both Indian and Pacific Oceans, but also succeeded in reproducing the striking feature of the different evolutions of the two indexes after 1985 (Figs. 1j–l). Thus, collectively, the changed IOB mode and IPO relationship around 1985 are robust in both observations and model. Why their relationship exhibits such a decadal change is a point we elaborate on further below. Our EOF analysis is not affected by excluding the western Pacific warm pool and ITF regions for EOF analysis, as the maximum signals for IPO are located in the eastern equatorial Pacific and North Pacific.
The Indian Ocean witnessed a rapid warming during recent decades, stronger than in other tropical oceans (Luo et al. 2012; Han et al. 2014a). The origin of this warming is still under debate but it is likely the result of external forcing, especially GHG forcing (Dong et al. 2014b; Dong and Zhou 2014). Thus, we might expect to see a consistent relationship between the two indexes after subtracting out external forcing from observations. To test this hypothesis, the impact of external forcing is removed from the observations by subtracting the ensemble mean of the three all-forcing runs from CESM1.2. The resulting internal component of the decadal IOB mode then follows the observed IPO variations closely, even after 1985, including the lower value in the early 1990s, the higher value in the mid-1990s, and the phase transition around 2000 (Fig. 2). The significantly improved correlation coefficients between the two indexes without external forcing during 1900–2012 are common to all the three observed datasets (Figs. 2a–c). Specifically, for the period of 1985–2012, the correlation with the IPO index changes from negative (−0.91, −0.96, and −0.97 for HadISST, ERSST, and Kaplan, respectively) in observations (red lines) to positive (0.69, 0.55, 0.71) with external forcing removed (blue lines). The results based on Hist_EP run from CESM1.2 can well reproduce such changes, with correlation coefficients improved after external forcing removed, from −0.51 to 0.70 for 1985–2012 (Fig. 2d). Furthermore, the decadal IOB mode also enters a negative phase after 2000 following the IPO as a result of internal variability (blue lines). Thus, we conclude that external forcing disrupted the relationship between the IPO and IOB during recent decades and induced a positive phase of the decadal IOB mode after 2000.
Time series of the IPO (black lines), the decadal IOB mode (red lines), and internal component of the decadal IOB mode with external forcing (obtained from ensemble mean of three all forcing runs based on CESM1.2) removed (blue lines) from (a) HadISST, (b) ERSST, (c) Kaplan, and (d) Hist_EP run of CESM1.2. All the time series are standardized. The correlation coefficients with the IPO during 1900–2012 and 1985–2012, respectively, are shown in parentheses.
Citation: Journal of Climate 30, 6; 10.1175/JCLI-D-16-0313.1
To clarify how external forcing changes the relationship between the decadal IOB mode and the IPO during recent decades, we further compare the time series of Indian Ocean–averaged SST anomalies in observations and under external forcing (Fig. 3). Observed Indian Ocean SST warming trends tend to increase in the late 1990s (Figs. 3a–c), consistent with a positive phase of IOB then and in step with an increase in external forcing (Fig. 3d). Note that external forcing based on CMIP5 models and CESM1.2 models are in general agreement with each other, especially for the period after 1960 (Fig. 3d). Thus, the ensemble mean of the three CESM1.2 all-forcing runs provides a robust measure of external forcing for our purposes. The differences between CMIP5 and CESM1.2 prior to 1960 do not affect the results significantly since we are mainly interested in the changed relationship between the IOB and IPO beginning around 1985. These results confirm that the positive phase of the decadal IOB mode after 2000 in Indian Ocean SSTs arises from the enhanced effect of external forcing, which overwhelms the effect of the IPO in determining the evolution of the IOB after 1985. It is worth noting that the rapid warming of the Indian Ocean SST under external forcing since the mid-1990s (Fig. 3d) is attributed to recovery from the cooling effect of the volcanic eruption of Pinatubo in 1991 (Smith et al. 2016) and a rapid increase of CO2 concentration since the 1960s (Wang et al. 2016a).
