INFLUENCE OF EURASIAN SPRING SNOW COVER ON ASIAN SUMMER RAINFALL

Transcription

1 INTERNATIONAL JOURNAL OF CLIMATOLOGY Int. J. Climatol. 22: (2002) Published online in Wiley InterScience ( DOI: /joc.784 INFLUENCE OF EURASIAN SPRING SNOW COVER ON ASIAN SUMMER RAINFALL XIAODONG LIU a and MICHIO YANAI b, * a Institute of Earth Environment, Chinese Academy of Sciences, Xi an, People s Republic of China b Department of Atmospheric Sciences, University of California, Los Angeles, USA Received 20 June 2001 Revised 21 January 2002 Accepted 23 January 2002 ABSTRACT The Eurasian snow cover anomaly in spring has been considered as one of the important factors affecting Asian summer monsoon variability. Using the long time series ( ) of Eurasian spring (March April) snow cover (ESSC) reconstructed by Brown (2000. Journal of Climate 13: 2339) and snow cover ( ) and depth ( ) data from satellite observation, the influences of ESSC on the all-india monsoon (June September) rainfall (AIMR) and the summer rainfall over all parts of Asia are examined. It is found that the statistical relation between AIMR and ESSC changes over a multi-decadal time scale. The negative correlation between them has increased markedly since the mid 1970s. The region where the summer rainfall has the strongest and most stable negative correlation with the preceding ESSC is located in northern Mongolia, south of Lake Baikal. The correlation between the summer rainfall and ESSC increases after the data are treated with a low-pass filter, showing that the impact of snow cover may be seen more clearly with the removal of the effect of El Niño southern oscillation. Comparative analyses for contrasting years with excessive and deficient snow cover show that the anomalies of ESSC occur mainly in northwestern Eurasia. In the years of excessive ESSC anomalies, cooling and a cyclonic circulation anomaly in the lower troposphere appear over the northern part of Eurasia, leading to a Rossby-wave-train-like circulation response, then a weakened East Asia summer monsoon and deficient rainfall with an anticyclonic circulation anomaly south of Lake Baikal. Anomalies with opposite signs occur in the years of deficient snow cover. Copyright 2002 Royal Meteorological Society. KEY WORDS: Eurasian snow cover; Asian summer rainfall; Asian monsoon; correlation 1. INTRODUCTION The Eurasian winter/spring snow cover has been considered an important factor influencing the land thermal conditions and then seasonal-to-interannualvariabilities of the succeeding Asian summer monsoon and rainfall. Observational studies using satellite-measured snow cover or depth data (Hahn and Shukla, 1976; Kripalani et al., 1996; Sankar-Rao et al., 1996; Bamzai and Shukla, 1999) and historical snow depth data (Kripalani and Kulkarni, 1999) showed that there generally exists an inverse relationship between the year-to-year variations of snow cover or depth over Eurasia and the Indian monsoon rainfall, indicating that excessive (deficient) Eurasian snow cover in winter/spring is followed by weak (strong) Indian monsoon rainfall. Moreover, the winter snow anomaly of western Eurasia seems to have the strongest inverse correlation with the Indian summer rainfall (Bamzai and Shukla, 1999; Kripalani and Kulkarni, 1999). The linkage between Eurasian snow cover and summer rainfall over China was shown to be more complicated (Yang and Xu, 1994). Additionally, the snow monsoon relation is influenced by El Niño southern oscillation (ENSO) events (Sankar-Rao et al., 1996; Yang, 1996). * Correspondence to: Michio Yanai, Department of Atmospheric Sciences, University of California, Los Angeles, 405 Hilgard Avenue, Los Angeles, CA , USA; yanai@atmos.ucla.edu Copyright 2002 Royal Meteorological Society

