Mechanism of the Spring Persistent Rains over southeastern China

Science in China Series D: Earth Sciences © 2007 Science in China Press Springer-Verlag Mechanism of the Spring Persistent Rains over southeastern China WAN RiJin1,2,3† & WU GuoXiong1 1 State Key Laboratory of Numerical Modeling for Atmospheric Sciences and Geophysical Fluid Mechanicals (LASG), Institute of Atmospheric Physics, Chinese Academy of Sciences, Beijing 100029, China; 2 Guangdong Climate and Agrometeorology Center, Guangzhou 510080, China; 3 Shanghai Typhoon Institute of China Meterological Administration, Shanghai 200030, China The Spring Persistent Rains (SPR) in the areas to the south of middle and lower reaches of the Yangtze River or over southeastern China (SEC) is a unique synoptic and climatic phenomenon in East Asia. This study reveals a possible mechanism responsible for the climatic cause of SPR formation through climatic mean data analysis and sensitive numerical model experiments. SEC is located at the downstream of the southwesterly velocity center (SWVC) which lies on the southeastern flank of the Tibetan Plateau (TP). As a result, there are strong southwesterly wind velocity convergence and moisture convergence over SEC. This is the immediate climatic cause of SPR formation. In spring, the seasonal evolution of the southwesterly velocity consists with the surface sensible heating over southeastern TP, indicating that the formation of SPR is related to not only the southwesterly wind of mechanical deflected flow of TP, but also the southwesterly wind of thermal-forced cyclonic low circulation. Sensitive numerical experiments demonstrate that, without TP, both SWVC and the SPR rain belt will disappear. The southwesterly wind velocity increases almost linearly with the amount of the total diabatic heating with TP rising. Therefore, SWVC is the result of the mechanical forcing and thermal forcing of TP. All these strongly suggest that the presence of TP plays a primary role in the climatic formation of SPR. SPR, climatic cause of formation, southwesterly, Tibetan Plateau, sensitive experiments. 1 Introduction The Asia summer monsoon is the most typical and important monsoon system in the world. It consists of two related and also independent members: the South Asia Monsoon system and the East Asia Monsoon system[1]. Under the influence of the Tibetan Plateau (TP), it is first established at Bay of Bengal (BOB). South China Sea (SCS) connects the India Ocean and the Pacific Ocean, and becomes an important region where the two monsoon systems interact. Therefore, more attention was paid to the studies of the monsoon onset and activities in both BOB and SCS[2,3]. However, the climatic features of the circulation and precipitation over East Asia in the transferring season of spring, before summer monsoon onset, are not well studied so far. www.scichina.com www.springerlink.com TP is the highest and the stiffest topography in the world. It is recognized that its mechanical and thermal forcing significantly influence the atmospheric circulation and the formation of climate of the Northern Hemisphere. Yeh et al.[4] and Zhao et al.[5] analyzed the heat source over TP respectively and pointed out that TP is an intense heat source in summer while a heat sink in winter. And the abnormal heating of TP can cause the large-scale circulation anomalies in the East Asia Summer Monsoon. Yanai et al.[6,7] found that the heating of TP is mainly sensible heating before monsoon on set Received January 24, 2006; accepted May 29, 2006 doi: 10.1007/s11430-007-2069-2 † Corresponding author (email: [email protected]) Supported by Chinese Academy Sciences (Grant No. ZKCX2-SW-210), the National Natural Science Foundation of China (Grant Nos. 40221503 and 40475027) and the National “973” Program (Grant No. 2006CB403600) Sci China Ser D-Earth Sci | January 2007 | vol. 50 | no. 1 | 130-144 while latent heating afterwards. With numerical simulation, Wu and Li[8] proved that sensible heating of TP drives the movement of the air over TP and brought forward the concept of the Sensible Heat Air Pump (SHAP). They found that SHAP not only affects the circulation over TP and its surrounding areas, but also disperses Rossby waves and thus affects the evolution of global circulation and climate. Wu and Zhang[9] found that the phase feature of the sensible heating of TP consists with that of the Asia Summer Monsoon. In addition, they proved that it is the mechanical and thermal forcing of TP that leads to the onset of the Asia Summer Monsoon firstly in BOB. In fact, in terms of sensible heating, TP changes from a heat sink to a heat source in early spring. What is the role of the mechanical and thermal forcing of TP in the formation of spring circulation and precipitation over East Asia subtropical area, and what is the mechanism? All these require further studies. This paper focuses on the Spring Persistent Rains (SPR), i.e., the sea-air-land interactions over SEC, which is located at mid-low latitudes of East Asia, the northwest flank of SCS subtropical high and the southeastern flank of TP, and occurs just before the onset of the SCS Summer Monsoon. Through analysis of the climatic features of SPR and the mechanism of SPR formation, the physical process of season transferring may be better understood, basic theories for the prediction of SPR may be provided, and even some valuable clues to the prediction of the Asia summer monsoon may be found. It is well known in China that spring persistent overcast and rain over most of SEC are disaster to agriculture and transportation. SPR is another rainy period besides Meiyu or the Plum rain period in early summer over the middle and lower reaches of the Yangtze River. Since the 1950s, the studies in the synoptic and midshort term forecast aspects have been widely empha― sized[10 13]. But as a climatological concept, till the late 1990s, SPR was first introduced by Tian and Yasunari[14] who proposed that the mechanism of the climatic formation of SPR is the west-land east-sea thermal contrast between the Indochina Peninsula and the western North Pacific to the east of the Philippines, or in other words, the effect of the time-lag in seasonal warming in spring. Furthermore, they suggested that SPR is not a phenomenon due to orographical effects because of the concurrent rapid increase of the rain over the region from SEC to the south of Japan in early spring. However, this study indicates that although there is a similar time-lag in spring over southern North America, no SPR-like rain belt exists over southeastern America, implying that the time-lag mechanism is questionable. Then what’s the essential mechanism of SPR climatic formation? This is the question that this studyis trying to answer . 2 Data and method Two set of rainfall data with different temporal span and spatial resolutions were used. One is the daily rainfall data set from 730 stations in China during the period of 1951－2000, newly released by the National Climate Center (NCC). In order to study climatic spatial distribution and seasonal evolution features of rainfall, the climatological pentad mean was calculated in the whole 50 years and then interpolated onto a 0.5°×0.5° longitude-latitude grid. Another data set is the pentad rainfall data from CMAP[15], which has a 2.5°×2.5° longitude-latitude grid, spanning from 1979 to 2004. Similarly, the climatological pentad mean was computed (in the whole 26 years) to show large-scale rainfall distribution over the Northern Hemisphere. The National Centers for Environmental Prediction/ National Center for Atmospheric Research (NCEP/ NCAR) reanalysis climatological daily mean data (1968 － 1996), including geopotential height, temperature, horizontal wind and specific humidity at 850 hPa, were used. All data were on a 2.5°×2.5° longitude-latitude grid. A five-day average was performed to get the climatological pentad data for the convenience of investigating the associated large-scale circulation distribution and seasonal variation. Moisture divergence ( ∇ ⋅ qV ) at 850 hPa was also computed from the above data. Climatological pentad mean of surface sensible heat flux (SH) averaged in 1968－1996 from NCEP/NCAR daily mean reanalysis was used to describe the seasonal evolution over southeastern TP and to examine the possible relationship to the southwesterly wind. All averaged data were further smoothed in drawing their seasonal evolution curves for they had still contained some short-term random fluctuations. 3 Climatic features of SPR Following Tian and Yasunari[14], the 12th－26th pen- WAN RiJin et al. Sci China Ser D-Earth Sci | January 2007 | vol. 50 | no. 1 | 130-144 131 tads was chosen to represent the SPR period. The mean rainfall over SEC in SPR is shown in Figure 1. The SPR rain belt lies to the south of the middle and lower reaches of the Yangtze River, roughly along the high Nanling, Wuyi Mountains, with high value region ranging about 24°N―30°N, 110°E―120°E, and with intensity of 6―7 northern pole and warm air from the tropical oceans. It is noticeable that there is a rain belt with center intensity exceeding 6 mm/d lying over East Asia continent, i.e., the SPR. But no similar rain belt appears over corresponding southeastern America. mm/d. The maximum rainfall centered at about 28°N prevails almost in a whole year except for being disturbed by summer monsoon rain belt (Figure 2), which is in accordance with the result of Tian and Yasunari[14]. Figure 2 Time-latitude section of pentad mean precipitation averaged over 110°E－120°E. Unit is mm/d. White arrows indicate movements of the rain belt. Figure 1 Spatial distribution of mean (1951－2000) rainfall over China in SPR (12th―26th pentad). Unit is mm/d. The shaded areas represent areas with mountains’ height exceeding 600 m. As shown in Figure 3, the spatial distributions