Time series of 13-yr low-pass filtered Indian Ocean-averaged (30°N–30°S, 40°–120°E) SST anomalies (in °C) from observations based on (a) HadISST, (b) ERSST, and (c) Kaplan, as well as (d) external forcing obtained from ensemble mean of three all forcing runs based on CESM1.2 (black line) and 45 historical runs based on eight CMIP5 models (red line). The dashed lines represent the long-term linear trend of each product.
Citation: Journal of Climate 30, 6; 10.1175/JCLI-D-16-0313.1
Spatial patterns of decadal anomalies for the negative IPO period of 2000–12 are calculated by averaging SST anomalies relative to the climatological state based on 1900–2012 (Fig. 4). The period from 2000 onward is chosen because it coincides with the global warming hiatus (Easterling and Wehner 2009; Meehl et al. 2011; Kosaka and Xie 2013) and the negative phase of the IPO (Dong et al. 2014a; Dai et al. 2015). For the observations, SST decadal variations after the long-term linear trend removed show warm anomalies over the whole Indian Ocean basin, which, although barely significant, are consistent with a signature of the positive phase of the IOB, while a negative IPO pattern covers the Pacific Ocean (Fig. 4a). The effect of external forcing during this period significantly favors warm SST anomalies over most of the global oceans (Fig. 4b), consistent with a period of accelerated warming in the observations. Removing the effect of external forcing derived from CESM1.2 in both observations (Fig. 4c) and Hist_EP run (Fig. 4d) reveals a negative IPO associated with weak marginally significant cool SST anomalies over most of the Indian Ocean, which is consistent with the influence of variability associated with the IPO on Indian Ocean SST decadal variations as described by Dong et al. (2016). Thus, the resultant warm anomalies over the Indian Ocean are weak in the observations (Fig. 4a), due to the competing effects of internal variability and external forcing. In conclusion, external forcing interferes with the ability of the IPO to induce a phase change in the decadal IOB mode to negative values after 2000.
Patterns of SST decadal anomalies (in °C) averaged for the negative IPO period of 2000–12, based on the unfiltered annual SST with the climatological state from 1900 to 2012 removed for (a) HadISST, (b) the external forcing obtained from the ensemble mean of three all forcing runs from CESM1.2, (c) the internal component of the HadISST observations with external forcing removed, (d) the internal component of Hist_EP. Note that the long-term (1900–2012) linear trend is removed from (a) and (b) to highlight the decadal signals. The dotted areas indicate the decadal changes are statistically significant at the 95% level of confidence.
Citation: Journal of Climate 30, 6; 10.1175/JCLI-D-16-0313.1
To address the question of what causes the enhanced external forcing (including anthropogenic forcing and natural forcing) in recent decades, the individual roles of GHG and AA, the two main components of anthropogenic forcing, as well as natural forcing (solar radiation and volcanic aerosols), are examined based on the historical simulations of CMIP5 models (Fig. 5; Table 1). We first remove the long-term trends from each ensemble to emphasize the decadal variations around those trends. Time series of Indian Ocean averaged SST under the influence of different detrended forcings indicate that the large negative values in external forcing during 1880–90, 1900–10, 1960–70, and 1990–2000 arise mainly from natural forcing. This natural forcing is predominantly the result of strong tropical volcanic eruptions (e.g., T. Wang et al. 2012; Zhang 2016), with variations in solar irradiance showing a regular small-amplitude 11-yr cycle (Zhang 2016). However, after the 1980s, enhanced external forcing can be primarily attributed to GHGs, while AA forcing and natural forcing shows relatively weak impacts (Fig. 5a).
(a) Time series of 13-yr low-pass filtered Indian Ocean-averaged SST (in °C) after the long-term linear trend is removed from the external forcing (black line), GHG-only forcing (red line), AA-only forcing (blue line), and natural-only forcing (orange line) based on eight CMIP5 models shown in Table 1. Patterns of SST anomalies (in °C) averaged for 2000–05, based on the unfiltered annual SST with the climatological mean and long-term linear trend during 1861–2005 removed from (b) external forcing, (c) GHG forcing, (d) AA forcing, (e) natural forcing. The dotted areas indicate the decadal changes are statistically significant at the 95% level of confidence.