2 1076 X. LIU AND M. YANAI A number of numerical experiments based on general circulation models (GCMs) (e.g. Barnett et al., 1989; Yasunari et al., 1991; Ose, 1996; Ferranti and Molteni, 1999) have been carried out to simulate the inverse snow monsoon relation found in the observational analyses. Most of the modelling results confirmed the statistical relationship between the snow and the monsoon. The effect of seasonal snow on the summer monsoon circulation is generally regarded as a result of two dominant feedback mechanisms (Shukla, 1987; Barnett et al., 1989; Yasunari et al., 1991; Meehl, 1994). Persistent Eurasian snow anomalies in winter and spring may induce changes in the surface and atmospheric temperatures with the surface albedo and soil hydrological effects, leading to variations in the large-scale land sea thermal contrast and the following summer monsoon. Despite the numerous studies on the relation between the Eurasian winter/spring snow anomaly and Asian summer rainfall, we still find some aspects of the relation worth investigating further. There is relatively little documentation of snow rainfall statistics on a multi-decadal time scale owing to a lack of comprehensive data. Most of the published observational studies, except Kripalani and Kulkarni (1999), were limited to the records of the recent 20 years or so for which the snow data from satellites are available. Moreover, these studies were mainly concerned with the relation of the winter/spring snow anomaly to the Indian or Chinese monsoon rainfall, without examining its possible relation with the summer rainfall over other parts of the Asian continent. In addition, most of the previous observation studies, except Sankar-Rao et al. (1996), focused on the effect of snow anomaly on the monsoon rainfall, and rarely examined the effect on the large-scale atmospheric circulation. In this paper we shall use a long time series ( ) of the Eurasian spring snow cover reconstructed by Brown (2000) and recent snow cover ( ) and depth ( ) data from satellite observations to examine further the influences of the snow anomaly on the Asian summer rainfall, atmospheric temperature and circulation over Eurasia. The data used in this study are described in Section 2. The correlation analyses on multi-decadal scales and case studies for years of excessive and deficient snow cover are detailed in Sections 3 and 4 respectively. Finally, in Section 5, we summarize the results and discuss briefly the importance, as well as limitation, of the snow anomaly in modulating the atmospheric circulation and rainfall variabilities. 2. DATA The monthly land precipitation data on a grid (Chen et al., 2002) are used as a basic dataset of precipitation for this paper. This gridded field of monthly precipitation is defined by interpolating gauge observations at over stations in the world using the optimum interpolation algorithm. Additionally, the all-india monsoon (June September) rainfall (AIMR) series for is also used. The gauge-based AIMR was prepared by the Indian Institute of Tropical Meteorology (Parthasarathy et al., 1987; Parthasarathy et al., 1994). The monthly mean tropospheric temperature and wind data at several levels with grid are extracted from the National Centers for Environment Prediction (NCEP) National Center for Atmospheric Research (NCAR) reanalysis dataset (Kalnay et al., 1996). We use these data to show possible impacts of the Eurasian snow cover anomaly on the temperature, circulation and, eventually, rainfall. Two time series related to the underlying surface forcing are employed: the June September Nino3 (5 N 5 S, W) sea surface temperature anomaly (SSTA) series is taken from Latif (2000); and a time series of the Eurasian land surface temperature anomaly (ELTA) averaged for March and April is calculated with monthly land surface air temperature on a 5 5 grid provided by the Climatic Research Unit, University of East Anglia, UK (Jones, 1994). Here, both SSTA and ELTA are calculated with as the base period. Three types of snow data are used in this study. The first is the monthly snow cover obtained from the National Oceanic and Atmospheric Administration (NOAA) satellites. The original data were described in Robinson et al. (1993). The satellite-measured snow-cover data used in the present paper are on a 2 2 Northern Hemisphere grid from 1973 to The second is the monthly 1 1 snow depth obtained by the National Aeronautics and Space Administration (NASA) with the Scanning Multi-channel