of precipitation over East Asia and North America in SPR are different. At extra-equator, the main rainfall belts are located at the western part of two oceans and close to the east coasts of the two continents. Without doubt, they are all rain-belts of polar front between cold air from Within mid-low latitudes, there exists comparability in land-sea distributions between East Asia (10°N－ 30°N, 90°E－120°E) and North America (10°N－30°N, 80°W－110°W) respectively shown as the two bold dashed rectangles in Figure 4. They all lie in subtropics, with continent to the north and vast ocean to the southeast. Both areas are influenced by connatural polar front rain belt. In Figure 4, regions A (12.5°N－22.5°N, 95°E －105°E ) and B (10°N－20°N, 130°E－140°E) which represent land and sea of southeast Asia respectively were used by Tian and Yasunari[14]. In this paper, according to the comparability, regions C (15°N－25°N, Figure 3 Spatial distribution of CMAP climatic mean (1979－2004) precipitation averaged over 12th－26th pentads. Unit is mm/d. 132 WAN RiJin et al. Sci China Ser D-Earth Sci | January 2007 | vol. 50 | no. 1 | 130-144 Figure 4 The comparing areas between East Asia and North America within mid-low latitude (bold dashed rectangle) and their low latitude representative areas: land (A, C) and sea (B, D). The shaded areas represent the regions where topography height exceeds 1500 meters. The topography data with 10′ ×10′ longitude-latitude grids resolution are from NCEP. Obviously, the strong thermal contrast between west land and east sea over southern North America does not produce SPR-like rainy belt. What is the reason? The next section will focus on the climatological mean features of large-scale atmospheric circulation associated with the rain belt in both East Asia and North America. The differences between them may be relevant to the climatic cause of formation of SPR. 4 Mean features of large-scale atmospheric circulation Figure 5 Seasonal variation of land-sea temperature difference over southeastern Asia (solid line, TA−TB) and southern North America ( dashed line, TC−TD) at 850 hPa. Unit is ℃. 95°W－105°W) and D (12.5°N－22.5°N, 60°W－70°W) which represent land and sea of southern North America respectively were selected. Though they are at higher latitude than regions A and D, it does not affect our results. Figure 5 shows the seasonal evolution of west-land east-sea temperature difference over southern East Asia and southern North America at 850 hPa. In spring, the temperature over west land increases faster than that over east sea. This phenomenon does not only exist in southern East Asia but also in southern North America. So the climatic formation mechanism of SPR, the effect of land-sea thermal contrast, or the time-lag in seasonal warming, is questionable. In addition, there is a significant difference between the two curves. The land-sea temperature contrast reverses completely in winter in southern East Asia but not in southern North America, which are related to the fact that monsoon is very strong in East Asia and very weak in North America. The mean spatial characteristics of the large-scale atmospheric circulation in SPR were examined from the fields of temperature, height, wind vector and moisture convergence ( ∇ ⋅ qV ) at 850 hPa over both East Asia and North America (Figure 6). It is obvious in Figure 6(a) and (b) that, over the south of the two continents, the temperature in the west land (A, C) is higher than that in the east sea (B, D) respectively. The result is consistent with Figure 5. Furthermore, the temperature difference between C and D is 4.1℃ and that of A and B is 3.2℃. That indicates that the temperature gradient in southern North America is greater than that in southern East Asia. The rain belt in southeastern America should have been stronger than the SPR rain belt in East Asia if the time-lag theory worked. In the height fields at 850 hPa (Figure 6(c) and (d)), the SCS subtropical high is very weak. At southeastern TP, the southwest warm low develops strongly and the pressure gradient between it and SCS subtropical high is very great. On the contrary, in North America, the Mexico subtropical high is very strong. The shallow topography trough of the Mexico Plateau has little influence WAN RiJin et al. Sci China Ser D-Earth Sci | January 2007 | vol. 50 | no. 1 | 130-144 133 Figure 6 Climatic mean fields of temperature ((a), (b)), height((c), (d)), wind vector ((e), (f))and moisture convergence ( ∇⋅ qV ) ((g), (h))at 850 hPa in SPR (12th－26th pentads) over East Asia (left panels) and North America (right panels). The elevation of the shaded areas exceeds 1500 m. The units of contours are ℃ for (a) and (b), gpm for (c) and (d), m/s for (e) and (f), and 10−5g·kg−1·s−1 for (g) and (h). on it. This remarkable difference between East Asia and North America comes from two factors. One is sea-land distribution. At near 25°N, there