Citation: Journal of Climate 30, 6; 10.1175/JCLI-D-16-0313.1
To examine the spatial pattern of SST anomalies in the Indian Ocean due to these forcings for the most recent period, we show the SST anomalies average over 2000–05 from the CMIP5 historical runs. (A longer interval after 2000 would have been desirable but unfortunately the historical runs end in 2005.) The results confirm that GHG is the main contributor to the accelerated Indian Ocean SST rise under external forcing apparent at the end of the record (Figs. 5b,c), whereas AA forcing favors a negative phase of the decadal IOB mode after 2000 (Fig. 5d). Natural forcing, mainly including solar radiation and volcanic aerosols, would lead to relatively weak positive anomalies in Indian Ocean SST due to the recovery from the cooling effect of the volcanic eruption of Pinatubo in 1991, but this effect is weak compared with GHG forcing and not statistically significant (Fig. 5e).
The effect of external forcing can be further examined by comparing experiments under three conditions based on CMIP5 models: 1) climate simulations with external forcing at historical levels, 2) RCP2.6 runs with external forcing under the weakest emission scenario for twenty-first-century projections, and 3) RCP8.5 runs with external forcing under the strongest emission scenario for twenty-first-century projections (Taylor et al. 2012). All the CCSM4 simulations produce IPO-like patterns, which, although they do not exactly replicate the observed IPO, nonetheless show a tropical maximum and midlatitude minima in the SST. The pattern correlation coefficients between simulations and observations (HadISST, ERSST, and Kaplan) are statistically significant at the 95% level of confidence (respectively, 0.50, 0.59, 0.57 for the historical run; 0.46, 0.62, 0.55 for the RCP2.6 run; and 0.58, 0.71, and 0.67 for the RCP8.5 run). Thus, we will consider these model structures as representative of the IPO observed in nature.
Under weak external forcing in the historical and RCP2.6 runs, the decadal IOB mode is well correlated with the IPO. Correlation coefficients are 0.76 and 0.88, respectively, which are statistically significant at the 95% level of confidence (Figs. 6g,h). In contrast, with stronger external forcing in the RCP8.5 run, the correlation of the decadal IOB mode with the IPO is degraded to 0.53 (Fig. 6i). Similarly, for the historical simulation, the correlation coefficient between the two indexes is much higher during 1850–1950 (r = 0.83) under weaker external forcing than that during 1951–2005 (r = 0.48) under stronger external forcing (Fig. 6g). Similar results are found in different realizations of CCSM4 and CanESM2 (not shown). We conclude therefore that for weak external forcing, the decadal IOB is highly and positively correlated with the IPO. However, as external forcing increases, these two modes of variability become less synchronized because external forcing competes with the IPO in modulating the decadal IOB mode. In particular, external forcing and a positive phase of the IPO have reinforcing effects on the positive IOB phase, while external forcing and a negative phase of the IPO have competing effects on the IOB.
(a)–(c) The first EOF patterns of decadal SST (in °C) for the Indian Ocean and (d)–(f) the IPO pattern (in °C) for the historical run, RCP2.6 run, and RCP8.5 run, respectively, from the CCSM4 model. (g)–(i) The associated standardized PCs based on the 13-yr low-pass filtered annual SST after long-term trend removed. The numbers at the top right denote the explained percentage variance by each mode for (a)–(f) and the correlation coefficient between the two indexes in (g)–(i).