3 SNOW COVER AND ASIAN RAINFALL ESSC (Left Y axis, 10 6 km 2 ) ELTA (Right Y axis, C) Figure 1. Year-to-year variations of the reconstructed ESSC anomaly and ELTA, averaged for March and April Microwave Radiometer (SMMR) on the Nimbus-7 satellite (Chang et al., 1992). The third is the reconstructed time series of the Eurasian spring (March April, the same in the following) snow cover (ESSC) for (Brown, 2000). This series of ESSC was derived from a snow-cover area index ( ) reconstructed with available in situ snow-depth data over western Eurasia and satellite-observed snow cover ( ) over the entire Eurasian continent. The method for reconstructing ESSC series was detailed in Brown (2000). The latest year of the original series was In this paper the value of 1998 is constructed from the regression of the original series with the satellite snow-cover data. Because of the poor coverage of in situ snow-depth data over eastern Eurasia, Brown (2000) reconstructed the ESSC before 1972 by using only a snow-cover area index for western Eurasia. It was found, however, using the satellite observations, that the variation of the snow-cover extent over western Eurasia is closely correlated with that over the entire continent, especially in March and April. Thus, the time series of reconstructed ESSC is considered sufficiently accurate for climate studies in reflecting fluctuations of the snow cover on a continental scale (Brown, 2000). Here, we can further confirm the reliability of the reconstructed ESSC data. By comparing the reconstructed ESSC anomaly with ELTA, the simultaneous land temperature for the entire Eurasian continent (Figure 1), we find a significant and stable inverse relationship between them, with a correlation coefficient of 0.67 for 77 years. This suggests that the reconstructed ESSC series is reliable even in the early years. 3. CORRELATION ANALYSES 3.1. Interdecadal change of the relation of Indian monsoon rainfall to the Eurasian spring snow cover Although the relationship between ESSC and AIMR has been investigated in previous studies, it has never been examined on a multi-decadal time scale. Here, we shall use the data from 1922 to 1998 to re-examine their relationship. Figure 2(a) shows year-to-year variations of AIMR and ESSC. For the sake of comparison, variations of AIMR and SSTA are shown in Figure 2(b). Note that the SSTA here means an average for June September, the same months as the AIMR uses. Thus we consider only the simultaneous relation between AIMR and SSTA here, because the AIMR has been found to be correlated with the simultaneous and time-lagged ENSO indices rather than time-lead ENSO indices (Elliott and Angell, 1987; Prasad and Singh, 1996). As shown in Figure 2, AIMR has a significant negative correlation with the simultaneous SSTA (correlation coefficient 0.51), but it has no clear correlation with the preceding ESSC ( 0.08) for

4 1078 X. LIU AND M. YANAI (a) AIMR (left Y axis) 19 ESSC (right Y axis) (mm) ( C) (b) AIMR (left Y axis) SSTA (right Y axis) Figure 2. (a) Year-to-year variations of the AIMR and ESSC. The thick solid lines represent the trends fitted with sixth-order polynomials. (b) Same as (a) but for AIMR and Nino3 summer (June September) SSTA the 77 year period. It can also be seen from Figure 2 that there exists a decreasing trend in ESSC, especially from the early 1970s. The shrinking ESSC may be related to the global warming. At the same time, SSTA shows a weak increasing trend. However, no clear trend is observed in AIMR during the past decades. It is well known that the Asian monsoon is a product of the land sea thermal contrast (e.g. Li and Yanai, 1996; Webster et al., 1998). The decreasing ESSC may intensify the heating from the Eurasian land surface, leading to strengthening of the Asian summer monsoon by increasing the thermal contrast between land and sea, whereas the increasing SSTA may restrain development of the monsoon by weakening the Walker circulation. On this basis, we speculate that no change in the trend of AIMR may be a result from the offsetting effects of the land and sea forcings on the Asian monsoon. The lack of significant correlation between AIMR and ESSC on the longer time scale is in sharp contrast to the snow monsoon inverse relation found in many previous studies for relatively short periods of recent years (e.g. Sankar-Rao et al., 1996; Bamzai and Shukla, 1999). After excluding possible unreliability of the reconstructed ESSC data in early years (Section 2), the lack of a clear relationship over the 77 years is likely attributed to the non-stationary nature of the correlation between AIMR and ESSC. It is found that the 11 year moving window correlations of AIMR with ESSC were generally weak or even positive before the 1970s, but became strongly negative after 1975 (not shown). This means that the relationship between ESSC and AIMR

