A tripole winter precipitation change pattern around the Tibetan Plateau in the late 1990s

Yali Zhu

a Institute of Atmospheric Physics, Chinese Academy of Sciences, Beijing, China

b Key Laboratory of Meteorological Disaster/Collaborative Innovation Center on Forecast and Evaluation of Meteorological Disasters, Nanjing University for Information Science and Technology, Nanjing, China

Keywords:Winter precipitation Tibetan plateau Interdecadal change East asian westerly jet stream Westerly—monsoon interaction

ABsTRACT Classical monsoon dynamics considers the winter/spring snow amount on the Tibetan Plateau (TP) as a major factor driving the East Asian summer monsoon (EASM) for its direct influence on the land—sea thermal contrast.Actually,the TP snow increased and decreased after the late 1970s and 1990s,respectively,accompanying the two major interdecadal changes in the EASM.Although studies have explored the possible mechanisms of the EASM interdecadal variations,and change in TP snow is considered as one of the major drivers,few studies have illustrated the underlying mechanisms of the interdecadal changes in the winter TP snow.This study reveals a tripole pattern of change,with decreased winter precipitation over the TP and an increase to its north and south after the late 1990s.Further analyses through numerical experiments demonstrate that the tropical Pacific SST changes in the late 1990s can robustly affect the winter TP precipitation through regulating the Walker and regional Hadley circulation.The cooling over the tropical central-eastern Pacific can enhance the Walker circulation cell over the Pacific and induce ascending motion anomalies over the Indo-Pacific region.These anomalies further drive descending motion anomalies over the TP and ascending motion anomalies to the north through regulating the regional Hadley circulation.Therefore,the positive—negative—positive winter precipitation anomalies around the TP are formed.This study improves the previously poor understanding of TP climate variation at interdecadal timescales.

1.Introduction

The Tibetan Plateau (TP) is often referred to as the “Third Pole ”due to it having the highest elevation in the world.It also has abundant glaciers,snow,lakes,and rivers,meaning it is also referred to as the“water tower ”of Asia for being home to the headwaters of many major rivers in Asia (Xu et al.,2008),including the Yangtze,Yellow,Ganges,and Indus.Therefore,the TP is critical for the supply of water resources in Asia.

The TP is one of the most sensitive regions to global change.Under global warming,its climate and phenology have undergone robust changes.Specifically,the TP has become warmer during the past few decades,at a rate of 0.46°C/10 yr during 1984—2009,which is obviously higher than the rate in the Northern Hemisphere (0.38°C/10 yr)and globally (0.32°C/10 yr) (Zhang et al.,2013 ;Kuang and Jiao,2016).Meanwhile,precipitation has increased over the TP in general,but with high spatial heterogeneity (Gao et al.,2014,2015 ;Yang et al.,2014 ;Han et al.,2017 ;Tang et al.,2020).There have also been significant changes in plant phenology over the TP.For instance,the start time of the growing season advanced during the 1980s and 1990s;whereas,during 2000—2011,it delayed over the southwest TP but advanced continuously over the northeast TP (Shen et al.,2015 ;Han et al.,2019).