Citation: Journal of Climate 30, 6; 10.1175/JCLI-D-16-0313.1
SST variability in the tropical Indo-Pacific Ocean has a large impact on global climate, including tropical precipitation (Xie et al. 2010), the global monsoons (B. Wang et al. 2012), the Walker circulation (Li and Ren 2012), and the east–west expansion of the western Pacific subtropical high and South Asian high (Zhou et al. 2009), in addition to regional rainfall changes (e.g., Arblaster et al. 2002; Giannini et al. 2003; Ashok et al. 2003; Li et al. 2010; Zhang and Zhou 2011). The out-of-phase relationship between the two basins during the recent hiatus is associated with unique SST patterns different from those during previous IPO negative phases. For example, SST anomalies averaged for 1947–75 show that cold SST anomalies in the Indian Ocean corresponded to the negative phase of the IPO in the Pacific Ocean, with weak easterly anomalies in the eastern Pacific Ocean and westerly anomalies in the west (Fig. 7a). In contrast, during 2000–12 a negative IPO phase corresponded with warm SST anomalies in the Indian Ocean. At the same time, stronger easterly anomalies covered the entire tropical Pacific Ocean, while westerly anomalies prevailed in the tropical Indian Ocean (Fig. 7b). Differences between these two IPO negative phases suggest that the anomalous warm SSTs in the Indian Ocean during 2000–12 give rise to stronger easterly wind anomalies (Luo et al. 2012) and a stronger Walker circulation over the tropical Pacific Ocean (Fig. 7c), leading to rapid sea level rise in the western tropical Pacific (Han et al. 2014a) and enhanced cooling SST anomalies in the eastern Pacific (Fig. 7c). Thus, while the negative phase of the IPO during 2000–12 resulted from natural internal variability, it was enhanced by the rapid Indian Ocean SST warming in response to external forcing (cf. Figs. 4c and 7c). Stronger Pacific easterlies also contributed to increased storage of heat in the Indo-Pacific and thus to the global warming hiatus (England et al. 2014; Lee et al. 2015).
Patterns of SST anomalies (in °C) and 1000-hPa wind anomalies (in m s−1) averaged for the negative IPO period of (a) 1947–75, (b) 2000–12, and (c) differences between 2000–12 and 1947–75 based on the unfiltered annual SST with both the climatological mean state and linear trend during the whole time period (1871–2012) removed for Hist_EP run from CESM1.2.
Citation: Journal of Climate 30, 6; 10.1175/JCLI-D-16-0313.1
In summary, we hypothesize that increased GHG forcing since 2000 induced stronger warming and a positive decadal IOB mode in the Indian Ocean, which enhanced the easterly wind anomalies over the tropical Pacific Ocean, contributing to an extreme phase of the negative IPO (England et al. 2014). This extreme and prolonged cooling in the eastern Pacific then played a dominant role in the recent global warming hiatus (Kosaka and Xie 2013; Dai et al. 2015), which appears to have ended as the IPO switched phase to positive (Thoma et al. 2015; Meehl et al. 2016), coincident with the onset of El Niño conditions in 2014–16 (McPhaden 2015). We infer that the relation between the IPO and the decadal IOB will be positive again after this IPO transition, as both a positive IPO and enhanced external forcing will favor a positive IOB mode.
4. Summary and discussion
The main motivation of the present study is to explore the reason for the changed relationship between the decadal IOB mode and IPO in recent decades. We have analyzed three different SST datasets plus the CESM1.2 and CMIP5 coupled global climate models to infer the potential role of internal variability and external forcing (namely greenhouse gases, anthropogenic aerosols, and natural forcing) in affecting the changing relationship. The results show that the evolution of the decadal IOB mode is dominated by the IPO before 1985. However, after 1985 the enhanced effect of external forcing overwhelms that of the IPO in determining the evolution of the IOB, altering its relationship with the IPO. Similar conclusions can be found by comparing the results of the CMIP5 experiments based on CCSM4 and CanESM2 with different external forcing levels in historical simulations, RCP2.6 runs and RCP8.5 runs. While the IPO enters a negative phase after 2000, the Indian Ocean stays warm and does not follow the IPO due to an acceleration in external forcing, especially from enhanced GHG forcing. The out-of-phase relationship between the IOB and IPO is associated with unique SST patterns in the Indo-Pacific region, featuring a warm Indian Ocean corresponding to a cool tropical Pacific since 2000. This anomalous Indo-Pacific SST pattern favored enhanced easterly wind anomalies over the tropical Pacific Ocean, resulting in a pronounced global warming hiatus in the early twenty-first century. Therefore, to the extent that the recent global warming hiatus was dominated by a negative IPO that arose from natural internal variability, it was further enhanced by rapid Indian Ocean warming in response to external forcing.