5 SNOW COVER AND ASIAN RAINFALL 1079 has undergone considerable interdecadal changes during the past 77 years, and that the inverse relation between the snow and the monsoon has intensified over the recent 20 years or so. The interdecadal change between ESSC and AIMR has also been noted by Kripalani and Kulkarni (1999). On the other hand, 11 year moving window correlations between AIMR and SSTA (not shown) indicate that the strong negative correlation between them has been interrupted since the late 1970s, as several investigators (Kumar et al., 1999; Miyakoda et al., 2000) have pointed out. Because of no long-term trend in AIMR and only a weakly increasing trend in SSTA, as shown in Figure 2, we speculate that the clear descending trend in ESSC with the recent global warming since the early 1970s may have made the Asian summer monsoon increasingly sensitive to the change in ESSC and eventually brought interdecadal changes of AIMR ESSC and even AIMR ENSO relationships Correlation between interannual variations of the Eurasian spring snow cover and summer rainfall over Asia In this section we shall examine the relation of ESSC with the summer (June September, the same in the following) rainfall everywhere over the Asian continent, including the Indian subcontinent. Figure 3(a) shows the correlation coefficients between ESSC and the summer rainfall at every grid-box for The correlation was not strong for the past half century as a whole. However, there is a systematic northeast southwest band with negative correlation from northeast China to northwest India. The strongest negative correlation (less than 0.2) is located in northern Mongolia, south of Lake Baikal. In contrast, summer rainfall over Asia, especially in north India and north China, shows much stronger negative correlation with the simultaneous SSTA (Figure 3(b)). When we divide the 49 years under discussion into two stages (Figure 4(a)) and (Figure 4(b)), it is found that the ESSC rainfall correlation pattern has changed and the overall correlation has intensified in the later stage (Figure 4(b)), especially in northwest India. At the same time, northern Mongolia is a region with relatively stable negative correlation, and the correlation there has also been strengthened in the later stage. Another interesting finding is that the correlation of the Asian summer rainfall with ESSC becomes stronger when the data of both the rainfall and ESSC are treated with a low-pass filter to retain only those variations with periods longer than 7 years (thus excluding the components of ENSO related to SSTA). In the patterns of the correlation between the low-pass filtered components of the rainfall and ESSC (not shown), it is seen Figure 3. Distributions of the correlations of the grid-box summer rainfall (a) with the ESSC and (b) with the Nino3 SSTA for The areas of negative correlation are shaded

6 1080 X. LIU AND M. YANAI Figure 4. Distributions of the correlations between the grid-box summer rainfall and the ESSC (a) for , and (b) for The areas of negative correlation are shaded ESSC (Left Y axis, 10 6 km 2 ) Rainfall (Right Y axis, mm) Figure 5. Year-to-year variations of the low-pass-filtered ESSC and summer rainfall rate averaged for a box ( N, E) over northern Mongolia that the negative correlations near Lake Baikal, with central values less than 0.6 for both stages, are much stronger than those before the filtering. This means that a closer relation between the summer rainfall and ESSC may exist at a time scale longer than that of ENSO. As seen in Figure 5, year-to-year variations of the low-pass filtered ESSC and the rainfall averaged for a box ( N, E) over northern Mongolia show a distinct inverse relationship. During the last 49 years, there are three major out-of-phase stages: the middle 1950s, late 1970s early 1980s and late 1980s early 1990s. Excluding the first stage, when satellite data were not available, we will choose 3 years each in the second stage ( ) and the third stage ( ) to analyse the characteristics and

7 SNOW COVER AND ASIAN RAINFALL 1081 impacts of ESSC. It should be pointed out that these 6 years are those without large anomalies in SSTA (see Figure 2(b)). This is important because the Asian monsoon variability is so closely linked to ENSO that erroneous conclusions may be drawn in exploring the impact and response of the Asian monsoon without excluding the effect of ENSO (Goswami and Jayavelu, 2001). 4. CASE STUDIES 4.1. Some characteristics of the Eurasian snow cover anomalies for typical years In order to show the annual cycle and persistence characteristics of the Eurasian snow cover, we examine monthly snow-cover anomalies (relative to ) from satellite observations since 1973 and find that the 3 years of ( ) are typical with extensive (deficient) spring snow cover (Figure 6). For example, the snow-cover anomaly averaged for April of 1979, 1980 and 1981 reaches km 2, but that for March of 1989, 1990 and 1991 is km 2. Moreover, the strong snow-cover anomalies show good persistence, maintaining the same sign at least until June. The climatological ( ) distribution of spring (March and April) snow depth on the Eurasian continent (Figure 7) shows that the seasonal snow mainly covers northern Eurasia north of about 45 N and the Tibetan Snow cover anomaly (10 6 km 2 ) (a) Mean -4.0 Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec Snow cover anomaly (10 6 km 2 ) (b) Mean -4.0 Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec Figure 6. Annual march of monthly snow cover anomalies for (a) three heavy-snow years ( ) and (b) three light-snow years ( ). The thick grey lines indicate the mean for the 3 years

