Start Submission Become a Reviewer

Reading: The Controlling Mechanisms of the Recent Global Warming Hiatus: A Focus on the Internal Vari...

Download

A- A+
Alt. Display

Original Research Papers

The Controlling Mechanisms of the Recent Global Warming Hiatus: A Focus on the Internal Variabilities

Authors:

Ruijian Gou ,

Key Laboratory of Physical Oceanography and Frontiers Science Center for Deep Ocean Multispheres and Earth System, Ocean University of China, Qingdao; College of Oceanic and Atmospheric Sciences, Ocean University of China, Qingdao, CN
X close

Yuhang Liu,

Institute for Marine and Antarctic Studies, University of Tasmania, Hobart, Tasmania, AU; College of Oceanic and Atmospheric Sciences, Ocean University of China, Qingdao, CN
X close

Chengcheng Wang

College of Oceanic and Atmospheric Sciences, Ocean University of China, Qingdao, CN
X close

Abstract

A flattening trend of global surface temperature from 1998 to 2013 is generally referred to global warming hiatus. In this review, its basics and controlling mechanisms are the focuses with the latter receiving more attention. The mechanisms for the hiatus are largely derived from internal climate variabilities which could be divided into sea surface temperature (SST) and energy variabilities. The major SST variabilities are Interdecadal Pacific Oscillation (IPO) and Atlantic Multi-decadal Oscillation (AMO). The negative phase of IPO during the hiatus is associated with strengthened Pacific trade winds that enhance subsurface heat uptake, which is generally thought to be the predominant mechanism for the hiatus. Furthermore, AMO influences the global surface temperature at decadal time scales, or even the IPO, which are considered to be the main drivers of hiatus. As for the energy variabilities, vertical uptake and interbasin redistribution of heat are also suggested to be capable of leading to a hiatus period. There was significant warming of global ocean below 700 m during the hiatus as an evidence for the storage of excessive heat in deep ocean. And various heat redistribution patterns like those through ITF or AMOC could also play a significant role in regulating the hiatus.

How to Cite: Gou, R., Liu, Y. and Wang, C., 2022. The Controlling Mechanisms of the Recent Global Warming Hiatus: A Focus on the Internal Variabilities. Tellus A: Dynamic Meteorology and Oceanography, 74(1), pp.172–186. DOI: http://doi.org/10.16993/tellusa.38
69
Views
30
Downloads
  Published on 08 Apr 2022
 Accepted on 18 Feb 2022            Submitted on 18 Feb 2022

The basics of the hiatus

The definitions of the hiatus

The global mean surface temperature (GMST) was relatively flat from the start of the 21st century to around 2013–2014, after a rise of 0.5°C beginning from the mid-1970s (Easterling & Wehner, 2009; Foster & Rahmstorf, 2011). The increasing rate is only one-third to one-half of the mean over the latter half of the 20th century, in contrast to most of the climate models of the fifth phase of the Coupled Model Intercomparison Project (CMIP5) simulating unabated strong warming (England et al., 2015; Fyfe et al., 2013; Kosaka & Xie, 2013). This review adopts the term “hiatus” to refer to this slowdown (also referred to as “pause”) in global surface warming.

According to the review by Easterling & Wehner (2009), three main definitions are used to characterize the hiatus: (i) the trend of GMST is negative, zero, or weakly positive; (ii) the preceding long-term warming trend is higher than the observationally estimated trend; (iii)the warming trend is lower than model simulated projections. Especially, definition (i) was the original definition that was common used in the literature.

The Latest occurrences and duration

The last notable slowdown in global warming occurred during the 1950s–1970s called “Big Hiatus” (Fyfe et al, 2016; Meehl et al., 2016). While this recent hiatus is unique in occurring under the scenario of continually increasing anthropogenic forcing (e.g. atmospheric greenhouse gas concentration) (Fyfe et al., 2016), which challenges the prevailing point that anthropogenic forcing drives global climate warming.

This global warming hiatus is generally thought to be from 1998 to 2013 and the global warming trend is suggested to have resumed in 2014 (Trenberth, 2015), with recently each year being the warmest year on record. The occurrence of an unusually intense El Niño in 1997/1998, as the beginning of the hiatus, leads to the increase of GMST up to 0.2°C compared with the predicted long-term warming trend (Hansen et al., 2006). Speculation arose about whether the global warming resuming is due to the El Niño event and a switch of the phases of the Interdecadal Pacific Oscillation (IPO) since early 2014. Besides, a recent study has made the case for this scenario whereby an El Niño event can trigger a transition of the IPO (Meehl et al., 2021a).

The hiatus period differs from the warming period in the cooler Southern Ocean and eastern tropical Pacific, while the rest part of the globe became warmer in the hiatus (Trenberth & Fasullo, 2013). Furthermore, both land and ocean in boreal winter show a decrease in the temperature by defining the hiatus as the above period (Cohen et al., 2012).

The main study focuses

The initial studies of this hiatus focused on explaining the difference between observed and modelling temperature trends in the 21st century. Recognized as a multi-decadal fluctuation superimposed on the overall warming trend (“rising staircase”; Kosaka & Xie, 2016), the hiatus cannot be simply attributed to the same reasons for global warming like the rise in greenhouse gases. In an analysis of the hiatus, various mechanisms have been proposed and mainly two schools of thought are proposed regarding the cause.

One school suggests external forcing change such as a slowdown in radiative forcing caused by stratospheric water vapor (Solomon et al., 2010) or aerosols caused by anthropogenic or volcanic activity (Kaufmann et al., 2011; Santer et al., 2014; Solomon et al., 2011), and reducing solar irradiance (Kaufmann et al., 2011; Schmidt et al., 2014). While since models simulate their own phasing, frequency and magnitude, of internal climate variability, the other regards the hiatus as a part of internal climate variabilities, such as Pacific decadal variability (e.g. Kosaka & Xie, 2013), Atlantic multi-decadal variability (e.g. Dai et al., 2015), increased vertical ocean heat uptake (e.g. Meehl et al., 2011), interbasin ocean heat redistribution (e.g. Nieves et al., 2015). Furthermore, the first two mentioned internal variabilities can be further clarified into “the SST view” with the latter two clarified into “the energy view” (Xie & Kosaka, 2017). The sea surface temperature (SST) and the energy views are not mutually exclusive, as the SST-related decadal variabilities could cause the uptake and redistribution of the heat in the climate system.

Comparing the external forcings and internal variabilities in forcing the hiatus

An trigger on the decadal timescale for the hiatus in the CMIP5 models is the increase of solar radiation reflection at deep-water formation sites, caused by an increased sea-ice cover. While a trigger on the interannual timescale is the increased reflection of solar radiation caused by the increase of cloud cover related to the La Niña phase and the subsequent reduction of the latent heat release (Drijfhout, 2018).

Using an observationally constrained model, Roberts et al. (2015) indicated that the warming hiatus may have lasted above 20 years relying on only internal variability, and speculated that if both external forcings and internal variability were important, this hiatus could have lasted even longer and be more extreme, offsetting more greenhouse-induced warming.

Models are capable of separating the external forcing from the internal climate variability, with the former given by the ensemble mean of simulations. It was proposed that external forcings are the dominant factor in GMST trend over a century while GMST trend on timescales shorter than a decade is more typically related to internal climate variability. The trend between these two timescales is primarily the result of the interaction between the external forcing and internal climate variabilities (Knutson et al., 2016; Watanabe et al., 2014; Yao et al., 2017).

Miller et al. (2020) utilized a model that makes use of the forcing data over the period from 1850 to 2016 to forecast the next hiatus, and found a multidecadal hiatus between 2023 and 2061, during which the GMST is predicted to grow by only 0.0001°C/yr. The year 2031 and 2061 are respectively identified of the next maximum and minimum of the global SST fluctuation, which is intimately linked to the AMO and caused by the variations of the AMOC (Miller et al., 2020).

The role of internal variabilities

The oceans are the source of low-frequency variability due to their higher heat capacity, compared to the land and atmosphere. Decadal-scale internal variability can generate GMST trends of about ± 0.25°C and sustain the deviations for decades (Hunt, 2010). Model simulations have shown that internal variability in heat uptake and ocean temperatures could mask the anthropogenic warming trend in GMST over a decade (Palmer et al., 2011).