is land in East Asia but 134 sea in North America. Moreover, as the west part of the Atlantic subtropical high, the Mexico subtropical high lies near three big continents, i.e., North America, South WAN RiJin et al. Sci China Ser D-Earth Sci | January 2007 | vol. 50 | no. 1 | 130-144 America and North Africa. The united thermal adaptation for air rising in land and sinking in sea makes the subtropical high extraordinarily strong. The other possible factor is the topography. The mechanical and thermal effects of giant TP in East Asia make low develop and subtropical high weaken. The effect of TP can be clearly seen in the following sensitive numerical model experiments in Section 5. Corresponding to the height fields, the wind vector fields are distinctly different as shown in Figure 6(e) and (f). Due to the mechanical diffluent effect of TP, the westerly jet belt in East Asia is split into two westerly jets, along the south and north sides of TP respectively, and then meet at about 30°N, 120°E. A southwesterly jet velocity center with maximum 7 m/s emerges at the southeastern flank of TP. Consequently, at the downstream of the center, there is strong southwesterly velocity convergence leading to strong moisture convergence over SEC (Figure 6(g)). The moisture convergence dominates most of SEC with its center value reaching –3×10–5 g/(kg · s). Only one westerly jet belt appears in North America, with its center located over the west of Atlantic (Figure 6(f)). No SWVC exists at upstream of southeastern America. And only strong wind divergence and weak moisture convergence exist in that area (Figure 6(h)). All above indicate that it is the unique SWVC at the southeastern flank of TP that leads to the formation of SPR. This can be seen more clearly in the latitude-time sec- tion of pentad mean fields of wind vector at 850 hPa averaged over East Asia (Figure 7(a)) and North America (Figure 7(b)). In Figure 7(a), the shaded regions indicate that wind velocity exceeds 3 m/s and the dashed rectangle roughly represents the temporal and spatial location of SPR. It can be seen that the southwesterly high velocity belt extends southward to 20°N and northward to 27.5°N and sustains for about 2.5 months before shifting southward, which coincides with the onset, developing and decaying of SPR exactly. Afterwards, the southwesterly wind expands to SCS and the SCS monsoon breaks out. In middle June, the southwesterly wind in SEC strengthens once more and Meiyu season appears in the middle-low reaches of the Yangtze River. In middle July, the strong southwesterly wind extends to 35°N and the rainy season begins in northern China. At the same time, the southwesterly wind in SCS gets intensified and South China enters a frequent typhoon season. The seasonal evolution of southwesterly consists with the seasonal evolution of rain belt (referred to Figure 2) in East China completely. In mid-low latitudes of North America (Figure 7(b)), the southeasterly wind prevails in almost a whole year. The southwesterly wind appears at about 30°N in early spring too. But it cannot persist, but withdraw along with the season to the north. So the unique southwesterly wind over SEC is crucial to SPR formation. What gives birth to SWVC at the southeastern flank of TP? It is supposed that it is the result of mechanical forcing and the thermal forcing of TP. This can be seen Figure 7 Time-latitude section of climate mean pentad wind vector at 850 hPa averaged over East Asia 110°E－120°E in (a) and North America 90°W－ 80°W in (b). The velocity of the vectors in shaded areas exceeds 3 m/s. WAN RiJin et al. Sci China Ser D-Earth Sci | January 2007 | vol. 50 | no. 1 | 130-144 135 in the following two aspects. The first is mechanical forcing of TP. It is known that the topography height of TP, even its southeast part, the Yunnan-Guizhou Plateau, exceeds the critical height of climbing or deflecting of flows. As a result, in SPR, the westerly jet in middle latitudes is split into two westerly jets by TP (Figure 6(e)). The southern southwesterly jet flows eastward along the south flank of Himalayas and encounters the Yunnan-Guizhou Plateau. The principle of vorticity conservation can be expressed as[17] ζθ + f Δp = constant , where ζ θ represents the mean vorticity of westerly flow between two isentropic surfaces, f is geostrophic vorticity and changes little in this case, and Δp is the difference of pressures between upper isentropic surface and the lowers. At the windward side (southwestern flank of the Yunnan-Guizhou Plateau), the Δp of westerly flow decreases, so ζθ decreases and the anti-cyclonic vorticity increases. Consequently, the flow deflects to the southeast and becomes northwesterly. Similarly, at the leeward side (southeastern flank of the Yunnan-Guizhou Plateau), when the Δp of westerly flow increases, ζ θ increases and cyclonic vorticity increases. As a result, the flow deflects to the northeast