The Break in the Mongolian Rainy Season and Its Relation to the Stationary Rossby Wave Along the Asian Jet

3394 JOURNAL OF CLIMATE VOLUME 19

The Break in the Mongolian Rainy Season and Its Relation to the Stationary Rossby Wave along the Asian Jet

HIROYUKI IWASAKI AND TOMOMI NII Faculty of Education, Gunma University, Maebashi, Japan

(Manuscript received 6 June 2005, in final form 22 November 2005)

ABSTRACT

Seasonal and interannual variation of rainfall over Mongolia was investigated using 10-day rainfall data from 92 stations during 1993–2001 and NCEP–NCAR reanalysis data from 1979 to 2001. The break in the rainy season was found in the middle of July, and the meteorological stations with a clear break period were concentrated in eastern Mongolia where the plain prevails relative to western Mongolia without regard to the difference in annual precipitation. Clear breaks in the rainy season were recognized in 5 yr among an analysis period of 9 yr. In the break period, the stationary Rossby wave trapped in the Asian jet was predominant at 200 hPa, and a barotropic ridge associated with the Rossby wave developed over Mongolia. Furthermore, interannual variation of the break also corresponded to the variation of the stationary Rossby wave. It is considered that the break of the Mongolian rainy season is caused by the stationary Rossby wave trapped in the Asian jet. The stationary Rossby wave was climatologically phase locked in seasonal evolution and, as a result, the break period was also concentrated around the middle of July.

1. Introduction rainfall are indispensable to understanding the variability of vegetation activity over the Mongolian grass- Grassland in Mongolia occupies 97% of the country. lands, there are only a few brief studies on the clima- These grasslands are the main source of forage for no- tology of rainfall due to the lack of accessible meteo- madic livestock, so that, vegetation activity in the grass- rological data. Annual precipitation over Mongolia is lands directly influence Mongolian society. Xue (1996) characterized by large variability from year to year showed that a decrease in vegetation activity over the (Slemnev et al. 1994; Hilbig 1995; Gunin et al. 1999). As Mongolian grasslands plays an important role in modi- for the seasonal change, it is well known that about fying the East Asian monsoon circulation and in pro- 70%–80% of annual precipitation falls as rain in the ducing a rainfall anomaly over China. It is suggested summer season and monthly rainfall amounts reach a that interannual variability of vegetation activity in the maximum in August (e.g., Hilbig 1995; Natsagdorj grasslands has a significant impact on not only Mongo- 2000; Miyazaki et al. 1999; Endo et al. 2006). However, lian society, but the East Asian climate as well. we could not find any studies on the detailed seasonal Iwasaki (2006) showed that the vegetation activity variation of rainfall over Mongolia since most of these over the Mongolian grasslands is impacted by interan- descriptions are based on monthly meteorological data nual variation of rainfall; there are significant positive and/or short-term analysis. In other words, there is not correlations between precipitation from June to July sufficient knowledge on the relationship between the and vegetation activity in the mature stage. Although seasonal evolution of large-scale weather patterns and knowledge on the interannual and seasonal variation of the seasonal variation of rainfall, to say nothing of interannual variations. In this paper, seasonal and interannual variation of rain over Mongolia will be analyzed using twice daily Corresponding author address: Hiroyuki Iwasaki, Faculty of Education, Gunma University, 4-2 Aramaki, Maebashi, Gunma precipitation data. As a result of the analysis, we found 371-8510, Japan. that there is a break in a Mongolian rainy season in the E-mail: [email protected] middle of July, which had not been previously reported.

Unauthenticated | Downloaded 10/01/21 05:27 PM UTC

JCLI3806 15 JULY 2006 I W A SAKI AND NII 3395

FIG. 1. Location of surface meteorological stations and their annual precipitation.