Nonetheless, the internal climate variability in the model can’t be in phase with the observed variability (Risbey et al., 2014), since it is hard to estimate from observations owing to the underlying forced signal and the short record. However, differences between the observations and models can be reconciled if considering the actual internal climate variabilities in the models (Maher et al., 2014; Risbey et al., 2014). But the point to be made here is that an ensemble average of model simulations will, by definition, average out the internal variability leaving only the externally forced response. Therefore, an internally generated signal, such as the strong negative IPO of the early 2000s hiatus, will not show up in an ensemble average of model simulations. However, Meehl et al. (2014, Fig. 1) showed for the CMIP5 model simulations that 10 ensemble members from various models actually, by chance, happened to have their internally generated variability match the randomly occurring internal variability of the observations, with the observed slowdown in global warming during the hiatus comparable to those 10 ensemble members, and the SST signature of those 10 ensemble members showing a negative phase of the IPO as observed. Therefore, the models are capable of simulating hiatus periods with negative IPO, but to actually match what was observed, the model internal variability in a fraction of all ensemble members has to match, by chance, that in the observations. As for the early-2000s hiatus, the primary cause is thought to be internal climate variability rather than external forcings that are quantitatively insufficient to explain the hiatus (Trenberth & Fasullo, 2013; Watanabe et al., 2014; Fyfe et al., 2016).

The hiatus could have continued for another few years due to internal climate variabilities indicated below (Knutson et al., 2016; Roberts et al., 2015). However, the hiatus would have ceased to exist eventually, since those variabilities would reverse phase. More specifically, the warm water subducted through the shallow overturning cells would ultimately re-emerge at the surface of the western Pacific, or the IPO would eventually turn to a positive phase as was the case in 2014.

In a word, the difference of the GMST trend during the hiatus period between CMIP5 models and observations is caused at least in part by internal climate variabilities (Guemas et al., 2013; Meehl et al., 2011; Meehl et al., 2013). And part of the difference is likely the result of the errors in the modeled radiative forcing or the model response to radiative forcing (Santer et al., 2014; Solomon et al., 2011).

SST-related mechanism: Pacific Decadal Variability

An introduction

Except for the global warming trend, leading modes of the unfiltered global SST are ENSO, Pacific Decadal Oscillation (PDO), and Atlantic Multi-decadal Oscillation (AMO). It has been indicated that decadal variability in the Pacific Ocean is largely responsible for this hiatus period in global warming, which is related to a persistent negative phase of the PDO, or more generally the IPO for the whole Pacific, as well as the increase in subsurface ocean heat uptake caused by intensified trade winds associated with the negative phase of the IPO (England et al., 2014).

For example, Dai et al. (2015) performed an EOF analysis for unforced surface temperature variability, including SST over the ocean and surface air temperature over the ocean for 1920 to 2013. The PDO is the leading mode and the fourth mode resembles the AMO.Therefore, the PDO dominates the GMST variability, with a considerable contribution from the AMO. The second and third modes explain more regional variability and therefore do not project onto GMST. The PDO and AMO modes have been used successfully for a reconstruction that reproduced observed GMST, including the hiatus period.

PDO, a mode sometimes called the IPO, is the leading principal component (PC) of low-pass filtered Pacific SST variability (Chen & Wallace, 2015; Zhang et al., 1997). In the following, we will refer to the IPO instead of the PDO. It could also be derived from the leading EOF of decadal global SST variability after the removal of radiatively forced change (Dai et al., 2015). Its tropical climatic signature is a low-frequency El Niño-like pattern that could persist for 2 to 3 decades but much broader in the meridional direction (Zhang et al., 1997). And it modulates the interannual ENSO on decadal time scales.

The impact from the La Niña-like decadal cooling

The ENSO cycle in the tropical Pacific, capable of influencing GMST, has favored the cold La Niña phase since the 1997/1998 El Niño event. The La Niña-like phase lifts cold equatorial thermocline water to the surface, producing cold SST anomalies in the eastern Pacific persistently.

Meehl et al. (2011) are the first to indicate that the hiatus is related to the La Niña-like cooling pattern representative of the negative phase of the IPO over the tropical Pacific in model simulations and periods of more rapid global warming trends during positive phases of the IPO (Meehl et al., 2013). Models fitting the observed hiatus have depended on La-Niña-like cooling to counteract greenhouse-induced warming (Foster & Rahmstorf, 2011; Lean & Rind, 2008), indicated by an increase of subsurface ocean heat uptake predominantly in the Pacific due to a combination of stronger Pacific subtropical cells, a reduction of Antarctic bottom water formation, and stronger AMOC (Balmaseda et al., 2013; Meehl et al., 2011; Meehl et al., 2013; Watanabe et al., 2013). While the ensemble mean of the CMIP5 models could not simulate a hiatus period because the ensemble mean only represents the externally forced part of the response, examination of each of the CMIP5 ensemble members showed that ten members actually did simulate the hiatus as observed because in those ensemble members the randomly-occurring internal decadal variability happened to match the randomly-occurring decadal variability in the observations (Meehl et al., 2014), though the decadal variability in the tropical Pacific is thought to be somewhat weaker than observed (Drijfhout, 2018).

By prescribing observed SST only over the central to eastern tropical Pacific in a model, Kosaka & Xie (2013) reproduced the GMST remarkably well for 1970–2012, indicating that this hiatus is driven by a La-Niña-like decadal cooling in the tropical Pacific. According to the extra-tropical response to tropical forcings (Seager et al., 2003; Trenberth, 2002), this cooling could have largely impacted GMST, and the phase of the IPO has been shown to directly impact the magnitude of GMST trends (Meehl et al., 2016).

Putting forward the IPO

The phase of the IPO switched from positive to negative around 1999 and began to reverse again after 2012 (Henley et al., 2015). In contrast to global warming periods, the two most recent hiatus both relate closely to the negative phase of the IPO, with strengthened winds and a cool tropical Pacific during its negative phase (Meehl et al., 2016).

Therefore, following the model results of Meehl et al. (2011) and Kosaka & Xie (2013), Trenberth & Fasullo (2013) confirmed with observations that the IPO modulates the global warming rate through changes in ocean heat uptake. The positive phase of the IPO strengthened the surface warming by reducing heat uptake in the deep ocean; and the negative phase of the IPO results in above 30% of heat deposit below 700 m depth mostly over the Pacific, which contributes to the cooling of the ocean surface but overall ocean warming (Balmaseda et al., 2013; Trenberth & Fasullo, 2013).

Observed and simulated global temperature fields have further shown that the IPO was largely responsible for GMST trend since 1920 (Dai et al., 2015; Meehl et al., 2016). And using a numerical simulation since 1900 where tropical Pacific SSTs are forced to follow the observed evolution, Kosaka & Xie (2016) revealed that the tropical Pacific part of IPO is the key pacemaker of the long-term GMST trend that drives the variability in warming rates.

The role of intensified trade winds

Associated with the negative phase of the IPO, England et al. (2014) revealed an unprecedented strengthening in the Pacific trade winds that is enough for explaining the cooling in the tropical Pacific during the hiatus. The increased winds caused the enhancing of the subtropical overturning cells, which strengthens the upwelling of the equatorial cold water that enables further cooling in other regions, and subducts a substantial amount of warm surface water into the thermocline thereby enhancing subsurface ocean heat uptake as simulated (Meehl et al., 2011; Meehl et al., 2013; von Känel et al., 2017). Therefore, they confirmed that the ocean heat uptake apart from that corresponding to the cooling, occurred through increasing subduction in the Pacific shallow overturning cells and enhancing heat convergence in the equatorial thermocline.

The La Niña-like decadal cooling is related to these intensified trade winds, indicative of the Bjerknes feedback as in ENSO. Considering this feedback, the pacemaking effect of the equatorial Pacific on the hiatus has been demonstrated by prescribing observed wind anomalies in models (England et al., 2014; Watanabe et al., 2014), instead of restoring SST. Both SST-forced and wind-forced pacemaker experiments reproduce the recent hiatus. The SST restoration generates the intensified trade winds (Watanabe et al., 2014), and similarly the prescribed intensification of the trade wind leads to the cooling in the tropical Pacific. But unlike the SST-restoring method, the wind-forcing method ensures a closed global ocean heat budget (Douville et al., 2015). The fact that the SST-restoring method reproduces the hiatus but could not conserve oceanic heat suggests that energy conservation is not a first-order constraint for the hiatus (Xie & Kosaka, 2017).

The mechanisms that modulates the IPO

A proposed mechanism for IPO is that the build-up of upper ocean heat content (OHC) in the off-equatorial western tropical Pacific could provide conditions for an ENSO event to reverse the tropical Pacific SSTs to the opposite IPO phase (Ding et al., 2013; Farneti et al., 2013; Giese, 2002; Meehl & Hu, 2006, Meehl et al., 2021a). This mechanism is a variant of the delayed-action oscillator (White, 2003) which depends on off-equatorial wind forcing (Wang et al., 2003a, 2003b).

Convective heating anomalies in the tropical Pacific generate anomalous sea level pressure signals in the mid-latitude North and South Pacific and Rossby waves in the atmosphere. These sea level pressure anomalies are correlated with surface wind stress curl anomalies that force anomalous ocean Rossby waves near 25°N and 25°S. And afterwards these Rossby waves generate off-equatorial OHC anomalies in the western tropical Pacific on decadal timescales. And this is the build-up of off-equatorial heat content essential for an IPO transition, which contributes to a 15- to 20-year timescale for the IPO (Farneti et al., 2013; Meehl & Hu, 2006). Combined with that, the interannual variability related to ENSO then produces equatorial ocean Kelvin waves that change the thermocline depth and thereby trigger a transition of the sign of equatorial Pacific SST anomalies to the opposite phase of the IPO.

