The Response of Subtropical Highs to Climate Change

Current Climate Change Reports https://doi.org/10.1007/s40641-018-0114-1

CLIMATE CHANGE AND ATMOSPHERIC CIRCULATION (R CHADWICK, SECTION EDITOR)

Annalisa Cherchi1 & Tercio Ambrizzi2 & Swadhin Behera3 & Ana Carolina Vasques Freitas4 & Yushi Morioka3 & Tianjun Zhou5

# Springer Nature Switzerland AG 2018

Abstract Purpose of Review Subtropical highs are an important component of the climate system with clear implications on the local climate regimes of the subtropical regions. In a climate change perspective, understanding and predicting subtropical highs and related climate is crucial to local societies for climate mitigation and adaptation strategies. We review the current understanding of the subtropical highs in the framework of climate change. Recent Findings Projected changes of subtropical highs are not uniform. Intensification, weakening, and shifts may largely differ in the two hemispheres but may also change across different ocean basins. For some regions, large inter-model spread representation of subtropical highs and related dynamics is largely responsible for the uncertainties in the projections. The understanding and evaluation of the projected changes may also depend on the metrics considered and may require investigations separating thermodynamical and dynamical processes. Summary The dynamics of subtropical highs has a well-established theoretical background but the understanding of its variability and change is still affected by large uncertainties. Climate model systematic errors, low-frequency chaotic variability, coupled ocean-atmosphere processes, and sensitivity to climate forcing are all sources of uncertainty that reduce the confidence in atmospheric circulation aspects of climate change, including the subtropical highs. Compensating signals, coming from a tug-of- war between components associated with direct carbon dioxide radiative forcing and indirect sea surface temperature warming, impose limits that must be considered.

Keywords Subtropical highs . Climate projections . Atmospheric circulation . Model biases

Introduction between 20 and 40° of latitude in each hemisphere. At the surface the subtropical high-pressure belt divides the easterly Subtropical highs (or subtropical anticyclones) are regions of trade winds from the mid-latitude westerly winds. This deter- semi-permanent high atmospheric pressure typically located mines, along with the subtropical jet in the upper troposphere, the poleward boundary of the tropical circulation, which moves equatorward of its annual-mean latitude during the This article is part of the Topical Collection on Climate Change and Atmospheric Circulation winter and poleward during the summer related to the seasonal migration of the Intertropical Convergence Zone [1]. ’ * Annalisa Cherchi Subtropical highs occupy about 40% of the Earth ssurface [email protected] and contribute, through the surface wind stress curl, to the maintenance of subtropical oceanic gyres with warm pole- 1 Fondazione Centro Euro-Mediterraneo sui Cambiamenti Climatici, ward western boundary currents and cool equatorward cur- Istituto Nazionale di Geofisica e Vulcanologia, Viale Berti Pichat 6/2, rents off the west coasts of the continents [2]. 40128 Bologna, Italy In the Northern Hemisphere, subtropical highs are located 2 University of São Paulo, São Paulo, SP, Brazil over the North Pacific and the North Atlantic Oceans, and in 3 Application Laboratory, Japan Agency for Marine Earth Science and the Southern Hemisphere, they are located in the South Technology, Yokohama, Japan Atlantic, southern Indian, and South Pacific Ocean (Fig. 1). 4 Federal University of Itajubá (UNIFEI), Itabira, MG, Brazil In both hemispheres, subtropical highs strongly influence the moisture transport from subtropical oceans, thus affecting re- 5 Institute of Atmospheric Physics, Chinese Academy of Sciences, Beijing, China gional precipitation over Caribbean [3], the USA [4, 5], East Curr Clim Change Rep

c d

e f

Fig. 1 Eight hundred fifty-hectopascal eddy streamfunction (106 m2/s) as differences between the end of the twenty-first century (2075–2100) in (a) JJA and (b) DJF climatology (1980–2005) computed from ERA- the CMIP5 RCP8.5 scenario and the CMIP5 historical simulation, in JJA Interim [138]. c and d are the same as a and b but computed as CMIP5 and DJF, respectively. In the differences, dotted shading denotes statistical multi-model mean of the twentieth century simulations. Eddy significance exceeding 99% confidence level based on a t test. The list of streamfunction is computed removing the zonal mean. e and f are the models used to compute the multi-model ensemble mean is as in [139]