8 1082 X. LIU AND M. YANAI Figure 7. Distribution of snow depth averaged for March April from 1979 to 1987 (units: cm) Figure 8. Difference in the March April snow-cover frequency (the average for three heavy-snow years minus the average for three light-snow years). Only the values greater than 1 week are shown Plateau. An area with snow depth more than 40 cm is found in the northeastern part of the continent. This pattern compares with that from historical snow-depth data (Kripalani and Kulkarni, 1999). However, the difference in snow-cover area between the extensive and deficient snow years occurs mainly in the northwestern part of the continent. For example, Europe experienced more snow cover in the first week of March 1979 than in the first week of March 1990 (not shown). Figure 8 shows the difference of spring snow-cover frequency (total number of weeks with snow cover during March and April) averaged for and for The snow cover in the heavy snow years lasted 2 weeks or longer than in the light snow years in eastern Europe, eastern Tibetan Plateau and northeast China. Thus, the snow-cover anomaly in northwestern Eurasia appears

9 SNOW COVER AND ASIAN RAINFALL 1083 to be a major factor in accounting for the interannual variability of ESSC. This may be a reason why the reconstructed ESSC series, even though only the snow cover data on western Eurasia were used for its reconstruction in the early years (Brown, 2000), can approximate the interannual variation of the whole Eurasian snow cover Impacts of the Eurasian snow cover anomalies on the atmospheric temperature, circulation and rainfall Previous observational (Sankar-Rao et al., 1996) and modelling (Walland and Simmonds, 1997) studies showed that there is a significant reduction in the lower tropospheric temperature with an increase in snow cover. Figure 9(a) (c) shows differences between the air temperatures averaged for March April of and those of at three levels. At 850 hpa, a centre of large negative values lower than 4 C appears Figure 9. The mean temperature differences (the average for three heavy-snow years minus the average for three light-snow years) for (a) March April at 500 hpa, (b) March April at 700 hpa, (c) March April 850 hpa, (d) May June at 500 hpa, (e) May June at 700 hpa and (f) May June at 850 hpa (units: C). The black areas indicate the Tibetan Plateau above the respective isobaric surfaces

10 1084 X. LIU AND M. YANAI over northern Europe and a band of negative values is present over the whole of northern Eurasia extending from Europe to Asia. The magnitude of temperature difference values is reduced with height. For example, the value at the centre over northern Europe is about 3 C at 700 hpa and 2 C at 500 hpa. This decreasing temperature difference with height suggests that the excessive snow cover occurring mainly in northwestern Eurasia (Figure 8) causes the cooling in the lower troposphere. The fact that the influence of the Eurasian snow cover is mainly limited in the lower troposphere is consistent with the finding of Liu and Yanai (2001); that there is little relationship between the upper tropospheric temperature and land surface temperature over Eurasia on an interannual time scale. In the following May June (Figure 9(d) (f)), the centre of negative values over northwestern Eurasia has shifted eastwards, with reduced values of temperature difference. The cooling over northeastern Eurasia still continues. The cooling over the Eurasian continent in spring will delay the development of the Asian summer monsoon. Figure 10. Same as Figure 9, but for the wind differences shown with streamlines