Interactions among tropical ocean basins also appear essential for IPO phase transition. For instance, the tropical Atlantic warming relevant to AMO phase transition may have contributed to the change of sign of the IPO to a negative phase (McGregor et al., 2014; Meehl et al., 2021b; Ruprich-Robert et al., 2017) and thereby the decadal tropical Pacific cooling from the 1990s contributed to the hiatus. The warming of the North Atlantic drives westerly wind anomalies over the eastern Pacific by Rossby waves and easterly wind anomalies over the Indo-Western Pacific by Kelvin waves (Liu & Xie, 2018). And the subsequent warming in the Indo-Western Pacific resulting from the WES feedback intensifies easterly trades, and through the Bjerkness feedback further enhances the La Niña-like response that contributes to the hiatus (Li et al., 2016). And the tropical Indian Ocean could be an important intermediary in this cross-basin interaction (Luo et al., 2012). Though the cross-basin interactions effect is not deterministic and could be mutually interactive with Pacific and Atlantic alternately affecting each other (Meehl et al., 2021b), if could modulate the phase of the IPO (Xie & Kosaka, 2017).

The interaction with the Indian Ocean

The Indian Ocean and the Pacific are believed to be a coupled system with the two can affecting each other (Han et al., 2014). Lee et al. (2015) noted that no substantial increase in tropical Pacific heat content is shown in observational estimates or their model during the hiatus, despite enhancing Pacific heat uptake. But there is a clear temperature increase in the upper 700 m of the Indian Ocean. It was then found that the Indian Ocean warms through an influx of warm water from the Pacific through the Indonesian archipelago, named Indonesian Throughflow (ITF).

As the intensified Pacific trade winds steepen the east–west tilt of the thermocline, they pile up warm water in the western Pacific and drive the ITF toward the Indian Ocean, thereby strengthening the transport of warm surface water from west Pacific to the Indian Ocean and transferring excess heat taken up by the Pacific to the Indian Ocean (Lee et al., 2015; Liu et al., 2016; Nieves et al., 2015). If the heat transport through ITF remains strong, heat will continue to accumulate in the Indian Ocean which may be projected into the Atlantic via Agulhas leakage, further increasing Atlantic heat content (Lee et al., 2015).

Driven by the increase of ITF transport, an interhemispheric gradient in SST trends exists in the Indian Ocean during the hiatus, with little warming trend in the northern Indian Ocean and an increased warming trend in the southern Indian Ocean (McPhaden & Dong, 2016). This caused a deepening of the thermocline in the south that facilitated SST warming possibly through a weakened vertical mixing, which accounts for above 70% of the observed increased warming trends in the southern Indian Ocean SST during the hiatus (McPhaden & Dong, 2016). It was also indicated that stronger warming in the tropical Indian Ocean than the Pacific favors strengthened trade winds over the Pacific (Han et al., 2014; Luo et al., 2012).

The questioning about the IPO

The ‘Big Hiatus’ was associated with a negative phase of the IPO and SST cooling in the tropical Pacific as well as in the Atlantic, partly offset by the Southern Ocean warming (Drijfhout et al., 2014). During the recent hiatus, Yao et al. (2017) argued that warming effects of other basins has largely compensated by the tropical Pacific-induced strong cooling. But SST changes in basins other than Pacific is believed to exert vital effects on multi-decadal GMST trend as well (Yao et al., 2017).

In the contrast to the demonstrated significant contribution from IPO to the hiatus, Meehl et al. (2016) showed that the contribution of IPO to multi-decadal GMST trends is the largest in its positive phase during the warming periods and much smaller in its transition phase from positive to negative during the recent hiatus accounting for possible contributions from volcanic eruptions during that time period (Fyfe et al., 2016). And von Känel et al. (2017) revealed that the eastern equatorial Pacific actually warms in 10% of all simulated decades with a hiatus, which suggests a 1/10 possibility that the hiatus could occur without the equatorial Pacific being the dominant mechanism, though the largest positive SST anomaly signature during positive IPO is in the central equatorial Pacific and off-equatorial tropical eastern Pacific (Power et al., 1999). Some studies using fully coupled models also suggested that such hiatus could occur as the IPO is transitioning into positive phase (Maher et al., 2014; Roberts et al., 2015). So other than Pacific Ocean variability, decadal variabilities in the Indian (Nieves et al., 2015), Atlantic and Southern Oceans (Dai et al., 2015) were proposed to contribute additionally to the hiatus.

SST-related mechanism: Atlantic Multi-decadal Variability

An introduction

The definition of AMO is the time series of low-pass filtered mean SST in the North Atlantic after removal of the trend because of anthropogenic and external forcing. It is referred to as the low-frequency oscillation of North Atlantic SST between alternating cold and warm phases (Enfield et al. 2001), with a peak-to-peak amplitude of around 0.3–0.4°C for its global-mean (Tung & Zhou, 2013). The warm phase of the AMO is referred to the warm anomalies in the sub-polar gyre and the Labrador Sea and a secondary warm anomalies in the northern tropical Atlantic (Kavvada et al., 2013).

As a recurrent multi-decadal internal variability, it is generally thought to exert considerable effect on global climate. Relying on how to account for the forced signal, different conclusions have been drawn about the effect of the AMO during the hiatus. Besides, further studies on the relationship between the IPO and AMO are required. Several studies find that the AMO decreased during the hiatus and therefore had a cooling rather than a warming impact on the GMST trend (Mann et al., 2014; Yao et al., 2015). This would be consistent with the result that the tropical Pacific can force a same-sign SST response in the tropical Atlantic (Meehl et al., 2021).

The effect of the AMO

At first, Wu et al. (2011) showed that the observational GMST record since 1850 contains two and a half cycles of a multi-decadal variability, lasting 65 years on average and consisting of an warming plus a slowdown period. They indicated that the North Atlantic is the primary location of this multi-decadal variability and the Pacific is only secondary. They also proposed that the AMO could be largely responsible for the GMST variability on a timescale of 60–70 year. Tung & Zhou (2013) also showed before that this multi-decadal cycle exists in a longer (353 years) observational record, with an average period of 70 years (could be as short as 40 years).

AMO has demonstrated sufficient amplitude to significantly contribute to global mean SST trends for periods of 30 years or shorter, about ±0.08 K per decade for a 30-yr trend compared to 0.1K per decade by external forcings (Tung & Zhou, 2013), accounting for about 40% of the observed GMST 50-year warming trend. Since it was stochastic, it did not contribute to trends on longer timescale and thus could not account for the overall warming trend observed in the 20th global mean SST trend. Besides, using the known radiative forcing and the ENSO index as explanatory variables, a linear regression analysis of GMST (1900–2012) accounted for 89% of the observed variance (Chylek et al., 2014). And it was found that the portion increases to 94% if the AMO index is further taken into account as explanatory variable. This study claimed that the positive phase of the AMO accounts for 1/3 of the post-1975 GMST trend with 2/3 being due to the anthropogenic effects.

The relation of the AMO with the IPO

The IPO actually consists of two distinct signals, one roughly bidecadal with a ~16- to 20-year period and the other multi-decadal with a ~50- to 70-year period (Minobe, 1997). The multi-decadal component of the IPO may partly be relevant to the AMO (centered in the Atlantic, but projecting at least weakly onto the Pacific (Knight, 2005)) and partially reflective of low-frequency variability related to the ENSO and its extratropical response (Marini & Frankignoul, 2013). d’Orgeville & Peltier (2007) indicated that the observed AMO and the multi-decadal component of the IPO were two components in phase quadrature of a same oscillation. As both the IPO and the AMO remove secular changes, their correlation remains as strong when the AMO leads the IPO but decreases when the IPO leads the AMO, suggesting that the AMO plays the dominant role in the interbasin connection. Nonetheless, in the Pacific pacemaker experiments of Kosaka & Xie (2013), lag correlations showed that the IPO leads the AMO on average by about 3 years (Meehl et al., 2016). However, Meehl et al. (2021b) analyzed pacemaker experiments and observations to show that the IPO and AMO are likely mutually and sequentially interactive, with the tropical Pacific forcing a same-sign SST response in the tropical Atlantic, and the tropical Atlantic forcing an opposite-sign SST response in the tropical Pacific.