Asia [6], southern and eastern Africa [7, 8], and southern precipitation likely driven by the intensification and poleward South America [9–11]. Subtropical highs also interact with expansion of both subtropical dry zones [23, 24]andHadley monsoon circulation [12, 13], tropical cyclone tracks cell [25, 26]. Considering their large contribution to global [14–16], marine stratus clouds with associated radiation bud- climate, this work intends to review the current knowledge on get [17–19], and Convergence Humidity Zones [10, 11]. the subtropical highs in the perspective of climate change and to Subtropical highs are important components of the atmo- highlight the main areas of future research for this topic. spheric circulation, and they play a major role in the formation of the world’s subtropical deserts and the zones of Mediterranean climate [2, 20]. In recent decades, in the context Background Framework of anthropogenic effects on climate and related global warming, changes have been observed in the position and intensity of the The subtropical highs can be distinguished into two main subtropical highs [21, 22], with decreased subtropical types: (i) zonally asymmetric highs, which are associated with Curr Clim Change Rep rotational flow and are strongest in summer, and (ii) zonally summer, a zonal wavenumber-3 component dominates the symmetric highs driven by Hadley cell descent, which are upper-level planetary wave field, showing some correspon- associated with divergent flow and are strongest in winter. dence with the three high-pressure cells at the surface [2]. This distinction is ruled by removal of the zonal mean and There, the intensity of the subtropical highs decreases from works well in describing the differences between subtropical austral winter to summer (Fig. 1). Possible reasons are related highs in the Northern and Southern Hemispheres, respective- with continents in the Southern Hemisphere being zonally ly, but also between the seasons. In fact, if the dynamic of the narrower than their counterpart in the Northern Hemisphere subtropical highs is considered similar in the two hemispheres [2], monsoons being less important generators of zonal [2], the dominant mechanisms responsible for their formation asymmetries in the mostly ocean-covered Southern vary with the season. In the summer hemisphere, the zonal Hemisphere, and topographic effects becoming stronger in asymmetry in diabatic heating associated with the land-sea winter as the flow intensifies [35]. Also, inter-hemispheric thermal contrast is dominant, whereas during winter, when teleconnections have been found to contribute to either main- convective activity over the continents is weaker, the presence taining or strengthening the southern subtropical anticyclones of the orography and the subsidence caused by the Hadley during the winter because of the heating and forcing from the circulation prevail [20, 27]. The observed mean position and summer monsoons and deep tropical convection over warm intensity of the subtropical highs in the two hemispheres and sea surface temperatures (SSTs) in the Northern Hemisphere in the different peak seasons is summarized in Fig. 1a, b in [35–37]. terms of eddy streamfunction at 850 hPa. Hadley circulation, radiation balances, and SSTs all con- Idealized numerical models with prescribed heating and tribute to the annual cycle of the subtropical highs [27, 38]. As realistic topography provided the hypothesis that the basic the transition to the early summer occurs, the wintertime zonal cause of the summer anticyclone formation is the latent heat band of high pressure underlying the subsiding branch of the released by the monsoons over the neighboring continents to Hadley cell starts to break when convection over land com- theeastorwest[28]. Remote diabatic heating in the monsoon mences. This is accomplished through vorticity balance, in- region induces a stationary Rossby wave pattern to the west cluding poleward low-level flow into the regions of deep con- that, interacting with the mid-latitude westerlies, produces adi- vection on the western flank of the subtropical high, while abatic descent with the aid of local orography. Also, local radiatively driven subsidence and equatorward flow dominate diabatic enhancement can lead to a strengthening of the de- on the eastern flank [22]. As the summer ends, the east-west scent [20]. With this monsoon-desert mechanism, the exis- asymmetry in SST tends to be damped by vertical flux of tence of the subtropical anticyclones in the Northern Pacific moist static energy. Hence, subtropical anticyclones ultimately and Atlantic sectors has been related to the Asian and North become more zonally symmetric when precipitation in the American monsoon, respectively, and that in the South subtropics weakens and eventually shifts to the other hemi- Atlantic Ocean to the South American monsoon [20]. sphere in winter [22]. Therefore, the monsoon-desert mechanism provides a link between tropical and subtropical climates through Rossby wave propagation. For South Asia and eastern Subtropical Highs: Observations, Modeling, Mediterranean, this mechanism is identified also at interannu- and Related Dynamics al time scale [29] with a precise cycle and seasonally locked anticyclones over the Middle East because of the local topog- Subtropical Highs of the Northern Hemisphere raphy effect [30]. Results from Atmospheric General Circulation Models The western North Pacific subtropical high (WNPSH) is an (AGCM) experiments confirm the conclusions based on linear important component of East Asian summer monsoon models in terms of the influence of monsoonal heating, local (EASM) system [39]. The WNPSH and the EASM have a diabatic enhancement, and land-atmosphere coupling process- significant effect on droughts, heat waves, and tropical cy- es for the formation of the subtropical highs [31, 32], mostly clone tracks over East Asia and the northwest Pacific for the Northern Hemisphere summer. Including an ocean [40–43] and can modulate summer rainfall extending from mixed layer, the importance of air-sea interaction to enhance the eastern China to Japan [44, 45]. Seasonal and intra- simulated subtropical anticyclones, to position them in the seasonal variations of the WNPSH are related with the onset eastern basins, and also to regulate their seasonal cycle has and withdrawal of the EASM [39], but also to remote ENSO been reported [27]. (El Niño Southern Oscillation)-SST forcing [46, 47]. In a zonal mean perspective, the Hadley circulation exists The interannual variability of the WNPSH is dominated by even in an idealized framework without the monsoon [33, 34], two leading modes with the positive phases featuring an and the formation of the subtropical highs in winter is associ- anomalous anticyclone over the western North Pacific ated with its descending branch. In the Southern Hemisphere (WNP). The first mode is closely connected with the SST Curr Clim Change Rep anomalies over the tropical Indian Ocean, the Maritime when the western ridge is located in the northwest, a precipi- Continent, and the equatorial central Pacific, while the second tation deficit prevails with downward motion dominating the mode is closely connected with the SST anomalies over the region [21]. The NASH variability, in terms of both intensity WNP [48]. Changes in the sensible heating over the Tibetan and position, is dominated by a quasi-biennial oscillation in Plateau also modulate the interannual variability of the winter and by a lower frequency (~ 7 years) in summer, likely WNPSH: when the heating is stronger in spring, it is usually forced by the North Atlantic Ocean variability as the correla- followed by an enhanced and westward-extended WNPSH in tions with ENSO is nonsignificant [62]. summer [49]. At both interannual and decadal time scales, the In the last 30 years, the variability of the precipitation over distributions of summer monsoon rainfall over East Asia are the eastern USA increased because of changes in the intensity dominated by the variations of WNPSH in intensity, structure, and position of the western ridge of the Bermuda high [4, 63]. and location [6, 39, 44]. Strong anomalous shallow cross-equatorial circulation en- Since the early 1990s, the interannual variability of the ables extreme cold surges from extra-tropical South America WNPSH is more strongly regulated by the SST anomalies to influence surface temperature and circulation over the trop- over the equatorial central Pacific and the Maritime ical and subtropical Atlantic Ocean and the southern edge of Continent. The early decays of El Niños in strong WNPSH the NASH, increasing its pressure and leading to equatorward years and the early development of El Niños in weak WNPSH expansion of its southern boundary, providing a potential years contributed to this change [50]. Both decaying El Niño source of predictability for lower tropospheric circulation and developing La Niña events accompanied by a warm anomalies during boreal summer [64]. Indian Ocean and cold central Pacific, respectively, are favor- The Azores high is the southern branch of the North able to hotter summers in the central eastern China because Atlantic Oscillation (NAO) [65] which is associated with these patterns strengthen and extend the WNPSH toward west changes in temperature and rainfall in Europe and North [51–53]. The Pacific Decadal Oscillation (PDO) accounts for America with day-to-day variability linked with weather sys- the low-frequency variability of the WNPSH through the me- tem and seasonal and longer term variability with some useful ridional shift of the subtropical jet [54]. In winter, tropical predictive skill [66]. Multi-decadal variations of the NAO can Atlantic warming associated with the positive phase of the induce multi-decadal variations in the Atlantic meridional Atlantic Meridional Oscillation (AMO) provides favorable overturning circulation and poleward ocean heat transport in conditions for the intensification and northwestward extension the Atlantic up to the Arctic. This is likely to contribute to loss of the WNPSH [55]. of Arctic sea-ice, warming of the Northern Hemisphere and Since the late 1970s, the WNPSH has extended westward, changes in the Atlantic tropical storm activities [67]. These which has resulted in a monsoon rain band shift over China, multi-decadal variations are superimposed on long-term an- with excessive rainfall along the middle and lower reaches of thropogenic forcing trends. the Yangtze River valley along ~ 30° N over eastern China, and deficient rainfall in north China [56, 57]. Numerical ex- Subtropical Highs of the Southern Hemisphere periments suggest that negative heating in central and eastern Pacific and increased convective heating in the equatorial The South Atlantic subtropical high (SASH) is very important Indian Ocean and Maritime Continent sector, associated with for the regional climate, influencing the rainfall variability in local warming, favor the westward extension of the WNPSH South America and southern Africa, whose orography has an [58]. important role on its localization over the ocean [36]. The The subtropical high located over the North Atlantic SASH is strongest and largest during the solstitial months, (NASH) is referred to as BBermuda high^ (because of the when its center is either closest to the equator and on the place where it is centered) but sometimes also BAzores high^ western side of the South Atlantic basin as in austral winter considering its eastern extension. The NASH has a major [68] or farthest poleward in the center of the South Atlantic influence on weather and climate over the eastern USA [21, basin as in late austral summer [22]. In austral summer, the 59], Western Europe, and northwestern Africa [60]. spatial variations of the SASH on the interannual time scale In boreal summer, the Azores anticyclone has high pressure are primarily in the meridional direction and are dominated by dominating the Atlantic basin, and in boreal winter, it extends ENSO [22]. During the austral winter, multiple physical over eastern North America and western Africa [61]. The mechanisms, like Southern Annular Mode (SAM), ENSO, movement of the NASH western ridge toward the American variations in the Asian-African monsoon, anomalous forcing continent regulates moisture transport and vertical motions from convection over the Arabian Sea and over the Coral Sea over the southeastern USA. In fact, when the NASH western to the northeast of Australia, as well as the variations of ridge is located southwest of its mean climate position, exces- Southern Hemisphere storm tracks, appear to impact the po- sive summer precipitation is observed over the southeastern sition of the SASH but none of them clearly dominate over the USA due to an enhanced moisture transport. On the contrary, anticyclone variability, both in terms of position and intensity Curr Clim Change Rep