The purpose of this paper is to report the existence of neighboring stations. Therefore, most of erroneous val- the break in Mongolian rainy season, and to discuss its ues in this dataset are missing values. Some outliers possible mechanism. naturally remain after this simple quality control check. However, these outliers did not affect the results of the present study, because the essential features of break in 2. Data the Mongolian rainy season can be well recognized by Data used in the present analysis are a surface me- the analysis when the questionable values are removed. teorological dataset provided by the Institute of Meteo- Any years that contained more than five erroneous rology and Hydorology, Mongolia, sonde data at Ulann- values, including missing values from April to Septem- baator, Mongolia, and the National Centers for Environ- ber, are omitted from the dataset. After the screening, mental Prediction–National Center for Atmospheric the number of stations in which 7, 8, and 9 yr of data Research (NCEP–NCAR) reanalysis data. survived are 8, 37, and 47, respectively. These 92 sta- The surface meteorological dataset contains 3-hourly tions are used for the present analysis (Fig. 1). Since air temperature, 3-hourly relative humidity, and twice- this paper will focus on the decrease in rainfall in July, daily precipitation values from 1993 to 2001 for 103 data quality in July should be examined. Figure 2 shows stations. Since there are more than a few erroneous number of erroneous data after screening every 10-day values in the dataset, including missing values, precipi- period. The number of erroneous data is relatively tation data are carefully screened before the analysis small in July, which means that data quality is good for each station. enough to discuss the decrease in rainfall in rainy sea- There are some obvious errors. The same twice-daily sons. precipitation are recorded for several consecutive days, Twice-daily precipitation values are compiled into and these values are discarded (two cases). Further- more, a simple quality control check is performed on precipitation extremes for each station and each month. In the first step, data that are 6 times greater than the standard deviation are flagged. Second, these questionable values are compared with data from neighboring stations and the Geostationary Meteorological Satellite (GMS) infrared imagery from 1995 to 2003. As a result of comparison, all questionable values are regarded as signals because they were associated with mesoscale to synoptic-scale disturbances on GMS infrared imagery FIG. 2. Time series of the number of erroneous data including and/or a certain precipitaion were recorded in the missing values after data screening.

Unauthenticated | Downloaded 10/01/21 05:27 PM UTC 3396 JOURNAL OF CLIMATE VOLUME 19

10-day rainfall for the analysis, which is a rather coarse temporal resolution. However, since a possible mechanism of the break in the rainy season is the stationary Rossby wave with a period of about 40 days (discussed in section 5), the 10-day rainfall period has enough resolution to compare with the stationary Rossby wave.

3. Definition of a break in the rainy season for each meteorological station a. A break in the Mongolian rainy season Mongolia is located in a transition zone for vegetation, where it ranges from taiga forest in the north to desert in the south. Furthermore, mean annual precipitation decreases from north to south (Fig. 1). Annual precipitation exceeds 350 mm in the taiga vegetation zone and is less than 100 mm in the desert vegetation zone. Figure 3 shows the seasonal variation of area- FIG. 4. Time series of mean 10-day rainfall at (a) Bayanchand- anima and (b) Khanbogd. Averaged periods of Bayanchandanima mean rainfall averaged for three different time scales. and Khanbogd are 8 and 9 yr, respectively. Mean rainfall increases rapidly from the beginning of

June and decreases rapidly in the middle of August. About 70%–80% of the annual precipitation amount occurs from June to August, which is referred to as the “Mongolian rainy season” in this paper. There is a local minimum in the middle of July in Figs. 3a–c, which is a “break” in the Mongolian rainy season. The frequency of rainy days also decreases in the middle of July regardless of rain intensity, however, the decrease is larger in intensive rain. Some previous climatological studies described seasonal variation of rainfall using monthly mean values (e.g., Hilbig 1995). However, the break in the rainy season had not been reported since a shorter-term average is required to see it. Figure 4 shows typical seasonal variation of mean 10-day rainfall at Bayanchandanima (BA in Fig. 1) with high annual precipitation in the forest steppe vegetation zone and Khanbogd (KB in Fig. 1) with low annual precipitation in the desert vegetation zone. Mean 10- day rainfall at Bayanchandanima in Fig. 4a reaches a maximun in the beginning of July (the first maximum), decreases to 1/3 the first maximum value in the middle of July (break), and recovers in the beginning of Au- gust (the second maximum). The rainfall at Khanbogd in the desert also has a clear break surrounded by two maxima. The break is unrelated to the difference in annual precipitation. Figure 5 shows the time series of twice-daily rainfall for these two stations in 1994, 1996, and 1999 when the FIG. 3. Seasonal change of (a) mean daily, (b) pentad, and (c) 10-day precipitation averaged for 92 stations. The arrow indicates break was well recognized along 110°–115°E. It is dif- the “break” in the Mongolian rainy season. ficult to determine the break period exactly for each