Terming PMO for the multi-decadal component of IPO, competition between a relatively flat, modestly positive peak in the AMO and a sharply negative-trending PMO are seen to produce the hiatus in the study of Marini & Frankignoul (2013). Dong & Zhou (2014) then indicated that this competition is between the anthropogenic warming signal, cooling effect of a strongly negative IPO and slight cooling (or warming) effect of the AMO. Yao et al. (2015) showed more evidence for that the hiatus is the natural result of the interactions between a secular global warming trend largely due to the increase of greenhouse-gas, and internal climate cooling caused by a cool phase of a quasi-60-year oscillation that is closely related to AMO and IPO. Or specifically, Yao et al. (2015) further demonstrated that the AMO and IPO could be recognized as an indicator and a harbinger of GMST trend on multi-decadal timescales, respectively. It was suggested these climate oscillations largely operate without driving longer-term heat sequestration into the deep ocean. Therefore, the drivers of the recent hiatus do not alter the century-scale global warming trend related to the increase of greenhouse gas.

Wu et al. (2019) indicated that both AMO and IPO have a significant effect on modulating GMST trend in the past a century and a half, contributing a combined 30% apart from the residual contributions from greenhouse-gas, with a potential higher contribution from AMO. Moreover, they found that a combined in-phase of IPO and AMO could contribute to global climate significantly larger than greenhouse-gases do, but a combined out-of-phase of IPO and AMO could minimize their contribution to the global climate (Su et al., 2017).

It has thought before that IPO is more responsible for the multi-decadal variation of GMST (Kosaka & Xie, 2013; Meehl et al., 2016; Trenberth, 2015). Nevertheless, it is more recently suggested that on multi-decadal timescale AMO could contribute more to GMST trend compared to IPO, but on decadal timescale IPO leads AMO with comparable contributions to GMST trend. Tung et al. (2019) indicated that IPO consists of different proportions of ENSO, PDO and AMO in the form of linear superposition, relying on the filter threshold. They argued that IPO is mostly AMO as the second EOF of the decadally filtered SST and therefore the contribution to the GMST trend on multi-decadal time scale usually ascribed to IPO is actually by AMO. And extracting the major climate modes with the method of pair-wise rotation of the PCs, Chen & Tung (2017) showed that the Pacific contributes to the GMST trend mostly on interannual timescale through ENSO and the Atlantic contributes significantly on multi-decadal timescale through AMO. However, if the IPO and AMO are mutually interactive as proposed by Meehl et al. (2021b), both would contribute in combination to GMST trends.

The period of the AMO

Both observations and reconstructions revealed one statistically significant spectral peak in the 50–70 year band of AMO before (Gray et al., 2004; Schlesinger & Ramankutty, 1994; Tung & Zhou, 2013). But a spectral analysis have showed an observed AMO index oscillations in the 70–80 year range in combination with oscillations in the 30–40 year range and shorter periods (Kavvada et al., 2013). It was concluded that models underestimate the period of AMO by increasing variability in the 10–20 year range so that it becomes more dominant than variability in the 70–80 year range.

Ice core data sets have provided observational evidence for two distinct time scales of AMO that are about 20 years and 45–85 year (Chylek et al., 2011). The former time scale is dominant and statistically significant while the latter one is not. It was suggested that the latter time scale may reflect a larger spatial scale mode related to a declined Atlantic meridional overturning circulation (AMOC) or a coupling of the Arctic and Atlantic circulations.

The mechanisms that modulates the AMO

In terms of internal climate variability, proposed crucial drivers of the AMO consist of density and salinity fluctuations caused by variations in AMOC (Latif et al., 2004; Medhaug & Furevik, 2011), changes in wind forcing and air–sea interactions (Huang et al., 2011). AMO exhibits a signature of perturbations involving subsurface OHC, SST, salinity and Arctic sea ice.

Long-term observations revealed that the thermohaline circulation variability in the Atlantic possibly causes the AMO (Tung & Zhou, 2013). And several model studies showed that AMOC intensification is followed by a positive AMO phase, albeit with a model dependent lag (e.g. Marini & Frankignoul, 2013). It was proposed that AMO depends on negative feedbacks between the Arctic ice melt responding to the warm SST and the strength of the AMOC that brings warm SST to the North Atlantic (Dima & Lohmann, 2007; Park et al., 2010). The former then slows AMOC after a delay of 20 year with reduced deep-water formation. It has also been simulated a 55 to 80 year AMO arising from the variability of the AMOC (Lohmann & Wei, 2012).

Over the Labrador Sea where deep water forms, fresh water anomalies stratify the ocean layer and thus contributes to the weakening of the deep water formation and AMOC, allowing for the cooling of surface because of the lack of vertical mixing with the warmer subsurface water. Because of this reduction in the heat lost to the atmosphere, this inhibits the deep convection further (Gelderloos et al., 2012). Therefore, generally there are fresh water anomalies and weaker AMOC during the cold phase of the AMO and vice versa. Sun et al. (2015) indicated further that the positive North Atlantic Oscillation (NAO) forces the strengthening of AMOC and induces a basin-wide SST warming that corresponds to the positive phase of AMO, while AMO in turn has a delayed negative effect on the NAO by producing a meridional SST gradient pattern. However, other studies have shown that much of the decadal timescale variability of the AMO is likely externally forced due to time-varying aerosol emission over North America (e.g., Booth et al., 2012).

Considering the role of AMOC in modulating the AMO, there are some other findings about the relation between AMO and IPO. Zhang & Delworth (2007) and Zhang et al. (2007) using models showed that with the northward heat transport by the AMOC increasing, the atmospheric heat transport decreases in compensation (and vice versa), offering a multi-decadal component for the IPO. Chen & Tung (2014) suggested that if the multi-decadal variability indicated by Wu et al. (2011) is related to the variations of the AMOC, the IPO component would have a time scale of 30 years between regime shifts. While Marini & Frankignoul (2013) showed that removing the IPO largely degrades the representation of AMOC changes by the AMO in model and tends to reduce the AMOC–AMO correlation, which refect the strong relationship between the AMO and the IPO.

Energy-related mechanism: vertical heat uptake

Earth’s energy imbalance

With increasing greenhouse gases in the atmosphere trapping more radiation and hence creating warming, global energy budgets demonstrated an ongoing excess of radiative energy input at the top of atmosphere (Trenberth & Fasullo, 2010). During the hiatus, this net energy flux into the climate system is supposed to generate warming somewhere in the system. Meehl et al. (2011) showed little difference of the global net radiative imbalance between hiatus and reference periods in their model, thus indicating that during periods of near-zero GMST warming trend, that energy must be deposited in the subsurface ocean. Hu et al. (2019) decomposed the global-mean surface energy budget respectively for the fast-warming period and the subsequent hiatus period. They showed that the increase of the downward longwave radiation since the 1980s is a result of the increase of air concentration of carbon dioxide (Dee et al. 2011). This could not explain the changes in the global warming rate between the two periods. While because of the increase of clouds during the hiatus, less downward shortwave heat absorbed by the ocean and more surface latent heat contributes to the slowdown of global warming (Hu et al., 2019).

The main energy reservoir is the ocean, sequestering energy as heat. Unless excess energy was absorbed by subsurface oceans, earth’s surface should have warmed. earth’s energy imbalance leads to global warming, and more than 90% of it is stored in the ocean over the past 50 years, increasing OHC. There is more heat staying near the surface during the rapid warming period in the latter part of the 20th century as shown in a modeling study for periods of positive IPO (Meehl et al., 2013). Ocean temperature data from the recent hiatus have revealed that each major ocean basin has warmed at nearly all latitudes and more heat moved into deeper oceans during the hiatus (Levitus et al., 2012). Several studies indicated that the Pacific Ocean absorbed more heat during the recent hiatus (England et al., 2014), while other studies suggested that the Southern (Lee et al., 2015), Indian, and Atlantic Oceans (Chen & Tung, 2014) also contributed to the increase of global ocean heat uptake. Since energy is exchanged between the ocean and the atmosphere, the absorbed heat can resurface afterwards to affect weather and climate globally. And because of the complexity of various climate forcings and natural variability, this imbalance is likely to change with time.

The related questions

Estimations of OHC have been achieved at near-global level thanks to the sustained observations from the Argo array of autonomous profiling floats (Abraham et al., 2013). But it is still difficult to get an accurate assessment of longer term trends in OHC since data coverage is insufficient and irregular. There are limited repeat observations in the deep (>2000 m) ocean and a large range of observation-based heat content estimates (Fasullo et al., 2014). Observation analyses vary from models and between datasets. Direct observation of whether the ocean heat uptake is increased is difficult. Besides, it was suggested that the uncertainties cause the underestimation of the long-term warming in the upper-ocean (Durack et al., 2014). So the heat flux in all ocean basins is still poorly constrained by measurements.

The SST change not only reflects the changes in atmospheric heat content, but also reflects the change in the mixed layer of the ocean. This underscores that the SST anomalies are not the only reason for the heat uptake anomalies. But the vertical heat uptake and interbasin heat redistribution in the ocean are unclear yet. So it is still vague whether the recent hiatus is due to the redistribution of OHC or changing ocean heat uptake. For example, studies have indicated an increased heat transport to deeper layers (Drijfhout et al., 2014; Meehl et al., 2011) or a shallower decadal heat redistribution between the Pacific and Indian Oceans (Lee et al., 2015; Nieves et al., 2015). Therefore, to what extent and by which mechanism heat was absorbed in the ocean is under debate (Su et al., 2017; Yan et al., 2016). Energy budget is not closed due to substantial uncertainties in the datasets and methodologies. So the different opinions in the following have highlighted the gap between models and observations, large uncertainties in OHC observations, and complicated relation between GMST and OHC (Palmer et al., 2011).