[22]. Intensification and poleward shift of the SASH in the cell widening [83]. Other anthropogenic forcing, like strato- recent decades have been explained in terms of intensification spheric ozone depletion, may likely play an important role of the westward mixed layer ocean currents and Benguela [88, 89]. upwelling region along the Namibian coast [69]. The South Pacific subtropical high (SPSH) is the main The SASH exhibits distinct decadal and multi-decadal var- atmospheric system that influences precipitation in southwest- iability. Once the SST anomalies are generated by the atmo- ern South America. The maintenance of the equatorward por- spheric variations through changes in meridional and vertical tion of the SPSH and the equatorward low-level jet along the heat transports, they initiate to migrate eastward as quasi- South American coast, during the austral winter, is due to the stationary oceanic Rossby waves along the eastward contribution of the adiabatic subsidence over the southeastern Antarctic Circumpolar Current [70]. This eastward migration tropical Pacific, produced by the warm pool convection [37]. further contributes to the decadal SST variability in the south- This interhemispheric response to heating in the Northern ern Indian Ocean [71, 72]. Since both the SST and sea level Hemisphere depends critically on the configuration of the pressure (SLP) anomalies exhibit distinct eastward propaga- mean zonal winds in the Southern Hemisphere, and it is more tion, highs over the South Atlantic and southern Indian sub- dramatic in the South Pacific, as numerical model simulations tropical oceans may have remote links on decadal time scale. indicate that the SPSH nearly disappears in the austral winter The subtropical high in the southern Indian Ocean is also without the influence from the Northern Hemisphere [35]. called BMascarene High.^ It is one major Southern Hemisphere circulation system that can also affect the Subtropical Ocean Dipoles EASM through changes in the Somali jet [39, 73]. Since the decadal variability in the Mascarene High is accompanied The interannual variations of the subtropical highs in the with basin-wide SST anomalies in the southern Indian Southern Hemisphere generate meridional dipole patterns of Ocean [74–76], local air-sea interaction may be involved in warm and cold SST anomalies, named Bsubtropical ocean sustaining the subtropical variations, both in intensity and po- dipoles.^ Three subtropical dipole SST modes have been, in sition, on decadal time scale. Also, the decadal variability of fact, identified in each of the three subtropical ocean basins: Mascarene High is likely responsible for low-frequency rain- Indian [7], Atlantic [90, 91], and Pacific [92, 93]. Subtropical fall variability over Southern Africa [77], through changes in dipoles develop during austral summer when the continental moisture transport [74, 75, 78]. regions in the subtropical Southern Hemisphere receive most As a potential source of the decadal SST variability, the of annual rainfalls and when the local atmospheric subtropical decadal modulation of the Indonesian Throughflow (ITF) highs are less intense. Generation mechanisms of subtropical may induce decadal changes in ocean heat transport from dipoles include air-sea interaction processes [91, 94, 95]and the tropical Pacific to the Indian Ocean [79]. Since the ITF remote teleconnection with the ENSO or the SAM [91, transport receives strong influences from the tropical Pacific 96–98]. These SST dipole events modulate the atmospheric condition, the decadal variability in the tropical Pacific Ocean circulation and convection with a clear impact on the rainfall may play indirect roles in inducing decadal variability of the over the subtropical continents [92, 99]. Mascarene High through modulation of ocean heat transport The subtropical South Atlantic experienced a cooling trend and hence air-sea interaction in the southern Indian Ocean. On in the last two decades [86] consistent with the response asso- the other hand, the decadal modulation of the SAM involving ciated with the positive phase of the South Atlantic subtropical ozone variability in the high latitudes is suggested to be the dipole [100, 101]. This suggests that the subtropical dipole other remote factor for the decadal variability of the pattern has likely been intensifying and/or become more Mascarene High [8]. prominent over the past three decades. In the southern Over the South Atlantic and southern Indian Oceans, the Indian Ocean, the positive phase of the subtropical dipole is subtropical anticyclones show a common tendency toward characterized by the occurrence of a cold (warm) SST anom- southward displacement [23, 80–82], consistently with the aly in the southeastern (southwestern) sector during the austral poleward expansion of the Hadley cell [80, 81, 83–85]. A summer [7]. During the mature phase, the subtropical high poleward shift in the subtropical anticyclones in both hemi- strengthens and shifts slightly southward. In the South spheres has been identified, consistently with the observed Pacific, the subtropical dipole was identified for the first time recent trend of positive polarity of the SAM and the observed as part of a global wavenumber-3 dipole SST mode [92], poleward expansion of the Hadley circulation and widening of linearly independent from both ENSO and the SAM. Later the tropical belt [81, 86]. Coupled Model Intercomparison on, it has been described as a mode associated with a Project Phase 5 (CMIP5, [87]) twentieth century simulations northeast-southwest-oriented dipole of positive and negative reproduce the observed Hadley cell expansion mostly in the SST anomalies in the central basin with the SST poles devel- Southern Atlantic and South Indian Oceans, but greenhouse oping during austral spring, peaking in austral summer, and, gases (GHGs) forcing alone cannot explain the local Hadley then, gradually decaying afterward [93]. The SLP anomalies Curr Clim Change Rep that generate the South Pacific subtropical dipole are linked more recent comparison is based on changes in subsidence, with the geopotential height anomalies in the upper tropo- low-level divergence, and rotational wind. In addition, eddy sphere and are associated with a stationary Rossby wave pat- geopotential height is used instead of the traditional tern along the westerly jet in the mid-latitudes, suggesting geopotential height to measure the intensity of the subtropical remotely induced signals in the generation of this dipole mode highs. This is because the latter is supposed to systematically [93]. increase with increased temperature, thus appearing less ap- propriate to be used in climate change studies [107, 108]. Diagnostic analyses and idealized simulations with linear Subtropical Highs in Future Warming baroclinic model suggest that the projected changes in the Scenarios and Related Climatic Consequences subtropical anticyclones are well-explained by the combined effect of increased tropospheric static stability and changes in The performance of CMIP5 models in representing mean po- diabatic heating. The pattern of change in diabatic heating is sition and intensity of the subtropical highs in the different dominated by latent heating associated with changes in pre- seasons, as described in the section above, is summarized in cipitation, which is enhanced over the WNP under the Brichest Fig. 1c, d in terms of eddy streamfunction at 850 hPa, to be get richer^ mechanism but is reduced over subtropical North compared with the observations. Atlantic and South Pacific due to a local minimum in the Alterations in the location or strength of the subtropical amplitude of SST warming. The change in the diabatic heating anticyclones have major implications for climate change at pattern substantially enhances the subtropical highs over the both regional and global scales. In the following, a review North Atlantic and South Pacific but weakens the North on the investigations of these changes is given based on stud- Pacific one [82]. ies using results from twenty-first century scenario experi- In climate projections, the weakening and the eastward ments performed in the framework of the 3rd and 5th phases retreat of the WNPSH are robust in the middle troposphere, of the Coupled Model Intercomparison Project (CMIP3 [102] while at low-levels, the projected changes in the WNPSH and CMIP5 [87], respectively), as well as on sensitivity intensity are approximately zero in the multi-model ensemble experiments. mean though with large inter-model spread [53]. The uncer- Unambiguous intensification of the subtropical highs in tainty of WNPSH projection adds uncertainty to the projection both Northern and Southern Hemispheres is reported based of future monsoon changes over East Asia. Under both on CMIP3- and CMIP5-coupled model projections [103, RCP4.5 and RCP8.5 scenarios, CMIP5 models with a signif- 104]. In the Northern Hemisphere, CMIP3 future warming icantly increased (decreased) WNPSH intensity have a signif- scenarios agree on the strengthening of the subtropical highs icant increase in the precipitation over the northern (southern) as primarily caused by enhanced diabatic heating over conti- part of eastern Asia and an enhanced (weakened) southerly nents and cooling over the ocean favoring a stronger near- wind [109]. surface anticyclonic circulation [103]. Same mechanisms Direct effect of the changing natural and anthropogenic have been found responsible for the intensification of the sub- forcing factors on atmospheric and land-surface temperature, tropical highs in the Southern Hemisphere summer comparing indirect effect on climate of these factors through SSTand sea- CMIP3- and CMIP5-coupled model future warming scenari- ice changes, and internal atmospheric and oceanic variability os, plus a positive feedback with the marine boundary layer all contribute to the climate change and variability observed so clouds mostly over the Atlantic and the Pacific subtropical far [110]. Combining CMIP5 results and AGCM sensitivity oceans [104]. In all CMIP5 scenarios, the WNPSH is experiments, the response of the Northern Hemisphere sum- projected to enlarge, strengthen, and extend westward with mertime subtropical highs can be linearly decomposed into a clear implications for the attribution and prediction of climate response to the carbon dioxide (CO2) direct effect and one to changes in East Asia. However, the ridge line of the high does SST warming [111]. In this framework, the response of the not evidence any long-term trend [105]. The WNPSH and the subtropical high in the North Pacific is weak because it repre-