11 SNOW COVER AND ASIAN RAINFALL 1085 The ESSC anomaly also exerts marked influences on the atmospheric circulation. For example, in the differences between the March April wind of and those for (Figures 10(a) (c)), there is a cyclonic anomaly over northwestern Eurasia at all three levels, suggesting a response of the wind field to the extensive snow-cover anomaly. The centre of the cyclonic anomaly generally corresponds to the cooling centres (see Figure 9(a) (c)) and is located on the northeastern side of the main ESSC anomaly in northwestern Eurasia (see Figure 8). From late spring to early summer (May June), a cyclonic circulation anomaly persists over central Russia (Figure 10(d) (f)). We notice that another cyclonic circulation anomaly appears near the eastern coast of Asia. Between the two cyclonic anomalies, there exists an anticyclonic anomaly on the northeast side of the Tibetan Plateau (south of Lake Baikal). In the wind field anomaly maps at all three levels, one can observe a chain of successive anticyclonic and cyclonic anomalies with a barotropic structure over the northern part of Eurasia. This pattern may be interpreted as a Rossby wave train, which makes atmospheric disturbances induced by the snow-cover anomaly propagate from Europe to Asia. We also notice that the March April temperature over the northwestern Tibetan Plateau is higher in the heavy ESSC years than in the light ESSC years (Figure 9(a) and (b)). A similar warming in April May surface temperature over the Tibetan Plateau has been recognized by Bamzai and Shukla (1999: figure 7). They suggested that the winter and spring anomaly in snow cover over the Tibetan Plateau tends to have a sign opposite to that in northern Eurasia. In the 6 years chosen for the present case studies, however, we observe more frequent snow cover over the eastern Tibetan Plateau, as well as over northwestern Eurasia (see Figure 8). It is difficult to investigate the reasons for the observed warming over the Tibetan Plateau owing to the lack of in situ observations on the western Tibetan Plateau. Warm horizontal temperature advection along the western periphery of the Tibetan Plateau inferred from the southerly circulation anomaly (Figure 10(a) and (b)) responding to the ESSC anomaly may, at least partly, explain the warming over the western Tibetan Plateau in the heavy-snow case. In summer, the most remarkable feature over Asia in the wind field at 850 hpa is the monsoonal circulations (Figure 11(a)), with prevailing westerly wind over South Asia and southerly wind over East Asia. In the field of the mean wind difference between the high snow-cover years and the low snow-cover years (Figure 11(b)), the anomalous patterns appearing in May June (Figure 10(f)) still exist in summer. The northerly wind anomaly over East Asia leads to a weaker summer monsoon there in the extensive snow years. The above anomalies in the atmospheric circulation will result in variations in the summer rainfall. In Figure 12, the average difference of summer rainfall rate between the heavy-snow and light-snow years is shown. An area of negative values is found in the region from northern Mongolia to northeastern China. The anticyclonic anomaly south of Lake Baikal (see Figure 11(b)) appears to be responsible for the decreased summer rainfall there for the heavy-snow years. Additionally, the summer rainfall also decreases clearly in Assam and the northernmost part of India. These variations of the summer rainfall shown in Figure 12 are consistent with the results of correlation analyses using multi-year data (see Figures 3(a) and 4) discussed in Section SUMMARY AND DISCUSSION In this work, we have examined the influence of ESSC on the Asian summer rainfall and large-scale temperature and circulation fields in the lower troposphere. Our main findings can be summarized as follows. (i) The relation between AIMR and ESSC is changing over a multi-decadal time scale. The negative correlation between them has markedly intensified since the middle 1970s. (ii) The region where the summer rainfall has the strongest and most stable negative correlation with the preceding ESSC is located in northern Mongolia, south of Lake Baikal. The correlation between the low-pass filtered summer rainfall and ESSC is stronger than that between the original data, showing that the effect of snow cover may be seen more clearly when the effect of ENSO is removed. (iii) Case studies of typical years with extensive and deficient snow cover show that the anomaly of ESSC occurs mainly in northwestern Eurasia and has significant simultaneous and delayed influences on the atmospheric conditions. Heavy ESSC will lead to cooling and a cyclonic circulation

12 1086 X. LIU AND M. YANAI Figure 11. (a) The June September 850 hpa streamlines averaged for the selected six typical years. (b) The difference of June September 850 hpa wind (the average for minus that for ). The black areas indicate the Tibetan Plateau above the 850 hpa isobaric surface anomaly in the lower troposphere over the northern part of Eurasia in spring, forming a Rossby-wave-trainlike circulation response, eventually leading to a weak East Asia summer monsoon and deficient rainfall with an anticyclonic circulation anomaly south of Lake Baikal. One interesting finding in this work is the clear interdecadal change of the relationship between ESSC and AIMR in the mid 1970s. Previous studies (Kumar et al., 1999; Krishnamurthy and Goswami, 2000; Miyakoda et al., 2000) reported that the connection of the Indian monsoon to ENSO has considerably weakened in recent decades. Miyakoda et al. (2000) further pointed out that the transition year was A number of possible mechanisms have been proposed to explain the interdecadal change of the monsoon ENSO connection. For example, Kumar et al. (1999) thought that a shift in the Walker circulation and increased surface temperature over Eurasia in winter and spring had disrupted the monsoon ENSO relation. Miyakoda et al. (2000) ascribed the recent weakening of the monsoon ENSO connection to the climate shift associated with interdecadal variations of global sea temperatures. Kripalani and Kulkarni (1997a,b) found that the impact of ENSO events on AIMR is modulated by the decadal behaviour of AIMR and depends on the prevailing epoch. In addition, a shift of the jet stream over the North Atlantic (Chang et al., 2001) and the change in