The TP can modulate the climate in the Northern Hemisphere robustly through dynamic and thermal effects (Trenberth and Chen,1988 ;Manabe and Broccoli,1990 ;Xie et al.,2005 ;Zhou et al.,2009 ;Lu et al.,2016 ;Wang et al.,2020).Specifically,many previous studies have found that TP snow is closely linked with Asian summer climate at interannual and interdecadal time scales (Zhao and Moore,2004 ;Zhu et al.,2007 ;Xu and Li,2010 ;Ren et al.,2016 ;Wang et al.,2017 ;Li et al.,2019 ;Liu et al.,2020).Accompanying the two interdecadal changes in the East Asian summer monsoon (EASM) system in the late 1970s and 1990s,the TP snow amount exhibited an abrupt increase and decrease,respectively (Ding et al.,2009 ;Hu and Liang,2013 ;Han et al.,2017).After the late 1970s,winter/spring TP snow increased,surface albedo and the hydrological effect of snowmelt weakened the TP heat source,and the air temperature over the TP decreased.Meanwhile,the tropical centraleastern Pacific got warmer,the land—sea thermal contrast decreased,the EASM weakened,and the “northern flood and southern drought ”rainfall pattern established over East China (Wang,2001 ;Wu and Qian,2003 ;Zhang et al.,2004 ;Zhu et al.,2007 ;Ding et al.,2009).In the past decade,new evidence has revealed another interdecadal change during the late 1990s —in the summer precipitation pattern in East China,with increased/decreased rainfall over the Huang-Huai/Yangtze River valley(Zhu et al.,2011,2015a,2015b,2016 ;Si and Ding,2013 ;Ding et al.,2018).Meanwhile,the East Asian westerly jet stream (EAWJS) weakened and moved northwards.The tropical central-eastern Pacific cooling and TP warming due to decreased snow are prominent driving factors for this interdecadal change.

Although numerous studies have revealed the critical impact of TP snow on Asian climate (Ma et al.,2014 ;Wang et al.,2017),only a handful of studies have focused on how the TP snow is affected and modulated.At the interannual time scale,the Arctic Oscillation,North Atlantic Oscillation,Indian Ocean Dipole,Southern Annular Mode,Asian westerly jet stream,and ENSO can all influence TP winter snow accumulation (Lü et al.,2008 ;Xin et al.,2010 ;Yuan et al.,2012 ;Dou and Wu,2018 ;Jiang et al.,2019).However,at interdecadal time scales,our understanding of the driving factors and dynamic processes of TP snow variation is still very limited.

Previous studies have found decreased TP winter snowfall after the late 1990s,which dominantly contributed to the decreased winter snow accumulation (Hu and Liang,2013,2014).The present study explores the possible linkage between the winter TP precipitation and regional SST,and the physical processes related with the change in the late 1990s.Researching this issue will help to elucidate the underlying physical mechanisms of the interdecadal variation in the TP climate.Section 2 describes the data and gives the details of the model experiment.Section 3 illustrates the features of winter precipitation and circulation changes in the late 1990s.The contributions of the regional(tropical) and global SST are presented in section 4.And finally,a summary and some further discussion are given in section 5.

2.Data and model experiment

Three sets of precipitation data are used to examine the winter precipitation change in the late 1990s over Asia: CRU TS (Harris et al.,2020),APHRODITE (Yatagai et al.,2012),and GPCP (Adler et al.,2003),provided by NOAA/OAR/ESRL PSL from their website at https://psl.noaa.gov/.The reanalysis data are from ERA-Interim,available from https://www.ecmwf.int/.SST data from the Met Office Hadley Centre are also used (Kennedy et al.,2011).

Several sets of experiments were carried out using an atmospheric general circulation model (AGCM) —namely,the Community Atmospheric Model,version 4 (CAM4).CAM4 has been proven to have reasonable capability in simulating Asian circulation,albeit with systematic errors (Neale et al.,2013),and has been widely used in recent Asian climate research (e.g.,Zhu et al.,2015a ;Ma et al.,2022).The control simulation was forced by the climatological mean global SST during 1979—2017.Five sensitivity simulations were performed,driven by the global (EXPG),tropical (EXPT,30°S—30°N),tropical Indian Ocean(EXP TIO,30°S—30°N,30°—110°E),tropical Pacific (EXP TP,30°S—30°N,110°E—70°W),and tropical Atlantic (EXPTA,30°S—30°N,70°W—30°E) SST difference between 2000—2017 and 1979—1998,added to the climatological mean global SST,respectively.Each experiment was integrated for 30 model years,and the last 20 years of data were analyzed.

Fig.1.(a) The winter precipitation difference percentage (relative to 1979—2017,units: %) between 2000—2017 and 1979—1998 in GPCP.(b—d) Time series of the winter precipitation (units: mm d —1) in the three bands covering 80°—120°E in 35°—45°N (NR),20°—35°N (CR),and 0°—20°N (SR).The values significant at the 0.1 level are shown as dotted areas.The three regions are denoted as blue boxes in (a).