East Asian Jet (EAJ) will remain the dominant systems sents the outcome of a tug-of-war between CO2 direct effect, influencing East Asian summer precipitation in the twenty- which strengthens the high, consistently with enhanced land- first century but the role of the WNPSH is projected to weaken ocean moist entropy contrast, and the SST warming which in compared to that of the EAJ [106]. turn weakens the high because of a weakened land-ocean A more recent comparison based on CMIP5 scenarios contrast. The opposing influences on the North Pacific sub- shows slightly different results, stating that subtropical highs tropical high also affect the response of the Pacific jet streams over the North Pacific, South Atlantic, and southern Indian [111]. Consistently, the dynamics component of the moisture Ocean are projected to become weaker, whereas the NASH transport over South and East Asia dominates near-term will intensify as well as the SPSH though there is uncertainty changes, while the thermodynamic component dominates in in its projections [82]. Differently from previous studies, this the long-term projections [112]. Curr Clim Change Rep

On the other hand, the response of the NASH and its west- Challenges and Future Prospects ward shift is dominated by CO2 direct effect [111], specifically for increased CO2 effect over land in the Eastern Hemisphere Subtropical highs undergo significant seasonal as well as in- [113]. Similar arguments, i.e., increased radiative forcing over terannual variations; hence, potential predictability skills of North Africa and land-sea heating contrast in response to the their variations are essential for local societies to implement forcing, have been used to explain the westward shift of measures to mitigate climate risks and for climate adaptation NASH documented in the Mid-Holocene driven by changes policies. The WNPSH may be highly predictable as it is pri- in the orbital parameters [114]. marily controlled by the central Pacific cooling/warming and a The Mediterranean climate with dry summers and wet win- positive feedback with the Indo-Pacific warm pool [42], en- ters, typical of the densely populated warm-temperate regions abling improved prediction skills for the EASM rainfall and of the world, is projected to shift northward and eastward in tropical storm activity in the western North Pacific. Skill for CMIP5 climate scenarios with the equatorward margins re- the predictability of the WNPSH also becomes important to placed by arid climate type [115]. These changes appear less the prediction of heatwaves over East Asia [52]. robust over California in winter. There, in fact, CMIP5 Most CMIP5 models are able to capture the spatial distri- models largely disagree in the projections, reflecting a bution and variability of the 500-hPa geopotential height and precarious balance between the subtropical highs zonal wind fields in the subtropical western North Pacific, but expanding from the south and the Aleutian low extend- they tend to underestimate the mean intensity of the WNPSH ing southeastward [116, 117]. [105]. This underestimation may be associated with the cold Impacts from changes in the intensity and characteristics of systematic bias of sea surface temperature in the tropical the subtropical highs extend to Australia over rangelands Indian and western Pacific oceans in the models. Nearly all [118] due to the expansion of subtropical dry zone and to CMIP3 and CMIP5 AGCMs exhibit northward shift in the the western coasts zone management [119] because of the mean state of the WNPSH ridge line [129]. Atmosphere- variability of the Southern Ocean storm belt related with the ocean coupling in the North Pacific remains a major source subtropical high-pressure ridge. Also, the health of coral of uncertainty for the relationship between the northward shift fauna over southwestern Atlantic [120] and in the sub- of WNPSH, the widening of Hadley cell, and related implica- tropics [121] as well as the urban climate in subtropical tions for winter precipitation in California, for example [117]. regions [122], including heat waves and related mortality The eastern South Atlantic is a region where most climate [123], may largely depend on subtropical high variability models typically exhibit serious errors in the form of a severe and change. warm bias in simulated SSTs, which tends to be accompanied In a warmer climate, changes in the strength and width of by an erroneous westward shift of the SASH [27, 130]. In the the Hadley cell may significantly alter stationary Rossby wave case of uncoupled atmospheric simulations with prescribed sources and characteristics of propagation [124], with impor- SSTs, serious deficiencies in the simulation of the SASH are tant consequences for winter conditions in South America and also documented [36]. These issues highlight the need to im- western Africa [125]. The recent expansion of the Hadley cell prove the representation of the SASH in the models, which in the Southern Hemisphere has been associated with the in- demands the nontrivial task of capturing the intricate loop of tensification and poleward shift of the subtropical highs [126], feedbacks involving the subtropical zonal SST gradient, land- likely followed by an increased aridity on the eastern flanks of sea heating contrasts, and ocean dynamics [131]. The interac- subtropical anticyclones under enhanced greenhouse gas forc- tion between the SASH and the continental thermal low needs ing [23, 127]. The poleward expansion has been found to be further investigation because of implications in present-day largest in autumn, raising questions as to whether it is a de- climate variability and future climate change [86]. The inabil- layed effect from ozone depletion that climate models fail to ity of the CMIP5 models to capture the co-variability between capture or entirely due to global warming, whose impact is SAM and ENSO substantially limits the confidence in their underestimated in CMIP3 models unless forced with observed future Southern Hemisphere rainfall projections, especially SSTs [23]. In austral summer, rainfall over the Southern regarding projections of extreme seasonal rainfall, that Hemisphere subtropics is projected to increase because of a may depend on the concurrence of SAM and ENSO events robust positive trend projected for the SAM [128]. [128, 132]. The fact that most models underestimate the in- The changes in mean intensity and position of the subtrop- tensification of subtropical ridge relates to the underestimation ical highs as described in the cited literature are summarized in of the Hadley cell expansion in present climate and in future Fig. 1e, f in terms of 850 hPa eddy streamfunction differences warming scenarios [124] and opens up the possibility that of the climatology at the end of the twenty-first century fol- declines of projected rainfall in the subtropical regions may lowing the RCP8.5 scenario and of the climatology at the end be underestimated [133]. of the twentieth century in historical simulations using CMIP5 It is also important to distinguish weak projected circula- model results. tion responses to global warming arising because of genuine Curr Clim Change Rep uncertainty and lack of model agreement, from weak re- model projections of the regional circulation responses to sponses arising because of competing effects that are robust global warming. and physically understood [111]. These situations are not un- common; for example, the future Southern Hemisphere sum- Acknowledgements We are grateful to the two anonymous reviewers mertime circulation is expected to involve a tug-of-war be- whose comments helped in improving the shape and content of the man- uscript. A special thank is due to Dr. X Chen for the help in redrawing tween ozone recovery and greenhouse gas increases [111]. Fig. 1 using CMIP5 model results.

Compliance with Ethical Standards

Conclusions Conflict of Interest On behalf of all authors, the corresponding author states that there is no conflict of interest. Semi-permanent high-pressure systems of the global atmospheric circulation and their influence on regional weather and climate are studied since the end of the nineteenth century, References when, for the first time, their centers of actions have been identified from maps of monthly mean sea-level pressure 1. Saha K. Monsoon over Australia (region – IV). In: Tropical cir- [134]. It is only in the mid-twentieth century that the charac- culation systems and monsoons. Berlin Heidelberg: Springer- Verlag; 2010. https://doi.org/10.1007/978-3-642-03373-5_7. teristics of these features and the relative variability have been 2. Miyasaka T, Nakamura H. Structure and mechanisms of the south- mathematically described [135].Thelackofanadequatenet- ern hemisphere summertime subtropical anticyclones. J Clim. work of stations in middle and high latitudes of the Southern 2010;23:2115–30. Hemisphere limited the knowledge of the circulation patterns. 3. Gamble DW, Parnell DB, Curtis S. Spatial variability of the As new stations were set up in these relatively unobserved Caribbean mid-summer drought and relation to North Atlantic high circulation. Int J Climatol. 2008;28:343–50. areas, attempts were conducted to describe synoptic patterns, 4. Li W, Li L, Fu R, Deng Y, Wang H. Changes to the North Atlantic although only for limited regions [136]. In late twentieth cen- subtropical high and its role in the intensification of summer rain- tury, the theoretical grounds setting the existence and charac- fall variability in the southeastern United States. J Clim. 2011;24: teristics of the subtropical highs as peculiar aspects of the 1499–506. climate system have been defined and recognized [27, 28, 5. Luchetti NT, Nieto Ferreira R, Rickenbach TM, Nissenbaum MR, McAuliffe JD. Influence of the North Atlantic subtropical high on 137]. Based on that background, efforts have been dedicated wet and dry sea-breeze events in North Carolina, United States. to study specific aspects of the subtropical highs addressing Investigaciones Geograficas. 2017;68:9–25. https://doi.org/10. different studies and communities for each sector of relevance 14198/INGEO2017.68.01. in the world, such as the Southern Hemisphere or the western 6. Zhou T, Gong D, Li J, Li B. Detecting and understanding the multi-decadal variability of the East Asian Summer Monsoon - North Pacific region. recent progress and state of affairs. Meteorol Z. 2009a;18(4):455– The current availability of global observations, 3-D atmo- 67. spheric re-analyses, and results from a large number of numer- 7. Behera SK, Yamagata T. Subtropical SST dipole events in the ical simulations, performed with sophisticated global-coupled southern Indian Ocean. Geophys Res Lett. 2001;28:327–30. models from international frameworks like CMIP, largely 8. Manatsa D, Morioka Y, Behera SK, Matarira CH, Yamagata T. Impact of Mascarene high variability on the east African Bshort helped the progresses in the understanding of the dynamics rains^. Clim Dyn. 2014;42(5–6):1259–74. of the subtropical highs and their representation. However, 9. Doyle ME, Barros VR. Midsummer low-level circulation and pre- there are still many dark or gray zones which are yet to be cipitation in subtropical South America and related sea surface explored. In particular, multi-model comparisons highlighted temperature anomalies in the South Atlantic. J Clim. 2002;15(23):3394–410. deficiencies in the representation of the subtropical highs, 10. Reboita MS, Gan MA, Da Rocha RP, Ambrizzi T. Precipitation mostly related to coupled ocean-atmosphere processes, low regimes in South America: a bibliographic review. J Meteor. frequency variability, sensitivity to climate forcing, and inter- 2010;25:185–204. Available in Portuguese relationships among different climate phenomena and their 11. Gilliland JM, Keim BD. Position of South Atlantic anticyclone variability. These reported biases and uncertainties open fur- and its impact on surface conditions across Brazil. J Appl Meteor Clim doi. 2018;57:535–53. https://doi.org/10.1175/ ther questions on the reliability of models used to perform JAMC-D-17-0178.1. future climate projections, thereby reducing the confidence 12. Sun S, Ying M. Subtropical high anomalies over the western pa- in atmospheric circulation aspects of climate change and the cific and its relations to the Asian monsoon and SSTanomaly. Adv corresponding regional patterns of precipitation change. Atmos Sci. 1999;16:559–68. Compensating signals, coming from a tug-of-war between 13. Ding YH, Chan JCL. The East Asian summer monsoon: an overview. Meteor Atmos Phys. 2005;89:117–42. components associated with direct CO radiative forcing and 2 14. Bannister AJ, Boothe MA, Carr LE, Elsberry RL (1997) Southern indirect SST warming, impose limits that must be considered hemisphere application of the systematic approach to tropical cy- in terms of the confidence in the attribution, detection, and clone track forecasting: part I: environmental structure Curr Clim Change Rep

