1. Introduction
The seasonal prediction of the Asian monsoon is one of the most challenging topics in short-term climate prediction. With the advance of multimodel ensemble forecasts, significant progress has been achieved (Krishnamurti et al. 2006; Yang et al. 2008). However, the progress has been mostly limited to the tropical region and southern midlatitudes because the predictability is rooted in the El Niño–Southern Oscillation (ENSO) (Wang et al. 2009), which does little to contribute to the predictability of the northern midlatitudes.
A most prominent feature of the East Asian winter monsoon (EAWM) is the Siberian high (SH), which has the highest sea level pressure in the world (Ding 1994; Chang et al. 2006, 2011). The SH has a distinct annual cycle, appearing in fall and disappearing in spring. The intensity of the SH has been used as an index to represent the EAWM strength (Gong and Ho 2004). The SH is not only linked to severe cold waves in northern Asia—including Siberia, Mongolia, northern China, Japan, and Korea (Ding and Krishnamurti 1987)—its movement is also closely associated with cold surges into the tropics, which are one of the most important mechanisms generating stormy weather in southern China, Indochina Peninsula, the Maritime Continent, and the Southern Hemisphere tropics (Chang et al. 2003, 2005; Chan and Li 2004). Thus, the prediction of SH is crucial for the seasonal forecasts of the EAWM for all latitudes. With little evidence of a significant relationship with ENSO, most research of SH variability has been concentrated in its relationship with the North Atlantic Oscillation (NAO)/Arctic Oscillation (AO) (Gong et al. 2001; Wu and Wang 2002; Park et al. 2011). The results suggest that a relationship may exist at the interdecadal scale but not the interannual scale.
2. Intraseasonal phase change
Recent research in intraseasonal oscillations (ISO), particularly the Madden–Julian oscillation (Wheeler and Hendon 2004), has suggested the possibility of improved forecast skill of the Asian monsoon due to the implied intraseasonal predictability. Since these ISOs are also mainly tropical phenomena, they offer hope for increased skill in forecasting tropical features, such as the onset of the summer monsoon, the large-scale tropical convection, and even tropical cyclogenesis (Goswami et al. 2011). There is as yet no evidence that suggests this promise can be extended to the northern latitudes (Kang and Kim 2011).
The seasonal forecasts made by all major operational centers are for rolling 3-month periods, which implies an expectation that the anomalies of a 3-month season are positively correlated with each of the three individual member months. To examine whether this is the case for the midlatitude EAWM, we used the 1979–2008 mean sea level pressure (MSLP) data from the NCEP–NCAR reanalysis (Kalnay et al. 1996) to examine running 3-month periods from October to February. The results were essentially unchanged using the National Centers for Environmental Prediction (NCEP)–Department of Energy (DOE) reanalysis that starts in 1979. Correlations of an SH index of daily MSLP anomalies averaged over a 3-month period with those averaged over each of its three member months for October–December (OND), November–January (NDJ), December–February (DJF), and January–March (JFM) are shown in Table 1. Here the SH area is defined as from 40° to 60°N, 70° to 120°E (Chang et al. 2006), and the MSLP anomalies are the departures from the 30-yr mean for each calendar day. All correlations remain virtually unchanged when detrended (not shown).
(top) Correlations between SH indexes of the individual months from October to March and each of the 3-month periods for the 1979–2008 winters. (bottom) Correlations among SH indexes of individual months from October to February for the 1979–2008 winters. DJ indicates the average of December and January time series. Boldface and italic numbers indicate significance at the 95% and 99% levels, respectively.
The NDJ period stands out as November is the only member month that is not significantly correlated with the 3-month values. Table 1 shows that the intercorrelation among individual months is generally very low, except for the correlations between November and the two following months. The November/December correlation is −0.30 (90% significance level), and the November–January correlation is −0.42 (98% significance level). When the December and January data are combined (DJ), the correlation coefficient becomes −0.51, which exceeds the 99% significance level. This negative correlation suggests that a positive SH anomaly in November tends to be followed by a negative SH anomaly in December/January and vice versa.
We next perform a singular value decomposition (SVD) analysis between November and the December/January MSLP to find the spatial structure of this negative relationship. The analysis covers 0°–60°N and 60°–150°E or most of Asia. The first two modes—SVD1 (Figs. 1a,b) and SVD2 (Figs. 1c,d)—are not appreciably changed when the domain is expanded to global. The SVD1 mode explains 67.9% of the covariance and has a largely in-phase pattern covering the entire tropics and most of the midlatitude domain that persists from November to December/January. The maximum amplitude is near the Philippines and the Maritime Continent. The correlation between the two coupled patterns is 0.68. This mode is clearly related to ENSO as the correlations with the DJF Niño-3.4 sea surface temperature (SST) is −0.60 for November and −0.75 for December/January.