13 SNOW COVER AND ASIAN RAINFALL 1087 Figure 12. Difference of summer rainfall rate (mm/day) (the average for three heavy-snow years minus the average for three light-snow years). Only contours of 2, 1, 0.5, 0, 0.5, 1 and 2 mm/day are shown. The shaded areas indicate the regions where the differences are negative and their absolute values exceed the standard deviation of interannual variations for solar irradiance forcing (Mehta and Lau, 1997) have been hypothesized to modulate the monsoon ENSO relationship on decadal multidecadal time scales. It is interesting to note the nearly simultaneous occurrences of the recent enhancing of ESSC AIMR correlation and the weakening of the monsoon ENSO connection. We speculate that the recent reduction in ESSC related with global warming (e.g. Brown, 2000) may have modulated the sensitivity of AIMR to ESSC and SSTA, and eventually induced the interdecadal change in correlations of AIMR with ESSC and ENSO. Therefore, understanding the change in the trend of ESSC may be important for clarifying the cause of the interdecadal monsoon ENSO relationship. Although we have shown some influences of ESSC on the atmospheric circulation and Asian summer rainfall, these results are mostly obtained by excluding the effects of ENSO related to the SSTA. ENSO events have complicated impacts on both the Eurasian snow cover in winter and spring (Ferranti and Molteni, 1999; Corti et al., 2000) and the Asian summer rainfall, including the monsoon rainfall (Elliott and Angell, 1987; Shukla, 1987). A number of numerical simulations (Lau and Bua, 1998; Shen et al., 1998; Yang and Lau, 1998) have suggested that variations in the sea surface temperature of the equatorial eastern Pacific exert a stronger control on the interannual variability of the Asian monsoon than those in the land surface processes related to the snow cover and soil moisture. However, the anomaly of ESSC, once established, will play a considerable role independent of the SSTA in modulating the circulation over Eurasia and the Asian rainfall. The interactions and feedbacks between the global sea surface temperatures, Eurasian snow cover, and Asian rainfall are worth investigating further in the future. ACKNOWLEDGEMENTS We are grateful to two anonymous reviewers for their constructive comments. We thank Dr R.D. Brown for providing the reconstructed Eurasian snow cover data and Dr M. Latif for the Nino3 SSTA data. We also acknowledge the NCEP, NOAA, the Indian Institute of Tropical Meteorology and the Climatic Research Unit, University of East Anglia, for providing other datasets used in this study. This work was supported by the NOAA Office of Global Programs under grant NA96GP0331. X. Liu was also supported by the CAS Innovation Projects (KZCX2-108, KZCX2-SW-118).