characteristics. Tech Rep NPS-MR-98-001 96 pp Naval anticyclones. J Clim. 2013;26:10193–204. https://doi.org/10. Postgraduate School, Monterey, CA. 1175/JCLI-D-13-00106.1. 15. Wu L, Wang B, Geng S. Growing typhoon influence on East Asia. 36. Richter I, Mechoso CR, Robertson AW. What determines the po- Geophys Res Lett. 2005;32:L18703. sition and intensity of the South Atlantic anticyclone in austral 16. Stowasser M, Wang Y, Hamilton K. Tropical cyclone changes in winter? - an AGCM study. J Clim. 2008;21:214–29. https://doi. the western North Pacific in a global warming scenario. J Clim. org/10.1175/2007JCLI1802.1. 2007;20:2378–96. 37. Wang C, Lee S-K, Mechoso CR. Interhemispheric influence of the 17. Klein SA, Hartmann DL. The seasonal cycle of low stratiform Atlantic warm pool on the southeastern Pacific. J Clim. 2010;23: clouds. J Clim. 1993;6:1587–606. 404–18. 18. Rahn DA, Garreaud R. Marine boundary layer over the subtropi- 38. Hastenrath S. Climate dynamics of the tropics: Kluwer; 1991. p. cal southeast Pacific during VOCALS-REx—part 1: mean struc- 488. ture and diurnal cycle. Atmos Chem Phys. 2010;10:4491–506. 39. Tao SY, Chen LX. A review of recent research on the East Asian 19. Wei W, Wenhong Li, Yi Deng, Song Yang, Jonathan H. Jiang, Lei summer monsoon in China. In: Chang CP, Krishnamurti TN, ed- Huang, W. Timothy Liu (2017) Dynamical and thermodynamical itors. Review of monsoon meteorology. London: Oxford Univ coupling between the North Atlantic subtropical high and the ma- Press; 1987. p. 353. rine boundary layer clouds in boreal summer. Clim Dyn DOI 40. Du Y, Yang I, Xie SP. Tropical Indian Ocean influence on https://doi.org/10.1007/s00382-017-3750-6, 50, 2457, 2469. Northwest Pacific tropical cyclones in summer following strong 20. Rodwell MJ, Hoskins BJ. Subtropical anticyclones and summer El Nino. J Clim. 2011;24:315–22. https://doi.org/10.1175/ monsoons. J Clim. 2001;14:3192–211. 2010JCLI3890.1. 21. Li L, Li W, Kushnir Y. Variation of the North Atlantic subtropical 41. Kosaka Y, Chowdary JS, Xie SP, Min YM, Lee JY. Limitations of high western ridge and its implication to southeastern US summer seasonal predictability for summer climate over East Asia and the precipitation. Clim Dyn. 2012a;39:1401–12. https://doi.org/10. northwestern Pacific. J Clim. 2012;25:7574–89. 1007/s00382-011-1214-y. 42. Wang B, Xiang BQ, Lee JY. Subtropical high predictability estab- 22. Sun X, Cook KH, Vizy EK. The South Atlantic subtropical high: lishes a promising way for monsoon and tropical storm predic- climatology and interannual variability. J Clim. 2017;30:3279–96. tions. Proc Nat Academy Sci. 2013;110:2718–22. 23. Cai W, Cowan T, Thatcher M. Rainfall reductions over southern 43. Zhang I, Zhou T. Drought over East Asia: a review. J Clim. hemisphere semi-arid regions: the role of subtropical dry zone 2015;28:3375–99. https://doi.org/10.1175/JCLI-D-14-00259.1. expansion. Sci Rep. 2012;2:702. https://doi.org/10.1038/ 44. Li T, Wang B, Wu B, Zhou T, Chang CP, Zhang R. Theories on srep00702. formation of an anomalous anticyclone in western North Pacific 24. Scheff J, Frierson D. Twenty-first-century multimodel subtropical during El Nino: a review. J Meteor Res. 2017;31:987–1006. precipitation declines are mostly midlatitude shifts. J Clim. https://doi.org/10.1007/s13351-017-7147-6. 2012;25:4330–47. https://doi.org/10.1175/JCLI-D-11-00393.1. 45. Nagata R, Mikami T. Changes in the relationship between summer 25. Hu Y, Zhou C, Liu J. Observational evidence for poleward expan- rainfall over Japan and the North Pacific subtropical high, 1901- sion of the Hadley circulation. Adv Atm Sci. 2011;28:33–44. 2000. Int J Climatol. 2017;37:3291–6. 26. Nguyen H, Hendon HH, Lim EP, Boschat G, Maloney E, Timbal 46. Dong X, Li R, Fan F. Comparison of the two modes of the Western B. Variability of the extent of the Hadley circulation in the south- Pacific subtropical high between early and late summer. Atm Sci ern hemisphere: a regional perspective. Clim Dyn. 2018;50:129– Lett. 2017;18:153–60. https://doi.org/10.1002/asl.737. 42. https://doi.org/10.1007/s00832-017-3592-2. 47. Qian W, Shi J. Tripole precipitation pattern and SST variations 27. Seager R, Murthugudde R, Naik N, Clement A, Gordon N, Miller linked with extreme zonal activities of the western Pacific subtrop- J. Air-sea interaction and the seasonal cycle of the subtropical ical high. Int J Climatol. 2017;37:3018–35. https://doi.org/10. anticyclones. J Clim. 2003;16:1948–66. 1002/joc.4897. 28. Rodwell MJ, Hoskins BJ. Monsoons and the dynamics of deserts. 48. He C, Zhou T. The two interannual variability modes of the west- Quart J Roy Meteor Soc. 1996;122:1385–404. ern North Pacific subtropical high simulated by 28 CMIP5-AMIP 29. Cherchi A, Annamalai H, Masina S, Navarra A. South Asian models. Clim Dyn. 2014;43:2455–69. https://doi.org/10.1007/ summer monsoon and the eastern Mediterranean climate: the s00382-014-2068-x. monsoon-desert mechanism in CMIP5 simulations. J Clim. 49. Duan A, Sun R, He J. Impact of surface sensible heating over the 2014;27:6877–903. https://doi.org/10.1175/JCLI-D-13-00530.1. Tibetan Plateau on the western Pacific subtropical high: a land-air- 30. Tyrlis E, Lelieveld J, Steil B. The summer circulation over the sea interaction perspective. Adv Atm Sci. 2017;34:157–68. eastern Mediterranean and the Middle East: influence of the 50. He C, Zhou T. Decadal change of the connection between summer South Asian monsoon. Clim Dyn. 2013;40:1103–23. https://doi. western North Pacific subtropical high and tropical SST in the org/10.1007/s00382-012-1258-4. early 1990s. Atm Sci Lett. 2015a;16:253–9. 31. Shaffrey LC, Hoskins BJ, Lu R. The relationship between the 51. Paek H, Yu JY, Zheng F, Lu MM. Impacts of ENSO diversity on North American summer monsoon, the Rocky Mountains and the western Pacific and North Pacific subtropical highs during the North Pacific subtropical anticyclone in HadAM3. Quart J boreal summer. Clim Dyn. 2016; https://doi.org/10.1007/s00382- Roy Meteor Soc. 2002;128:2607–22. 016-3288-z. 32. Liu Y, Wu G, Ren R. Relationship between the subtropical anti- 52. Chen X, Zhou T. Relative contributions of external SST forcing cyclone and diabatic heating. J Clim. 2004;17:682–98. and internal atmospheric variability to July-August heat wave over 33. Held IM, Hou AY. Nonlinear axially symmetric circulations in a the Yangtze River valley during 1979-2014. Clim Dyn. 2017; nearly inviscid atmosphere. J Atm Sci. 1980;37:515–33. https://doi.org/10.1007/s00382-017-3871-y. 34. Held IM (2000) The general circulation of the atmosphere. Woods 53. He C, Zhou T, Lin A, Wu B, Gu D, Li C, et al. Enhanced or Hole Oceanographic Institute Geophysical Fluid Dyamics weakened western North Pacific subtropical high under global Program, Woods Hole, Mass (available at https://www.gfdl.noaa. warming? Sci Rep. 2015b;5:16771. https://doi.org/10.1038/ gov/wp-content/uploads/files/user_files/ih/lectures/woods_hole. srep16771. pdf). 54. Matsumura S, Horinouchi T. Pacific Ocean decadal forcing of 35. Lee SK, Mechoso CR, Wang C, Neelin JD. Interhemispheric in- long-term changes in the western Pacific subtropical high. Sci fluence of the northern summer monsoons on southern subtropical Rep. 2016;6:37765. https://doi.org/10.1038/srep37765. Curr Clim Change Rep