3.Winter precipitation and circulation changes in the late 1990s

Previous studies have revealed a decreased winter snow amount and snowfall after the late 1990s (Hu and Liang,2013,2014).Our study presents a sandwich-like/tripole pattern of changes in the winter precipitation around the TP covering about 80°—120°E (Fig.1).Three bands,with positive—negative—positive values,appear over 0°—20°N,20°—35°N,and 35°—45°N,respectively.The three regions are thus named as the southern (SR),central (CR),and northern region (NR),correspondingly.SR includes southeastern Asia and the adjacent ocean.CR locates over the southern TP,India,and southern China.NR covers northwestern China and the northern edge of the TP.The increase/decrease in precipitation can reach more than 50% of the climatological mean value.The time series of the three regional means of winter precipitation also exhibit obvious changes after the late 1990s (Fig.1 (b—d)).A movingt-test (through a window width of 11 years) was applied to test the point of shift in the three precipitation indices (Fig.S1).The points of shift in the SR,CR,and NR precipitation indices were found to be 1996—1998,1998,and 2000—2002,respectively.Therefore,the two periods of 1979—1998 and 2000—2017 were selected as a reasonable compromise among the three regions.The two centers over CR and NR of this tripole pattern partly resemble the moisture dipole over the northern and southern TP,as revealed by five-century tree-ring chronologies (Zhang et al.,2015).Results using other precipitation data such as CRU (Fig.S2) and APHRODITE (not shown) present a similar pattern and magnitude as those from GPCP.

Fig.2.Difference in the Walker circulation averaged over 20°S—20°N (a) between 2000—2017 and 1979—1998 in the reanalysis data and (b—f) between the sensitivity and control experiments.The values significant at the 0.1 level are shown as green dotted areas.The contours are the zonal wind difference,with a contour interval of 1 m s -1.Blue,black,and red contours denote negative,zero,and positive values,respectively.Arrows denote the zonal—vertical wind difference.The maximum vertical velocity in (a—f) is 22.1,8.6,7.1,13.1,12.2,and 6.9 ×10 -3 Pa s -1,respectively.

In addition,a large-scale rainfall deficit occurs over the eastern tropical Pacific (ETP).Such negative rainfall anomalies correspond to a cooler SST,trade wind acceleration,and anomalous descending motion there (England et al.,2014).The global SST difference also exhibits robust changes in the late 1990s (Fig.S3),as revealed by previous studies(Zhu et al.,2011,2015a).The Pacific SST changes show a La Ni?a-like pattern,with cooler SST over the ETP and warmer SST over the western tropical Pacific (WTP).Besides,the Indian Ocean gets warmer and the North Atlantic much warmer.

The cooler/warmer SST over the ETP/WTP accompanies strengthened Walker circulation in the Pacific Ocean (Fig.2 (a)).In the Walker circulation averaged over 20°S—20°N,descending motion anomalies happen over the ETP (120°—180°W),and ascending motion anomalies over the Indo-Pacific region (80°—150°E).The significant anomalies over the Indo-Pacific region can further affect the regional Hadley circulation.In the regional Hadley circulation averaged over 80°—120°E,ascending,descending,and ascending motion anomalies are observed from south to north,covering three zonal bands at about 0°—15°N,20°—35°N,and 35°—45°N,respectively (Fig.3 (a)).These ascending—descending—ascending motion anomalies in the vertical direction present signals consistent with the positive—negative—positive precipitation changes in Fig.1.

On the other hand,the EAWJS,a major climate system controlling winter climate in Asia,also shows robust changes (Fig.4 (a)).The core of the EAWJS usually locates over (30°—35°N,120°—160°E) in winter.After the late 1990s,in the upper-level wind,a northward wave train pattern exists from the ETP to the North Pacific (Fig.4 (a)).A dipole pattern appears to the north and south of 30°N over East Asia in the zonal wind difference at the 200 hPa level (Fig.4 (a)),which consists of an anomalous anticyclone over Asia.The westerly and easterly anomalies to the north and south of 30°N indicate a northwestward shift of the EAWJS.Changes in the intensity and location of the EAWJS can exert significant influence on Asian climate (Huang et al.,2015 ;Wu and Sun,2017).