Unauthenticated | Downloaded 10/01/21 05:27 PM UTC 15 JULY 2006 I W A SAKI AND NII 3397

FIG. 5. Time series of twice-daily rainfall at (a) Byanchandanima and (b) Khanbogd in 1994, 1996, and 1999. Open bars indicate the break periods. station using data from one station, so that the period of station. Although many stations have a minimum in the break was defined by referring to other stations July, the degree of the decrease in the rainfall are very along 110°–115°E. The break periods in 1994 and 1996 diverse because data contain high-frequency “noise” were about 10 days in the middle of July, when precipi- due to the short analysis period (the solid line in Fig. 6). tation was not observed over the wide area. The break The break was defined using smoothed rainfall with a period in 1999 had a length of about 20 days. three 10-day period running mean (the dotted line in Although there are first and second maxima in pre- Fig. 6) so as to eliminate this noise. Biner (BI in Fig. 1) cipitation at Byanchandanima (Fig. 5a), neither maxi- has a minimum in the rainy season, and it is considered ma were observed in some years for Khanbogd in the that Biner has a break in the rainy season (Fig. 6a). On arid regions (Fig. 5b). Two maxima were usually ob- the other hand, since Byuant-Ukhaa (BU in Fig. 1) served in northern Mongolia where annual precipita- does not have a decrease in the smoothed 10-day pre- tion is relatively large, and were not always observed in cipitation, hence, the minimum in mean 10-day precipi- the arid regions of southern Mongolia. tation in the middle of July is regarded as noise (Fig. 6c). The break extracted by this simple criterion is referred to as Type I. b. Definition of a break in the rainy season The second criterion is adopted for another type of To extract the stations with a clear break in the rainy the break. Although the smoothed rainfall does not have season, we adopt two empirical criteria. Mean 10-day minimum in the rainy season, Khahgol (KH in Fig. 1) has rainfall values are calculated from raw data for each a clear decrease in a mean 10-day rainfall (Fig. 6b).

FIG. 6. Time series of mean 10-day rainfall for stations (a),(b) with a typical break in the rainy season and (c) without a break. Solid lines and dotted lines indicate 10-day rainfall and smoothed rainfall with three 10-day periods running mean, respectively.

Unauthenticated | Downloaded 10/01/21 05:27 PM UTC 3398 JOURNAL OF CLIMATE VOLUME 19

FIG. 7. Distribution of stations with a break in the rainy season. Closed circles and closed stars indicate the location of stations with the break of Type I and Type II, respectively. An oval with an “H” indicates the location of the Khenty Mountains.

Such stations are extracted using the following stan- cent of the stations have a break, and their locations dards: are shown in Fig. 7. There are two features in Fig. 7. One is that the break in the rainy season is well recog- 1) Both differences between the minimum and the two nized over eastern Mongolia where the plain prevails maxima (D1 and D2) exceed 10 mm, and relative to western Mongolia. The other is that five 2) The sum of the two differences (D1 ϩ D2) exceeds stations around the Khenty mountains (H in Fig. 7) do 30 mm. not have a break even if neighboring stations have a The stations that satisfied these two criteria are also clear break. However, two stations (Ulaanbaator and considered to have a break in rainy season, and this Byuant-Ukhaa) are hidden behind the large closed break is referred to as Type II. However, we could not circles. The effect of these mountains will be discussed elucidate the difference between Type I and Type II in in section 5b. a meteorological process, and the two types will be As shown in Fig. 8, the timing of the break and the treated as the same. second maximum of the 10-day rainfall is concentrated in the middle of July and the beginning of August, re- 4. Results spectively. As is well known, interannual variability in rainfall is large in Mongolia, however, the timing of the a. Features of a break in the rainy season break and the second maximum are rather fixed. On Thirty-two and nine stations satisfied the criteria for the other hand, the timing of the first maximum fluc- Type I and Type II breaks, respectively. Forty-five per- tuates to a greater extent.

FIG. 8. Timing of the break and the two maxima of the 10-day rainfall.