We note that the out-of-phase relationship between the decadal IOB mode and IPO after 1985 may indicate an active role for Indian Ocean in modulating the Indo-Pacific climate. In particular, the enhanced tropical warming in the Indian Ocean since 1950 relative to the Pacific Ocean favors stronger Pacific trade winds (Luo et al. 2012) and a negative IPO pattern (England et al. 2014). Therefore, we can infer that Pacific Ocean dominates Indo-Pacific climate variations on decadal time scales before 1985, whereas the Indian Ocean makes a substantial contribution after 1985. The relative importance of SST variations in the two basins to the recent global warming hiatus is an open question beyond the scope of this study, but one that deserves further attention.
Several studies (Li et al. 2015; Kucharski et al. 2016) have argued that the positive phase of the AMO in recent decades led to the formation of Indo-Pacific SST anomalies with cooling in the eastern Pacific and warming in the western Pacific and Indian Ocean. Based on these studies, one could argue that the positive AMO favored a positive phase of the decadal IOB mode as observed. Even so, our analyses stress the critical role of external forcing in inducing a positive IOB mode, especially since 2000. We thus infer that, even if the AMO affects the IOB, the AMO’s impact is weaker than that of external forcing. However, this issue requires further research to quantitatively determine the relative importance of external forcing and the AMO on the decadal IOB mode.
Acknowledgments
This research was performed while the first author held a National Research Council Research Associateship Award at NOAA/PMEL. We thank Dr. Fengfei Song and Dr. Bo Wu of LASG/Institute of Atmospheric Physics, Chinese Academy of Sciences, who together completed the CESM1.2 experiments. We also thank three anonymous reviewers for their thoughtful and constructive comments on an earlier version of this manuscript.
REFERENCES
Arblaster, J., G. Meehl, and A. Moore, 2002: Interdecadal modulation of Australian rainfall. Climate Dyn., 18, 519–531, doi:10.1007/s00382-001-0191-y.
Ashok, K., Z. Guan, and T. Yamagata, 2003: Influence of the Indian Ocean dipole on the Australian winter rainfall. Geophys. Res. Lett., 30, 1821, doi:10.1029/2003GL017926.
Bonfils, C., and D. Benjamin, 2011: Investigating the possibility of a human component in various Pacific decadal oscillation indices. Climate Dyn., 37, 1457–1468, doi:10.1007/s00382-010-0920-1.
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.
Deser, C., A. S. Phillips, and M. A. Alexander, 2010: Twentieth century tropical sea surface temperature trends revisited. Geophys. Res. Lett., 37, L10701, doi:10.1029/2010GL043321.
Dong, L., and T. Zhou, 2014: The Indian Ocean sea surface temperature warming simulated by CMIP5 models during the 20th century: Competing forcing roles of GHGs and anthropogenic aerosols. J. Climate, 27, 3348–3362, doi:10.1175/JCLI-D-13-00396.1.
Dong, L., T. Zhou, and X. Chen, 2014a: Changes of Pacific decadal variability in the twentieth century driven by internal variability, greenhouse gases, and aerosols. Geophys. Res. Lett., 41, 8570–8577, doi:10.1002/2014GL062269.
Dong, L., T. Zhou, and B. Wu, 2014b: Indian Ocean warming during 1958–2004 simulated by a climate system model and its mechanism. Climate Dyn., 42, 203–217, doi:10.1007/s00382-013-1722-z.
Dong, L., T. Zhou, A. Dai, F. Song, B. Wu, and X. Chen, 2016: The footprint of the inter-decadal Pacific oscillation in Indian Ocean sea surface temperatures. Sci. Rep., 6, 21251, doi:10.1038/srep21251.