Top 700 m heat uptake

Global warming signal remains in the upper ocean mostly (Hansen et al., 2005; Levitus, 2005). It was suggested that the upper 2000 m of the ocean encompasses numerous variability and trends of the OHC over time. Guemas et al. (2013) retrospectively predicted the hiatus up to 5 years ahead and attributed the onset of this hiatus to heat uptake of the upper 700 m of the ocean, which accounts for most of the excess top-of-atmosphere net energy input (65% in the tropical Pacific and Atlantic oceans). Additionally, decadal prediction model simulations initialized with ocean conditions close to observations were able to predict the hiatus (Meehl et al., 2014).

Nonetheless, observational datasets have revealed robust warming but a slowdown of OHC increase (but not as much as the full-depth OHC) with a flattening trend of the global upper ocean over 0–700 m depth in the early 21st century (Levitus et al., 2012; Lyman et al., 2010). An observational estimate from Fasullo et al. (2014) showed that throughout the hiatus, the OHC of the upper 700 m started increasing in the 1970s, increased to the largest in the start of the hiatus and then started decreasingly increasing after 2005. In the global oceans and compared to the deeper layers, Hu et al. (2013) found from model results that heat content trends of the upper 300 m warms less during the hiatus but more and rapidly during accelerated warming decades. Since the ocean surface layer is not heating up during the hiatus, the deeper layer beneath should absorb more heat.

Heat uptake from 700 m to 2000 m

It has been proposed that the hiatus was accompanied by an increase of the ocean heat uptale that increases the ocean temperature below 700 m (Balmaseda et al., 2013; Drijfhout et al., 2014; Watanabe et al., 2013), and thereby a substantial amount of heat stored in the subsurface ocean on the decadal time scale (e.g. Balmaseda et al., 2013; Guemas et al., 2013; Levitus et al., 2012; Meehl et al., 2011; Palmer et al., 2011; Watanabe et al., 2013).

It was shown that since 1865, approximately 50% of the OHC has been accumulated after the mid-1990s, including 35% at a depth larger than 700 m (Gleckler et al., 2016). Ocean temperature data has revealed that more heat is sequestrated in the deep ocean during the recent hiatus (Balmaseda et al., 2013). For instance, Su et al. (2017) demonstrated a rapid increase of oceanic heat content in the world’s subsurface and deep ocean of 300–2000 m depth during 1998–2013. Furthermore, the model study by Meehl et al. (2011) showed that the hiatus is highly correlated with the accelerated warming of deep layers in all the ocean basins, below 750 m in the Southern and Atlantic Oceans and below 300 m in the Indian and Pacific Oceans.

With multiple observational datasets, a major observational study from Levitus et al. (2012) demonstrated that the world OHC increased with a rate of 0.39 W·m2 and a volume mean warming of 0.09°C for the 0–2000 m layer during the hiatus. While the world OHC increased with a rate of 0.27 W·m2 and a volume mean warming of 0.18°C for the 0–700 m layer. It is shown in this study that 700–2000 m ocean layer made up around one-third of the warming in the 0–2000 m layer. Temperature anomalies in the deeper ocean are important for global climate but their contribution to the net integral of OHC is relatively small (Levitus, 2000). They also indicated that heat may have been stored below 2000 m during the period of observation in their study and may be important for the earth’s heat balance.

Model results have revealed an accelerated increase of OHC during the hiatus, with a substantial contribution from 700 m to the bottom (Balmaseda et al., 2013). Watanabe et al. (2013) found a weakening of ocean heat uptake efficiency in models and enhancing in nature implying the reason for the overestimation of the GMST trend, but the strengthened heat uptake efficiency in nature has not been clearly explained yet. In contrast to the expectation from the simple global energy balance models, von Känel et al. (2017) suggested that OHC changes showed a decreasing trend during hiatus periods. They found a generally reduced rate of OHC changes in the top 100 m and enhanced rate below during hiatus decades, consistent with previous studies (e.g. Meehl et al., 2011).

Heat uptake below 2000 m

Despite that the upper oceans (above the 2000 m) continue to sequester heat, there is no measurable warming of the bottom (below 2000 m) during the last decade (Llovel et al., 2014). It was speculated that the excess energy is possible to be absorbed by the deep ocean where few measurements for verification exist (Loeb et al., 2012; Trenberth & Fasullo, 2010). On the other hand, Meehl et al. (2011) indicated that the warming rate of the deep ocean layers is higher than the surface layers during the hiatus in a model.

Energy-related mechanism: interbasin heat redistribution

An introduction

During the hiatus, surface latent and sensible heat fluxes force GMST trends while the solar radiation fluxes and net surface longwave have a reduced effect, suggesting that the unforced GMST trend is due to ocean heat redistribution (Drijfhout, 2018). The hiatus has been proposed to be compensated by enhanced warming in either the Pacific (Balmaseda et al., 2013; England et al., 2014; Kosaka & Xie, 2013; Trenberth & Fasullo, 2013) or the Atlantic (Chen & Tung, 2014) or the Indian Ocean (Lee et al., 2015), or a combination of the Southern, Atlantic, and Indian Oceans (Drijfhout et al., 2014).

From The Pacific to the Indian Ocean

Considering the effect of the IPO, a significant amount of excess heat is thought to be taken up by the Pacific. Nonetheless, in situ data indicated that heat content in the Pacific has been decreasing (Levitus et al., 2009). Observational data have revealed that the warming in the 100 to 300 m layer of the Western Pacific and Indian Oceans, with the largest contribution in the tropics, is compensated by the cooling in the top 100-m layer of the Pacific (Nieves et al., 2015). Lee et al. (2015) further found that the strengthened heat uptake in the Pacific has been compensated by the increased heat transport from the Pacific to the Indian Ocean through the ITF. They showed heat content in the Indian Ocean has increased significantly, accounting for above 70% of the global ocean heat uptake in the top 700 m over the past ten year. Moreover, Su et al. (2017) demonstrated that the Indian Ocean accounted for around 30% of global ocean heat uptake, playing a particularly important role during the hiatus.

From the Pacific to the Southern Ocean

While numerical simulations, on the other hand, suggested a subsurface warming during the initial phase of the hiatus in the equatorial Pacific (Oka & Watanabe, 2017), enhancing equatorial Ekman transport to subtropical regions in the later phase of the hiatus (after 2002) that causes enhancing the subtropical Ekman downwelling, which further accelerated heat uptake below 700 m in the subtropical Southern Ocean that contributes to the post-2002 hiatus period. These anomalies in the subtropical Southern Ocean have been considered before to have a crucial impact on generating a multi-decadal IPO cycle (Luo & Yamagata, 2001).

Moreover, it was also suggested before that the cold SST anomalies and relevant cooling of the atmosphere in the equatorial Pacific, owing to the negative phase of the IPO, could force extra-tropical stationary waves to strengthen the heat fluxes into the Atlantic and Southern oceans remotely (Trenberth et al., 2014).

To the Atlantic and the Southern Ocean

Several studies suggested a westward pathway for the warming signal from the Atlantic to the Pacific through teleconnection, directly intensifying warming in the Atlantic since the early 1990s, with strengthening anomalies of the Walker circulation and La Niña-like Pacific (McGregor et al., 2014; Ruprich-Robert et al., 2017). The robust mechanisms controlling the Atlantic variabilities are still being discussed, salinity driven mechanism is one of the explanations for the increased heat uptake in the Atlantic (Chen & Tung, 2014).

Using updated OHC estimates, Cheng et al. (2017) recently found that the greatest warming during the hiatus is in the Southern Ocean, followed by the tropical/subtropical Pacific Ocean and tropical/subtropical Atlantic Ocean respectively. And by Argo observations, much of the global heat uptake has been accounted for in the Southern Ocean. Besides, as global ocean heat uptake accounted for 75%–99% south of the equator, the Southern Hemisphere Ocean played a significant role in the warming of subsurface layers over the past decade (Roemmich et al., 2015; Wijffels et al., 2016). Specifically, the Southern Ocean is demonstrated to play a secondary role in warming the 100-m to 300-m layer with a steady pace over the past two decades (Nieves et al., 2015).