Unauthenticated | Downloaded 10/01/21 05:27 PM UTC 15 JULY 2006 I W A SAKI AND NII 3399

FIG. 9. Seasonal change of the mean 10-day rainfall for western and eastern Mongolia. The gray line and the black line indicate rainfall of western and eastern Mongolia, respectively. Arrows indicate the break period for each year. b. Interannual variation of the break in the rainy rainfall was observed in the middle of July over both season western and eastern Mongolia. These three years will be discussed from the point of view of the possible Since it is well known that interannual variation of mechanism of the break in the next section. rainfall in Mongolia is large, the break did not always occur. Figure 9 shows the seasonal change of 10-day rainfall for each year averaged in stations located west TABLE 1. Properties of the break in each rainy season. Exis- and east of 100°E. The property of the break is sum- tence of the break for western and eastern Mongolia and the stationary Rossby wave are described in four categories [very marized in Table 1. However, the averaged rainfall is clear (VC), clear (C), rather clear (RC), and not clear (NC)]. heavily weighted by the stations with high annual precipitation. To classify the years with and without the The stationary Rossby break, the seasonal variation of rainfall in the arid and Existence wave around 85°–110°EinJul of break semiarid region is also considered. In the results, 5 yr Existence of (1994, 1996, 1999, 2000, and 2001) are classified as the Yr West East Rossby wave Phase year with the break (break years), and the others are 1993 N C N — the year without the break (nonbreak years). 1994 C VC VC In phase There are three characteristic years. The break in 1995 N N C Out of phase 1996 RC C RC In phase 1994 had the strongest signal in the 9-yr period over all 1997 N N VC Out of phase of Mongolia. Although the break in 2000 was also clear, 1998 N N N — the minimum of rainfall was observed at the end of 1999 C VC C In phase July; it was 10 days later when compared with the cli- 2000 RC C VC 10 days late matological mean. In 1997 (a nonbreak year) a lot of 2001 RC C N —

Unauthenticated | Downloaded 10/01/21 05:27 PM UTC 3400 JOURNAL OF CLIMATE VOLUME 19

FIG. 10. The seasonal change of the 3-hourly mixing ratio of water vapor at the surface for (a) Bayanchandanima and (b) Khanbogd from 1993 to 2001. Data in 1994 for Khanbogd are missing.

5. A possible mechanism of the break in the 0.5–1.0 K than in June and August, indicating that la- Mongolian rainy season tent instability (minimum of SSI) in the break period is caused by high water vapor mixing ratio values in lower a. Seasonal change of water vapor atmosphere, and the stratification in the break period is In this subsection, the seasonal change of water va- rather stable. This increase in the lapse rate in July por, which is a source of rain, will be investigated substantially attributes to higher temperatures at 500 around the break period. Figure 10 shows the time se- hPa (not shown). ries of the surface water vapor mixing ratio at Bayan- chandanima and Khanbogd. Water vapor at the surface b. Effect of the large-scale circulation reached a maximum in the middle of July at both sta- 1) DEVELOPMENT OF A RIDGE AT 500 HPAIN tions. There are no apparent minimum during the break THE BREAK PERIOD period in the middle of July, even though a very clear break is recognized at both stations as shown in Fig. 4. The increase of the temperature lapse rate in the Figure 11 shows the time series of 31-yr mean pre- middle of July in Fig. 11c suggests the development of cipitable water (PW) and Showalter stability index a ridge over Mongolia. Here, the relationship between (SSI) at Ulaanbaator located 35 km to the southeast of the development of a ridge and the break in the rainy Bayanchandanima. The distance between the two sta- season will be described using NCEP–NCAR reanalysis tions is not very long, so the difference in stratification data. Figure 12a shows a composite map of geopoten- between the stations is negligible. To eliminate the af- tial height at 500 hPa (Z500) for the middle of July in fect of diurnal variation, upper-air sounding data at the break years. A wave pattern is predominant from 8 MST (Mongolia Standard Time: 0000 UTC) were used. 20° to 140°E along 45°N, and a weak ridge exists over In addition to water vapor at the surface, mean PW Mongolia. Yatagai and Yasunari (1998) studied the fea- also reaches a maximum and mean SSI reaches a mini- tures of the seasonal variation of rainfall in the Takli- mum in the middle of July (Figs. 11a,b). The mixing makan Desert (around 40°N, 90°E). In Fig. 4 of Yatagai ratio of water vapor at the surface, PW, and SSI for and Yasunari (1998), no rain periods similar to the each break year fluctuated from their respective mean Mongolian break period are seen for the middle of July values, however, their variations do not always corre- in 1981, 1983, and 1984, and a similar wave pattern also spond to the break period. The maxima of water vapor developed around the middle of July for these 3 yr. and the latent instability are frequently recognized even This wave pattern can be seen more clearly in the in a break year. Thus, the break period in the Mongo- difference of mean Z500 in the break years and mean lian rainy season still has moist air with large latent Z500 in the nonbreak years for the middle of July (Fig. instability. 12b). Differences in geopotential height associated with The seasonal change of the mean temperature lapse the ridge R1 and trough T1 are extremely large because rate from 850 to 500 hPa is shown in Fig. 11c. It should a weak trough and a weak ridge exist to the east of the be noted that the mean lapse rate in July is larger by Black Sea and the Caspian Sea, respectively, in the non-

Unauthenticated | Downloaded 10/01/21 05:27 PM UTC 15 JULY 2006 I W A SAKI AND NII 3401

FIG. 11. Seasonal change of (a) mean daily precipitable water, (b) mean daily Sho- walter stability index, and (c) mean daily temperature lapse rate from 850 to 500 hPa at Ulaanbaator (1306 m ASL). Upper-air sounding data at 0800 MST (0000 UTC) from 1973 to 2004 were used. Sounding data in the warm season of 1994 and 1999 are missing. break year. Westerly circulation in the break years over zonal direction. Both the wavelength and the phase- the Eurasian continent is quite different with that in the locked feature are similar to those of Fig. 12 in the nonbreak years. Thus, these figures indicate that a ridge present study. The stationary Rossby wave trapped in R2 develops over Mongolia in the break year. the winter Asian jet had been discussed using a theo- retical model by Hoskins and Ambrizzi (1993). Am- 2) THE STATIONARY ROSSBY WAVE AS A brizzi et al. (1995) showed the Rossby wave activity can POSSIBLE MECHANISM propagate along the Asian jet even in the summer, and Terao (1998, 1999) showed that the stationary although the summer waveguide for the stationary Rossby waves trapped in the subtropical jet predomi- Rossby wave was weaker, it shifted northward and ex- nated over the Eurasian continent during summer tended more eastward than in the winter. through statistical and model analysis. The dominant Teleconnection over the Eurasian continent in the mode of the Rossby waves is characterized by a a baro- summer season reported by Krishnan and Sugi (2001) tropic structure, a time scale of 40 days, and a wave- and Wakabayashi and Kawamura (2004) has similar length of 50° in longitude (wavenumber is 7). Further- features to the propagation of the stationary Rossby more, the Rossby wave train is phase locked in the wave. Recently, Enomoto et al. (2003) and Enomoto

Unauthenticated | Downloaded 10/01/21 05:27 PM UTC 3402 JOURNAL OF CLIMATE VOLUME 19

FIG. 12. (a) Composite 500-hPa height map for the middle of July (11–20 July) in the break years. Contours are every 100 m. (b) Difference of 500-hPa height between the break years and nonbreak years for the middle of July (break years minus nonbreak years). Contours are every 20 m, but values from Ϫ10 to 10 m are not plotted.

(2004) extended this theory to the formation mechanism of the Bonin high and interannual variation of the baiu-front activity. We also adopt this theory to inter- pret the mechanism of the break in the Mongolian rainy season. Through a bandpass-filtered meridional wind speed analysis in the manner of Terao (1998), the possible mechanism of the break in the Mongolian rainy season will be discussed. Figure 13 shows a time–longitude cross section of mean meridional wind speed at 200 hPa along the climatic Asian jet (40°–45°N) calculated from NCEP– NCAR reanalysis data from 1979 to 2001. A 40-day bandpass-filtered period (half-amplitude periods are 25 and 60 days) was applied for the meridional wind so as to extract the dominant mode of the stationary Rossby FIG. 13. Time–longitude cross section of 40-day bandpass- wave pointed out by Terao (1998, 1999). The stationary filtered 23-yr mean meridional wind speed at 200 hPa along Ϫ1 Rossby wave can be seen clearly from day-of-year the climatic Asian jet (40°–45°E). Contours are every 0.5 m s , but values from Ϫ1to1msϪ1 are not plotted. Negative areas (DOY) 180 to 225 (from 28 July to 15 August) and less than Ϫ1msϪ1 are shaded. Vertical dashed lines indicate extends eastward to westernmost Mongolia (90°E). The the location of Mongolia and the horizontal line is 200 in DOY origin of the wave train is near 30°W far from the en- (18 July).