Significant changes also happened in the horizontal wind and moisture flux (Fig.S4).In the upper level (Fig.S4(a)),robust large-scale anticyclonic anomalies appear over southern Asia,with westerly anomalies over the northern TP and NR,indicating a weakened EAWJS.Significant westerlies and easterlies happen over the tropical Pacific and Indian Ocean,representing converging and diverging anomalous flow over the central-eastern tropical Pacific and Indo-Pacific region,respectively.The upper-level anomalous convergence and divergence correspond to the descending and ascending motion anomalies over the ETP and Indo-Pacific region (Fig.S4(a) and Fig.2 (a)).The difference in horizontal wind at 500 hPa is shown considering the topography of the TP (Fig.S4(b)),which resembles the upper-level wind but with much weaker amplitude.The low-level wind difference resembles the vertically integrated water vapor flux difference (Fig.S4(c,d)),with the largest water vapor flux anomalies converging over SR,which has the largest rainfall changes among the three regions.

4.Regional (tropical) and global SST contributions

As mentioned above,the global SST has shown obvious changes since the late 1990s.But could these SST anomalies have influenced the winter precipitation and circulation change pattern over Asia in the late 1990s? To answer this question,we compare the control and sensitivity experiment.

Fig.3.Difference in the regional Hadley circulation averaged over 80°—120°E (a) between 2000—2017 and 1979—1998 in the reanalysis data and (b—f) between the sensitivity and control experiments.The values significant at the 0.1 level are shown as green dotted areas.The magenta lines denote the profile of TP.The contours are the zonal wind difference,with a contour interval of 1 m s -1.Blue,black,and red contours denote negative,zero,and positive values,respectively.Arrows denote the zonal—vertical wind difference.The maximum vertical velocity in (a—f) is 15.6,1.37,1.26,1.87,2.2,and 1.21 ×10 -3 Pa s -1,respectively.

The respective ascending and descending motion anomalies over the Indo-Pacific and ETP,which represents strengthened Walker circulation,can be reproduced in EXP TP (Fig.2 (e)),rather than EXP G and EXP T.Anomalous descending motion over the ETP also appears in EXPTA(Fig.2 (f)).However,both EXP TA and EXP TIO cannot capture the anomalous ascending motion over the Indo-Pacific region (Fig.2 (d,f)).In the regional Hadley circulation,the respective ascending and descending anomalies over SR and CR can be seen in EXPTP(Fig.3 (e)),while neither signal emerges in the other experiments (Fig.3 (b,c,d,f)).The failure of the experiments other than EXPTPprobably suggests a robust influence of the tropical Pacific SST,rather than SST forcing in other regions.It should also be noted that although EXPTPcan reproduce the vertical motion changes over SR and CR,the ascending motion anomalies over NR are not shown in EXP TP.One possible reason is that the winter precipitation anomalies over NR are probably more affected by factors from middle and high latitudes (e.g.,sea ice) rather than the SST anomalies.

For the upper-level zonal wind,EXP TIO and EXP TA (Fig.4 (d,f)) can cause a dipole pattern to the southwest of TP,which locates mainly west of 80°E and south of 40°N.Although this dipole pattern somewhat resembles the observed one (Fig.4 (a)),its location is much more to the west than observed,and the zero line has moved southwards to about 20°N compared to the observed one at 30°N.The wave-train pattern over the Pacific Ocean cannot be correctly reproduced in EXPTIOor EXP TA (Fig.4 (d,f)) either.EXP TP induces a more eastward dipole pattern over Asia and a wave-train pattern over the Pacific with similar location but weaker magnitude (Fig.4 (e)) compared to the observation.The combined effect of the tropical Pacific,Indian,and Atlantic SST (EXPT;Fig.4 (c)) induces a pattern more similar to the observation(Fig.4 (a)).However,the global SST only results in a less consistent pattern (EXP G ;Fig.4 (b)),which may suggest a secondary forcing effect of the extratropical SST on the circulation pattern in the late 1990s’ interdecadal change.