Chen & Tung (2014) showed by observational data that the hiatus is largely due to heat transported to deeper layers in the Atlantic and the Southern oceans, initiated by a periodic salinity anomaly in the subpolar North Atlantic that affects AMOC. And they showed that the hiatus related to this mechanism of deeper heat-sequestration lasted 20 to 35 years historically. Besides, the increase of OHC (300–1500 m) accelerates in the Atlantic and Southern Oceans but changes little in the Pacific and Indian Oceans. Therefore, their observational results are against the Pacific-centric view (e.g. Hu et al., 2013; Meehl et al., 2011). Chen & Tung (2014) argued that compared with observations, there are too little variability with too high frequency in their model’s Atlantic. However, also by observations, Nieves et al. (2015) showed that the Atlantic switched from warming to cooling during the hiatus, but its area was suggested to be too small to contribute to the hiatus significantly.

Using the same observational dataset and analyzing the ocean heat uptake the same way as Chen & Tung (2014) did, Liu et al. (2016) demonstrated that the ocean heat uptake in deep Atlantic and Southern Oceans is not unique to the hiatus but showed the downward penetration of anthropogenic heat through AMOC. It is shown to occur at a similar rate no matter during the hiatus or not and is hence suggested not the dominant factor affecting the GMST trend during the hiatus. Instead, their findings support the Indo-Pacific heat redistribution mechanism (Lee et al., 2015; Nieves et al., 2015).

Summary

The global warming hiatus from 1998 to 2013 is suggested to be largely due to the natural variabilities in the climate system. So this review focused on the proposed mechanisms for the hiatus associated with internal climate variabilities. From the perspective of SST variabilities, the IPO and AMO related mechanisms stand out for the hiatus. In terms of energy variabilities, vertical uptake and interbasin redistribution of heat are related to the mechanisms that regulate the hiatus.

The IPO is in its negative phase during the hiatus, with anomalously strengthened trade winds and thereby La Niña-like cooling. And the intensifying of the trade winds facilitates ocean heat uptake by enhancing subtropical overturning cells, potentially leading to the hiatus. This mechanism of the IPO is generally the most prevailing view. The AMO is demonstrated to be or at least highly related to the multi-decadal component of the IPO. Other than the prevalent voice of domination of IPO, there are supports for that both the IPO and AMO contributed to the hiatus, or even AMO could have exceeded the IPO in the effects on the hiatus.

During the hiatus, the global OHC over 0–700 m only increased with a relatively flattening trend while below 700 m rapidly increased implying the sequestration of excessive heat. And the global OHC below 700 m is shown to be highly correlated with the hiatus and attributed to the weakening of the AMOC, negative IPO and wind modulation. On the other hand, heat redistribution, such as Pacific-Indo heat transport through ITF, heat transport to the deep Atlantic and the Southern Ocean through AMOC, is believed to exert a significant impact on the hiatus.

Despite the debate about the causes of the hiatus, the continuous studies and debates of the hiatus have tremendously promoted the understanding of the complex climate system.

Acknowledgements

This review is supported by the Fundamental Research Funds for the Central Universities (202161058). We gratefully thanks Gerald Meehl for helping editing the paper.

Competing Interests

The authors have no competing interests to declare.

References

  1. Abraham, JP. 2013. A review of global ocean temperature observations: Implications for ocean heat content estimates and climate change. Reviews of Geophysics, 51: 450–483. DOI: https://doi.org/10.1002/rog.20022 

  2. Balmaseda, MA. 2013. Distinctive climate signals in reanalysis of global ocean heat content. Geophysical Research Letters, 40(9): 1754–1759. DOI: https://doi.org/10.1002/grl.50382 

  3. Booth, B. 2012. Aerosols implicated as a prime driver of twentieth-century North Atlantic climate variability. Nature, 484: 228–232. DOI: https://doi.org/10.1038/nature10946 

  4. Chen, X. 2014. Varying planetary heat sink led to global-warming slowdown and acceleration. Science, 345(6199): 897–903. DOI: https://doi.org/10.1126/science.1254937 

  5. Chen, X. 2015. ENSO-Like Variability: 1900–2013. Journal of Climate, 28(24): 9623–9641. DOI: https://doi.org/10.1175/JCLI-D-15-0322.1 

  6. Chen, X. 2017. Global-mean surface temperature variability: space–time perspective from rotated EOFs. Climate Dynamics, 51(5–6): 1719–1732. DOI: https://doi.org/10.1007/s00382-017-3979-0 

  7. Cheng, L. 2017. Improved estimates of ocean heat content from 1960 to 2015. Science Advances, 3(3): e1601545. DOI: https://doi.org/10.1126/sciadv.1601545 

  8. Chylek, P. 2011. Ice-core data evidence for a prominent near 20 year time-scale of the Atlantic Multi-decadal Oscillation. Geophysical Research Letters, 38(13): L13704. DOI: https://doi.org/10.1029/2011GL047501 

  9. Chylek, P. 2014. The Atlantic Multi-decadal Oscillation as a dominant factor of oceanic influence on climate. Geophysical Research Letters, 41(5): 1689–1697. DOI: https://doi.org/10.1002/2014GL059274 

  10. Cohen, JL. 2012. Asymmetric seasonal temperature trends. Geophysical Research Letters, 39(4): L04705. DOI: https://doi.org/10.1029/2011GL050582 

  11. d’Orgeville, M. 2007. On the Pacific Decadal Oscillation and the Atlantic Multi-decadal Oscillation: Might they be related? Geophysical Research Letters, 34(23): L23705. DOI: https://doi.org/10.1029/2007GL031584 

  12. Dai, A. 2015. Decadal modulation of global surface temperature by internal climate variability. Nature Climate Change, 5(6): 555–559. DOI: https://doi.org/10.1038/nclimate2605 

  13. Dima, M. 2007. A Hemispheric Mechanism for the Atlantic Multi-decadal Oscillation. Journal of Climate, 20(11): 2706–2719. DOI: https://doi.org/10.1175/JCLI4174.1 

  14. Ding, H. 2013. Hindcast of the 1976/77 and 1998/99 Climate Shifts in the Pacific. Journal of Climate, 26(19): 7650–7661. DOI: https://doi.org/10.1175/JCLI-D-12-00626.1 

  15. Dong, L. 2014. The formation of the recent cooling in the eastern tropical Pacific Ocean and the associated climate impacts: A competition of global warming, IPO, and AMO. Journal of Geophysical Research: Atmospheres, 119(19): 11, 272–211, 287. DOI: https://doi.org/10.1002/2013JD021395 

  16. Douville, H. 2015. The recent global warming hiatus: What is the role of Pacific variability? Geophysical Research Letters, 42(3): 880–888. DOI: https://doi.org/10.1002/2014GL062775 

  17. Drijfhout, S. 2018. The relation between natural variations in ocean heat uptake and global mean surface temperature anomalies in CMIP5. Scientific Reports, 8(1): 7402. DOI: https://doi.org/10.1038/s41598-018-25342-7 

  18. Drijfhout, SS. 2014. Surface warming hiatus caused by increased heat uptake across multiple ocean basins. Geophysical Research Letters, 41(22): 7868–7874. DOI: https://doi.org/10.1002/2014GL061456 

  19. Durack, PJ. 2014. Quantifying underestimates of long-term upper-ocean warming. Nature Climate Change, 4(11): 999–1005. DOI: https://doi.org/10.1038/nclimate2389 

  20. Easterling, DR. 2009. Is the climate warming or cooling? Geophysical Research Letters, 36(8): L08706. DOI: https://doi.org/10.1029/2009GL037810 

  21. England, MH. 2014. Recent intensification of wind-driven circulation in the Pacific and the ongoing warming hiatus. Nature Climate Change, 4(3): 222–227. DOI: https://doi.org/10.1038/nclimate2106 

  22. England, MH. 2015. Robust warming projections despite the recent hiatus. Nature Climate Change, 5(5): 394–396. DOI: https://doi.org/10.1038/nclimate2575 

  23. Farneti, R. 2013. Pacific interdecadal variability driven by tropical–extratropical interactions. Climate Dynamics, 42(11–12): 3337–3355. DOI: https://doi.org/10.1007/s00382-013-1906-6 

  24. Fasullo, JT. 2014. Earth’s Energy Imbalance. Journal of Climate, 27(9): 3129–3144. DOI: https://doi.org/10.1175/JCLI-D-13-00294.1 

  25. Foster, G. 2011. Global temperature evolution 1979–2010. Environmental Research Letters, 6(4): 044022. DOI: https://doi.org/10.1088/1748-9326/6/4/044022 

  26. Fyfe, JC. 2013. Overestimated global warming over the past 20 years. Nature Climate Change, 3(9): 767–769. DOI: https://doi.org/10.1038/nclimate1972 

  27. Fyfe, JC. 2016. Making sense of the early-2000s warming slowdown. Nature Climate Change, 6(3): 224–228. DOI: https://doi.org/10.1038/nclimate2938 

  28. Gelderloos, R. 2012. Mechanisms behind the Temporary Shutdown of Deep Convection in the Labrador Sea: Lessons from the Great Salinity Anomaly Years 1968–71. Journal of Climate, 25(19): 6743–6755. DOI: https://doi.org/10.1175/JCLI-D-11-00549.1 