Unauthenticated | Downloaded 10/01/21 05:27 PM UTC 15 JULY 2006 I W A SAKI AND NII 3403 trance region of the Asian jet, which is consistent with Terao (1998, 1999). These features are also well recognized from 35° to 50°N, implying the enhancement of the stationary Rossby wave in July was not attributed to the northward propagation of the Asian jet, but to the seasonal evolution of the Rossby wave train itself. It is noted that the Rossby wave trains are phase locked not only in the zonal direction, but in seasonal evolution as well, and the amplitude of mean meridional wind speed reaches a maximum around 200 in DOY (18 July). This period corresponds to the break in Mongolian rainy season, and weak northerly and southerly winds can be seen even in 23-yr mean data around 80° and 110°E, respectively, indicating the existence of a weak ridge over Mongolia. The relationship between the interannual variation of the break and the stationary Rossby wave will be examined for 3 yr (1994, 1997, and 2000). A distinct break was found in 1994 as shown in Fig. 9. Figure 14 shows the time–longitude cross section of Z500 over Mongolia for these 3 yr. The ridge developed at 500 hPa over Mongolia during the break period in the middle of July. This ridge corresponds to R2 in Fig. 12b. The FIG. 15. Time–longitude cross section of 40-day bandpass- development of the ridge is consistent with the increase filtered meridional wind speed averaged from 40° to 45°Nat200 Ϫ1 Ϫ of the temperature lapse rate in Fig. 11c, and convective hPa. Contours are every 2.5 m s , but values from 2.5 to 2.5 msϪ1 are not plotted. Negative areas less than Ϫ5msϪ1 are clouds would be suppressed under the developed ridge. shaded. Vertical dashed lines indicate the location of Mongolia Figure 15 shows the time–longitude cross section of 40- and horizontal lines are 200 in DOY (18 July). day bandpass-filtered meridional wind speed at 200 hPa along the climatological Asian jet. The stationary wave trains are predominant until 200 in DOY, and they are are considerable southerly and northerly components in phase with the climatological mean in Fig. 13. There around 80° and 110°E, respectively, in the middle of July, implying that the ridge developed at 200 hPa over Mongolia. The distinct break in 1994 and the developed ridge at 500 hPa were coincident with the development of the ridge at 200 hPa due to the stationary Rossby wave. Because of the barotopic structure of the stationary Rossby wave (Terao 1998, 1999), a ridge at 500 hPa should be considered as a part of the stationary Rossby wave. The break in 2000 was observed at the end of July, which is 10 days later than the climatological mean (Fig. 9). The phase of the ridge at 500 hPa and the stationary Rossby wave at 200 hPa were also late compared to the climatological mean (Figs. 14c and 15c). Therefore the break period in 2000 was also synchronized with the stationary Rossby wave. As for the middle of July in 1997, the break was not only observed, but 10-day rainfall reached a maximum (Fig. 9) and a trough prevailed at 500 hPa over Mon- golia (Fig. 14b). Although the stationary Rossby wave and wave trains were clear from July to August (Fig. FIG. 14. Time–longitude cross section of geopotential height at 500 hPa averaged from 45° to 50°N over Mongolia. Contours are 15b), they are out of phase in comparison with the cli- every 50 m and the dashed line is 5700 m. matological mean. The precipitation in the middle of

Unauthenticated | Downloaded 10/01/21 05:27 PM UTC 3404 JOURNAL OF CLIMATE VOLUME 19