and becomes southwesterly. Therefore, the deflected southwesterly flow is one main part of the southwesterly wind. The second is thermal forcing of TP. Figure 8 shows a few regions over Eurasia continent (left panel) and the seasonal evolutions of the related variables averaged over these regions (right panel). The temperature difference between south (G region) and north (E region) at 850 hPa (TG-E) decreases quickly from 30℃ to 18℃ in SPR. And the mean westerly at 850 hPa (UF) over middle latitudes (F region) decreases from above 3 m/s to below 2 m/s in SPR. It is known that it results from the effect of thermal wind. The decrease of mean westerly means that both the inflow and the deflected flow of TP are decreasing in SPR. It is surprising that the southwesterly wind at the southeastern flank of TP (I region) at 850 hPa (Vsw) increases remarkably from winter to spring before it reaches its peak (above 6 m/s) in middle March. If the southwesterly wind is only regarded as the deflected flow of TP, why is the deflected flow increasing while the inflow is decreasing? The SH over the southeastern TP (J region) is focused on now. As a result of locating at low latitudes, the SH over J region is uniformly positive in the first half year. It increases slowly in January, quickly in February, until reaching its peak in about middle March. Then it persists for about 1 month and begins to decrease in late April due to more overcast over J region. As a whole, it consists with the Vsw except leading decrease in 2 pentads. The consistency strongly demonstrates the southwesterly wind over the southeastern flank of TP highly relate to the SH over southeastern TP. Wu et al.[17] put forward the thermal adaptation theory of diabatic heating: near the surface of ground it is the positive poten- Figure 8 Regions and corresponding seasonal evolutions of related pentad mean variables in the first half of a year. TG-E: the mean temperature difference between G region (5°N－25°N, 40°E－140°E) and E region (45°N－65°N, 40°E－140°E) at 850 hPa. UF: westerly velocity averaged over F region (25°N －45°N, 40°E－140°E) at 850 hPa. VSW: southwesterly wind velocity averaged over I region (20°N－25°N, 105°E－110°E) at 850 hPa. SH: surface sensible heat flux averaged over J region (22.5°N－32.5°N, 95°E－105°E). SPR region: H region (24°N－30°N, 110°E－120°E). Main part of TP: K region (27.5°N－40°N, 70°E－105°E). The shaded areas represent the regions where elevation exceeds 1500 m. 136 WAN RiJin et al. Sci China Ser D-Earth Sci | January 2007 | vol. 50 | no. 1 | 130-144 tial vorticity source produced by the increasing diabatic heating with height, which maintains the cyclonic circulation. In addition to causing surface air inner energy increasing, surface sensible heating can cause pressure decreasing, isentropic surface sunken and warm cyclonic vortex forming. So it is understandable that there are a temperature ridge and deep low over the southeastern TP as shown in Figure 6(a) and (c). The low is named as the Southwest Warm Low (SWWL). At the southeastern part of SWWL southwesterly wind prevails, indicating that the thermal-forced southwesterly wind is another main part of the southwesterly wind over I region. As season advances, warm and damp ascent flow intensify, the instable energy releases with clouds increasing and SH decreasing in late April. But then latent heating intensifies and it can maintain SWWL. In early May, the pressure gradient between SWWL and the SCS subtropical high decreases quickly when the latter retreats eastward. And then Vsw turns to decrease. But no sensible heating over the main part of TP (K region) in early spring contributes apparently to the seasonal evolution of Vsw, which indicates that SWWL is partially a local thermal-forced hot low. In late June and early July, Vsw strengthens once more and reaches its second peak and causes Meiyu in Central China. The peak is different from the former. It corresponds to the peak of the latent heating over the main part of TP (figure omitted), which agrees well with the theory that latent heating produces cyclonic vortex in lower levels[18]. Therefore, it is TP that gives birth to SWVC. The mechanical-forced deflected flow and the thermal-forced low circulation of TP play a primary role in the climatic formation of SPR. In next section, sensitive numerical model experiments were conducted to verify the results above. 5 Sensitive numerical model experiments There have been lots of studies on the effects of TP by ― utilizing numerical model experiments[19 24]. Qian and Qian[25] used a zonal domain numerical model to simulate the variation of atmospheric circulation in summer while TP rising. The results showed that the effect of the heating of TP on atmosphere strengthens with the height of TP, especially the latent heating is primary. Wu et al.