Easterling, D. R., and M. F. Wehner, 2009: Is the climate warming or cooling? Geophys. Res. Lett., 36, L08706, doi:10.1029/2009GL037810.
Enfield, D., A. Mestas-Nuñez, and P. 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. H., 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.
Feng, M., M. J. McPhaden, and T. Lee, 2010: Decadal variability of the Pacific subtropical cells and their influence on the southeast Indian Ocean. Geophys. Res. Lett., 37, L09606, doi:10.1029/2010GL042796.
Giannini, A., R. Saravanan, and P. Chang, 2003: Oceanic forcing of Sahel rainfall on interannual to interdecadal time scales. Science, 302, 1027–1030, doi:10.1126/science.1089357.
Goddard, L., W. Baethgen, B. Kirtman, and G. A. Meehl, 2009: The urgent need for improved models and predictions. Eos, Trans. Amer. Geophys. Union, 90, 343, doi:10.1029/2009EO390004.
Hamming, R. W., 1989: Digital Filters. Prentice Hall, 284 pp.
Han, W., and Coauthors, 2010: Patterns of Indian Ocean sea-level change in a warming climate. Nat. Geosci., 3, 546–550, doi:10.1038/ngeo901.
Han, W., and Coauthors, 2014a: Intensification of decadal and multi-decadal sea level variability in the western tropical Pacific during recent decades. Climate Dyn., 43, 1357–1379, doi:10.1007/s00382-013-1951-1.
Han, W., J. Vialard, M. J. McPhaden, T. Lee, Y. Masumoto, M. Feng, and W. P. M. de Ruijter, 2014b: Indian Ocean decadal variability: A review. Bull. Amer. Meteor. Soc., 95, 1679–1703, doi:10.1175/BAMS-D-13-00028.1.
Hurrell, J. W., G. A. Meehl, D. Bader, T. Delworth, B. Kirtman, and B. Wielicki, 2009: A unified approach to climate system prediction. Bull. Amer. Meteor. Soc., 90, 1819–1832, doi:10.1175/2009BAMS2752.1.
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.
Jha, B., Z. Hu, and A. Kumar, 2014: SST and ENSO variability and change simulated in historical experiments of CMIP5 models. Climate Dyn., 42, 2113–2124, doi:10.1007/s00382-013-1803-z.
Kaplan, A., M. A. Cane, Y. Kushnir, A. C. Clement, M. B. Blumenthal, and B. Rajagopalan, 1998: Analyses of global sea surface temperature 1856–1991. J. Geophys. Res., 103, 18 567–18 589, doi:10.1029/97JC01736.
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.
Krishnamurthy, L., and V. Krishnamurthy, 2016: Decadal and interannual variability of the Indian Ocean SST. Climate Dyn., 46, 57–70, doi:10.1007/s00382-015-2568-3.
Kucharski, F., and Coauthors, 2016: The teleconnection of the tropical Atlantic to Indo-Pacific sea surface temperatures on inter-annual to centennial time scales: A review of recent findings. Atmosphere, 7, 29, doi:10.3390/atmos7020029.
Lee, S.-K., W. Park, M. O. Baringer, A. L. Gordon, B. Huber, and Y. Liu, 2015: Pacific origin of the abrupt increase in Indian Ocean heat content during the warming hiatus. Nat. Geosci., 8, 445–449, doi:10.1038/ngeo2438.
Lee, T., and M. J. McPhaden, 2008: Decadal phase change in large-scale sea level and winds in the Indo-Pacific region at the end of the 20th century. Geophys. Res. Lett., 35, L01605, doi:10.1029/2007GL032419.
Li, G., and B. Ren, 2012: Evidence for strengthening of the tropical Pacific Ocean surface wind speed during 1979–2001. Theor. Appl. Climatol., 107, 59–72, doi:10.1007/s00704-011-0463-3.