55. Lyu K, Yu JY,Paek H. The influences of the Atlantic multidecadal 76. Yamagami Y, Tozuka T. Interdecadal changes of the Indian Ocean oscillation on the mean strength of the North Pacific subtropical subtropical dipole mode. Clim Dyn. 2015;44(11–12):3057–66. high during boreal winter. J Cli. 2017;30:411–24. https://doi.org/ 77. Tyson PD, Dyer TG, Mametse MN. Secular changes in south 10.1175/JCLI-D-16-0525.1. African rainfall: 1880 to 1972. Q J R Meteorol Soc. 56. Hu Z, Yang S, Wu R. Long-term climate variations in China and 1975;101(430):817–33. global warming signals. J Geophys Res. 2003;108:4614. https:// 78. Malherbe J, Landman WA, Engelbrecht FA. The bi-decadal rain- doi.org/10.1029/2003JD003651. fall cycle, Southern Annular Mode and tropical cyclones over the 57. Yu R, Zhou T. Seasonality and three-dimensional structure of the Limpopo River Basin, southern Africa. Clim Dyn. 2014;42(11– interdecadal change in East Asian monsoon. J Clim. 2007;20: 12):3121–38. 5344–55. 79. Reason CJC, Allan RJ, Lindesay JA. Evidence for the influence of 58. Zhou T, Yu R, Zhang J, Drange H, Cassou C, Deser C, et al. Why remote forcing on interdecadal variability in the southern Indian the western Pacific subtropical high has extended westward since Ocean. J Geophys Res Oceans. 1996;101(C5):11867–82. the late 1970s. J Clim. 2009b;22:2199–215. 80. Choi J, Son SW, Lu J, Min SK. Further observational evidence of 59. Katz RW, Parlange MB, Tebaldi C. Stochastic modeling of the Hadley cell widening in the southern hemisphere. Geophys Res effects of large-scale circulation on daily weather in the southeast- Lett. 2014;41:2590–7. https://doi.org/10.1002/2014GL059426. ern US. Clim Ch. 2003;60:189–21. 81. Lucas C, Nguyen H. Regional characteristics of tropical expansion 60. Nicholson SE. The West African Sahel: a review of recent studies and the role of climate variability. J Geophys Res Atm. 2015;120: on the rainfall regime and its interannual variability. ISRN Meteor 6809–24. https://doi.org/10.1002/2015JD023130. doi. 2013;2013:1–32. https://doi.org/10.1155/2013/453521. 82. He C, Wu B, Zou L, Zhou T. Responses of the summertime sub- 61. Davis RE, Hayden BP, Gay DA, Phillips WL, Jones GV. The tropical anticyclones to global warming. J Clim. 2017;30:6465– North Atlantic subtropical anticyclone. J Clim. 1997;10:728–44. 79. 62. Hasanean HM. Variability of the North Atlantic subtropical high 83. Kim Y-H, Min S-K, Son S-K, Choi J. Attribution of local Hadley and associations with tropical sea surface temperature. Int J cell widening in the southern hemisphere. Geophys Res Lett. Climatol. 2004;24:945–57. https://doi.org/10.1002/joc.1042. 2017;44:1015–24. https://doi.org/10.1002/2016GL072353. 63. Diem JE. Influences of the Bermuda high and atmospheric moist- 84. Lu J, Vecchi GA, Reichler T. Expansion of the Hadley cell under ening on changes in summer rainfall in the Atlanta, Georgia re- global warming. Geophys Res Lett. 2007;34:L06805. https://doi. gion, USA. Int J Climatol. 2013;33:160–72. https://doi.org/10. org/10.1029/2006GL028443. 1002/joc.3421. 85. Tao L, Hu Y, Liu J. Anthropogenic forcing on the Hadley circu- 64. Bowerman AR and co-authors (2017) An influence of extreme lation in CMIP5 simulations. Climate Dyn. 2016;46:3337–50. southern hemisphere cold surges on the North Atlantic subtropical https://doi.org/10.1007/s00382-015-2772-1. high through a shallow atmospheric circulation. J Geophys Res 86. Vizy EK, Cook KH. Understanding long-term (1982–2013) multi- Atm 122: 10,135 - 10,148. decadal change in the equatorial and subtropical South Atlantic 65. Walker GT, Bliss EW. World weather. V Mem Roy Meteor Soc. climate. Clim Dyn. 2016;46:2087–113. 1932;4:53–83. 87. Taylor KE, Stouffer RJ, Meehl GA. An overview of CMIP5 and 66. Scaife AA, et al. Skillful long-range prediction of European and the experiment design. Bull Am Meteor Soc. 2012;93:485–98. North American winters. Geophys Res Lett. 2014;41:2514–9. https://doi.org/10.1175/BAMS-D-11-00094.I. https://doi.org/10.1002/2014GL059637. 88. Min SK, Won SW. Multimodel attribution of the southern hemi- 67. Delworth TL, Zheng F, Vecchi GA, Yang X, Zhang L, Zhang R. sphere Hadley cell widening: major role of ozone depletion. J The North Atlantic oscillation as a driver of rapid climate change Geophys Res Atm. 2013;118:3007–15. in the northern hemisphere. Nat Geosci. 2016;9:509–12. https:// 89. Waugh DW, Garfinkel CI, Polvani LM. Drivers of the recent trop- doi.org/10.1038/ngeo2738. ical expansion in the southern hemisphere: changing SSTs or 68. Machel H, Kapala A, Flohn EH. Behaviour of the centres of action ozone depletion? J Clim. 2015;28:6581–6. above the Atlantic since 1881. Part I: characteristics of seasonal 90. Venegas SA, Mysak LA, Straub DN. Atmosphere–ocean coupled and interannual variability. Int J Climatol. 1998;18:1–22. variability in the South Atlantic. J Clim. 1997;10(11):2904–20. 69. Vizy EK, Cook KH, Sun X. Decadal change of the South Atlantic 91. Fauchereau N, Trzaska S, Richard Y,Roucou P, Camberlin P. Sea- Ocean Angola-Benguela frontal zone since 1980. Clim Dyn. surface temperature co-variability in the Southern Atlantic and 2018; https://doi.org/10.1007/s00382-018-4077-7. Indian Oceans and its connections with the atmospheric circula- 70. Morioka, Y., Taguchi, B., & Behera, S. K. (2017). Eastward- tion in the southern hemisphere. Int J Climatol. 2003;23(6):663– propagating decadal temperature variability in the South Atlantic 77. and Indian Oceans. Journal of Geophysical Research: Oceans. 92. Wang F. Subtropical dipole mode in the southern hemisphere: a 71. Morioka Y, Engelbrecht F, Behera SK. Potential sources of decad- global view. Geophys Res Lett. 2010;37:L10702. https://doi.org/ al climate variability over southern Africa. J Clim. 2015b;28(22): 10.1029/2010GL042750. 8695–709. 93. Morioka Y,Ratnam JV,Sasaki W, Masumoto Y.Generation mech- 72. Le Bars D, Viebahn JP, Dijkstra HA. A Southern Ocean mode of anism of the South Pacific subtropical dipole. J Clim. multidecadal variability. Geophys Res Lett. 2016;43(5):2102–10. 2013;26(16):6033–45. 73. Xue F and co-authors (2015) Recent advances in monsoon studies 94. Suzuki R, Behera SK, Iizuka S, Yamagata T. Indian Ocean sub- in China. Adv Atm Sci 32: 206–229 doi: https://doi.org/10.1007/ tropical dipole simulated using a coupled general circulation mod- s00376-014-0015-8. el. J Geophys Res: Oceans. 2004;109(C9) 74. Allan RJ, Lindesay JA, Reason CJ. Multidecadal variability in the 95. Morioka Y, Tozuka T, Yamagata T. Climate variability in the climate system over the Indian Ocean region during the austral southern Indian Ocean as revealed by self-organizing maps. summer. J Clim. 1995;8(7):1853–73. Clim Dyn. 2010;35(6):1059–72. 75. Reason CJC. Warm and cold events in the southeast Atlantic/ 96. Colberg F, Reason CJC, Rodgers K. South Atlantic response to El southwest Indian Ocean region and potential impacts on circula- Niño–Southern Oscillation induced climate variability in an ocean tion and rainfall over southern Africa. Met Atm Phys. 1998;69(1– general circulation model. J Geophys Res: Oceans. 2):49–65. 2004;109(C12) Curr Clim Change Rep