July in 1997 was also coincident with the development On the other hand, the anomalous high R2 at 500 hPa of a trough at 200 hPa due to the stationary Rossby is predominant over western Mongolian in the break wave. years in Fig. 12b, which is a contradiction to the result These facts indicate that the break in the Mongolian that most of stations over western Mongolia did not rainy season is caused by the stationary Rossby wave exhibit the break as shown in Fig. 7. The effect of the trapped in the Asian jet, and the interannual variation thermally induced local circulation is one of mecha- of the break is also caused by variation of the stationary nisms used to solve the contradiction. That is, the local Rossby wave. Since the stationary Rossby wave was circulation around the mountainous region in western climatologically phase locked in seasonal evolution and Mongolia would be the cause of considerable rain even the mean amplitude reached a maximum around the in the period when the ridge developed. In other words, middle of July as shown in Fig. 13, the break period was if there were more mountains in eastern Mongolia, the also concentrated around the middle of July. signal of the break period would be weaker. On the other hand, the anomalous high R2 at 500 hPa does not spread to the easternmost area of Mongolian 6. Summary in the break year as shown in Fig. 12b, which is a contradiction to the result that break is apparent over east- Seasonal and interannual variation of rainfall over ern Mongolia as shown in Fig. 7. According to Terao Mongolia was analyzed using 9-yr twice-daily rainfall (1998), although the Rossby wave has a barotropic data and NCEP–NCAR reanalysis data. The 9-yr mean structure, the vertical structure, extracted using a 40- rainfall decreased in the middle of July, even though day bandpass filter, exhibits eastward phase tilt toward the middle of July has air with a high moisture content the lower levels, and the phase difference between 500 and latent instability according to upper-air soundings and 850 hPa is up to 6° of longitude. That is, the ridge at Ulaanbaator. We called this period “the break in the in the lower troposphere should be located about 500 Mongolian rainy season.” The break was apparent over km to the east of the ridge at 500 hPa. The vertical eastern Mongolia, where the plain prevails relative to structure of the Rossby wave is one possible mechanism western Mongolia, without regard to the difference in used to solve the contradiction. annual precipitation. The break in the rainy season was not observed every c. Effect of a local scale circulation year. Clear breaks were recognized in 5 yr among the Three stations without the break period around the analysis period of 9 yr. In the 5 yr with the break peri- Khenty Mountains (H in Fig. 7) are seen in Fig. 7. In ods, the stationary Rossby waves trapped in the Asian addition to these stations, Ulaanbaator and Byuant- jet were predominant at 200 hPa and a barotropic ridge Ukhaa do not have the break, and these stations are associated with the stationary Rossby wave developed hidden behind the closed circles. Several stations lo- over Mongolia. These conditions suppressed deep con- cated in the east and south of the Khenty Mountains do vection even in the moist environment with latent in- not have a break even if stations in their neighborhood stability. Furthermore, interannual variation of the have a clear break. In these stations, heavy rain had break also corresponded to the variation of the station- sometimes occurred in July even in the break year, so ary Rossby wave. Thus, it is concluded that the break of that these stations were not classified as stations with Mongolian rainy season results from the stationary the break. Rossby wave trapped in the Asian jet. The stationary As shown in Fig. 11b, mean SSI at 0800 MST at Ul- Rossby wave was climatologically phase locked in sea- aanbaator reached a minimum in the middle of July. sonal evolution and the mean amplitude reached a The latent instability should be enhanced around noon maximum around the middle of July, as the result, the due to surface heating in summer sunny days. Further- timing of the break in the Mongolian rainy season was more, a valley wind should develop due to a thermally also concentrated around the middle of July. induced local circulation. As a consequence of the local circulation, latent instability is released and cumulon- Acknowledgments. This research has been supported imbus clouds develop over the mountains (e.g., Astling by a CREST Project (The Rangelands Atmosphere– 1984; Iwasaki 2004; Sato and Kimura 2003). This is a Hydrosphere–Biosphere Interaction Study Experiment possible mechanism to cause heavy rainfall in “the in Northeastern Asia of the Japan Science and Tech- break period” around the Khenty Mountains. In other nology Agency (JST), and partially supported by the words, thermally induced local circulations around Ministry of Education, Sports, Culture, Science, and mountainous regions would be the cause of consider- Technology of Japan. The authors wish to express able rain even in the break period. thanks to a staff member of Institute of Meteorology

Unauthenticated | Downloaded 10/01/21 05:27 PM UTC 15 JULY 2006 I W A SAKI AND NII 3405 and Hydorology, Mongolia, for the supply of the table water and convective activity in summer season around dataset of the meteorological stations, and Dr. Sato of Mt. Tanigawa in the north Kanto District, Japan. J. Meteor. JST for useful advice for the data analysis. Soc. Japan, 82, 805–816. ——, 2006: Impact of interannual variability of meteorological The NCEP–NCAR reanalysis data were provided by parameters on vegetation activity over Mongolia. J. Meteor. the NOAA–CIRES Climate Diagnostics Center (CDC), Soc. Japan, in press. Boulder, Colorado, from their Web site (http://www. Krishnan, R., and M. Sugi, 2001: Baiu rainfall variability and as- cdc.noaa.gov/). GMS infrared imagery from 1995 to sociated monsoon teleconnection. J. Meteor. Soc. Japan, 79, 2003 was provided by the Institute of Industrial Science 851–860. (IIS), University of Tokyo (http://www.tkl.iis.u-tokyo. Miyazaki, S., T. Yasunari, and T. Adyasuren, 1999: Abrupt seasonal changes of surface climate observed in northern Mon- ac.jp/SIAIIS/index-j.html). The authors also wish to ex- golia by an automatic weather station. J. Meteor. Soc. Japan, press thanks to a staff member of CDC and ISS. 77, 583–593. Natsagdorj, L., 2000: Climate change. Climate Change and Its Im- REFERENCES pacts in Mongolia, P. Batima and D. Dagvadovj, Ed., JEMR Publishing, 13–42. Ambrizzi, T., B. J. Hoskins, and H.-H. Hsu, 1995: Rossby wave Sato, T., and F. Kimura, 2003: A two-dimensional numerical study propagation and teleconnection patterns in the austral win- on diurnal cycle of mountain lee precipitation. J. Atmos. Sci., ter. J. Atmos. Sci., 52, 3661–3672. 60, 1992–2003. Astling, E. G., 1984: On the relationship between diurnal meso- Slemnev, N. N., P. D. Gunin, and T. I. Kazantseva, 1994: On the J. scale circulations and precipitation in a mountain valley. problem of natural restitution of dominating plants in the Climate Appl. Meteor., 23, 1635–1644. ecosystems of the desert zone of Mongolia (in Russian). Veg. Endo, N., T. Kadota, J. Matsumoto, B. Ailikun, and T. Yasunari, Resour., 4, 1–14. 2006: Climatology and trend in summer precipitation charac- Terao, T., 1998: Barotropic disturbance on intraseasonal time teristics in Mongolia for the period 1960–98. J. Meteor. Soc. scales observed in the midlatitude over the Eurasian conti- Japan, in press. nent during the northern summer. J. Meteor. Soc. Japan, 76, Enomoto, T., 2004: Interannual variability of the Bonin high as- 419–436. sociated with the propagation of Rossby wave along the Asian jet. J. Meteor. Soc. Japan, 82, 1019–1034. ——, 1999: The zonal wavelength of the quasi-stationary Rossby ——, B. J. Hoskins, and Y. Matsuda, 2003: The formation mecha- wave trapped in the westerly jet. J. Meteor. Soc. Japan, 77, nism of the Bonin high in August. Quart. J. Roy. Meteor. Soc., 687–699. 129, 157–178. Wakabayashi, S., and R. Kawamura, 2004: Extraction of major Gunin, P. D., E. A. Vostokva, N. I. Dorofeyuk, P. E. Tarasov, and teleconnection patterns possibly associated with anomalous C. C. Black, 1999: Natural and anthropogenic factors and the summer climate in Japan. J. Meteor. Soc. Japan, 82, 1577– dynamics of vegetation distribution in Mongolia. Vegetation 1588. Dynamics of Mongolia, Kluwer Academic, 7–43. Xue, Y., 1996: The impact of desertification in the Mongolian and Hilbig, W., 1995: Introduction to the country. The Vegetation in the Inner Mongolian grassland on the regional climate. J. Mongolia, SPB Academic Publishing, 13–32. Climate, 9, 2173–2189. Hoskins, B. J., and T. Ambrizzi, 1993: Rossby wave propagation Yatagai, A., and T. Yasunari, 1998: Variation of summer water on a realistic longitudinally varying flow. J. Atmos. Sci., 50, vapor transport related to precipitation over and around the 1661–1671. arid region in the interior of the Eurasian continent. J. Me- Iwasaki, H., 2004: Diurnal variation with dual maxima of precipi- teor. Soc. Japan, 76, 799–815.

Unauthenticated | Downloaded 10/01/21 05:27 PM UTC