It should be noted that the SST anomaly—forced responses are much weaker than in the reanalysis data,which is a common failing of AGCMs due to the lack of air—sea interaction processes (Mcgregor et al.,2014).

5.Summary and discussion

The recent strengthening of the Walker circulation has been shown to be directly linked with tropical central-eastern Pacific cooling and amplified by Atlantic warming (Mcgregor et al.,2014).However,no previous study has explored the possible linkage between the Pacific SST/Walker circulation and the winter precipitation over the TP.This study found a tripole winter precipitation change pattern around the TP during the late 1990s.The tripole pattern is suggested to result from the cooperation of the Walker and regional Hadley circulation changes.After the late 1990s,the equatorial Pacific exhibits cooling over the central-eastern part,and warming over the western part.The related Walker circulation shows descending motion anomalies over the ETP and ascending motion over the Indo-Pacific region.The ascending motion anomalies over the Indo-Pacific region can regulate the regional Hadley circulation over the zonal band of 80°—120°E and induce descending motion anomalies over the southern TP and northern India(Fig.5).

Fig.4.Difference in the 200 hPa zonal wind (a) between 2000—2017 and 1979—1998 in the observation and (b—f) between the sensitivity and control experiments.The red and blue lines denote the main body of the EAWJS before and after the late 1990s.The solid and dashed lines represent the westerly and easterly anomalies,respectively.The shading shows significant values at the 0.1 level.

Fig.5.Schematic diagram illustrating the possible mechanism linking the tropical Pacific SST and Asian winter precipitation in the late 1990s.The two arrows forming an oval represent the upper-level westerly and easterly anomalies to the north and south of the EAWJS.

The ascending motion anomalies over the north of the TP seem to relate more closely with the EAWJS.Zhang et al.(2015) suggested that the moisture dipole over the TP is very likely to arise from stochastic,unforced interaction between the westerlies and monsoon system rather than from a steady dominant mechanism.However,the detailed processes involved in this interaction and the underlying mechanisms remain unclear and still need more in-depth exploration.Besides,the dipole zonal wind anomalies to the north and south of the TP imply a northwestward shifting of the EAWJS in the late 1990s.This change in the EAWJS is associated with the increased/decreased precipitation over NR/CR,which can induce a strengthened/weakened meridional temperature gradient and westerly/easterly anomalies over NR/CR.Highlatitude climate systems,such as the Arctic sea ice and related atmospheric circulation,may also influence the EAWJS and the precipitation over NR/CR (e.g.,Wu et al.,2016).

The connection between the tropical Pacific SST and winter Asian precipitation changes in the late 1990s are unlikely to have been “oneway ”.The precipitation change could also have stimulated circulation anomalies and feedback to the SST.Positive feedback may exist between the equatorial Pacific SST and winter Asian precipitation to help sustain such decadal climate phases.In addition,the late 1990s’ interdecadal change of the East Asian summer rainfall (e.g.,Zhu et al.,2011,2015a)seems to appear as just a small part of the large-scale rainfall change pattern,which requires more in-depth investigation for the summer season.Besides,the impact of anthropogenic forcing,including greenhouse gases and aerosols,on the TP precipitation change needs further exploration (Qiu,2008).

Funding

This study was jointly supported by the Second Tibetan Plateau Scientific Expedition and Research (STEP) program [grant number 2019QZKK0102] and the National Natural Science Foundation of China[grant numbers 41675083 and 41991281].

Declaration of Competing Interest

The author declares no conflict of interest.

Acknowledgements

The author would like to thank Dr.Jun Wang and Dr.Jiehua Ma for their help in tuning the AGCM.

Supplementary materials

Supplementary material associated with this article can be found,in the online version,at doi: 10.1016/j.aosl.2022.100223.