Li, H., A. Dai, T. Zhou, and J. Lu, 2010: Responses of East Asian summer monsoon to historical SST and atmospheric forcing during 1950–2000. Climate Dyn., 34, 501–514, doi:10.1007/s00382-008-0482-7.
Li, X., S.-P. Xie, S. T. Gille, and C. Yoo, 2015: Atlantic-induced pan-tropical climate change over the past three decades. Nat. Climate Change, 6, 275–279, doi:10.1038/nclimate2840.
Li, Y., and W. Han, 2015: Decadal sea level variations in the Indian Ocean investigated with HYCOM: Roles of climate modes, ocean internal variability, and stochastic wind forcing. J. Climate, 28, 9143–9165, doi:10.1175/JCLI-D-15-0252.1.
Luo, J.-J., W. Sasakia, and Y. Masumoto, 2012: Indian Ocean warming modulates Pacific climate change. Proc. Natl. Acad. Sci. USA, 109, 18 701–18 706, doi:10.1073/pnas.1210239109.
Mantua, N., S. Hare, Y. Zhang, J. Wallace, and R. Francis, 1997: A Pacific interdecadal oscillation with impacts on salmon production. Bull. Amer. Meteor. Soc., 78, 1069–1079, doi:10.1175/1520-0477(1997)078<1069:APICOW>2.0.CO;2.
McPhaden, M. J., 2015: Playing hide and seek with El Niño. Nat. Climate Change, 5, 791–795, doi:10.1038/nclimate2775.
Meehl, G., A. Hu, and B. Santer, 2009a: The mid-1970s climate shift in the Pacific and the relative roles of forced versus inherent decadal variability. J. Climate, 22, 780–792, doi:10.1175/2008JCLI2552.1.
Meehl, G., and Coauthors, 2009b: Decadal prediction: Can it be skillful? Bull. Amer. Meteor. Soc., 90, 1467–1484, doi:10.1175/2009BAMS2778.1.
Meehl, G., 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., and Coauthors, 2012: Climate system response to external forcings and climate change projections in CCSM4. J. Climate, 25, 3661–3683, doi:10.1175/JCLI-D-11-00240.1.
Meehl, G., and Coauthors, 2014: Decadal climate prediction: An update from the trenches. Bull. Amer. Meteor. Soc., 95, 243–267, doi:10.1175/BAMS-D-12-00241.1.
Meehl, G., A. Hu, and H. Teng, 2016: Initialized decadal prediction for transition to positive phase of the Interdecadal Pacific Oscillation. Nat. Commun., 7, 11718, doi:10.1038/ncomms11718.
Nieves, V., J. K. Willis, and W. C. Patzert, 2015: Recent hiatus caused by decadal shift in Indo-Pacific heating. Science, 349, 532–535, doi:10.1126/science.aaa4521.
Power, S., T. Casey, C. Folland, A. Colman, and V. Mehta, 1999: Inter-decadal modulation of the impact of ENSO on Australia. Climate Dyn., 15, 319–324, doi:10.1007/s003820050284.
Rayner, N. A., D. E. Parker, E. B. Horton, C. K. Folland, L. V. Alexander, D. P. Rowell, E. C. Kent, and 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, 2014: Volcanic contribution to decadal changes in tropospheric temperature. Nat. Geosci., 7, 185–189, doi:10.1038/ngeo2098.
Santer, B. D., and Coauthors, 2015: Observed multivariable signals of late 20th and early 21st century volcanic activity. Geophys. Res. Lett., 42, 500–509, doi:10.1002/2014GL062366.
Smith, D. M., and Coauthors, 2016: Role of volcanic and anthropogenic aerosols in the recent global surface warming slowdown. Nat. Climate Change, 6, 936–940, doi:10.1038/nclimate3058.
Smith, T. M., R. W. Reynolds, T. C. Peterson, and J. Lawrimore, 2008: Improvements to NOAA’s historical merged land–ocean surface temperature analysis (1880–2006). J. Climate, 21, 2283–2296, doi:10.1175/2007JCLI2100.1.