  29. Giese, BS. 2002. Southern Hemisphere Origins of the 1976 Climate Shift. Geophysical Research Letters, 29(2): 1014. DOI: https://doi.org/10.1029/2001GL013268 

  30. Gleckler, PJ. 2016. Industrial-era global ocean heat uptake doubles in recent decades. Nature Climate Change, 6(4): 394–398. DOI: https://doi.org/10.1038/nclimate2915 

  31. Gray, ST. 2004. A tree-ring based reconstruction of the Atlantic Multi-decadal Oscillation since 1567 A.D. Geophysical Research Letters, 31(12): L12205. DOI: https://doi.org/10.1029/2004GL019932 

  32. Guemas, V. 2013. Retrospective prediction of the global warming slowdown in the past decade. Nature Climate Change, 3(7): 649–653. DOI: https://doi.org/10.1038/nclimate1863 

  33. Han, W. 2014. Intensification of decadal and multi-decadal sea level variability in the western tropical Pacific during recent decades. Climate Dynamics, 43: 1357–1379. DOI: https://doi.org/10.1007/s00382-013-1951-1 

  34. Hansen, J. 2005. Earth’s energy imbalance: confirmation and implications. Science, 308(5727): 1431–1435. DOI: https://doi.org/10.1126/science.1110252 

  35. Hansen, J. 2006. Global temperature change. PNAS, 103(39): 14288–14293. DOI: https://doi.org/10.1073/pnas.0606291103 

  36. Henley, BJ. 2015. A Tripole Index for the Interdecadal Pacific Oscillation. Climate Dynamics, 45(11–12): 3077–3090. DOI: https://doi.org/10.1007/s00382-015-2525-1 

  37. Hu, X. 2019. Decadal evolution of the surface energy budget during the fast warming and global warming hiatus periods in the ERA-interim. Climate Dynamics, 52: 2005–2016. DOI: https://doi.org/10.1007/s00382-018-4232-1 

  38. Huang, B. 2011. Influences of tropical–extratropical interaction on the multi-decadal AMOC variability in the NCEP climate forecast system. Climate Dynamics, 39(3–4): 531–555. DOI: https://doi.org/10.1007/s00382-011-1258-z 

  39. Hunt, BG. 2010. The role of natural climatic variation in perturbing the observed global mean temperature trend. Climate Dynamics, 36(3–4): 509–521. DOI: https://doi.org/10.1007/s00382-010-0799-x 

  40. Kaufmann, RK. 2011. Reconciling anthropogenic climate change with observed temperature 1998–2008. PNAS, 108(29): 11790–11793. DOI: https://doi.org/10.1073/pnas.1102467108 

  41. Kavvada, A. 2013. AMO’s structure and climate footprint in observations and IPCC AR5 climate simulations. Climate Dynamics, 41(5–6): 1345–1364. DOI: https://doi.org/10.1007/s00382-013-1712-1 

  42. Knight, JR. 2005. A signature of persistent natural thermohaline circulation cycles in observed climate. Geophysical Research Letters, 32(20): L20708. DOI: https://doi.org/10.1029/2005GL024233 

  43. Knutson, TR. 2016. Prospects for a prolonged slowdown in global warming in the early 21st century. Nature Communications, 7: 13676. DOI: https://doi.org/10.1038/ncomms13676 

  44. Kosaka, Y. 2013. Recent global-warming hiatus tied to equatorial Pacific surface cooling. Nature, 501(7467): 403–407. DOI: https://doi.org/10.1038/nature12534 

  45. Kosaka, Y. 2016. The tropical Pacific as a key pacemaker of the variable rates of global warming. Nature Geoscience, 9(9): 669–673. DOI: https://doi.org/10.1038/ngeo2770 

  46. Latif, M. 2004. Reconstructing, Monitoring, and Predicting Multi-decadal-Scale Changes in the North Atlantic Thermohaline Circulation with Sea Surface Temperature. Journal of Climate, 17(7): 1605–1614. DOI: https://doi.org/10.1175/1520-0442(2004)017<1605:RMAPMC>2.0.CO;2 

  47. Lean, JL. 2008. How natural and anthropogenic influences alter global and regional surface temperatures: 1889 to 2006. Geophysical Research Letters, 35(18). DOI: https://doi.org/10.1029/2008GL034864 

  48. Lee, SK. 2015. Pacific origin of the abrupt increase in Indian Ocean heat content during the warming hiatus. Nature Geoscience, 8(6): 445–449. DOI: https://doi.org/10.1038/ngeo2438 

  49. Levitus, S. 2000. Warming of the World Ocean. Science, 287(5461): 2225–2229. DOI: https://doi.org/10.1126/science.287.5461.2225 

  50. Levitus, S. 2005. Warming of the world ocean, 1955–2003. Geophysical Research Letters, 32(2): L02604. DOI: https://doi.org/10.1029/2004GL021592 

  51. Levitus, S. 2009. Global ocean heat content 1955–2008 in light of recently revealed instrumentation problems. Geophysical Research Letters, 36(7): L07608. DOI: https://doi.org/10.1029/2008GL037155 

  52. Levitus, S. 2012. World ocean heat content and thermosteric sea level change (0–2000 m), 1955–2010. Geophysical Research Letters, 39(10): L10603. DOI: https://doi.org/10.1029/2012GL051106 

  53. Liu, W. 2016. Tracking ocean heat uptake during the surface warming hiatus. Nature Communications, 7: 10926. DOI: https://doi.org/10.1038/ncomms10926 

  54. Llovel, W. 2014. Deep-ocean contribution to sea level and energy budget not detectable over the past decade. Nature Climate Change, 4(11): 1031–1035. DOI: https://doi.org/10.1038/nclimate2387 

  55. Loeb, NG. 2012. Observed changes in top-of-the-atmosphere radiation and upper-ocean heating consistent within uncertainty. Nature Geoscience, 5(2): 110–113. DOI: https://doi.org/10.1038/ngeo1375 

  56. Lohmann, G. 2012. Simulated Atlantic Multi-decadal Oscillation during the Holocene. Journal of Climate, 25(20): 6989–7002. DOI: https://doi.org/10.1175/JCLI-D-11-00667.1 

  57. Luo, JJ. 2001. Long-term El Niño-Southern Oscillation (ENSO)-like variation with special emphasis on the South Pacific. Journal of Geophysical Research, 106(C10): 22211–22227. DOI: https://doi.org/10.1029/2000JC000471 

  58. Luo, JJ. 2012. Indian Ocean warming modulates Pacific climate change. PNAS, 109(46): 18701–18706. DOI: https://doi.org/10.1073/pnas.1210239109 

  59. Lyman, JM. 2010. Robust warming of the global upper ocean. Nature, 465(7296): 334–337. DOI: https://doi.org/10.1038/nature09043 

  60. Maher, N. 2014. Drivers of decadal hiatus periods in the 20th and 21st centuries. Geophysical Research Letters, 41(16): 5978–5986. DOI: https://doi.org/10.1002/2014GL060527 

  61. Mann, ME. 2014. On forced temperature changes, internal variability, and the AMO. Geophysical Research Letters, 41(9): 3211–3219. DOI: https://doi.org/10.1002/2014GL059233 

  62. Marini, C. 2013. An attempt to deconstruct the Atlantic Multi-decadal Oscillation. Climate Dynamics, 43(3–4): 607–625. DOI: https://doi.org/10.1007/s00382-013-1852-3 

  63. McGregor, S. 2014. Recent Walker circulation strengthening and Pacific cooling amplified by Atlantic warming. Nature Climate Change, 4(10): 888–892. DOI: https://doi.org/10.1038/nclimate2330 

  64. McPhaden, MJ. 2016. Interhemispheric SST Gradient Trends in the Indian Ocean prior to and during the Recent Global Warming Hiatusa. Journal of Climate, 29(24), 9077–9095. DOI: https://doi.org/10.1175/JCLI-D-16-0130.1 

  65. Medhaug, I. 2011. North Atlantic 20th century multi-decadal variability in coupled climate models: sea surface temperature and ocean overturning circulation. Ocean Science, 7(3): 389–404. DOI: https://doi.org/10.5194/os-7-389-2011 

  66. Meehl, GA. 2006. Megadroughts in the Indian Monsoon Region and Southwest North America and a Mechanism for Associated Multi-decadal Pacific Sea Surface Temperature Anomalies. Journal of Climate, 19(9): 1605–1623. DOI: https://doi.org/10.1175/JCLI3675.1 

  67. Meehl, GA. 2011. Model-based evidence of deep-ocean heat uptake during surface-temperature hiatus periods. Nature Climate Change, 1(7): 360–364. DOI: https://doi.org/10.1038/nclimate1229 

  68. Meehl, GA. 2013. Externally forced and internally generated decadal climate variability associated with the Interdecadal Pacific Oscillation. Journal of Climate, 26: 7298–7310. DOI: https://doi.org/10.1175/JCLI-D-12-00548.1 