14 1088 X. LIU AND M. YANAI REFERENCES Bamzai AS, Shukla J Relation between Eurasian snow cover, snow depth and the Indian summer monsoon: an observational study. Journal of Climate 12: Barnett TP, Dumenil L, Schlese U, Roeckner E, Latif M The effect of Eurasian snow cover on regional and global climate variations. Journal of the Atmospheric Sciences 46: Brown RD Northern Hemisphere snow cover variability and change, Journal of Climate 13: Chang ATC, Foster JL, Hall DK, Powell HW, Chien YL Nimbus-7 SMMR derived global snow depth data set. The pilot land data system. NASA/Goddard Space Flight Center: Greenbelt, MD; Chang CP, Harr P, Ju J Possible roles of Atlantic circulations on the weakening Indian monsoon rainfall ENSO relationship. Journal of Climate 14: Chen M, Xie P, Janowiak JE, Arkin PA Global land precipitation: a 50-yr monthly analysis based on gauge observations. Journal of Hydrometeorology 3: Corti S, Molteni F, Brankovic E Predictability of snow-depth anomalies over Eurasia and associated circulation patterns. Quarterly Journal of the Royal Meteorological Society 126: Elliott WP, Angell JK Relation between Indian monsoon rainfall, the southern oscillation, and hemispheric air and sea temperatures: Journal of Climate and Applied Meteorology 26: Ferranti L, Molteni F Ensemble simulations of Eurasian snow-depth anomalies and their influence on the summer Asian monsoon. Quarterly Journal of the Royal Meteorological Society 125: Goswami BN, Jayavelu V On possible impact of the Indian summer monsoon on the ENSO. Geophysical Research Letters 28: Hahn DG, Shukla J An apparent relationship between Eurasian snow cover and Indian monsoon rainfall. Journal of the Atmospheric Sciences 33: Jones PD Hemispheric surface air temperature variations: a reanalysis and an update to Journal of Climate 7: Kalnay E, Kanamitsu M, Kistler R, Collins W, Deaven D, Gandin L, Iredell M, Saha S, White G, Woollen J, Zhu Y, Chelliah M, Ebisuzaki W, Higgins W, Janowiak J, Mo KC, Ropelewski C, Wang J, Leetmaa A, Reynolds R, Jenne R, Joseph D The NCEP/NCAR 40-year reanalysis project. Bulletin of the American Meteorological Society 77: Kripalani RH, Kulkarni A. 1997a. Climatic impact of El Nino/La Nina on Indian monsoon: a new perspective. Weather 52: Kripalani RH, Kulkarni A. 1997b. Rainfall variability over South-East Asia connections with Indian monsoon and ENSO extremes: new perspectives. International Journal of Climatology 17: Kripalani RH, Kulkarni A Climatology and variability of historical Soviet snow depth data: some new perspectives in snow Indian monsoon teleconnections. Climate Dynamics 15: Kripalani RH, Singh SV, Vernekar AD, Thapliyal V Empirical study on Nimbus-7 snow mass and Indian summer monsoon rainfall. International Journal of Climatology 16: Krishnamurthy V, Goswami BN Indian monsoon ENSO relationship on interdecadal timescale. Journal of Climate 13: Kumar KK, Rajagopalan B, Cane MA On the weakening relationship between the Indian monsoon and ENSO. Science 284: Latif M On the interpretation of climate change in the tropical Pacific. Report No Max-Planck Institut für Meteorologie: Hamburg, Germany; 13 pp. Lau KM, Bua W Mechanisms of monsoon southern oscillation coupling: insights from GCM experiments. Climate Dynamics 14: Li C, Yanai M The onset and interannual variability of the Asian summer monsoon in relation to land sea thermal contrast. Journal of Climate 9: Liu X, Yanai M Relationship between the Indian monsoon rainfall and the tropospheric temperature over the Eurasian continent. Quarterly Journal of the Royal Meteorological Society 127: Meehl GA Coupled land ocean atmosphere processes and south Asian monsoon variability. Science 266: Mehta VM, Lau KM Influence of solar irradiance on the Indian monsoon ENSO relationship at decadal multidecadal time scales. Geophysical Research Letters 24: Miyakoda K, Kinter J, Yang S Analysis of the connection from the south Asian monsoon to ENSO by using precipitation and circulation indices. COLA Technical Report No. 90. Center for Ocean Land Atmosphere Studies: Calverton, MD 20705, USA; Ose T The comparison of the simulated response to the regional snow mass anomalies over Tibet, eastern Europe, and Siberia. Journal of the Meteorological Society of Japan 74: Parthasarathy A, Munot A, Kothawale DR All-India monthly and seasonal rainfall series Theoretical and Applied Climatology 49: Parthasarathy B, Sontakke NA, Munot AA, Kothawale DR Droughts/floods in the summer monsoon rainfall season over different meteorological subdivisions of India for the period Journal of Climatology 7: Prasad KD, Singh SV Seasonal variations of the relationship between some ENSO parameters and Indian rainfall. International Journal of Climatology 16: Robinson DA, Dewey KF, Heim RR Global snow cover monitoring an update. Bulletin of the American Meteorological Society 74: Sankar-Rao M, Lau KM, Yang S On the relationship between Eurasian snow cover and the Asian summer monsoon. International Journal of Climatology 16: Shen X, Kimoto M, Sumi A Role of land surface processes associated with interannual variability of broad-scale Asian summer monsoon as simulated by the CCSR/NIES AGCM. Journal of the Meteorological Society of Japan 76: Shukla J Interannual variability of monsoons. In Monsoons, Fein JS, Stephens PL (eds). John Wiley & Sons: Walland DJ, Simmonds I Modelled atmospheric response to changes in Northern Hemisphere snow cover. Climate Dynamics 13:

15 SNOW COVER AND ASIAN RAINFALL 1089 Webster PJ, Magana VO, Palmer TN, Shukla J, Tomas RA, Yanai M, Yasunari T Monsoons: processes, predictability, and the prospects for prediction. Journal of Geophysical Research 103: Yang S ENSO snow monsoon associations and seasonal interannual predictions. International Journal of Climatology 16: Yang S, Lau KM Influences of sea surface temperature and ground wetness on Asian summer monsoon. Journal of Climate 11: Yang S, Xu L Linkage between Eurasian winter snow cover and regional Chinese summer rainfall. International Journal of Climatology 14: Yasunari T, Kitoh A, Tokioka T Local and remote responses to excessive snow mass over Eurasia appearing in the northern spring and summer climate a study of the MRI GCM. Journal of the Meteorological Society of Japan 69: