1. Introduction
The South Asian high (SAH) is a planetary-scale anticyclonic system stably located in the upper troposphere and lower stratosphere in the Northern Hemisphere during boreal summer (e.g., Mason and Anderson 1963; Tao and Zhu 1964). Many studies described its intensity variations and location shifts at 100 hPa (Mason and Anderson 1963; Chen and Liao 1990; Zhang and Qian 2000; Zhang and Wu 2001; Zhang et al. 2002; Huang and Qian 2003); however, Wei et al. (2012, 2014, 2015) studied the interannual variation of the SAH at 200 hPa. Only a few studies used the variables at 150 hPa to describe the SAH activities (Liu et al. 2009, 2013; Li et al. 2011). All studies employed a fixed level mainly based on experience.
In the zonal direction, the longitudinal shift of the SAH was first pointed out by Mason and Anderson (1963) and has been treated as the most significant feature of the SAH activities ever since then. More precisely, the SAH was classified into the east pattern, the west pattern, and the zonal (transitional) pattern according to its central location in total geopotential height (Luo et al. 1982). The three patterns have been widely used in previous studies from then on. Zhang et al. (2002) documented the longitudinal bimodality of the SAH including the Tibetan mode (TM) and the Iranian mode (IM) in a climatological perspective and its differences from the east–west shift in synoptic scale. Recently, Wei et al. (2012, 2014, 2015) constructed three SAH indices to elaborate the movements in zonal and meridional directions as well as the combined southeast–northwest movement. Relationships between these indices and summer rainfall anomalies in China and India have also been revealed in their studies.
In early years, Tao and Zhu (1964) point out that the east–west shift of the SAH has a quasi-biweekly time scale. The maintaining period for one mode is about 10–13 days (Luo et al. 1982), which is close to the 13–15-day oscillation period of the streamfunction over the Tibetan Plateau (TP) revealed by Krishnamurti et al. (1973). In recent decades, however, many studies have illustrated the east–west shift of the SAH using datasets with coarse temporal resolution. For example, monthly mean data were employed in some studies (e.g., Huang and Qian 2003; Peng et al. 2010), while seasonal-mean results were also a common choice (e.g., Tan et al. 2005; Wei et al. 2014, 2015). These studies treated the SAH seasonal (or monthly) features as a whole and did not consider the specific activities of the SAH within a month. Recently, on the intraseasonal time scales, Ren et al. (2015) investigated the subseasonal eastward extension of the SAH based on an index defined near its eastern flank. They attributed this movement to the joint role of diabatic heating feedback over eastern Asia and midlatitude wave train. Further, they diagnosed the evolution of the SAH movement. Based on previous studies, the SAH variation near its western flank and its associated relationship with the surface air temperature (SAT) and precipitation patterns over the broader Eurasia need to be explored.
In application, many studies have documented that the summer rainfall anomalies in China have a close relationship with the SAH positions (Chen and Liao 1990; Zhang and Wu 2001; Huang and Qian 2003; Hu et al. 2010a,b; Wei et al. 2012, 2014, 2015; Ren et al. 2015). Furthermore, the variation of the SAH has been proven to exert great influence on Asian–Pacific climate during summer (e.g., Jiang et al. 2011). Some have suggested that the East Asian summer monsoon and the South Asian summer monsoon are connected through the SAH (Wei et al. 2015). On the other hand, the influence of SAH on the Asian–Pacific regional climate is not isolated. In lower latitudes, the western Pacific subtropical high (WPSH) is also a crucial system for Asian–Pacific regional climate. Tao and Zhu (1964) reported that the zonal shift of the SAH leads that of the WPSH for a few days. To a larger extent, the two large systems interact with each other and play an important role in the Asian–Pacific regional climate (Zhao et al. 2000; Zhang et al. 2005; Liu et al. 2006; Ren et al. 2007; Jiang et al. 2011; Ren et al. 2013, 2015). In higher latitudes, some point out the importance of the polar vortex and its circulation characteristics associated with the SAH (Chen and Li 2007, 2008a,b).
Based on previous studies, several issues need to be clarified. First, which level is the best to describe the SAH zonal activities? Second, how many patterns or modes are there to describe the SAH zonal variabilities and how long are their durations? Third, is there a connection between the SAH and other crucial systems nearby at the same levels and what is the relationship between these systems and the anomalous climate pattern in Eurasia? To illuminate these issues, the paper is organized as follows. A brief description of datasets and methods is given in section 2. Two optimum domains and extreme events in each domain are defined and selected in section 3. Four SAH modes are also described in this section. Composite analyses for these events and their respective connections with the climate anomalies over Eurasia are illustrated in section 4. Section 5 gives conclusions and discussion.
2. Data and methods
Two reanalysis products and one precipitation dataset are used in this study. To well capture the zonal features of the SAH, daily mean data are used here. The first reanalysis dataset is the National Centers for Environmental Prediction (NCEP) Reanalysis-1 (NCEP R1) product from 1948 to the present (Kalnay et al. 1996, updated afterward). It has a horizontal resolution of 2.5° × 2.5° in latitude and longitude with 17 standard pressure levels vertically. Geopotential height H, temperature T, and SAT are used in this study. Also, considering the accuracy of the data, only the period from 1980 to 2014 is used here. The second reanalysis dataset is the European Centre for Medium-Range Weather Forecasts (ECMWF) interim reanalysis (ERA-Interim) product (Dee et al. 2011). We used the version with a horizontal resolution of 0.75° × 0.75° with 37 vertical pressure levels. The data are available every 6 h, at 0000, 0600, 1200, and 1800 UTC of each day and are averaged to obtain daily mean values. This dataset is only used to verify the results from the NCEP R1. The precipitation dataset is the Climate Prediction Center (CPC) unified gauge-based analysis of global daily precipitation (Xie et al. 2007; Chen et al. 2008). This product combines all information sources available at CPC and takes advantage of the optimal interpolation (OI) objective analysis technique. Its horizontal resolution is 0.5° × 0.5° and it covers the global land. The retrospective version ranges from 1979 to 2005 and uses about 30 000 stations. The real-time version spans from 2006 to present and uses about 17 000 stations.
3. Extreme events and four SAH modes
a. Optimum 3D domain
The following calculations have been conducted on both of the reanalysis products, and similar SAH features are obtained. This is not surprising because the SAH is the most intense anticyclone in the upper troposphere and lower stratosphere during boreal summer. Thus, only the results from NCEP R1 are illustrated here. One is able to reproduce the same results using the ERA-Interim dataset.
The SAH acts with great seasonality. In May, it moves onto Indo-China from the western equatorial Pacific, and then moves northwestward in June and July (Fig. 1a). The location of the SAH in August is similar to that in July. It shrinks southeastward in September. So, the SAH is predominant from May to September, with the strongest intensity and widest coverage in boreal summer [June–August (JJA)]. So, we focus on the SAH features in JJA.
(a) Climatic 14 300-gpm contours at 150 hPa from May to September (solid lines) and the dashed lines indicating the seasonal track of maximum centers (colored dots) of the SAH. (b) Vertical geopotential height SD peaks (white contours; 10-gpm contour interval) between 400 and 70 hPa and corresponding pressure levels (shading) for JJA from 1981 to 2010. (b) Climatic geopotential heights (black contours; 40-gpm contour interval) at 150 hPa are averaged over JJA; the white shading denotes the area with no SD peaks between 400 and 70 hPa; the red line denotes the mean ridge of the SAH in JJA; the two white dashed lines denote the extension lines of maximal SDs from higher latitudes. The two black boxes denote domain E (25°–30°N, 100°–120°E) and domain W (25°–30°N, 45°–65°E).
Citation: Journal of Climate 29, 22; 10.1175/JCLI-D-15-0823.1
To find the best domain to reveal the SAH zonal activities, standard deviations (SDs) of geopotential height anomalies are obtained in the summers during 1981–2010. Higher SDs indicate larger variability of the SAH. As the SAH is prominent in the upper troposphere and lower stratosphere, Fig. 1b plots the spatial distribution of gridded maximal SDs of height anomalies within 400 and 70 hPa over the whole summer (JJA). All maximal SDs are located between 300 and 100 hPa depending mainly on latitude. The area with white shading in the lower latitudes denotes that the SD peak does not exist between 400 and 70 hPa. In fact, over this area SDs increase with height (figures not shown). In other words, there is no SD peak but rather a monotonic increase within those certain pressure levels. As shown in Fig. 1b, the oval-shaped SAH as outlined by the 14 340-gpm contour occurs above most of the southern part of Asia, with the ridge line (red line) climatologically along 27.5°N. Therefore, three standard latitudes that include 25°, 27.5°, and 30°N are chosen to reveal the zonal activities of the SAH. The body of the SAH mainly covers the longitudes from 30° to 120°E, with its center at around 75°E (Fig. 1b). The primary domain of the SAH is mainly covered by orange and yellow shading, denoting the peak levels of SDs for geopotential height at 150 and 100 hPa, respectively.
Taking the standard latitudes as a reference, the SD peaks only exist in the two flanks of the SAH (Fig. 1b), with no SD peak near its central location. In addition, two SD extension lines (white dashed lines in Fig. 1b) extend to the two SAH flanks from midlatitudes. So, the two SAH flanks are emphasized in this study. To determine the exact locations of the two disturbance centers with relatively higher SDs, daily variations of maximal SDs of geopotential height over 1981–2010 and the corresponding longitudes and pressure levels along 27.5°N are shown in Fig. 2. The SDs of height anomalies in summer are smaller than those in other seasons (figures not shown). Most daily SD centers of geopotential height anomalies (black lines with black dots) are stably located at 150 or 100 hPa (Figs. 2a,b).
Daily variations of maximal geopotential height SD (blue lines, gpm) over 1981–2010 along 27.5°N with the corresponding longitudes (red lines, °E) and pressure levels (black lines, hPa) for the (a) eastern and (b) western centers of the SAH from 1 June to 31 August.
Citation: Journal of Climate 29, 22; 10.1175/JCLI-D-15-0823.1
For the eastern center (Fig. 2a), the maximal SD values (blue line) are stable with small oscillation and the longitudinal shift of its central location (red line) is relatively small. Similar results are obtained at 25° and 30°N in terms of maximal SDs and corresponding longitudes and levels (figures not shown). To cover more positional information, a horizontal grid over 25°–30°N, 100°–120°E is chosen for the following study. For vertical pressure levels, 100 and 150 hPa are nearly of the same importance along 27.5°N (Fig. 2a), so they are both considered with the same weight. Moreover, 100 hPa (150 hPa) is the main level of maximal SDs at 25°N (30°N) (figures not shown), which is also consistent with Fig. 1b. In summary, the proportion for different longitudes and latitudes is equal while the levels are chosen at 100 hPa (150 hPa) along 25°N (30°N) and both 100 and 150 hPa along 27.5°N for all nine zonal grids from 100° to 120°E. The zone above composes the crucial eastern domain (domain E) to describe the SAH zonal activities [the eastern box in Fig. 1b; see Eq. (3) for reference].
The SD peaks of geopotential height near the western flank mainly range from 20 to 80 gpm with a larger east–west shift of its peak longitudes (Fig. 2b). Some missing points exist in Fig. 2b because the SD extrema near the SAH western flank disappear occasionally. Most height SD centers are located within 45° and 65°E (nine zonal grids) in longitude and within 150 and 100 hPa vertically. The results along 25° and 30°N (figures not shown) for the western flank are nearly the same, except the most important level at 30°N is only 150 hPa. Based on the above, a horizontal grid over 25°–30°N, 45°–65°E is chosen for the following study. The proportion for different longitudes and latitudes is equal. In the vertical direction, pressure levels 100 and 150 hPa are chosen for 25° and 27.5°N, whereas only 150 hPa is chosen for 30°N [see Eq. (4) for reference]. So, the western domain (domain W; the western box in Fig. 1b) is determined to measure the SAH zonal activities in its western part.
The two 3D domains defined above are hereafter referred to as domain E and domain W. They are optimum to capture the SAH zonal variability and not fixed on a certain level. The zonal range of domain W is larger than that of the IM defined by Zhang et al. (2002), which lies between 55° and 65°E, and domain E is located to the east of their TM range over 82.5°–92.5°E. Recently, two indices occupying the longitudes 55°–75°E and 85°–105°E as well as 50°–80°E and 85°–115°E were defined by Wei et al. (2014, 2015). The zonal ranges defined by Wei et al. (2014, 2015) were based on the SAH locations from total geopotential height while our study determines the two domains by height SDs, which indicate the SAH variability. Similarly, the zonal range of domain E is the same as that of the SAH eastward extension index defined by Ren et al. (2015), only with a smaller meridional range compared with theirs (22.5°–32.5°N).
b. Extreme events
Persistent (equal to or longer than 10 days) and extreme [larger (smaller) than 1.0 (−1.0) SD] events including their years, start dates, durations, average intensities, and order for summers (JJA) from 1981 to 2014 for domain E. The orders are based on their average NAs. The strongest events are in boldface.
For domain E (Table 1), 13 positive events lasted for 10–23 days with average intensities of 1.26–1.79 SD, while 7 negative events occurred with average intensities from −1.40 to −1.98 SD. The longest event lasted for 36 days. In the western domain, 10 positive events persisted for 10–22 days with average intensities of 1.21–1.98 SD (Table 2). The durations of the 11 negative events ranged from 10 to 24 days with higher intensities than those of the positive events.
In domain E (Table 1), positive events were less common in the 1990s, with the same number in the 1980s and 2000s. There were more negative cases in the 2000s while no negative event occurred in the 1980s. In domain W (Table 2), five positive events but only one negative event occurred in the 1980s. In the 1990s, however, there were three positive events and seven negative events. The number of events in the 2000s was similar between positive and negative ones. The distribution for different months is uniform. In addition, most SAH events lasted for less than one month, leading to their intraseasonal feature of circulation anomalies (Tables 1 and 2). Therefore, these extreme events are random phenomena and it is reasonable to use the daily mean variables in this study instead of monthly mean or seasonal-mean variables.
The strongest positive and negative events in each domain are highlighted in Tables 1 and 2. Their time series are plotted in Fig. 3. In domain E, the strongest positive case (3.03 SD; Fig. 3b) and the strongest negative case (−2.62 SD; Fig. 3d) peaked on 25 August 1981 and 24 June 1997, respectively. For domain W, the strongest positive case (Fig. 3a) and the strongest negative case (Fig. 3c) occurred with their peaks (2.85 and −3.32 SD) on 29 June 1998 and 27 July 2000, respectively.
Normalized series of daily geopotential height anomalies in JJA for the four strongest events. (a) The strongest positive case for domain W in 1998. (c) The strongest negative case for domain W in 2000. (b),(d) As in (a),(c), but for corresponding cases for domain E in 1981 and 1997, respectively. The thick solid lines with dots indicate the event processes. The numbers in parentheses in boldface and lightface font denote the strongest and average intensities of the events, respectively. In each panel, the two thick dashed horizontal lines denote ±1.0 SD as the criteria for event identification.
Citation: Journal of Climate 29, 22; 10.1175/JCLI-D-15-0823.1
c. Four modes of the SAH
Composite total, climatic, and anomalous geopotential heights for all the positive and negative cases in the two domains are presented in Fig. 4. For the positive composite in domain W (Fig. 4a), a positive center (white H) of height anomalies is identified near this domain and the SAH strengthens with its center (black H) moving westward over the Iranian Plateau (IP) compared with its climatic location (red H). Moreover, most area plotted is covered by positive anomalies over the subtropical region, indicating the consistent variations of geopotential heights in planetary scale. The SAH center shows a preference for the largest positive anomalies. In the negative composite for domain W (Fig. 4c), the SAH center (black H) located over the TP is much weaker in intensity, compared to its climatic center (red H). The SAH shrinks and moves over the TP as a result of the negative center (white L) near domain W (Fig. 4c). Similarly in domain E, an anomalous positive center (white H) superimposes the climatic heights and forms a strong SAH with its center over the TP for the positive composite (Fig. 4b). The SAH (black H) shifts eastward and is located near the anomalous high center (white H). Conversely, a negative center (white L) near domain E weakens the climatic SAH (red H) and forms a weak SAH (black H) over the IP in the negative composite for domain E (Fig. 4d). In fact, the anomalies in the two active domains act as disturbances and take major responsibilities for SAH variation. In summary, four SAH modes are detected to reveal zonal variability of the SAH, showing different characteristics in location and intensity. They are named the strong IM (SIM; Fig. 4a), strong TM (STM; Fig. 4b), weak TM (WTM; Fig. 4c), and weak IM (WIM; Fig. 4d), respectively. The four SAH modes are persistent zonal modes, which are similar to the east and west patterns of Luo et al. (1982) in location and duration.
Composites of geopotential height anomalies (shading, 10-gpm interval), total geopotential heights (black contours; 20-gpm contour interval, 14 360-gpm contour labeled), and corresponding climatic geopotential heights (red contours; 20-gpm contour interval, 14 360-gpm contour labeled) denoting the SAH at 150 hPa. Composites of (a) positive and (c) negative events in domain W. (b),(d) As in (a),(c), but for domain E. Black letter H denotes the center of the SAH in original (total) field, and the red letter H is for the climatic field. White letter H (L) indicates the anomalous high (low) centers.
Citation: Journal of Climate 29, 22; 10.1175/JCLI-D-15-0823.1
4. Composite analyses and climatic connection over Eurasia
a. Composite analyses for domain E
It is conspicuous that the four SAH modes are actually associated with the four centers of anomalous geopotential height at the upper troposphere near domain E and domain W (Fig. 4). We further examine the composite vertical structures of geopotential height and temperature anomalies for domain E, from the troposphere to the lower stratosphere (Fig. 5). It should be noted that all the composite analyses in the following are the results averaged over the whole duration of each case, with no emphasis on their evolution. Vertical structures along 27.5°N for the positive and negative composites in domain E are shown in Figs. 5a and 5c, respectively. When a STM occurs, a warm center (W1) below and a cold center (C1) above are separated by a positive center (H1) of height anomalies at the upper troposphere around 150–100 hPa (Fig. 5a). The H1 mainly ranges from 100° to 120°E, indicating the maximum height anomalies are over the domain E. Positive height anomalies are prevalent over the whole hemisphere in the planetary scale. Opposite features for both height and temperature anomalies can be observed in the negative composite (WIM; Fig. 5c). Composite results for geopotential height and temperature anomalies along 110°E are also shown in Figs. 5b and 5d. A dipolar height pattern including a positive center (H1) near 30°N and a negative center (L2) near 50°N is noted in the meridional direction (Fig. 5b). The H1 separates a warm center (W1) below and a cold center (C1) above, while the L2 separates a cold center (C2) below and a warm center (W2) above. The vertical location of H1 is around 150 hPa, while L2 is lower and located near 300 hPa. A reversed and significant dipole is evident in the negative composite for domain E (Fig. 5d). The height and temperature anomalies satisfy the hydrostatic balance. It means that the temperature anomalies from the troposphere to stratosphere can be derived from the height anomalies. It can be concluded that the SAT anomalies are the downward extension of temperature anomalies below the height anomalies at the upper troposphere. Generally, regression patterns against daily height NA series in domain E reconfirm the results above in both vertical sections (Figs. 5e,f). In addition, geopotential height and temperature anomalies in the negative composite are stronger than those in the positive composite for domain E, which is consistent with the intensities of extreme events (Table 1).
Composite vertical cross sections of geopotential height anomalies (red contours for positive and blue contours for negative; 20-gpm contour interval) and temperature anomalies (white solid contours with shading; 1°C contour interval) along (a) 27.5°N and (b) 110°E for positive events in domain E. (c),(d) As in (a),(b), but for the composites for negative events. Black slashes for height and dots for temperature in (a)–(d) indicate the composite differences exceeding 95% confidence level. (e) Regressed geopotential height anomalies (red contours for positive and blue contours for negative; 10-gpm contour interval) and temperature anomalies (shading with white contours; 1°C contour interval) along 27.5°N against daily height NA series in domain E. (f) As in (e), but along 110°E. Black slashes for height and dots for temperature indicate the areas exceeding 95% confidence level. Letters H and L (with numbers) denote the high and low centers of geopotential height anomalies, and letters W and C (with numbers) denote the warm and cold centers of temperature anomalies, respectively.
Citation: Journal of Climate 29, 22; 10.1175/JCLI-D-15-0823.1
Composite horizontal distribution of geopotential height anomalies for cases in domain E is drawn for 150 hPa (Figs. 6a,b), which corresponds to the STM (Fig. 4b) and the WIM (Fig. 4d) of the SAH. As shown in the positive composite (Fig. 6a), three positive centers (Ap, Bp, and Dp) of height anomalies are separated by a negative center (Cn). The positive centers Ap, Bp, and Dp are located mainly over western Russia and Kazakhstan, far eastern Russia, and southern China, respectively, whereas the negative center (Cn) mainly covers Mongolia and northern China. The center Dp captures well the center H1 in Figs. 5a and 5b, leading to the STM of the SAH. Inverse distribution of anomalous centers of geopotential height appears in the negative composite (Fig. 6b). Three negative centers (An, Bn, and Dn) separated by a positive center (Cp) are clearly observed in Fig. 6b, and Dn is the crucial reason for the WIM of the SAH. The regression pattern (Fig. 6c) shows great consistency with the composite results (Figs. 6a,b). The wave train in the midlatitudes and the dipolar pattern in the meridional direction over East Asia are clear in Fig. 6, connecting the SAH with systems in higher latitudes. The wave train pattern was also revealed in Ren et al. (2015). Differently, we have not plotted the composite propagation of the wave train because the three anomalous systems in the wave train vary largely among different cases (figures not shown).
(a) Positive composite of geopotential height anomalies (red contours for positive and blue contours for negative; 20-gpm contour interval) at 150 hPa for domain E. (b) As in (a), but for negative composite. Dots in (a) and (b) denote the composite differences exceeding 95% confidence level. (c) Regressed geopotential height anomalies (red contours for positive and blue contours for negative; 10-gpm contour interval) at 150 hPa against daily height NA series in domain E. Dots in (c) indicate the areas exceeding 95% confidence level. Letters A, B, C, and D denote four areas, appended with p (n) for positive (negative) height anomalies. The dashed lines are along 110°E.
Citation: Journal of Climate 29, 22; 10.1175/JCLI-D-15-0823.1
Figures 7a and 7b show the composite SAT distributions of all the positive and negative cases for domain E using the NCEP R1, respectively. When the STM occurs (Fig. 7a) three warm centers (Aw, Bw, and Dw) are separated by a cold center (Cc) on the surface that corresponds well to the anomalous geopotential height pattern in Fig. 6a. In the negative composite (Fig. 7b) a warm center (Cw) is surrounded by three cold centers (Ac, Bc, and Dc) that also matches the horizontal pattern in Fig. 6b. In China, a dipolar pattern with a warm center in the south and a cold center in the north is observed in the STM (Fig. 7a), and this is reversed in the WIM (Fig. 7b). The opposite SAT patterns result from the opposite extensions of temperature anomalies in the troposphere (Figs. 5 and 6). The SAT anomalies in southern China are directly influenced by the SAH variations in intensity and location. Other anomalous SAT centers are directly affected by the corresponding local temperature anomalies aloft in the troposphere. Figure 7c plots the SAT differences between the composites of positive and negative cases. The four anomalous temperature centers (Aw, Bw, Cc, and Dw) all reach the 95% confidence level in Fig. 7c, showing significant SAT differences between positive (Fig. 7a) and negative (Fig. 7b) composites. To further verify the spatial patterns of SAT anomalies, correlations between daily geopotential height NA series in domain E and daily SAT anomalies in summers from 1981 to 2010 are calculated (Fig. 7d). The four centers in Fig. 7d (Ap, Bp, Cn, and Dp) are robust and significant exceeding 95% confidence level.
Composites of SAT anomalies (red contours with yellow–orange shading for positive and blue contours with green shading for negative; 0.4°C contour interval) at sigma level 0.995 for (a) positive and (b) negative events in domain E. (c) The differences between (a) and (b) (shading and contours; 0.8°C contour interval). Black and magenta dots in (c) denote the areas exceeding 95% confidence level of Student’s t test. (d) Correlation coefficients (shading) between daily geopotential height NA series in domain E and SAT anomalies for JJA from 1981 to 2010. Shaded areas in (d) indicate correlations exceeding the 95% confidence level. Letters A, B, C, and D denote four areas with warm (w) or cold (c) and positive (p) or negative (n) correlations. Thick gray contours indicate the Tibetan Plateau region with topography exceeding 3000 m.
Citation: Journal of Climate 29, 22; 10.1175/JCLI-D-15-0823.1
Composite distributions of precipitation anomalies for positive and negative cases in domain E are shown in Figs. 8a and 8b, respectively. For the positive composite (STM of the SAH), three dry centers (Ad, Bd, and Dd) isolated by a rainy center (Cr) are observed in Fig. 8a at the same locations as the ones in Figs. 6a and 7a. In addition, another dry center (Ed) appears over the southern foothills of the TP. For the WIM, a reversed pattern of precipitation anomalies can be found (Fig. 8b). Further, differences of the precipitation anomalies between Figs. 8a and 8b are calculated in Fig. 8c. Five precipitation centers (Ad, Bd, Cr, Dd, and Ed) can also be identified by the significant points exceeding the 95% confidence level. Again, a correlation map reconfirms the spatial pattern of precipitation anomalies along with the SAH variations simultaneously (Fig. 8d). The correlation with precipitation is weaker than that of temperature. An important reason is that the former is affected by other factors, such as moisture. Interestingly but not surprisingly, the three centers of precipitation features (C, D, and E in Fig. 8) resemble the dipolar distribution over eastern China and the distribution over the southern foothills of the TP, as shown in Figs. 5b and 5e of Ren et al. (2015).
Composites of precipitation percentage anomalies (red contours with red shading for negative and blue contours with blue shading for positive; 20% contour interval) for (a) positive and (b) negative events in domain E. (c) The differences between (a) and (b) (shading and contours; 40% contour interval). Black dots in (c) denote the areas exceeding 95% confidence level of Student’s t test. (d) Correlation coefficients (shading) between daily geopotential height NA series in domain E and precipitation percentage anomalies for JJA from 1981 to 2010. Shaded areas in (d) indicate correlations exceeding the 95% confidence level. Letters A, B, C, D, and E denote five areas with dry (d) or rainy (r) and positive (p) or negative (n) correlations. Thick gray contours indicate the Tibetan Plateau region with topography exceeding 3000 m.
Citation: Journal of Climate 29, 22; 10.1175/JCLI-D-15-0823.1
b. Composite analyses for domain W
Similar composite analyses are conducted for domain W. Figures 9a and 9c display the vertical structures along 27.5°N for the positive and negative composites in domain W, respectively. For the SIM, a warm center (W1) and a cold center (C1) of temperature anomalies are located vertically below and above the positive center of geopotential height anomalies (H1) at the upper troposphere around 100 hPa (Fig. 9a). Positive height anomalies are prominent between 500 and 50 hPa, with only one positive height center over IP at 100 hPa. Conversely, the opposite configuration is observed from the negative composite of geopotential height and temperature anomalies (Fig. 9c). Moreover, the amplitude of geopotential height and temperature anomalies is stronger with an anomalous low center (L1) near 150 hPa.
As in Fig. 5, but for domain W, and (b),(d),(f) are along 55°E.
Citation: Journal of Climate 29, 22; 10.1175/JCLI-D-15-0823.1
In the latitude–pressure cross section along 55°E (Fig. 9b), the positive center (H1) and a negative center (L2) form a dipolar pattern in the positive composite of domain W. The H1 of geopotential height anomalies near 32.5°N separates the warm center (W1) below and the cold center (C1) above. To the contrary, an anomalous low center (L2) exists around 55°N with opposite temperature anomalies. The vertical location of H1 is around 150 hPa, while L2 is near 300 hPa. A reversed but comparable configuration is clear in the negative composite for domain W (Fig. 9d). All centers of geopotential height and temperature anomalies satisfy the hydrostatic balance. Figures 9a–d clearly show that the SAT anomalies are the downward extension of temperature anomalies below the height anomalous centers at the upper troposphere. To further confirm this vertical structure, regression patterns against daily height NA series in domain W are calculated (Figs. 9e,f). It is clear that most areas near domain W are statistically significant in the troposphere and lower stratosphere, showing similar structure as the extreme events demonstrated.
Figure 10 depicts the composite horizontal structures at 150 hPa for domain W. When the SIM occurs, four large-scale systems of geopotential height anomalies can be found (Ap, Bp, Cn, and Dp; Fig. 10a). The three positive centers, Ap located in western Europe, Bp located mainly over central–eastern Russia, and Dp located in the vicinity of domain W, surround a negative center (Cn) over Kazakhstan and western Siberia. They display a wave train over Eurasia in the midlatitudes and a dipolar pattern near 55°E in the meridional direction, indicating the simultaneous relationship between the geopotential height anomalies in the subtropics and midlatitudes. Once again, an opposite pattern can be discovered in the negative composite for domain W (Fig. 10b) that matches the WTM (Fig. 4c) above. The four anomalous centers (An, Bn, Cp, and Dn) with opposite signs exist at similar locations. In addition, the regression pattern against the daily height NA series in domain W is significant over the same areas (Fig. 10c). Therefore, the anomalous pattern at 150 hPa is robust, including the wave train in the midlatitudes and a dipolar pattern near 55°E in the meridional direction.
As in Fig. 6, but for domain W, and the dashed lines are along 55°E.
Citation: Journal of Climate 29, 22; 10.1175/JCLI-D-15-0823.1
Again, to identify the possible connection between the above geopotential height and temperature anomalies and the surface temperature extremes, the anomalous SAT patterns corresponding to the positive and negative composites in domain W are illustrated (Figs. 11a,b). For the positive composite (SIM; Fig. 11a), three warm centers (Aw, Bw, and Dw) and one cold center (Cc) can be found easily. The four temperature centers are consistent with the four centers of anomalous geopotential height in Fig. 10. As indicated by Figs. 9a and 9b, the warm surface center (Dw; Fig. 11a) is the downward extension of the warm center (W1) in the upper troposphere, which matches well with H1. The anomalous geopotential height–temperature configurations of other three temperature centers in Fig. 11a also agree with this vertical structure (figures not shown). In contrast, three cold centers (Ac, Bc, and Dc) and one warm center (Cw) exist in the negative composite of domain W (WTM; Fig. 11b). The differences (Fig. 11c) confirm the robustness of the four centers by Student’s t test, consistent with the wave train and dipole in the anomalous geopotential height field at 150 hPa (Fig. 10). Generally, three positive centers (Ap, Bp, and Dp) and one negative center (Cn) stand out significantly in the correlation map between daily height NAs in domain W and daily SAT series over Eurasia (Fig. 11d).
As in Fig. 7, but for domain W.
Citation: Journal of Climate 29, 22; 10.1175/JCLI-D-15-0823.1
Further, the relationship between the SAH zonal variation in domain W and precipitation anomalies over Eurasia is illustrated using composite and correlation analyses (Fig. 12). The positive composite shows three centers (Ad, Bd, and Cr) in the midlatitudes (Fig. 12a), which correspond well with the centers of anomalous geopotential height (Ap, Bp, and Cn in Fig. 10a) and SAT (Aw, Bw, and Cc in Fig. 11a). Moreover, another rainy center (Er) is obvious over the southern foothills of the TP due to the negative height anomalies in this region (Fig. 9a). For negative composite (Fig. 12b), opposite centers (Ar, Br, Cd, and Ed) appear in the same locations as those in the positive composite (Fig. 12a). These four centers of precipitation anomalies are statistically significant under the Student’s t test of the differences between positive and negative composites (Fig. 12c). For further detection, the four centers (An, Bn, Cp, and Ep) can also be distinguished in the correlation map (Fig. 12d). However, the centers in the midlatitudes (An, Bn, and Cp) are much weaker, while the Ep in South Asia becomes dominant. It is interesting that a negative center (Dn) only appears in the correlation map. Dry climatic features over this region (the Middle East) may lead to the disappearance of this center in Figs. 12a–c, although it has a significant negative correlation with geopotential height anomalies.
As in Fig. 8, but for domain W.
Citation: Journal of Climate 29, 22; 10.1175/JCLI-D-15-0823.1
c. Original signals of anomalous SAH systems
The centers of geopotential height anomalies can be traced day by day through anomalous geopotential height maps until they cannot be identified. This is a back-tracing approach following the daily movement of maximum (or minimum) centers of anomalous geopotential height (Qian et al. 2015, 2016). Based on this, the original locations and leading days of anomalous geopotential height centers as well as their locations on the peak days for each SAH mode are illustrated in Fig. 13. For each positive (negative) event, the peak day is defined by the day with the largest (smallest) geopotential height NAs. For the 10 positive events in the domain W, both original locations and locations on peak days were scattered (Fig. 13a). Two distant cases originated near Spain and the Baltic Sea leading at 10 and 16 days, respectively. However, most negative centers of geopotential height anomalies finally concentrated near the domain W (Fig. 13c). They emerged mainly from the west and the north. Two remote cases initiated disturbances from northern Europe at 6 and 4 days before the peak days of each case. For cases over domain E, most centers of anomalous geopotential height are concentrated in central and eastern China on the peak days (Figs. 13b,d). Most signals came from the west or the north and they originated from China or the countries nearby. One exception arose east of Japan leading at 9 days and matured over the East China Sea, while the other originated near Turkey leading at 12 days and moved over China (Fig. 13b). Further, in order to clarify the anomalous systems related to the SAH in detail, case enumeration is used on the strongest positive and negative cases in domain E (Figs. 13e,f). It is observed that all the anomalous systems showed complexity in their tracks. Some were stable and motionless, while others originated from distant areas and moved in different ways. Their leading days spanned 3–13 days. The tracks of anomalous systems in other cases were also varied (figures not shown). Therefore, based on the particularity of each case and each anomalous system, composite event evolution is not carried out here.
The lines (solid for the strongest case and dashed for other cases) connecting the original central locations (dots) of height anomalies and their central locations on the peak days (circles) with leading days (numbers beside dots) for all the events for the (a) positive and (b) negative centers related to the SAH near domain W. (b),(d) As in (a),(c), but for domain E. The geopotential height anomalies (solid contours for positive and dashed contours for negative; 40-gpm contour interval) at 150 hPa on (e) 25 Aug 1981 and (f) 24 Jun 1997, the peak days during the strongest positive and negative events in domain E, respectively. The largest solid circles (squares) indicate the original locations of positive (negative) height centers. The tracks of anomalous centers near domain E are denoted by the lines with solid circles in (e) and solid squares in (f) with black edges, respectively. The lines with hollow circles (squares) indicate the central tracks of other positive (negative) centers. The small numbers beside each large solid circles or squares indicate the leading days. Letters H and L indicate the high and low anomalous centers, respectively.
Citation: Journal of Climate 29, 22; 10.1175/JCLI-D-15-0823.1
5. Conclusions and discussion
a. Conclusions
Through this study, several issues about the SAH zonal variability and surface climate anomalies over Eurasia have been clarified. First, the optimal vertical levels to describe the SAH zonal activities are 150 and 100 hPa, while the maximum height anomaly (MHA; Qian et al. 2016) related to the SAH appears at 150 hPa more frequently. Based on the analysis of variance, domain E (25°–30°N, 100°–120°E) over central-eastern China and domain W (25°–30°N, 45°–65°E) over the Middle East at the above two pressure levels are chosen. Considering the spatial and temporal scales of the SAH activities, a positive event is defined when the height NAs are greater than 1.0 (less than −1.0) SD for at least 10 days. They are ranked according to their average NAs. In domain E, 13 positive and 7 negative cases are selected and centered mainly over China. In domain W, 10 positive and 11 negative cases are detected. The ranges of duration are similar in both domain E and domain W, lasting for 10–36 days. However, the intensities of negative events are stronger than those for the positive ones.
Second, it is well known that the SAH is an intense anticyclone stably situated in the upper troposphere and lower stratosphere in the Northern Hemisphere during the boreal summer. According to the locations of total geopotential heights, different modes have been named and widely documented in previous studies. Based on the analyses of daily anomalies, four SAH modes can be classified for both location and intensity: the strong Iranian mode (SIM), strong Tibetan mode (STM), weak Tibetan mode (WTM), and weak Iranian mode (WIM). Different zonal modes are primarily caused by anomalous geopotential height centers near the two flanks of the SAH. In short, for the second issue, four SAH modes exist in nature and their durations are mostly shorter than one month based on analyses of extreme SAH events.
For the third issue, a zonal wave train in the midlatitudes and a meridional dipole of geopotential height anomalies at 150 hPa are statistically significant and robust in both domain E and domain W, establishing a connection between the SAH and the wave train in the midlatitudes. For domain E, the three anomalous highs (lows) at 150 hPa denoting their surface warm and dry (cold and wet) anomalies situated mainly over western Russia and Kazakhstan, far eastern Russia, and southern China are separated by the opposite center over Mongolia and northern China. For domain W, three anomalous highs (lows) at 150 hPa indicating their surface warm and dry (cold and wet) anomalies, located over western Europe, central–eastern Russia, and the vicinity of domain W, surround the anomalous low (high) with its surface cold and wet (warm and dry) anomalies over Kazakhstan and western Siberia. Taking all cases into consideration, each anomalous center displays individual features in the locations and tracks. For each positive (negative) center of geopotential height anomalies in the upper troposphere, a cold center and a warm center are always located above and below it, respectively. The temperature and geopotential height anomalies satisfy the hydrostatic balance vertically, meaning that the temperature anomalies in the troposphere and stratosphere can be derived from the geopotential height anomalies. The precipitation anomalies over the southern foothills of the TP (center E in Figs. 8 and 12) are also remarkable in the two domains. One exception is that the anomalous precipitation center near domain W is not understood (Fig. 12) and is probably a result of the dry climatology there. The SAT anomalies are directly associated with temperature anomalies aloft in the mid-to-upper troposphere while the precipitation anomalies are closely related to the geopotential height anomalies vertically. Then, the third issue is clarified, emphasizing the wavelike circulation related to the SAH and the joint connection with the anomalous climate pattern over Eurasia.
b. Discussion
In current practice and previous studies, atmospheric variables are limited to certain pressure levels based on a forecasters’ experience. For example, temperature at 850 hPa and geopotential height at 500 hPa are widely used from analyses and forecasts. However, as shown in this study, the anomalous patterns in the upper troposphere are crucial in indicating not only the circulation anomalies but also the surface climate anomalies. So the forecasts of variables in the upper troposphere should be treated with more consideration.
In the present study, we only illustrate the simultaneous connection between the SAH and the wave train in the midlatitudes, as well as their joint connection with the anomalous climate pattern over Eurasia. However, the causality and underlying mechanisms among them are not well demonstrated here. Some instructive studies reported the relationship between the El Niño–Southern Oscillation (ENSO) phenomenon and SAH variation (Peng et al. 2009, 2010; Xue et al. 2015). Interestingly, the strongest positive case over domain W occurred in 1998 over a strong La Niña period, whereas a strong negative case over domain W appeared in 1997 over the strongest El Niño on record. But taking all cases in domain E into consideration, four positive cases were during El Niño conditions, four positive cases were during La Niña conditions, and five positive cases were during neutral conditions. For negative cases, five were during El Niño condition, one occurred during a La Niña period, and one occurred during neutral condition. Therefore, there may be some connection between the anomalous SAH events and the ENSO phenomenon, but this needs further investigation. Moreover, the sea surface temperature anomalies over the Indian Ocean also prove to be crucial for the SAH variation (Yang and Liu 2008; Wei et al. 2012; Huang et al. 2011; Xue et al. 2015). On the other hand, it is likely that the winter–spring snow over Eurasia impacts the climate pattern locally via the SAH (Kripalani and Kulkarni 1999; Mamgain et al. 2010). These are crucial topics and deserve deeper investigation in the future. Based on this insight, numerical simulations and longer data are necessary for further studies.
Acknowledgments
The authors thank the three anonymous reviewers for their constructive comments for this manuscript, which lead to the great improvements of it. We also appreciate the Physical Sciences Division (PSD), CPC, and ECMWF for providing the valuable data. This work is supported by the National Natural Science Foundation of China (41375073 and 41221064), the Strategic Priority Research Program of the Chinese Academy of Sciences (XDA05090407), and the Global Change and Air–Sea Interaction Program (GASI-03-02-01-02).
REFERENCES
Chen, G., and Q. Liao, 1990: Relationship between the location features of 100 hPa South Asia high and mid-summer rainfall in China (in Chinese). Plateau Meteor., 9, 432–438.
Chen, M., W. Shi, P. Xie, V. B. S. Silva, V. E. Kousky, R. W. Higgins, and J. E. Janowiak, 2008: Assessing objective techniques for gauge-based analyses of global daily precipitation. J. Geophys. Res., 113, D04110, doi:10.1029/2007JD009132.
Chen, Y., and Y. Li, 2007: Relationships between polar vortex and oscillation of South Asia high in east and west directions (in Chinese). Plateau Meteor., 26, 1067–1076.
Chen, Y., and Y. Li, 2008a: Advances in researches on activities and influences of South Asia high on weather and climate (in Chinese). Mater. Sci. Technol., 36, 174–179.
Chen, Y., and Y. Li, 2008b: The variations of polar vortex and South Asia high and characters of northern atmosphere circulations at 100 hPa (in Chinese). J. Trop. Meteor., 24, 519–526.
Dee, D. P., and Coauthors, 2011: The ERA-Interim reanalysis: Configuration and performance of the data assimilation system. Quart. J. Roy. Meteor. Soc., 137, 553–597, doi:10.1002/qj.828.
Hu, J., L. Tao, and B. Zhou, 2010a: Characteristic of South Asia high activity and its relation with the precipitation of east China in summer (in Chinese). Plateau Meteor., 29, 128–136.
Hu, J., B. Zhou, and L. Tao, 2010b: Comparative analysis of the relation between characteristic parameters of South Asia high and summer precipitation of China (in Chinese). Meteor. Mon., 36, 51–56.
Huang, G., X. Qu, and K. Hu, 2011: The impact of the tropical Indian Ocean on South Asian high in boreal summer. Adv. Atmos. Sci., 28, 421–432, doi:10.1007/s00376-010-9224-y.
Huang, Y., and Y. Qian, 2003: Relationships between South Asian high and summer rainfall in North China (in Chinese). Plateau Meteor., 22, 602–607.
Jiang, X., Y. Li, S. Yang, and R. Wu, 2011: Interannual and interdecadal variations of the South Asian and western Pacific subtropical highs and their relationships with Asian-Pacific summer climate. Meteor. Atmos. Phys., 113, 171–180, doi:10.1007/s00703-011-0146-8.
Kalnay, E., and Coauthors, 1996: The NCEP/NCAR 40-Year Reanalysis Project. Bull. Amer. Meteor. Soc., 77, 437–471, doi:10.1175/1520-0477(1996)077<0437:TNYRP>2.0.CO;2.
Kripalani, R. H., and A. Kulkarni, 1999: Climatology and variability of historical Soviet snow depth data: Some new perspectives in snow–Indian monsoon teleconnections. Climate Dyn., 15, 475–489, doi:10.1007/s003820050294.
Krishnamurti, T. N., S. M. Daggupaty, J. Fein, M. Kanamitsu, and J. D. Lee, 1973: Tibetan high and upper tropospheric tropical circulation during northern summer. Bull. Amer. Meteor. Soc., 54, 1234–1239.
Li, C., L. Li, and Y. Tan, 2011: Structure of South Asia high in the stratosphere and influence of ENSO. J. Trop. Meteor., 17, 293–301, doi:10.3969/j.issn.1006-8775.2011.03.011.
Liu, B., J. He, and L. Wang, 2009: Characteristics of the South Asia high establishment processes above the Indo-China Peninsula from April to May and their possible mechanism (in Chinese). Chin. J. Atmos. Sci., 33, 1319–1332.
Liu, B., G. Wu, J. Mao, and J. He, 2013: Genesis of the South Asian high and its impact on the Asian summer monsoon onset. J. Climate, 26, 2976–2991, doi:10.1175/JCLI-D-12-00286.1.
Liu, H., S. Zhao, C. Zhao, and Z. Lu, 2006: Weather abnormal and evolutions of western Pacific subtropical high and South Asian high in summer of 2003 (in Chinese). Plateau Meteor., 25, 169–178.
Luo, S., Z. Qian, and Q. Wang, 1982: The climatic and synoptical study about the relation between the Qinghai-Xizang high pressure on the 100-mb surface and the flood and drought in east China in summer (in Chinese). Plateau Meteor., 1, 1–10.
Mamgain, A., S. K. Dash, and P. P. Sarthi, 2010: Characteristics of Eurasian snow depth with respect to Indian summer monsoon rainfall. Meteor. Atmos. Phys., 110, 71–83, doi:10.1007/s00703-010-0100-1.
Mason, R. B., and C. E. Anderson, 1963: The development and decay of the 100-mb summertime anticyclone over southern Asia. Mon. Wea. Rev., 91, 3–12, doi:10.1175/1520-0493(1963)091<0003:TDADOT>2.3.CO;2.
Peng, L., Z. Sun, D. Ni, H. Chen, and G. Tan, 2009: Interannual variation of summer South Asia high and its association with ENSO (in Chinese). Chin. J. Atmos. Sci., 33, 783–795.
Peng, L., Z. Sun, H. Chen, and D. Ni, 2010: The persistent anomaly of the South Asia high and its association with ENSO events (in Chinese). Acta Meteor. Sin., 68, 855–864.
Qian, W., Y. Chen, M. Jiang, and Q. Hu, 2015: An anomaly-based method for identifying signals of spring and autumn low-temperature events in the Yangtze River valley, China. J. Appl. Meteor. Climatol., 54, 1216–1233, doi:10.1175/JAMC-D-14-0240.1.
Qian, W., T. Yu, and J. Du, 2016: A unified approach to trace surface heat and cold events by using height anomaly. Climate Dyn., 46, 1647–1664, doi:10.1007/s00382-015-2666-2.
Ren, R., Y. Liu, and G. Wu, 2007: Impact of South Asia high on the short-term variation of the subtropical anticyclone over western Pacific in July 1998 (in Chinese). Acta Meteor. Sin., 65, 183–197.
Ren, X., X.-Q. Yang, and X. Sun, 2013: Zonal oscillation of western Pacific subtropical high and subseasonal SST variations during Yangtze persistent heavy rainfall events. J. Climate, 26, 8929–8946, doi:10.1175/JCLI-D-12-00861.1.
Ren, X., D. Yang, and X.-Q. Yang, 2015: Characteristics and mechanisms of the subseasonal eastward extension of the South Asian high. J. Climate, 28, 6799–6822, doi:10.1175/JCLI-D-14-00682.1.
Tan, J., H. Yang, S. Sun, and P. Wang, 2005: Characteristics of the longitudinal oscillation of South Asia high during summer (in Chinese). J. Nanjing Inst. Meteor., 28, 452–460.
Tao, S., and F. Zhu, 1964: The 100-mb flow patterns in southern Asia in summer and its relation to the advance and retreat of the west-Pacific subtropical anticyclone over the Far East (in Chinese). Acta Meteor. Sin., 34, 385–396.
Wei, W., R. Zhang, and M. Wen, 2012: Meridional variation of South Asian high and its relationship with the summer precipitation over China (in Chinese). J. Appl. Meteor. Sci., 23, 650–659.
Wei, W., R. Zhang, M. Wen, X. Rong, and T. Li, 2014: Impact of Indian summer monsoon on the South Asian high and its influence on summer rainfall over China. Climate Dyn., 43, 1257–1269, doi:10.1007/s00382-013-1938-y.
Wei, W., R. Zhang, M. Wen, B.-J. Kim, and J.-C. Nam, 2015: Interannual variation of the South Asian high and its relation with Indian and East Asian summer monsoon rainfall. J. Climate, 28, 2623–2634, doi:10.1175/JCLI-D-14-00454.1.
Xie, P., A. Yatagai, M. Y. Chen, T. Hayasaka, Y. Fukushima, C. Liu, and S. Yang, 2007: A gauge-based analysis of daily precipitation over East Asia. J. Hydrometeor., 8, 607–626, doi:10.1175/JHM583.1.
Xue, X., W. Chen, S. Chen, and D. Zhou, 2015: Modulation of the connection between boreal winter ENSO and the South Asian high in the following summer by the stratospheric quasi-biennial oscillation. J. Geophys. Res., 120, 7393–7411, doi:10.1002/2015JD023260.
Yang, J., and Q. Liu, 2008: The “charge/discharge” roles of the basin-wide mode of the Indian Ocean SST anomaly–influence on the South Asian high in summer (in Chinese). Acta Oceanol. Sin., 30, 12–19.
Zhang, P., S. Yang, and V. E. Kousky, 2005: South Asian high and Asian–Pacific–American climate teleconnection. Adv. Atmos. Sci., 22, 915–923, doi:10.1007/BF02918690.
Zhang, Q., and Y. Qian, 2000: Interannual and interdecadal variations of the South Asia high (in Chinese). Chin. J. Atmos. Sci., 24, 67–78.
Zhang, Q., and G. Wu, 2001: The large area flood and drought over Yangtze River valley and its relation to the South Asia high (in Chinese). Acta Meteor. Sin., 59, 569–577.
Zhang, Q., G. Wu, and Y. Qian, 2002: The bimodality of the 100-hPa South Asia high and its relationship to the climate anomaly over East Asia in summer. J. Meteor. Soc. Japan, 80, 733–744, doi:10.2151/jmsj.80.733.
Zhao, G., B. Liu, and S. Qi, 2000: Application of period relation between SAH and WPSH to forecasting Shaanxi rainy spells summer (in Chinese). Plateau Meteor., 19, 172–178.