Solomon, S., K. H. Rosenlof, R. W. Portmann, J. S. Daniel, S. M. Davis, T. J. Sanford, and G. Plattner, 2010: Contributions of stratospheric water vapor to decadal changes in the rate of global warming. Science, 327, 1219–1223, doi:10.1126/science.1182488.
Taylor, K. E., R. J. Stouffer, and G. A. Meehl, 2012: An overview of CMIP5 and the experiment design. Bull. Amer. Meteor. Soc., 93, 485–498, doi:10.1175/BAMS-D-11-00094.1.
Thoma, M., R. J. Greatbatch, C. Kadow, and R. Gerdes, 2015: Decadal hindcasts initialized using observed surface wind stress: Evaluation and prediction out to 2024. Geophys. Res. Lett., 42, 6454–6461, doi:10.1002/2015GL064833.
Timmermann, A., S. McGregor, and F.-F. Jin, 2010: Wind effects on past and future regional sea level trends in the southern Indo-Pacific. J. Climate, 23, 4429–4437, doi:10.1175/2010JCLI3519.1.
Trenberth, K. E., J. T. Fasullo, G. Branstator, and A. S. Phillips, 2014: Seasonal aspects of the recent pause in surface warming. Nat. Climate Change, 4, 911–916, doi:10.1038/nclimate2341.
Wang, B., J. Liu, H.-J. Kim, P. J. Webster, and S.-Y. Yim, 2012: Recent change of the global monsoon precipitation (1979–2008). Climate Dyn., 39, 1123–1135, doi:10.1007/s00382-011-1266-z.
Wang, H., S.-P. Xie, and Q. Liu, 2016a: Comparison of climate response to anthropogenic aerosol versus greenhouse gas forcing: Distinct patterns. J. Climate, 29, 5175–5188, doi:10.1175/JCLI-D-16-0106.1.
Wang, H., S.-P. Xie, H. Tokinaga, Q. Liu, and Y. Kosaka, 2016b: Detecting cross-equatorial wind change as a fingerprint of climate response to anthropogenic aerosol forcing. Geophys. Res. Lett., 43, 3444–3450, doi:10.1002/2016GL068521.
Wang, T., O. H. Otterå, Y. Gao, and H. Wang, 2012: The response of the North Pacific decadal variability to strong tropical volcanic eruptions. Climate Dyn., 39, 2917–2936, doi:10.1007/s00382-012-1373-5.
Xie, S.-P., C. Deser, G. A. Vecchi, J. Ma, H. Y. Teng, and A. T. Wittenberg, 2010: Global warming pattern formation: Sea surface temperature and rainfall. J. Climate, 23, 966–986, doi:10.1175/2009JCLI3329.1.
Xie, S.-P., B. Lu, and B. Xiang, 2013: Similar spatial patterns of climate responses to aerosol and greenhouse gas changes. Nat. Geosci., 6, 828–832, doi:10.1038/ngeo1931.
Zhang, L., 2016: The roles of external forcing and natural variability in global warming hiatuses. Climate Dyn., 47, 3157–3169, doi:10.1007/s00382-016-3018-6.
Zhang, L. X., and T. J. Zhou, 2011: An assessment of monsoon precipitation changes during 1901–2001. Climate Dyn., 37, 279–296, doi:10.1007/s00382-011-0993-5.
Zhang, R., and Coauthors, 2013: Have aerosols caused the observed Atlantic multidecadal variability? J. Atmos. Sci., 70, 1135–1144, doi:10.1175/JAS-D-12-0331.1.
Zhou, T. J., and R. Yu, 2006: Twentieth-century surface air temperature over China and the globe simulated by coupled climate models. J. Climate, 19, 5843–5858, doi:10.1175/JCLI3952.1.
Zhou, T. J., and Coauthors, 2009: Why the western Pacific subtropical high has extended westward since the late 1970s. J. Climate, 22, 2199–2215, doi:10.1175/2008JCLI2527.1.