97. Morioka Y, Masson S, Terray P, Prodhomme C, Behera SK, and other Mediterranean climate regions. Sci Rep 7: 10783 doi: Masumoto Y. Role of tropical SST variability on the formation https://doi.org/10.1038/s41598-017-11285-y. of subtropical dipoles. J Clim. 2014;27(12):4486–507. 117. Choi J, et al. Uncertainty in future projections of the North Pacific 98. Rodrigues RR, Campos EJ, Haarsma R. The impact of ENSO on subtropical high and its implication for California winter precipi- the South Atlantic subtropical dipole mode. J Clim. 2015;28(7): tation change. J Geophys Res Atm. 2016;121:795–806. https:// 2691–705. doi.org/10.1002/2015JD023858. 99. Reason CJC. Subtropical Indian Ocean SST dipole events and 118. Eldridge DJ, Beecham G (2018) The impact of climate variability southern African rainfall. Geophys Res Lett. 2001;28(11):2225–7. on land use and livelihoods in Australia’s rangelands. In: Gaur 100. Morioka Y,Tozuka T, Yamagata T. On the growth and decay of the MK, Squires VR (eds) BClimate variability impacts on land use subtropical dipole mode in the South Atlantic. J Clim. and livelihoods in drylands^ Springer. 2011;24(21):5538–54. 119. Wandres M, Pattiaratchi C, Wijeratne EMS, Hetzel Y. The influ- 101. Yuan C, Tozuka T, Luo JJ, Yamagata T. Predictability of the sub- ence of the subtropical high-pressure ridge on the western tropical dipole modes in a coupled ocean atmosphere model. Australian wave climate. J Coast Res. 2016;75:567–71. https:// Climate Dyn. 2014;42:1291–308. https://doi.org/10.1007/ doi.org/10.2112/SI75-114.1. s00382-013-1704-1. 120. Cassola GE, et al. Decline in abundance and health state of an 102. Meehl GA and co-authors (2007) The WCRP CMIP3 multimodel Atlantic subtropical gorgonian population. Mar Poll Bull. – dataset: a new era in climate change research. Bull Am Meteor Soc 2016;104:329 34. 88: 1383–1394 doi: https://doi.org/10.1175/BAMS-88-9-1383. 121. Xie JY and co-authors (2017) The 2014 summer coral bleaching 103. Li W, Li L, Ting M, Liu Y. Intensification of northern hemisphere event in subtropical Hong Kong. Mar Poll Bull doi: https://doi.org/ subtropical highs in a warming climate. Nat Geosci. 2012b;5:830– 10.1016/j.marpolbull.2017.03.061, 124, 653, 659. 4. https://doi.org/10.1038/NGEO1590. 122. Tan Z, Lau KKL, Ng E. Planning strategies for roadside tree 104. Li W, Li L, Ting M, Deng Y,Kushnir Y,Liu Y,et al. Intensification planting and outdoor comfort enhancement in subtropical high- – of the southern hemisphere summertime subtropical anticyclones density urban areas. Build Environ. 2017;120:93 109. https://doi. in a warming climate. Geophys Res Lett. 2013;40:5959–64. org/10.1016/j.buildenv.2017.05.017. https://doi.org/10.1002/2013GL058124. 123. Son JY, et al. The impact of temperature on mortality in a subtropical city: effects of cold, heat and heatwaves in Sao Paulo, Brazil. 105. Liu Y, Li W, Zuo J, Hu ZZ. Simulation and projection of the Int J Biometeorol. 2016;60:113–21. western Pacific subtropical high in CMIP5 models. J Meteor ’ Res. 2014;28:327–40. 124. Freitas ACV, Frederiksen JS, O Kane TJ, Ambrizzi T. Simulated austral winter response of the Hadley circulation and stationary 106. Ren Y, Zhou B, Song L, Xiao Y. Interannual variability of western Rossby wave propagation to a warming climate. Clim Dyn. North Pacific subtropical high, East Asian jet and East Asian sum- 2017;49:521–45. https://doi.org/10.1007/s00382-016-3356-4. mer precipitation: CMIP5 simulation and projection. Quart Int. 125. Freitas ACV, Ambrizzi T. Changes in the austral winter Hadley 2017;440:64–70. circulation and the impact on stationary Rossby waves propaga- 107. He C, Zhou T, Wu B. The key oceanic regions responsible for the tion. Adv Meteorol. 2012;2012:1–15. https://doi.org/10.1155/ interannual variability of the western North Pacific subtropical 2012/980816. high and associated mechanisms. J Meteor Res. 2015a;29(4): 126. Nguyen H, Evans A, Lucas C, Smith I, Timbal B. The Hadley 562–75. circulation in re-analyses: climatology, variability, and change. J 108. He C, Lin A, Gu D, Li C, Zheng B, Wu B, et al. Using eddy Clim. 2013;26:3357–76. https://doi.org/10.1175/JCLI-D-12- geopotential height to measure the western North Pacific subtrop- 00224.1. ical high in a warming climate. Theor Appl Climatol. 2018;131: – 127. Seager R, Naik N, Vecchi GA. Thermodynamic and dynamic 681 91. https://doi.org/10.1007/s00704-016-2001-9. mechanisms for large-scale changes in the hydrological cycle in 109. He C, Zhou T. Responses of the western North Pacific subtropical response to global warming. J Clim. 2010;23:4651–68. https://doi. high to global warming under RCP4.5 and RCP8.5 scenarios org/10.1175/2010JCLI3655.1. projected by 33 CMIP5 models: the dominance of tropical 128. Lim EP, et al. The impact of the southern annular mode on future Indian Ocean - Tropical Western Pacific SST Gradient. J Clim. changes in southern hemisphere rainfall. Geophys Res Lett. – 2015b;28:365 80. 2016;43:7160–7. https://doi.org/10.1002/2016GL069453. 110. Folland CK, Sexton DMH, Karoly DJ, Johnson CE, Rowell DP, 129. Song F, Zhou T. Interannual variability of East Asian summer Parker DE. Influence of anthropogenic and oceanic forcing on monsoon simulated by CMIP3 and CMIP5 AGCMs: skill depen- – recent climate change. Geophys Res Lett. 1998;25:353 6. dence on Indian Ocean-western Pacific anticyclone 111. Shaw TA, Voigt A. Tug of war on summertime circulation be- teleconnections. J Clim. 2014a;27:1679–97. tween radiative forcing and sea surface warming. Nat Geosci. 130. Richter I, Xie S-P, Wittenberg AT, Masumoto Y. Tropical Atlantic – 2015;8:560 6. https://doi.org/10.1038/NGEO2449. biases and their relation to surface wind stress and terrestrial pre- 112. Li X, Ting M, Li C, Henderson N. Mechanisms of Asian summer cipitation. Clim Dyn. 2012;38:985–1001. https://doi.org/10.1007/ monsoon changes in response to anthropogenic forcing in CMIP5 s00382-011-1038-9. models. J Clim. 2015;28:4107–25. 131. Cabos W, Sein DV,Pinto JG, et al. The South Atlantic anticyclone 113. Shaw TA, Voigt A. Land dominates the regional response to CO2 as a key player for the representation of the tropical Atlantic cli- direct radiative forcing. Geophys Res Lett. 2016;43:11,383–91. mate in coupled climate models. Clim Dyn. 2017;48:4051–69. 114. Kelly P, Kravitz B, Lu J, Leung LR. Remote drying in the North https://doi.org/10.1007/s00382-016-3319-9. Atlantic as a common response to precessional changes and CO2 132. Hendon HH, Lim EP, Arblaster J, Anderson DTL. Causes and increase over land. Geophys Res Lett. 2018;45:3615–24. https:// predictability of the record wet spring over Australia in 2010. doi.org/10.1002/2017GL076669. Clim Dyn. 2014;42:1155–74. https://doi.org/10.1007/s00382- 115. Alessandri A and co-authors (2014) Robust assessment of the 013-1700-5. expansion and retreat of Mediterranean climate in the 21st century. 133. Nguyen H, et al. Expansion of the southern hemisphere Hadley Sci Rep 4: 7211 doi: https://doi.org/10.1038/srep07211. cell in response to greenhouse forcing. J Clim. 2015;28:8067–77. 116. Polade SD and co-authors (2017) Precipitation in a warming 134. Teissereng de Bort L (1883) Etude sur l’hiver de 1879–80 et world: assessing projected hydro-climate changes in California recherches sur la position des centres d’action de l’atmosphere Curr Clim Change Rep

dans les hivers anormaux. Bureau Central Meteor. de la France, 138. Dee DP, Uppala SM, Simmons AJ, Berrisford P, Poli P, Kobayashi Annales, 1881, 4, 17–62. S, et al. The ERA-interim re-analysis: configuration and perfor- 135. Stewart HJ. Periodic properties of semi-permanent atmospheric mance of the data assimilation system. Quart J Roy Meteor Soc. pressure systems. Q Appl Math. 1943;1:276–7. 2011;137:553–97. 136. Rubin MJ, van Loon H. Aspects of the circulation of the southern 139. Chen X, Zhou T. Distinct effects of global mean warming and hemisphere. J Meteor. 1954;11:68–76. regional sea surface warming pattern on projected uncertainty in 137. Chen PC, Hoerling MP, Dole RM. The origin of the subtropical the South Asian summer monsoon. Geophys Res Lett. 2015;42: anticyclones. J Atmos Sci. 2001;58:1827–35. 9433–9.