  69. Meehl, GA. 2014. Climate model simulations of the observed early-2000s hiatus of global warming. Nature Climate Change, 4: 898–902. DOI: https://doi.org/10.1038/nclimate2357 

  70. Meehl, GA. 2016. Contribution of the Interdecadal Pacific Oscillation to twentieth-century global surface temperature trends. Nature Climate Change, 6(11): 1005–1008. DOI: https://doi.org/10.1038/nclimate3107 

  71. Minobe, S. 1997. A 50–70 year climatic oscillation over the North Pacific and North America. Geophysical Research Letters, 24(6): 683–686. DOI: https://doi.org/10.1029/97GL00504 

  72. Nieves, V. 2015. Recent hiatus caused by decadal shift in Indo-Pacific heating. Science, 349(6247): 532–535. DOI: https://doi.org/10.1126/science.aaa4521 

  73. Oka, A. 2017. The post-2002 global surface warming slowdown caused by the subtropical Southern Ocean heating acceleration. Geophysical Research Letters, 44(7): 3319–3327. DOI: https://doi.org/10.1002/2016GL072184 

  74. Palmer, MD. 2011. Importance of the deep ocean for estimating decadal changes in Earth’s radiation balance. Geophysical Research Letters, 38(13): L13707. DOI: https://doi.org/10.1029/2011GL047835 

  75. Park, W. 2010. The Impact of North Atlantic–Arctic Multi-decadal Variability on Northern Hemisphere Surface Air Temperature. Journal of Climate, 23(21): 5668–5677. DOI: https://doi.org/10.1175/2010JCLI3347.1 

  76. Power, S. 1999. Inter-decadal modulation of the impact of ENSO on Australia. Climate Dynamics, 15: 319–324. DOI: https://doi.org/10.1007/s003820050284 

  77. Risbey, JS. 2014. Well-estimated global surface warming in climate projections selected for ENSO phase. Nature Climate Change, 4(9): 835–840. DOI: https://doi.org/10.1038/nclimate2310 

  78. Roberts, CD. 2015. Quantifying the likelihood of a continued hiatus in global warming. Nature Climate Change, 5(4): 337–342. DOI: https://doi.org/10.1038/nclimate2531 

  79. Roemmich, D. 2015. Unabated planetary warming and its ocean structure since 2006. Nature Climate Change, 5(3): 240–245. DOI: https://doi.org/10.1038/nclimate2513 

  80. Ruprich-Robert, Y. 2017. Assessing the Climate Impacts of the Observed Atlantic Multidecadal Variability Using the GFDL CM2.1 and NCAR CESM1 Global Coupled Models. Journal of Climate, 30(8): 2785–2810. DOI: https://doi.org/10.1175/JCLI-D-16-0127.1 

  81. Santer, BD. 2014. Volcanic contribution to decadal changes in tropospheric temperature. Nature Geoscience, 7(3): 185–189. DOI: https://doi.org/10.1038/ngeo2098 

  82. Schlesinger, ME. 1994. An oscillation in the global climate system of period 65–70 years. Nature, 367(6465): 723–726. DOI: https://doi.org/10.1038/367723a0 

  83. Schmidt, GA. 2014. Reconciling warming trends. Nature Geoscience, 7(3): 158–160. DOI: https://doi.org/10.1038/ngeo2105 

  84. Seager, R. 2003. Mechanisms of Hemispherically Symmetric Climate Variability. Journal of Climate, 16(18): 2960–2978. DOI: https://doi.org/10.1175/1520-0442(2003)016<2960:MOHSCV>2.0.CO;2 

  85. Solomon, S. 2010. Contributions of stratospheric water vapor to decadal changes in the rate of global warming. Science, 327(5970): 1219–1223. DOI: https://doi.org/10.1126/science.1182488 

  86. Solomon, S. 2011. The persistently variable “background” stratospheric aerosol layer and global climate change. Science, 333(6044): 866–870. DOI: https://doi.org/10.1126/science.1206027 

  87. Su, H. 2017. Inconsistent Subsurface and Deeper Ocean Warming Signals During Recent Global Warming and Hiatus. Journal of Geophysical Research: Oceans, 122(10): 8182–8195. DOI: https://doi.org/10.1002/2016JC012481 

  88. Sun, C. 2015. A delayed oscillator model for the quasi-periodic multi-decadal variability of the NAO. Climate Dynamics, 45(7–8): 2083–2099. DOI: https://doi.org/10.1007/s00382-014-2459-z 

  89. Trenberth, KE. 2002. Evolution of El Niño–Southern Oscillation and global atmospheric surface temperatures. Journal of Geophysical Research, 107(D8): 4065. DOI: https://doi.org/10.1029/2000JD000298 

  90. Trenberth, KE. 2010. Tracking Earth’s energy. Science, 328(5976): 316–317. DOI: https://doi.org/10.1126/science.1187272 

  91. Trenberth, KE. 2013. An apparent hiatus in global warming? Earth’s Future, 1(1): 19–32. DOI: https://doi.org/10.1002/2013EF000165 

  92. Trenberth, KE. 2014. Seasonal aspects of the recent pause in surface warming. Nature Climate Change, 4(10): 911–916. DOI: https://doi.org/10.1038/nclimate2341 

  93. Trenberth, KE. 2015. Has there been a hiatus? Science, 349(6249): 691–692. DOI: https://doi.org/10.1126/science.aac9225 

  94. Tung, KK. 2013. Using data to attribute episodes of warming and cooling in instrumental records. PNAS, 110(6): 2058–2063. DOI: https://doi.org/10.1073/pnas.1212471110 

  95. Tung, KK. 2019. Interdecadal variability in pan-Pacific and global SST, revisited. Climate Dynamics, 52(3): 2145–2157. DOI: https://doi.org/10.1007/s00382-018-4240-1 

  96. von Känel, L. 2017. Hiatus-like decades in the absence of equatorial Pacific cooling and accelerated global ocean heat uptake. Geophysical Research Letters, 44(15): 7909–7918. DOI: https://doi.org/10.1002/2017GL073578 

  97. Watanabe, M. 2013. Strengthening of ocean heat uptake efficiency associated with the recent climate hiatus. Geophysical Research Letters, 40(12): 3175–3179. DOI: https://doi.org/10.1002/grl.50541 

  98. Watanabe, M. 2014. Contribution of natural decadal variability to global warming acceleration and hiatus. Nature Climate Change, 4(10): 893–897. DOI: https://doi.org/10.1038/nclimate2355 

  99. White, WB. 2003. A delayed action oscillator shared by biennial, interannual, and decadal signals in the Pacific Basin. Journal of Geophysical Research, 108(C3): 3070. DOI: https://doi.org/10.1029/2002JC001490 

  100. Wijffels, S. 2016. Ocean temperatures chronicle the ongoing warming of Earth. Nature Climate Change, 6(2): 116–118. DOI: https://doi.org/10.1038/nclimate2924 

  101. Wu, T. 2019. New insights into natural variability and anthropogenic forcing of global/regional climate evolution. npj Climate and Atmospheric Science, 2(1): 18. DOI: https://doi.org/10.1038/s41612-019-0075-7 

  102. Wu, Z. 2011. On the time-varying trend in global-mean surface temperature. Climate Dynamics, 37(3–4): 759–773. DOI: https://doi.org/10.1007/s00382-011-1128-8 

  103. Xie, SP. 2017. What Caused the Global Surface Warming Hiatus of 1998–2013? Current Climate Change Reports, 3(2): 128–140. DOI: https://doi.org/10.1007/s40641-017-0063-0 

  104. Yan, XH. 2016. The global warming hiatus: Slowdown or redistribution? Earth’s Future, 4(11): 472–482. DOI: https://doi.org/10.1002/2016EF000417 

  105. Yao, SL. 2015. The global warming hiatus – a natural product of interactions of a secular warming trend and a multi-decadal oscillation. Theoretical and Applied Climatology, 123(1–2): 349–360. DOI: https://doi.org/10.1007/s00704-014-1358-x 

  106. Yao, SL. 2017. Distinct global warming rates tied to multiple ocean surface temperature changes. Nature Climate Change, 7(7): 486–491. DOI: https://doi.org/10.1038/nclimate3304 

  107. Zhang, R. 2007. Impact of the Atlantic Multi-decadal Oscillation on North Pacific climate variability. Geophysical Research Letters, 34(23): L23708. DOI: https://doi.org/10.1029/2007GL031601 

  108. Zhang, R. 2007. Can the Atlantic Ocean drive the observed multi-decadal variability in Northern Hemisphere mean temperature? Geophysical Research Letters, 34(2): L02709. DOI: https://doi.org/10.1029/2006GL028683 

  109. Zhang, Y. 1997. ENSO-like Interdecadal Variability: 1900–93. Journal of Climate, 10(5): 1004–1020. DOI: https://doi.org/10.1175/1520-0442(1997)010<1004:ELIV>2.0.CO;2 

comments powered by Disqus