Abstract

Utilizing winter (November–March) accumulated snow depth data at 60 stations over the Tibetan Plateau (TP) for the period 1960–98, three typical patterns of the TP snow anomaly's spatial distribution were objectively classified by means of empirical orthogonal function (EOF) analysis. They are characterized by light snow over the entire Tibet region (LS pattern), by an eastern Tibet heavy snow (ETHS pattern), and by a southwestern Tibet heavy snow (SWTHS pattern), respectively.

The possible relations between various patterns of the Tibet winter snow anomaly and subsequent summer monsoon and rainfall over south, southeast, and east Asia are investigated using composite analysis. In ETHS and SWTHS years, the south and southeast Asian summer monsoon becomes weak and there is less summer rainfall over south and southeast Asia than in normal years. In LS years, the anomalies of the subsequent summer monsoon and rainfall are opposite to those in ETHS and SWTHS years. The physical mechanism is, in part, attributed to the impact of heavy snow on Tibet's atmospheric temperature, on the land–sea meridional thermal contrast, and also on the strength of the summer monsoon. The variation of summer rainfall over China associated with the preceding winter TP snow anomaly is also analyzed. There is a clear positive correlation between the Tibetan winter snow and the subsequent summer rainfall over the middle and lower reaches of the Yangtze River valley (central China). In contrast to the previous studies that use snow cover averaged over all of the Tibetan Plateau as a single number, the association between the winter snow and the subsequent summer precipitation over east China is much clearer.

1. Introduction

Over the Northern Hemisphere, the mean monthly snow cover ranges from about 7% to over 40% of the land area, making snow the most rapidly varying natural surface feature (Chang et al. 1990). Snow is a sensitive indicator of climate change. Its existence depends on temperature, precipitation, and solar radiation. Once covering the ground, snow on a large scale will, in turn, have a feedback on weather and climatic change. In comparison with bare soil, snow has high albedo, high thermal emissivity (Sellers 1965), small roughness length (Garratt 1992), and low thermal conductivity, making snow an important influencing factor in modifying regional, and possibly remote, climate through changes in the surface energy balance (e.g., Yeh et al. 1983; Namias 1985; Walsh et al. 1982; Barnett et al. 1989; Groisman et al. 1993, 1994), in the hydrological cycle via snowmelt (e.g., Aguado 1985), and in the atmospheric circulation (e.g., Barnett et al. 1989; Yasunari 1991; Vernekar et al. 1995).

Blanford (1884) suggested that the varying extent and thickness of the Himalayan snow cover exerted some influence on the climatic conditions and weather over the plains of India. He associated the increased winter snow cover in the northwest Himalayas with a decreased rainfall over the plains of western India. Walker (1910) extended this result and found a negative correlation between the accumulated snow depth at the end of May and the amount of summer monsoon rainfall over India during the period 1876–1908. Using the National Oceanic and Atmospheric Administration (NOAA) satellite-derived snow cover (SC) data, the negative Himalayan snow–Indian monsoon relationship was reexamined (Dey and Bhanukumar 1983; Dey et al. 1985; Dey and Kathuria 1986). The connection between Himalayan snow and the Indian summer monsoon was further extended to Eurasian winter snow cover (e.g., Hahn and Shukla 1976; Dey and Bhanukumar 1982; Dickson 1984; Ropelewski et al. 1984; Khandekar 1991; Yang 1996; Parthasarathy and Yang 1995; Sankar-Rao et al. 1996; Bamzai and Shukla 1999). Although weekly NOAA SC over the Northern Hemisphere has been charted since 1966, it is recognized that early observations prior to 1972 underestimate the snow cover extent (Kukla and Robinson 1981; Ropelewski et al. 1985). The values of the correlation between Eurasian winter snow cover anomalies and Indian summer rainfall show that studies that have included the 1967–71 data (e.g., Hahn and Shukla 1976; Dey and Bhanukumar 1982) report higher correlation. Dickson (1984) observed a larger correlation between Eurasian snow cover and Indian monsoon rainfall when the period 1967–71 was excluded. Based on the previous findings, some GCM studies (e.g., Barnett et al. 1989; Yasunari et al. 1991; Zwiers 1993; Vernekar et al. 1995; Ose 1996) also investigated possible physical mechanisms responsible for the influence of Eurasian snow cover on the Asian summer monsoon. But in the recent studies of Bamzai and Shukla (1999), no significant relation is found between the Himalayan seasonal snow cover and the subsequent monsoon rainfall; and western Eurasia is the only geographical region for which a significant inverse correlation exists between the winter snow cover and the subsequent summer monsoon rainfall.

The east Asian monsoon is somewhat different from the south Asian monsoon (Tao and Chen 1987). Accumulated empirical evidences indicate that the snow–rainfall connection in China is rather complex. Chen and Yan (1979) reported that there is an in-phase relation between the winter snow anomaly in the central Tibetan Plateau (TP) and subsequent May and June rainfall over south China, based on data from 1957–74. The relationship between the TP or Eurasian snow and Chinese rainfall, and the possible influence of winter snow on the subsequent summer atmospheric circulation over east Asia has been investigated in numerous research papers (e.g., Chen and Yan 1981; Xu et al. 1994; Wei and Luo 1994; Yang and Xu 1994; Zhai and Zhou 1997; Chinese National Climate Center 1998; Wu et al. 1998; Wu and Qian 2000; Chen et al. 2000). Recent studies (e.g., Wu and Qian 2000; Chen et al. 2000) suggest a positive correlation between Tibetan winter snow and subsequent summer precipitation over the middle and lower reaches of the Changjiang River but a negative correlation over southern China and northern China. There are some disagreements between those findings and the early results reported by Chen and Yan (1979).

The purpose of this paper is to reexamine the TP snow–China rainfall connection and to explore the influence of the TP snow on local and regional climate. In previous observational studies, the authors commonly used satellite-derived snow cover, ground-station-observed snow depth, or days with snow cover taking the entire Tibetan Plateau as a whole. In this paper, the geographical distribution of features of the TP snow anomaly and their associations with the Asian monsoon and rainfall are emphasized.

2. Data and methodology

Five databases were used:

  1. A 39-yr (1960–98) daily ground-observed snow depth (SD) from 60 stations covering the Tibetan Plateau (cf. Fig. 1a) where there was a good temporal continuity of snow depth observations. Among the 60 Tibetan stations, there are 51 stations established in the late 1950s or early 1960s and 9 stations after 1962 (Fig. 1a). Those stations established after 1962 were excluded when we analyzed the geographic distribution of features of the TP winter snow anomaly using empirical orthogonal function (EOF) analysis. Although the earliest observations of snow began in 1956, temporal continuity of data was not enough until 1960. So, only the 39-yr snow observations after 1960 are used in this work. The daily snow depth at all stations is regularly measured at 0800 Beijing time (Tan et al. 1980). If there is no snow cover on the ground in the morning but a snowfall in the afternoon, the snow depth must be measured again at 2000 Beijing time. The unit of measured snow depth is cm. [These datasets can be obtained by corresponding with the authors.]

  2. Thirty-nine years (1960–98) of monthly precipitation at 160 Chinese rain gauge stations (Fig. 1b) where the data are without any discontinuities.

  3. The global 2.5° × 2.5° gridded monthly precipitation rate for the period 1979–98 estimated by Xie and Arkin (1996), obtained by merging gauge measurements and five kinds of satellite-derived rainfall [GOES precipitation index (GPI), OLR-based precipitation index (OPI), Special Sensor Microwave Imager (SSM/I) scattering, SSM/I emission, and Microwave Sounding Unit (MSU)] rainfall.

  4. The global 2.5° × 2.5° gridded monthly NOAA outgoing longwave radiation (OLR) for the period 1978–91 from Arkin's Climate Analysis Center (CAC) monthly OLR archive (Gruber and Krueger 1984).

  5. The National Centers for Environmental Prediction–National Center for Atmospheric Research (NCEP–NCAR) reanalyses (1958–97) of the global 2.5° × 2.5° gridded monthly data of zonal and meridional wind, temperature, and geopotential height at various pressure levels (Kalnay et al. 1996).

Fig. 1.

The geographic distributions of (a) 60 ground-observing snow stations over the Tibetan Plateau and (b) 160 Chinese rain gauge stations marked by small circles. In (a), there are 51 stations established before 1962 marked by • and 9 stations after 1962 marked by . The contour lines show the large-scale Tibetan orographic altitude above sea level at intervals of 500 m. The dashed, thick lines denote the approximate locations of the major mountain ranges

Fig. 1.

The geographic distributions of (a) 60 ground-observing snow stations over the Tibetan Plateau and (b) 160 Chinese rain gauge stations marked by small circles. In (a), there are 51 stations established before 1962 marked by • and 9 stations after 1962 marked by . The contour lines show the large-scale Tibetan orographic altitude above sea level at intervals of 500 m. The dashed, thick lines denote the approximate locations of the major mountain ranges

According to Li (1993), the TP snow cover begins to build up in mid-September and develops very quickly after mid-October or early November. Its ablation period lasts from late February to June. The accumulated snow depth (ASD) for all days from November (of the previous year) to March can be regarded as a good representative of snow mass variation in each year. The ASD and its anomaly were constructed for 38 winters (1960/61–1997/98) for 60 stations over Tibet.

The EOF analysis method was used to explore the spatial distribution of the TP winter ASD anomaly. EOF loadings derived from the normalized ASD anomaly series show a correlation between the temporal fluctuation of the EOF eigenvector and the local snow anomaly. Based upon the EOF loadings, typical spatial patterns of the TP winter snow anomaly were objectively classified.

A composite analysis was undertaken to examine the connection between each spatial pattern of the TP winter snow anomaly and the Asian summer monsoon and rainfall. Although variations of the Asian summer monsoon and rainfall in Tibet snow anomaly years may simultaneously be influenced by other important factors such as Eurasian snow cover, ENSO event, etc., the composite analysis could make the signal of the Tibetan snow influence on the Asian summer monsoon and rainfall much clearer. The statistical significance of the composite anomaly was tested using the Student's t test (with the null hypothesis that the anomaly is not significantly different from zero). A large area with t values above a specified threshold suggests that a coherent signal may have been identified.

3. Results

a. Climatology of Tibetan winter snow

To explore the Tibetan snow variation, first of all, we must understand the climatology of the snow over the Tibetan Plateau. Figure 2 depicts winter (November–March) snow over the Tibetan Plateau averaged for many years. The ground ASD in the Himalayas is larger than that in east Tibet. To the east of 80°E, there are three snowy regions where the orographic altitude is, on average, 4000–5000 m above sea level (cf. Fig. 1a). One region is located on the southern slope of the Himalayas, another is between Nyainqentanglha and Tanggula Mountains, and the third is between A'nymaqen and Bayan Har Mountains. There is a possible reason to account for snow forming over the former two regions. In winter, there is a nearly steady trough over both the Arabian Sea and the Bay of Bengal. Ahead of this trough, humid southwesterly airflow undergoes orographic uplift and then brings rich snowfall to the windward side of the mountains. Over the third snowy region, between A'nymaqen and Bayan Har Mountains, the snow formation may be attributed to cold and dry northwesterly flow coming from higher latitudes meeting with warm and wet southwesterly flow from lower latitudes. In Fig. 2, to the northern part of the Himalayan mountain range, there is a large area of less snow where the orographic altitude is higher than 5000 m above sea level (cf. Fig. 1a). Although the two factors of high topography and low air temperature are favorable for the persistent maintenance of snow cover on the ground, there is less water vapor content in the air column, a factor which is not beneficial to bringing rich snowfall in the area.

Fig. 2.

The climatology of Nov–Mar accumulated snow depth (cm month−1) over the Tibetan Plateau averaged for 39-yr (1960–98) ground-observed data at 60 Tibetan stations

Fig. 2.

The climatology of Nov–Mar accumulated snow depth (cm month−1) over the Tibetan Plateau averaged for 39-yr (1960–98) ground-observed data at 60 Tibetan stations

b. Year-to-year variation of Tibetan winter snow

The interannual variation of the Tibetan winter snow has large inhomogeneities in its spatial distribution. The typical distribution patterns were explored using EOF analysis. The first two eigenvectors have obvious separation from the others. In the following, we only analyze the loadings of the first two EOF principal components. They are shown in Fig. 3. The loadings of the first eigenvector (denoted EOF1) are negative values nearly over the entire area of Tibet. This indicates that the winter snow over the entire area of Tibet tends to vary in phase. This pattern accounts for 15.4% of the total interannual variance of normalized winter snow anomalies. In the loading map of the second eigenvector (EOF2), there are two regions with opposite sign, that is, positive loading over southeast Tibet and negative loading over northern and central Tibet. These reveal that the winter snow anomalies over these two areas tend to vary out of phase. This pattern accounts for 10.3% of the total variance of normalized winter snow anomalies.

Fig. 3.

The loadings (multiplied by 100) of (a) the first and (b) the second EOF eigenvectors of Nov–Mar accumulated Tibetan snow anomalies, based on the normalized series of 36 winters (1962/63–1997/98) at 51 Tibetan stations. The percentages of variance explained by the first two modes are 15.4 and 10.3, respectively

Fig. 3.

The loadings (multiplied by 100) of (a) the first and (b) the second EOF eigenvectors of Nov–Mar accumulated Tibetan snow anomalies, based on the normalized series of 36 winters (1962/63–1997/98) at 51 Tibetan stations. The percentages of variance explained by the first two modes are 15.4 and 10.3, respectively

In order to test the distribution of EOF loadings in Fig. 3 and elucidate that the EOF1 and EOF2 patterns are not artifacts of the eigenvector analysis technique, we calculated one-point correlation coefficients of winter ASD between three base stations and every other station. The three base stations are selected at the minimum center (36.18°N, 98.06°E) of the EOF1 loadings and at the two opposite sign maxima centers (29.15°N, 91.46°E and 37.22°N, 97.22°E) of the EOF2 loadings. The resulting correlation maps are shown in Fig. 4. They enhance the domain-scale spatial structures of the EOF loadings and do not incorporate any variance maximizing constraints. The magnitude of the correlation coefficient in Fig. 4 only represents the correlation significance for an individual pair of points. Similar spatial patterns between Figs. 3a and 4a and between Figs. 3b and 4b,c reconfirm the existence of the main characteristics of the EOF1 and EOF2 spatial distributions mentioned above.

Fig. 4.

The one-point correlation coefficients (multiplied by 100) between the 39-yr (1960–98) Nov–Mar accumulated snow depth at three base stations (marked by solid circles) and those at every other station. The three base stations are located at (a) 36.18°N, 98.06°E, (b) 29.15°N, 91.46°E and (c) 37.22°N, 97.22°E. The correlation coefficient of 0.25 is significant at the 90% confidence level

Fig. 4.

The one-point correlation coefficients (multiplied by 100) between the 39-yr (1960–98) Nov–Mar accumulated snow depth at three base stations (marked by solid circles) and those at every other station. The three base stations are located at (a) 36.18°N, 98.06°E, (b) 29.15°N, 91.46°E and (c) 37.22°N, 97.22°E. The correlation coefficient of 0.25 is significant at the 90% confidence level

Reviewing the distribution of the 38-winter (1960/61–1997/98) Tibetan snow anomaly, we find a clear characteristic. The less-snow anomaly over Tibet easily appears over a large extent of the area but the heavy-snow anomaly has strong regional features. Based on the loadings of the first two EOF principal components, we classify the Tibetan winter snow anomaly into three typical patterns of spatial distribution. The main features may be seen in Fig. 5. The first pattern (Fig. 5a) is characterized by light snow (a negative anomaly, denoted LS) over the whole of Tibet which is constructed with respect to the loadings of EOF1. The second pattern (Fig. 5b) is characterized by heavy snow (a positive anomaly, denoted ETHS) over east Tibet, whose construction is based upon the opposite phase of EOF1 loadings and partially upon EOF2 loadings. The third pattern (Fig. 5c) is characterized by a heavy snow over northern and central Tibet (denoted NCTHS), based upon the opposite phase of the EOF2 loadings.

Fig. 5.

The composite of Nov–Mar accumulated snow depth anomalies (cm month−1) at 60 stations for the (a) 14 LS, (b) 7 ETHS, (c) 5 NCTHS, and (d) 3 SWTHS winters. Shaded areas are where the t test is significant at the 90% level

Fig. 5.

The composite of Nov–Mar accumulated snow depth anomalies (cm month−1) at 60 stations for the (a) 14 LS, (b) 7 ETHS, (c) 5 NCTHS, and (d) 3 SWTHS winters. Shaded areas are where the t test is significant at the 90% level

The Himalayas are located on the southern edge of the Tibetan Plateau where there is rich annual snowfall and a very thick snowpack on the ground. However, inadequacy in the number of ground observation stations and in the length of observation must influence our EOF analysis. Considering the significance of Himalayan snow, a fourth pattern is included (Fig. 5d), characterized by a southwest Tibet heavy snow (hereafter SWTHS). There is an ASD positive anomaly of deeper than 400 cm month−1 at the southern slope of the Himalayas. This spatial distribution pattern is similar to the loadings of the fifth EOF principal component (omitted).

According to the time series of the first two EOF principal components and the spatial distributions of the ASD anomaly of 38 winters (1960/61–1997/98, November–March), we have picked out 14 LS, 7 ETHS, 5 NCTHS, and 3 SWTHS winters. These are listed in Table 1. Under the limitation of sparse stations in the western part of the Tibet (Fig. 1a), there are only three typical SWTHS cases selected. Although the analysis of only three SWTHS cases lacks statistical significance, it is important to help explore the influence of Himalayan snow anomalies.

Table 1.

The four spatial distribution patterns of the Tibetan winter snow anomaly in the period 1960–98

The four spatial distribution patterns of the Tibetan winter snow anomaly in the period 1960–98
The four spatial distribution patterns of the Tibetan winter snow anomaly in the period 1960–98

In Fig. 5c, a positive snow anomaly in an NCTHS winter is located over northern and central Tibet. However, in the map of the climatological mean (Fig. 2), northern and central Tibet is an area of less snow. Analyzing the variation of the subsequent summer south and southeast Asian monsoon, we found the NCTHS pattern has a weak association with the south Asia summer monsoon. However, it has important influence on the climate over north Tibet. Therefore, in the following sections, we will only focus on the connections between the other three patterns (LS, ETHS, and SWTHS) of Tibetan winter snow and the subsequent summer monsoon and rainfall.

c. Thermal influence of Tibetan winter snow anomalies

Figure 6 shows the composite of the mean tropospheric atmospheric temperature anomaly averaged over the 27.5°–37.5°N latitude belt for the period of November–May of 14 LS, 6 ETHS, and 3 SWTHS years, respectively. The areas of less and more Tibetan snow on a large scale correspond well to the positive and negative tropospheric atmospheric temperature anomalies, respectively. On the composite map of the 14 LS cases (Fig. 6a), the center of the largest positive temperature anomaly is over the central Tibetan Plateau where the atmospheric temperature of the air column below 250 mb is 0.1°–0.2°C higher than the climatological mean. Although a magnitude of 0.1°C is small, it has statistical significance at the level of 90% or greater. In contrast to this, the composite map of the six ETHS cases (Fig. 6b) contains a large expanse of negative temperature anomalies over the whole Tibetan Plateau. The center of the negative temperature anomaly on the composite map of the three SWTHS cases (Fig. 6c) is positioned to the west of 85°E. This must be due to the Himalayan winter snow anomalies.

Fig. 6.

The height–longitude sections of the composite atmospheric temperature anomalies averaged for 27.5°–37.5°N, for the period of Nov–May, and for (a) 14 LS, (b) 6 ETHS, and (c) 3 SWTHS years that fall into the period 1958–97. The contour interval is 0.1°C. The Tibetan Plateau terrain is shown by the dark-shaded area. The lightly shaded area indicates where the temperature anomaly is significant at the 90% confidence level of the t test

Fig. 6.

The height–longitude sections of the composite atmospheric temperature anomalies averaged for 27.5°–37.5°N, for the period of Nov–May, and for (a) 14 LS, (b) 6 ETHS, and (c) 3 SWTHS years that fall into the period 1958–97. The contour interval is 0.1°C. The Tibetan Plateau terrain is shown by the dark-shaded area. The lightly shaded area indicates where the temperature anomaly is significant at the 90% confidence level of the t test

In Figs. 6a–c, the signals of the atmospheric temperature anomalies at the upper troposphere are opposite to those at the middle and lower troposphere. In fact, there exists a quasi-stationary trough at the upper troposphere off the east coasts of the Eurasian continent in winter and spring seasons. The plateau is located behind this major trough where the northwesterly wind is prevailing. In LS winters, the negative temperature anomaly at the upper troposphere is possibly attributed to stronger dry and cold northwesterly wind than that in the normal year, which brings poor snowfall over the plateau. Less snow cover would result in the decrease of surface albedo, which is favorable for warmer ground surface and positive anomaly of the atmospheric temperature over the lower and middle troposphere. They are opposite to those in ETHS and SWTHS winters.

d. The influence of Tibetan winter snow on the land–sea thermal contrast

The south Asian summer monsoon is a dynamically stable system (Charney and Shukla 1981). Its onset and retreat are characterized by two reversals of the midtropospheric meridional temperature gradient between the equator and 30°N, respectively (Flohn 1957, 1960; Fu and Fletcher 1985). The variation of the atmospheric temperature over the Tibetan Plateau bears the greatest responsibility for the reversals of the meridional temperature gradient. It is an important factor in driving the south Asian summer monsoon. Figure 7 shows the annual cycle of the mean atmospheric temperature of the longitude belt 65°–105°E along 32.5°N (the central latitude of the TP), along the equator, and the difference between them. The intra-annual variation of the midtropospheric meridional temperature gradient between 32.5°N and the equator is dominated by the variation of the atmospheric temperature over the TP (Fig. 7b) because the atmospheric temperature over the equator at any given pressure level (Fig. 7a) has a very small magnitude of intra-annual variation. As shown in Fig. 7c, there are two reversals of the meridional temperature gradient between 32.5°N and the equator, namely the transition from negative to positive values during the period May–June and the transition from positive to negative values during the period September–October. According to the theory of thermal wind, these two reversals must cause two reversals of the zonal wind direction over south Asia per year. Figure 7d shows the annual cycle of the zonal wind averaged over the Tibetan longitudes (60°–105°E) along 15°N, which happens to be the middle latitude of the south Asian monsoon area (5°–20°N, 40°–110°E). The first transition from westerly to easterly wind above the 400-mb level occurs during April–May, a clear mark of the onset of the south Asian summer monsoon. The second transition from easterly to westerly wind occurs during September–October, representing the retreat of the summer monsoon and the reestablishment of the winter monsoon. Thus the period from June–September (JJAS) is the prevailing time of the south Asian summer monsoon circulation.

Fig. 7.

The monthly climatology vs height of (a) the monthly mean temperature of 65°–105°E averaged along the equator, (b) along 32.5°N, (c) the meridional temperature gradient (32.5°N–0°), and (d) the zonal wind along 15°N. The results are based on the 30-yr (1961–90) NCEP–NCAR reanalyses data. The contour interval is 5 K in (a) and (b), 2 K in (c), and 2 m s−1 in (d).

Fig. 7.

The monthly climatology vs height of (a) the monthly mean temperature of 65°–105°E averaged along the equator, (b) along 32.5°N, (c) the meridional temperature gradient (32.5°N–0°), and (d) the zonal wind along 15°N. The results are based on the 30-yr (1961–90) NCEP–NCAR reanalyses data. The contour interval is 5 K in (a) and (b), 2 K in (c), and 2 m s−1 in (d).

As one of the important impacting factors on the tropospheric atmospheric temperature variation over the TP, the anomaly of Tibetan snow would result in the year-to-year variation (or anomaly) of the land–sea thermal contrast between the TP and the equator. Figure 8 shows a composite of the meridional temperature gradient anomalies (from 32.5°N to the equator) averaged over the Tibetan 65°–105°E longitude belt for the 14 LS, 6 ETHS, and 3 SWTHS snow years (defined here as from November of the previous year to October). The negative and positive anomalies in Fig. 8 denote a weak and a strong land–sea thermal contrast, respectively. It is easy to understand that more (less) snow weakens (intensifies) the land–sea thermal contrast. On the composite map of the 14 LS (6 ETHS) cases, less (more) snow over the TP corresponds well to the persistent positive (negative) anomaly of the meridional temperature gradient. Especially for the ETHS pattern (Fig. 8b), the meridional temperature gradient possesses the same sign all year round. The center of negative anomaly is slowly moving upward with time from November at 400 mb to next October at 100 mb. Is this influenced by the snow anomaly or the results of the internal low-frequency disturbance of atmospheric circulation in ETHS years? It is somewhat difficult to distinguish them, because snowfall is the result of atmospheric circulation but reacts to the atmospheric thermal variation and the atmospheric circulation as well. On the composite map of the three SWTHS cases (Fig. 8c), the atmospheric temperature over the whole troposphere gradually decreases and becomes a persistent negative anomaly from April. This may be attributed to a later melting time over the Himalayas than over east Tibet. In Figs. 8a–c, the largest amount of meridional temperature gradient anomalies is centralized between 200 and 300 mb. There is a clear characteristic of wavelike variation with time. It may be attributed to low-frequency fluctuation of the atmosphere in itself.

Fig. 8.

The height–time sections of the composite meridional temperature gradient anomalies from 32.5°N to the equator averaged over the Tibetan 65°–105°E belt for (a) 14 LS, (b) 6 ETHS, and (c) 3 SWTHS snow years (from Nov of previous year to Oct during 1958–97. The contour interval is 0.2°C. Shaded areas are where the t test is significant at the 90% level

Fig. 8.

The height–time sections of the composite meridional temperature gradient anomalies from 32.5°N to the equator averaged over the Tibetan 65°–105°E belt for (a) 14 LS, (b) 6 ETHS, and (c) 3 SWTHS snow years (from Nov of previous year to Oct during 1958–97. The contour interval is 0.2°C. Shaded areas are where the t test is significant at the 90% level

Tibetan snow is an important influencing factor upon the local atmospheric temperature variation and the land–sea thermal contrast. It will indirectly influence the variation of the Asian summer monsoon.

e. Influence of Tibetan winter snow on the Asian summer monsoon

The Asian summer monsoon involves at least two subsystems, the Indian monsoon (Krishnamurti and Bhalme 1976) and the east Asian monsoon (Tao and Chen 1987). The period from June to September is the prevailing time of the south Asian summer monsoon circulation. Considering the fact that the retreat of the summer monsoon over the northern part of the east Asian monsoon area is earlier than that over the south Asian summer monsoon region, we only explore the features of Asian monsoon variation for June, July, and August (JJA).

1) The 850-mb wind anomalies

On the map of the JJA climatological mean horizontal wind vector at 850 mb (Fig. 9a), there is a stream of flow between 40° and 60°E, crossing the equator from the Southern Hemisphere. It forms the Somali jet, which turns eastward around the Indian peninsula and then gradually divides into two branches. The northern branch turns northeastward surrounding the TP, crosses the southwestern and central parts of China, and then reaches north China. The southern branch continues to flow eastward, rotates counterclockwise around a monsoon trough over the area from the South China Sea to the western Pacific (which is commonly called the southeast Asian monsoon trough), and then crosses the southeastern and eastern coast of China and Japan. To the east of 105°E, there is another stream of cross-equatorial flow from the Southern Hemisphere. It flows northeastward and merges with the southeast Asian monsoon trough.

Fig. 9.

(a) Mean JJA 850-mb wind vectors and streamlines climatology, and wind vector anomaly composites for the (b) 14 LS, (c) 6 ETHS, and (d) 3 SWTHS years during 1958–97. The TP is the dark-shaded region. Lightly shaded areas in (a) stand for where the magnitude of zonal wind is greater than 5 m s−1 and those in (b), (c), and (d) show where the t test of anomalies is significant at the 90% level

Fig. 9.

(a) Mean JJA 850-mb wind vectors and streamlines climatology, and wind vector anomaly composites for the (b) 14 LS, (c) 6 ETHS, and (d) 3 SWTHS years during 1958–97. The TP is the dark-shaded region. Lightly shaded areas in (a) stand for where the magnitude of zonal wind is greater than 5 m s−1 and those in (b), (c), and (d) show where the t test of anomalies is significant at the 90% level

Figures 9b–d show the composites of mean JJA 850-mb wind anomaly for the 14 LS, 6 ETHS, and 3 SWTHS years, respectively. On the composite maps of the 14 LS (6 ETHS and 3 SWTHS) cases, we can see that there are clear westerly (easterly) wind anomaly vectors over south Asia, a cross-equatorial southerly wind between 80° and 100°E, and a cyclonic wind over the South China Sea. From the point of view of the anomaly, they represent a strong (weak) summer monsoon circulation for the LS (ETHS and SWTHS) patterns. In order to ascertain the veracity of the 850-mb anomalous flow composite in Figs. 9b–d, the magnitude of the composite wind speed anomaly is tested using the Student's t test with the null hypothesis that the anomaly is not significantly different from zero. As shown in Figs. 9b–d, the anomalies over southeast and south Asia are noticeable and have statistical significance at the level of 90% or greater.

2) OLR anomalies

The vertical shear between the 850-mb westerly wind and the 200-mb easterly wind averaged over 5°–20°N, 40°–110°E is widely regarded as a monsoon index (Webster and Yang 1992). A large vertical shear will lead to a strong vertical convection and a strong monsoon rainfall. The magnitude of the outgoing longwave radiation over the tropical region is a good representative of convection activity intensity. Low (high) OLR corresponds well to strong (weak) vertical convection. In addition, OLR has a significant correlation with vertical shear over the broad-scale monsoon region (Webster and Keller 1975; Webster and Yang 1992).

In the composite map of the seven LS cases (omitted), there is a zonal belt of negative OLR anomaly over south Asia and a zonal belt of positive OLR anomaly over the region from the middle and lower reaches of the Yangtze River, from China to Japan. In contrast to this, there are opposite OLR anomalies over south and east Asia for six ETHS and three SWTHS cases (not shown). This provides additional important evidence to support the foregoing results that the summer monsoon over south Asia tends to strengthen (weaken) following LS (ETHS and SWTHS) winters.

3) 500-mb geopotential height

On the map of the climatological mean of the JJA 500-mb geopotential height (not shown), there is a subtropical high over the western Pacific and a low over India and Bengal Bay. The western Pacific subtropical high (WPSH) is one of the important members of the east Asian summer monsoon system (Tao and Chen 1987) and the monsoon low over India and Bengal Bay is one of the important members of the Indian summer monsoon system (Krishnamurti and Bhalme 1976).

Figure 10 illustrates the composite of the JJA mean 500-mb geopotential height averaged for the 14 LS, 6 ETHS, 5 NCTHS, and 3 SWTHS years, respectively. In order to analyze the variation of the mean geographical location of the WPSH, we regard the contour line of 586 dm as its western boundary. The WPSH in ETHS (Fig. 10b) and SWTHS years (Fig. 10c) extends more westward than in LS years (Fig. 10a). In the area of the South China Sea and southeast Asian summer monsoons, there must be a stronger anticyclonic circulation in the midtroposphere and a weaker vertical convection in ETHS and SWTHS years than in LS years. The variation in the extent of the contour isoline 584 dm may, in part, reflect the variation in the intensity of the Indian summer monsoon strength. For example, a larger extent in LS years (Fig. 10a) than in ETHS and SWTHS years represents a deeper Indian monsoon low and a stronger summer monsoon system. They all coincide with the anomalies of the 850-mb wind (Figs. 9b–d).

Fig. 10.

The composite of the mean JJA 500-mb geopotential height for (a) 14 LS, (b) 6 ETHS, and (c) 3 SWTHS years that fall into the period 1958–97. The contour interval is 2 dm. Shaded areas are where the t test is significant at the 90% level

Fig. 10.

The composite of the mean JJA 500-mb geopotential height for (a) 14 LS, (b) 6 ETHS, and (c) 3 SWTHS years that fall into the period 1958–97. The contour interval is 2 dm. Shaded areas are where the t test is significant at the 90% level

f. Association between Tibetan winter snow and summer rainfall

The summer rainfall anomaly over south Asia is closely associated with the interannual variation of summer monsoon intensity. Figure 11 shows a composite of the JJA total precipitation anomaly following various Tibetan winter snow anomalies, using the 20-yr (1979–98) monthly global rainfall data estimated by Xie and Arkin (1996). In the composite map of the seven LS years, there is a zonal belt of positive rainfall anomaly (Fig. 11a) from the Indian peninsula to the western Pacific. Its maximum (>20%) is centered over the South China Sea and the Philippines, which must be the result of strong vertical convection and a deep monsoon trough (Fig. 9b). In contrast to this, the composite maps of the six ETHS (Fig. 11b) and three SWTHS (Fig. 11c) years show anomalies of less rainfall over the whole of south Asia, which is partly attributable to weak monsoon circulation (Figs. 9c,d). In Fig. 11, there are some small-scale characteristics in the rainfall anomaly over south Asia. They possibly involve the influence of orographic rainfall or data errors, because Xie–Arkin's rainfall data are merged from values of rain gauge and five kinds of satellite-derived rain data (Xie and Arkin 1996). In east Asia, the summer precipitation anomaly is rather complicated. As shown in Fig. 11, the rainfall anomaly over the belt from the middle and lower reaches of Yangtze River valley, China, to Japan is opposite to the one over south and southwest China and southeast Asia.

Fig. 11.

The composite of the JJA total precipitation anomaly percentages (%) for (a) 5 LS, (b) 6 ETHS, and (c) 3 SWTHS years during 1979–98. Results are based on the Xie and Arkin (1996) estimated rainfall data. The contour interval is 10%. Shaded areas are where the t test is significant at the 90% level

Fig. 11.

The composite of the JJA total precipitation anomaly percentages (%) for (a) 5 LS, (b) 6 ETHS, and (c) 3 SWTHS years during 1979–98. Results are based on the Xie and Arkin (1996) estimated rainfall data. The contour interval is 10%. Shaded areas are where the t test is significant at the 90% level

The reliability of the composite picture in Fig. 11 needs to be verified using ground observations. It is a pity that we only have at hand, Chinese rain gauge observations. Figure 12 illustrates the composites of the normalized JJA total rainfall anomaly percentage over China averaged for the 14 LS, 7 ETHS, and 3 SWTHS years. Comparing Fig. 12 with Fig. 11, there are similar spatial distributions of the mean JJA rainfall anomaly over east China. In the composite maps of the seven ETHS and three SWTHS years, there is a clear more-rainfall belt (positive anomaly) along the middle and lower reaches of the Changjiang River valley (central China). The mean JJA rainfall anomalies over east China are positive–negative–positive (in Fig. 12a) or negative–positive–negative (in Figs. 12b,c) sandwich distributions from south to north.

Fig. 12.

The composite of the JJA total precipitation anomaly percentages (%) over China for (a) 14 LS, (b) 7 ETHS, and (c) 3 SWTHS years during 1960–98. Results are based on 160 Chinese rain gauge stations' rainfall data. The contour interval is 10%

Fig. 12.

The composite of the JJA total precipitation anomaly percentages (%) over China for (a) 14 LS, (b) 7 ETHS, and (c) 3 SWTHS years during 1960–98. Results are based on 160 Chinese rain gauge stations' rainfall data. The contour interval is 10%

4. Summary and discussion

In the paper, the results of a detailed analysis of the climatology and interannual variation of Tibetan accumulated winter snow have been shown. The eastern and southwestern parts of the Tibetan Plateau are snowy areas. Utilizing EOF analysis, we have discussed the spatial distribution of Tibetan winter (November–March) ASD anomalies and classified them into three typical spatial patterns (LS, ETHS, and NCTHS). In order to emphasize the significance of the snow anomaly in the Himalayas, the fourth pattern (SWTHS) is also included. The mechanism behind the four typical patterns of snow anomaly will be discussed in another paper. According to the spatial distributions of the November–March Tibetan ASD anomaly in 38 winters (1960/61–1997/98), we picked out typical LS, ETHS, NCTHS, and SWTHS winters.

Potential associations of Tibetan snow with local and regional variations of land–sea thermal contrast, atmospheric circulation, and precipitation have been explored. A more (less) Tibetan snow anomaly linked to the formation of negative (positive) tropospheric atmospheric temperature anomaly over the Tibetan Plateau, the weak (strong) land–sea thermal contrast over south Asia, and the weak (strong) south Asian summer monsoon circulation. In south and southeast Asia, summer rainfall anomalies are closely associated with the variation of summer monsoon intensity. In LS (ETHS and SWTHS) years, summer rainfall of south Asia, as a whole, tends to increase (decrease). Over east Asia, the relation between summer rainfall and the preceding winter Tibetan snow varies with different regions. There is an obvious positive correlation between summer rainfall from the middle and lower reaches of Yangtze River valley, China, to Japan and the preceding winter's Tibetan snow variation. This variation of precipitation is opposite to that over the south Asian monsoon region and that over north China.

The signals of associations between the NCTHS pattern of winter snow and the subsequent summer monsoon and precipitation over south Asia and east Asia are weak (details were not analyzed in the text). In fact, the NCTHS pattern of snow is mainly characterized by positive anomaly of winter snow over the northern and central Tibetan Plateau, where snow cover is ephemeral within the snow season. Only the snow variations in Tibet's snowy areas, such as the LS, ETHS, and SWTHS snow anomalies, have relatively evident connection with the subsequent south and southeast Asian summer monsoon.

The connection between the Tibetan snow anomaly and south Asian monsoon are easy to understand. The wavelike pattern of anomalous summer precipitation from the south to the north in east Asia is possibly associated with the teleconnection of atmospheric circulation. Based on the results of Li and Ji (1995), the region to the south of the Tibetan Plateau is one of the important areas where atmospheric disturbances have a strong possibility to absorb the energy from the basic-state flow and to form geographically fixed teleconnection of atmospheric circulation at specific locations. As an external forcing, Tibetan snow has an important role in the persistent anomaly of south Asian summer monsoon circulation, which could stimulate Rossby wave dispersion along the coast of east Asia to bring forth a wavelike pattern of anomalous precipitation.

The speculation of the associations above is only one possibility to explain the physical mechanism of the influence of Tibetan snow on monsoons and precipitation. The variations of summer rainfall or winter snow itself could be the response to an additional, independent forcing, most likely sea surface temperature anomalies or perhaps to a less degree a continental-wide forcing such as Eurasian snow. So, comparative numerical experiments with realistic models of the climate system using realistic snow anomaly are essential to understand the mechanism of the influence of Tibetan snow on Asian summer monsoon and precipitation.

However, what is the physical mechanism to maintain the lasting anomalies of atmospheric thermal variation over the Tibetan Plateau and summer monsoons over south Asia? Is it the ground moisture anomaly that works, which is retained when the spring snow melts, or the snow-covered ground albedo that matters? Or is there any memory such as low-frequency changes of the planetary-scale circulation retained in the atmosphere itself from winter to summer? This needs to be analyzed in the future.

It is clear that this study has suggested a clearer picture of the connection between the Tibetan snow and the Chinese regional summer rainfall than was reported in previous studies. The relationship between the TP snow anomaly and the subsequent summer Indian monsoon rainfall variation is not so clear yet. A possible explanation is that there is less information available on the variation of the snow over the west TP due to the sparse ground stations there. The obvious negative (positive) summer rainfall anomaly over India in the SWTHS (LS) composite map (Figs. 11a and 11c) suggests that winter snow anomalies in the Himalayas have a clearer reverse relationship with Indian monsoon precipitation than that in the east Tibetan Plateau. This needs to be verified in the future by additional observational and modeling studies.

Acknowledgments

We would like to thank the National Environmental Satellite Data and Information Services (NESDIS) of the U.S. National Oceanic and Atmospheric Administration (NOAA), for providing the NOAA snow cover and SMMR snow depth data. We would also like to thank NCAR for access to the OLR and the Xie–Arkin rainfall data from NCAR dataset archives and for providing the NCEP–NCAR reanalysis data. Thanks also go to reviewers of J. Climate, Prof. Guoxiong Wu, and Prof. Liren Ji (LASG/IAP) for their helpful comments and to Mr. D. P. Griffith for correcting the English expression. This work was conducted under the joint support of Chinese Academy Project ZKCX2-SW-210 and the Natural Sciences Foundation of China under Grants 40231005, 40005008, and 40135020.

REFERENCES

REFERENCES
Aguado
,
E.
,
1985
:
Radiation balances of melting snow covers at an open site in the central Sierra Nevada, California.
Water Resour. Res.
,
21
,
1649
1654
.
Bamzai
,
A. S.
, and
J.
Shukla
,
1999
:
Relation between Eurasian snow cover, snow depth, and the Indian summer monsoon: An observational study.
J. Climate
,
12
,
3117
3132
.
Barnett
,
T. P.
,
L.
Dümenil
,
V.
Schlese
,
E.
Eoeckner
, and
M.
Latif
,
1989
:
The effect of Eurasian snow cover on regional and global climate variations.
J. Atmos. Sci.
,
46
,
661
685
.
Blanford
,
H. F.
,
1884
:
On the connexion of Himalayan snowfall with dry winds and seasons of drought in India.
Proc. Roy. Soc. London
,
37
,
3
22
.
Chang
,
A. T. C.
,
J. L.
Foster
,
D. K.
Hall
,
H. W.
Powell
, and
Y. L.
Chien
,
1990
:
Nimbus-7 SMMR derived global snow cover and snow depth data set.
The Pilot Land Data System. NASA Goddard Space Flight Center, Greenbelt, MD, 40 pp
.
Charney
,
J. G.
, and
J.
Shukla
,
1981
:
Predictability of monsoons.
Monsoon Dynamics, Sir J. Lighthill and R. P. Pearce, Eds., Cambridge University Press, 99–109
.
Chen
,
L-T.
, and
Z-X.
Yan
,
1979
:
Impact of Himalayan winter-spring snow cover on atmospheric circulation and on southern Chinese rainfall during the pre-rainy period.
Collected Papers on Medium- and Long-term Hydrologic and Meteorological Forecasts (1) (in Chinese). Water Conservancy and Power Press, 185–194
.
Chen
,
L-T.
, and
Z-X.
Yan
,
1981
:
A statistical study of the impact of Himalayan winter-spring snow cover anomalies on the pre-summer monsoon.
Collected Papers on Medium- and Long-term Hydrologic and Meteorological Forecasts (2) (in Chinese). Water Conservancy and Power Press, 133–141
.
Chen
,
Q-J.
,
B.
Gao.
, and
W-J.
Wi
,
2000
:
Studies on relationships among winter snow cover over the Tibetan Plateau and droughts/floods during Meiyu season in the middle and lower reaches of the Changjiang river as well as in the atmosphere/ocean system (in Chinese).
Acta Meteorol. Sin.
,
58
,
582
595
.
Chinese National Climate Center
,
1998
:
The 1998 Flood in China and Climatic Abnormality (in Chinese).
Meteorology Press, 137 pp
.
Dey
,
B.
, and
O. S. R. U.
Bhanukumar
,
1982
:
An apparent relationship between Eurasian snow cover and the advance period of the Indian summer monsoon.
J. Appl. Meteor.
,
21
,
1929
1932
.
Dey
,
B.
, and
O. S. R. U.
Bhanukumar
,
1983
:
Himalayan winter snow cover area and summer monsoon rainfall over India.
J. Geophys. Res.
,
88
,
5471
5474
.
Dey
,
B.
, and
S. N.
Kathuria
,
1986
:
Himalayan snow cover area and onset of summer monsoon over Kerala, India.
Mausam
,
37
,
193
196
.
Dey
,
B.
,
S. N.
Kathuria
, and
O. B.
Kumar
,
1985
:
Himalayan summer snow cover and withdrawal of the Indian summer monsoon.
J. Appl. Meteor.
,
24
,
865
868
.
Dickson
,
R. R.
,
1984
:
Eurasian snow cover versus Indian monsoon rainfall—An extension of the Hahn–Shukla results.
J. Appl. Meteor.
,
23
,
171
173
.
Flohn
,
H.
,
1957
:
Large-scale aspects of the “summer monsoon” in South and East Asia.
J. Meteor. Soc. Japan, 75th Ann. Vol., 180–186
.
Flohn
,
H.
,
1960
:
Recent investigations on the mechanism of the “summer monsoon” of southern and eastern Asia.
Symp. on Monsoons of the World. New Delhi, India, India Meteorology Department, 75–88
.
Fu
,
C.
, and
J. O.
Fletcher
,
1985
:
The relationship between Tibet-tropical ocean thermal contrast and interannual variability of Indian monsoon rainfall.
J. Appl. Meteor.
,
24
,
841
847
.
Garratt
,
J. R.
,
1992
:
The Atmospheric Boundary Layer.
Cambridge University Press, 334 pp
.
Groisman
,
P. Y.
,
T. R.
Karl
, and
R. W.
Knight
,
1993
:
Observed impact of snow cover on the heat balance over the continental spring temperature.
Science
,
263
,
198
200
.
Groisman
,
P. Y.
,
T. R.
Karl
,
R. W.
Knight
, and
G. L.
Stenchikov
,
1994
:
Changes of snow cover, temperature, and radiative heat balance over the Northern Hemisphere.
J. Climate
,
7
,
1633
1656
.
Gruber
,
A.
, and
A. F.
Krueger
,
1984
:
The status of the NOAA outgoing longwave radiation data set.
Bull. Amer. Meteor. Soc.
,
65
,
958
962
.
Hahn
,
D. J.
, and
J.
Shukla
,
1976
:
An apparent relationship between Eurasian snow cover and Indian monsoon rainfall.
J. Atmos. Sci.
,
33
,
2461
2462
.
Kalnay
,
E.
, and
Coauthors
.
1996
:
The NCEP/NCAR 40-Year Reanalysis Project.
Bull. Amer. Meteor. Soc.
,
77
,
437
471
.
Khandekar
,
M. L.
,
1991
:
Eurasian snow cover, Indian monsoon and El Niño/Southern Oscillation—A synthesis.
Atmos.–Ocean
,
29
,
636
647
.
Krishnamurti
,
T. N.
, and
H. N.
Bhalme
,
1976
:
Oscillation of a monsoon system. Part I. Observational aspects.
J. Atmos. Sci.
,
33
,
1937
1954
.
Kukla
,
G.
, and
D. A.
Robinson
,
1981
:
Accuracy of snow and ice monitoring.
Snow Watch 1980, Glaciological Data Report GD-5
.
Li
,
P-J.
,
1993
:
Characteristics of snow cover in western China.
Acta Geograph. Sin.
,
48
,
505
514
.
Li
,
Z-J.
, and
L-R.
Ji
,
1995
:
The favorable forcing mode in barotropic atmosphere, its responding, and patterns of atmospheric teleconnection.
Sci. China
,
25B
,
532
539
.
Namias
,
J.
,
1985
:
Some empirical evidence for the influence of snow cover on temperature and precipitation.
Mon. Wea. Rev.
,
113
,
1542
1553
.
Ose
,
T.
,
1996
:
The comparison of the simulated response to the regional snow mass anomalies over Tibet, Eastern Europe, and Siberia.
J. Meteor. Soc. Japan
,
74
,
845
866
.
Parthasarathy
,
B.
, and
S.
Yang
,
1995
:
Relationships between regional Indian summer monsoon rainfall and Eurasian snow cover.
Adv. Atmos. Sci.
,
12
,
143
150
.
Ropelewski
,
C. F.
,
A.
Robock
, and
M.
Matson
,
1984
:
Comments on “An apparent relationship between Eurasian spring snow cover and the advance period of the Indian summer monsoon.”.
J. Appl. Meteor.
,
23
,
341
342
.
Ropelewski
,
C. F.
,
J. E.
Janowiak
, and
M. S.
Halpert
,
1985
:
The analysis and display of real time surface climate data.
Mon. Wea. Rev.
,
113
,
1101
1106
.
Sankar-Rao
,
M.
,
K. M.
Lau
, and
S.
Yang
,
1996
:
On the relationship between Eurasian snow cover and the Asian summer monsoon.
Int. J. Climatol.
,
16
,
605
616
.
Sellers
,
W. D.
,
1965
:
Physical Climatology.
The University of Chicago Press, 272 pp
.
Tan
,
H.
,
Z.
Wang
,
P.
Yu
, and
J.
Zhang
,
1980
:
Observations of precipitation, snow, and evaporation.
Meteorological Observations on Land Surface (in Chinese). Meteorology Press, 146–163. 1980
.
Tao
,
S. Y.
, and
L. X.
Chen
,
1987
:
A review of recent research on the East Asian summer monsoon in China.
Monsoon Meteorology, C. P. Chang and T. N. Krishnamurti, Eds., Oxford University Press, 60–92
.
Vernekar
,
A. D.
,
J.
Zhou
, and
J.
Shukla
,
1995
:
The effect of Eurasian snow cover on the Indian monsoon.
J. Climate
,
8
,
248
266
.
Walker
,
G. T.
,
1910
:
Correlations in seasonal variations of weather.
Memo. India. Meteor. Dept.
,
21
,
22
45
.
Walsh
,
J. E.
,
D. R.
Tucek
, and
M. R.
Peterson
,
1982
:
Seasonal snow cover and short-term climatic fluctuations over the United States.
Mon. Wea. Rev.
,
110
,
272
286
.
Webster
,
P. J.
, and
J.
Keller
,
1975
:
Strong long-period tropospheric and stratospheric rhythm in the Southern Hemisphere.
Nature
,
248
,
212
213
.
Webster
,
P. J.
, and
S.
Yang
,
1992
:
Monsoon and ENSO: Selectively interactive systems.
Quart. J. Roy. Meteor. Soc.
,
118
,
877
926
.
Wei
,
Z-G.
, and
S-W.
Luo
,
1994
:
Influence of snow cover in western China on precipitation in the flood period in China (in Chinese).
Plateau Meteor.
,
14
,
347
354
.
Wu
,
T-W.
, and
Z-A.
Qian
,
2000
:
Further analyses of the linkage between winter and spring snow depth anomaly over Qinghai-Xizang Plateau and summer rainfall of eastern China (in Chinese).
Acta Meteorol. Sin.
,
58
,
570
581
.
Wu
,
T-W.
,
Z-A.
Qian
,
P-J.
Li
,
M-H.
Song
, and
X-B.
Ma
,
1998
:
Some comparative analyses of precipitation over the Northwest China drought area after the Qinghai-Xizang Plateau heavy- and light-snow years (in Chinese).
Plateau Meteor.
,
17
,
364
372
.
Xie
,
P.
, and
P. A.
Arkin
,
1996
:
Analyses of global monthly precipitation using gauge observations, satellite estimates, and numerical model predictions.
J. Climate
,
9
,
840
858
.
Xu
,
G-C.
,
S.
Li
, and
B.
Hong
,
1994
:
The influence of the abnormal snow cover over the Qinghai-Tibet Plateau on Chinese precipitation and atmospheric circulation (in Chinese).
Quart. J. Appl. Meteor.
,
5
,
62
67
.
Yang
,
S.
,
1996
:
ENSO-snow-monsoon associations and seasonal-interannual predictions.
Int. J. Climatol.
,
16
,
125
134
.
Yang
,
S.
, and
L.
Xu
,
1994
:
Linkage between Eurasian winter snow cover and regional Chinese summer rainfall.
Int. J. Climatol.
,
14
,
739
750
.
Yasunari
,
T.
,
A.
Kitoh
, and
T.
Tokioka
,
1991
:
Local and remote responses to excessive snow mass over Eurasia appearing in the northern spring and summer climate—A study with the MRI GCM.
J. Meteor. Soc. Japan.
,
69
,
473
487
.
Yeh
,
T-C.
,
R.
Wetherald
, and
S.
Manabe
,
1983
:
A model study of the short-term climatic and hydrological effects of sudden snow-cover removal.
Mon. Wea. Rev.
,
111
,
1013
1024
.
Zhai
,
P-M.
, and
Q-F.
Zhou
,
1997
:
The change of Northern Hemisphere snow cover and its impact on summer rainfall in China (in Chinese).
Quart. J. Appl. Meteor.
,
8
,
231
235
.
Zwiers
,
F. W.
,
1993
:
Simulation of the Asian summer monsoon with CCC GCM-1.
J. Climate
,
6
,
470
486
.

Footnotes

Corresponding author address: Dr. Tong-Wen Wu, State Key Laboratory of Atmospheric Sciences and Geophysical Fluid Dynamics (LASG), Institute of Atmospheric Physics, Chinese Academy of Sciences, P.O. Box 9804, Lanzhou 100029, China. Email: twwu@lasg.iap.ac.cn