SVD of SH between November and December/January for 1979–2008. The fractional covariance for (a),(b) SVD1 is 67.9% and for (c),(d) SVD2 is 19.3%. The correlation between the two subperiods in (a) and (b) is 0.68 for SVD1 and for the two subperiods in (c) and (d) is 0.59 for SVD2, both with 99% significance. The correlations with NIÑO-3.4 SST are given in the lower corner of each panel; red numbers indicate significance at 99%.
Citation: Journal of Climate 25, 5; 10.1175/JCLI-D-11-00500.1
The SVD2 mode explains 19.3% of the covariance and has a clearly out-of-phase pair of patterns with maximum amplitudes in the midlatitudes. This mode describes the negative correlation between November and December/January, with a correlation coefficient of 0.59 (exceeding 99% significance). The November pattern has maximum amplitude in central Asia with a meridional dipole whose northern center is between 45° and 50°N and southern center around 20°N. The main pattern covers the entire northern half of the domain, with a secondary center near Lake Baikal. The pattern appears to extend from this center eastward to northern Japan and southward to southern China. The December/January pattern resembles the winter mean pattern of cold air outbreaks with the SH situated along 50°N adjacent to Lake Baikal and migrating anticyclonically southward to bring cold surges to the southern China coast. The low correlations with the DJF Niño-3.4 SST indicate that this mode is not related to ENSO.
The phase reversal between November and December/January that is manifest by SVD2 is quite remarkable. It implies an intraseasonal predictability such that a useful forecast of the December/January EAWM can be made in late November based on the sign of the November anomalies.
3. Possible interdecadal changes
The cause of the phase reversal is unclear. The effect of snow cover has been a frequently mentioned parameter for forecasting seasonal variations of the Asian monsoon (Ding 1994; Ding and Ma 2007), but snow cover persists on a seasonal scale, so its role is a persistence effect that tends to contribute to a positive correlation between adjacent months. In addition to the lack of a correlation with ENSO, the correlations between SVD2 and NAO/AO are also weak.
The dataset used in computing the SVD2 starts in 1979, which is the year when the increase of global surface temperature becomes more evident (Trenberth et al. 2007). It also coincides with the beginning of the transition period of the interdecadal-scale phase change of NAO/AO from negative (cold) to positive (warm). To examine the EAWM evolution during the earlier decades, the November versus December/January SVD analysis was repeated for the period 1958–78 when the decadal NAO/AO phase was mostly negative. None of the resultant SVD modes exhibit a significant out-of-phase relationship (not shown). That the phase reversal mode was absent in the two earlier decades is also confirmed in the insignificant positive correlations between November, December, and January. During 1958–78 the correlation between November and DJ is 0.31 instead of −0.51 during 1979–2008. This interdecadal change is unlikely due to the known MSLP errors in the NCEP–NCAR reanalysis (Yang et al. 2002; Wu et al. 2005), as the 40-yr European Centre for Medium-Range Weather Forecasts (ECMWF) Re-Analysis data also show a positive November–DJ correlation (0.40) for 1958–78 and a negative November–DJ correlation (−0.49, 95% significance level) for 1979–2002.
In the decades since the late 1970s, the AO phase was positive and the strength of EAWM, as measured by the SH intensity, declined noticeably from the negative AO phase of the earlier decades. There was also a general decrease of higher frequency oscillations, particularly at the synoptic scale (Gong and Ho 2004), with a corresponding decrease in the occurrence of cold surges (Wang and Ding 2006; Park et al. 2011). Nevertheless, the intraseasonal variation of the SH remained significant. Takaya and Nakamura (2005a,b) pointed out that the SH is influenced by blocking patterns on both sides. To the east, retrogression of the Pacific blocking reinforced by the strong feedback from the Pacific storm track forces the SH. To the west, sustained forcing is associated with a quasi-stationary Rossby wave train propagating across the Eurasian continent, which is facilitated by Atlantic and Ural blockings. Since blocking activities are more active during negative AO when the circumpolar flow is less zonal (Thompson and Wallace 2001; Luo 2005), the SH may have experienced less frequent forcing during the recent decades than in the earlier decades. This leads us to propose a hypothesis to explain the intraseasonal phase reversal in the recent decades.
We will composite the blocking events based on the time series of the November and DJ SH index (Fig. 2). During the 30 winters of 1979–2008, 18 phase reversals are identified, in which 11 have a negative November followed by positive DJ, and 7 are opposite. This compares with the earlier decades of 21 winters (1958–78) when only 7 phase reversals (5:2) are identified. Figure 3a shows the composite of Pacific blocking during 1979–2008. For the November positive polarity, a strong Pacific blocking event tends to occur in November and followed by a lull that lasts several weeks. For the December positive polarity, a long period of suppressed blocking condition tends to start 2 months earlier and lasts at least 7 weeks. This behavior is reconfirmed in the calendar-based composite of Pacific blocking days per year (Fig. 3b), where the November positive polarity has higher values for most of early winter and lower values for most of late winter. Therefore, if a blocking event occurred in early winter, the tendency during the recent three decades is for the SH to be stronger in November and weaker in December/January. If not much blocking occurred in early winter, then the SH tended to be weaker and was followed by increased blocking activities in December/January and a stronger SH.
SH index for (a) November and (b) December/January from 1958 to 2008. Vertical dashed line indicates 1979—the beginning of the more recent multidecadal period. The horizontal lines indicate +0.5 and −0.5 standard deviation of the magnitude of the difference between November and December/January SH. Years with opposite polarity for the two multidecadal periods are listed at the bottom.
Citation: Journal of Climate 25, 5; 10.1175/JCLI-D-11-00500.1
Composite time series of blocking events for 1979–2008. (a) 90-day time series of composite blocking indices (Lu and Chang 2009; Tibaldi and Molteni 1990) averaged over the Pacific blocking longitudes (160°E–172.5°W). For the November positive polarity the data are plotted from 20 days prior to the maximum value of the blocking index GHGS (Lu and Chang 2009) in November until 69 days afterward. For the November negative polarity the data are plotted from 60 days prior to the maximum value in December until 29 days afterward. (b) Composite Pacific blocking days from 1 Oct to 24 Jan for the two polarities, and their fourth-order polynomial fit. A blocking day is defined when the blocking index in all five longitudinal points are nonzero.
Citation: Journal of Climate 25, 5; 10.1175/JCLI-D-11-00500.1
To compare the Pacific blocking activity of the recent decades (1979–2008) with the earlier decades (1958–78), we constructed a calendar-based composite of the modified (Lu and Chang 2009) Pacific blocking indices (Tibaldi and Molteni 1990) of the two periods from 1 October to 24 January based solely on the sign of the November anomaly. (The blockings were active in the last week of January for all cases.) During the earlier decades, the sign of the composite anomaly persisted through late January (Fig. 4a). During the recent decades, the sign changed within the same season (Fig. 4b), so a positive anomaly in November changed to negative after early December and vice versa.
Composite Pacific blocking index with respect to the sign of the November SH anomaly only, for (a) 1958–78 and (b) 1979–2008.
Citation: Journal of Climate 25, 5; 10.1175/JCLI-D-11-00500.1
Lu and Chang (2009) showed that the Atlantic blocking does not correlate well with the strength of SH except in extreme cases. The composites of the Atlantic blocking in 1979–2008 indeed do not show the pattern of intraseasonal and interdecadal changes indicated by the Pacific blocking. The blocking tends to be stronger from November to mid-December for the November negative–DJ positive polarity. After mid-December the blocking for the two polarities are similar (not shown). Therefore, a prolonged Atlantic blocking in early winter may lead to a stronger SH intensity in midwinter, but cannot explain the phase reversal.
Wang et al. (2010) reported that a pronounced relationship between the Ural blocking (30°–90°E) and the SH emerged after the mid-1970s, when the frequency of blocking decreased. Figure 5 shows the composites of the Ural blocking. During 1979–2008 the blocking index for the November positive polarity is stronger before mid-December and weaker afterward. During 1958–78 the blocking remains stronger during most of NDJ for the November positive composite. These tendencies are similar to those of the Pacific blocking and suggest a similar effect to force the SH phase reversal in the more recent decades. The Ural blocking region is close to the SH center, and the two features may be linked by the large westward vertical tilt in strong baroclinic systems. Wang et al. (2010) showed that the stronger Ural blocking–SH relationship after the mid-1970s may be the result of an eastward shift of the averaged blocking position from closer to Europe to closer to Asia.
As in Fig. 4 except for the Ural blocking index averaged over 30°–90°E (Wang et al. 2010).
Citation: Journal of Climate 25, 5; 10.1175/JCLI-D-11-00500.1
The above results suggest that the lower frequency of the Pacific and Ural blockings during the recent positive AO decades may contribute to an intraseasonal variation of the SH. A stronger forcing of blocking events in November may be followed by a period of inactive forcing in December/January, which then results in the phase reversal of the SH anomaly.
4. Concluding remarks
The phase reversal of SH is rather unique in the seasonal march of the atmosphere. We computed the correlations between MSLP of all adjacent months over the global domain during the last three decades (not shown) and found the SH to be the only large area with a significant out-of-phase feature. Thus, the EAWM may be the only known major circulation system whose intensity can be predicted to reverse from the previous month. This relationship calls into question the practice of making rolling 3-month forecasts of the EAWM for November–January, but raises the possibility that a forecast of November may be extended into December and January and thus extend the intraseasonal predictability. We present evidence that this intraseasonal phase reversal may be related to the less frequent Pacific and Ural blockings during a positive AO period. Therefore, this predictability may diminish if the phase of AO is reversed, but it may also become more prevalent if the prediction of more frequent positive AO-like patterns in a warming world forced by greenhouse gases is borne out (Hori et al. 2007; Wu et al. 2007; Choi et al. 2010).
Acknowledgments
This research was supported in part by the National Science Council of Taiwan, R.O.C. Ms. Yonghua Ji provided critical help in data processing.
REFERENCES
Chan, J. C. L., and C. Li, 2004: The East Asia winter monsoon. East Asian Monsoon, C.-P. Chang, Ed., World Scientific Series on Meteorology of East Asia, Vol. 2, World Scientific, 54–106.
Chang, C.-P., C.-H. Liu, and H.-C. Kuo, 2003: Typhoon Vamei: An equatorial tropical cyclone formation. Geophys. Res. Lett., 30, 1150, doi:10.1029/2002GL016365.
Chang, C.-P., P. A. Harr, and H.-J. Chen, 2005: Synoptic disturbances over the equatorial South China Sea and western Maritime Continent during boreal winter. Mon. Wea. Rev., 133, 489–503.
Chang, C.-P., Z. Wang, and H. Hendon, 2006: The Asian winter monsoon. The Asian Monsoon, B. Wang, Ed., Springer, 89–127.
Chang, C.-P., M.-M. Lu, and B. Wang, 2011: The East Asian winter monsoon. The Global Monsoon System: Research and Forecast, C.-P. Chang et al., Eds., World Scientific Series on Asia-Pacific Weather and Climate, Vol. 5, World Scientific, 99–109.
Choi, D.-H., J.-S. Kug, W.-T. Kwon, F.-F. Jin, H.-J. Baek, and S.-K. Min, 2010: Arctic Oscillation responses to greenhouse warming and role of synoptic eddy feedback. J. Geophys. Res., 115, D17103, doi:10.1029/2010JD014160.
Ding, Y. H., 1994: Monsoons over China. Atmospheric Sciences Library, Vol. 16, Kluwer Academic Publishers, 432 pp.
Ding, Y. H., and T. N. Krishnamurti, 1987: Heat budget of the Siberian high and the winter monsoon. Mon. Wea. Rev., 115, 2428–2449.
Ding, Y. H., and X. Ma, 2007: Analysis of isentropic potential vorticity for a strong cold wave in 2004/2005 winter. Acta Meteor. Sin., 65, 695–707.
Gong, D.-Y., and C.-H. Ho, 2004: Intra-seasonal variability of wintertime temperature over East Asia. Int. J. Climatol., 24, 131–144.
Gong, D.-Y., S.-W. Wang, and J.-H. Zhu, 2001: East Asian winter monsoon and Arctic Oscillation. Geophys. Res. Lett., 28, 2073–2076, doi:10.1029/2000GL012311.
Goswami, B. N., M. C. Wheeler, J. C. Gottschalck, and D. E. Waliser, 2011: Intraseasonal variability and forecasting: A review of recent research. The Global Monsoon System: Research and Forecast, C.-P. Chang et al., Eds., World Scientific Series on Asia-Pacific Weather and Climate, Vol. 5, World Scientific, 389–408.
Hori, M. E., D. Nohara, and H. L. Tanaka, 2007: Influence of Arctic Oscillation towards the Northern Hemisphere surface temperature variability under the global warming scenario. J. Meteor. Soc. Japan, 85, 847–859.
Kalnay, E., and Coauthors, 1996: The NCEP/NCAR 40-Year Reanalysis Project. Bull. Amer. Meteor. Soc., 77, 437–471.
Kang, I. S., and H. M. Kim, 2011: Intraseasonal prediction and predictability for boreal winter. The Global Monsoon System: Research and Forecast, C.-P. Chang et al., Eds., World Scientific Series on Asia-Pacific Weather and Climate, Vol. 5, World Scientific, 409–419.
Krishnamurti, T. N., A. K. Mitra, T. S. V. Vijaya Kumar, W. T. Yun, and W. K. Dewar, 2006: Seasonal climate forecasts of the South Asian monsoon using multiple coupled models. Tellus, 58A, 487–507, doi:10.1111/j.1600-0870.2006.00184.x.
Lu, M.-M., and C.-P. Chang, 2009: Unusual late-season cold surges during the 2005 Asian winter monsoon: Roles of Atlantic blocking and central Asian anticyclone. J. Climate, 22, 5205–5217.
Luo, D., 2005: Why is the North Atlantic block more frequent and long-lived during the negative NAO phase? Geophys. Res. Lett., 32, L20804, doi:10.1029/2005GL022927.
Park, T.-W., T. Won, C.-H. Ho, and S. Yang, 2011: Relationship between the Arctic Oscillation and cold surges over East Asia. J. Climate, 24, 68–83.
Takaya, K., and H. Nakamura, 2005a: Geographical dependence of upper-level blocking formation associated with intraseasonal amplification of the Siberian high. J. Atmos. Sci., 62, 4441–4449.
Takaya, K., and H. Nakamura, 2005b: Mechanisms of intraseasonal amplification of the cold Siberian high. J. Atmos. Sci., 62, 4423–4440.
Thompson, D. W. J., and J. M. Wallace, 2001: Regional climate impacts of the Northern Hemisphere annular mode. Science, 293, 85–89.
Tibaldi, S., and F. Molteni, 1990: On the operational predictability of blocking. Tellus, 42A, 343–365.
Trenberth, K. E., and Coauthors, 2007: Observations: Surface and atmospheric climate change. Climate Change 2007: The Physical Science Basis, S. Solomon et al., Eds., Cambridge University Press, 235–336.
Wang, B., and Coauthors, 2009: Advance and prospectus of seasonal prediction: Assessment of the APCC/CliPAS 14-model ensemble retrospective seasonal prediction (1980–2004). Climate Dyn., 33, 93–117, doi:10.1007/s00382-008-0460-0.
Wang, L., W. Chen, W. Zhou, J. C. L. Chan, D. Barriopedro, and R. Huang, 2010: Effect of the climate shift around mid 1970s on the relationship between wintertime Ural blocking circulation and East Asian climate. Int. J. Climatol., 30, 153–158, doi:10.1002/joc.1876.
Wang, Z.-Y., and Y.-H. Ding, 2006: Climate change of the cold wave frequency of China in the last 53 years and possible reasons. Adv. Atmos. Sci., 30, 1068–1076.
Wheeler, M. C., and H. H. Hendon, 2004: An all-season real-time multivariate MJO index: Development of an index for monitoring and prediction. Mon. Wea. Rev., 132, 1917–1932.
Wu, A., W. W. Hsieh, G. J. Boer, and F. W. Zwiers, 2007: Changes in the Arctic Oscillation under increased atmospheric greenhouse gases. Geophys. Res. Lett., 34, L12701, doi:10.1029/2007GL029344.
Wu, B., and J. Wang, 2002: Winter Arctic Oscillation, Siberian high and East Asian winter monsoon. Geophys. Res. Lett., 29, 1897, doi:10.1029/2002GL015373.
Wu, R., J. L. Kinter III, and B. P. Kirtman, 2005: Discrepancy of interdecadal changes in the Asian region among the NCEP–NCAR reanalysis, objective analyses, and observations. J. Climate, 18, 3048–3067.
Yang, S., K.-M. Lau, and K.-M. Kim, 2002: Variations of the East Asian jet stream and Asian–Pacific–American winter climate anomalies. J. Climate, 15, 306–325.
Yang, S., Z. Zhang, V. Kousky, R. W. Higgins, S.-H. Yoo, J. Liang, and Y. Fan, 2008: Simulations and seasonal prediction of the Asian summer monsoon in the NCEP Climate Forecast System. J. Climate, 21, 3755–3775.