[26,27] reviewed the studies on TP affecting climates and pointed out that it is the cyclonic vortex source pro- duced by TP heating that maintains the low-level low stably. Nevertheless, most model users pay little attention to the effect of TP in spring, especially on SPR before the SCS monsoon onset. And no study has been found about how TP affects the circulation and SPR in spring with TP rising. In this section, sensitive numerical model experiments were used to study the effects of TP in spring. 5.1 Numerical model system The model system used in this paper is a newly developed global atmospheric spectral model (SAMILR42L9) in the State Key Laboratory of Numerical Modeling for Atmospheric Sciences and Geophysical fluid mechanicals of the Institute of Atmospheric Physics of Chinese Academy of Sciences (LASG/IAP/CAS). The model has 42 rhomboidally truncated spectral waves in horizontal with resolution equivalent to 2.8125° (lon) ×1.66° (lat). In vertical, it has 9 levels in σ-coordination system. The model uses a unique mechanical frame by introducing a reference atmosphere and using semiimplicit time integrate scheme[28]. The physics procedure of the model is quite self-contained, including K-distribution radiation scheme[29], Slingo’s cloud diagnosis scheme[30], the diurnal variation of solar radiation[31], and the simplified simple biosphere model (SSiB)[32,33]. Its performance was then evaluated and compared with the original R15L9/LASG model by simulating mainly the global climatological mean states[34,35]. They showed that the R42L9 model well reproduces the observed basic patterns and significantly improves the simulation compared to the original R15L9/LASG model. Two sets of sensitive experiments with R42L9 model were performed. The first set experiments consisted of simulations with and without TP respectively. The topography was set as real topography in control experiment with TP (hereafter referred to as CTL). In the experiment without TP, in order to highlight the effect of the main part of TP, the topography height was set at 200 m in K region (as shown in Figure 8), reserving the surrounding topography, such as the Yunnan-Guizhou Plateau, the Iranian Plateau, the Tianshan Mountains and the Artai Mountains. This experiment was called no TP experiment (hereafter referred to as NTP). The two experiments integrated 15 years respectively and the mean fields of March and April of the latter 10 years are analyzed as the SPR period for purpose of convenience. WAN RiJin et al. Sci China Ser D-Earth Sci | January 2007 | vol. 50 | no. 1 | 130-144 137 Figure 9(a) and (b) are the topography in control experiment CTL and sensitive experiment NTP, respectively. The second set experiments were to lift TP stepwise (hereafter referred to as LFTP). The details are as follows. The solar elevation angle was fixed as that at 10th April (then SPR strengthens remarkably). First, Eurasia continent was leveled to 0 km (Figure 9(c)). Then the model integrated and monthly data of 3 years (1095 days) were output. And then Eurasia was undercut to 1 km, the model integrated and output monthly data of 3 years one more time. And so on, Eurasia was undercut step by step at intervals of 1 km till 5 km. When the elevation exceeded 5 km, only the elevation of the region where elevation exceeds 2500 m in K region was lifted, till 9 km. The main part of TP was set different elevations from 0 km to 9 km, respectively, thus there were 10 ex- periments. Figure 9(d)－(f) shows the topography when TP elevations are 2 km, 4 km and 6 km, respectively. The figures of the other levels are not shown here. 5.2 Results of the CTL and NTP cases Figure 10 shows the results of the CTL run and NTP run. The model simulated the two westerly jet belts at the south and the north of TP at 850 hPa very well (Figure 10(a)). There are strong southwesterly wind on the southeastern flank of TP and strong northwesterly wind on the far northeastern flank. The East Asia big trough is deep while the subtropical high anti-cyclonic circulation in mid-low latitudes is weak, basically agreeing with the observations (Figure 6(e)). But in NTP case (Figure 10(c)), there is only one westerly jet belt over Eurasia. A strong easterly jet belt emerges over South Asia sub-continent. Between the two jets there is an obvious Figure 9 The topography fields of model in CTL (a), NTP (b) and LFTP 0 km (c), 2 km (d), 4 km (e), and 6 km (f). The elevation of the main part of TP in NTP (b) is 200 m. The contour unit is m. 138 WAN RiJin et al. Sci China Ser D-Earth Sci | January 2007 | vol. 50 | no. 1 | 130-144 Figure 10 The fields of circulation and average precipitation of CTL and NTP in spring (March－April). (a) Wind vector at 850 hPa in CTL; (b) precipitation in CTL; (c) wind vector at 850 hPa in NTP; (d) precipitation in NTP; (e) the differential wind field of CTL−NTP at 850 hPa; (f) the differential temperature field of CTL-NTP at 850 hPa. The shaded areas represent where elevation exceeds 2500 m. The units are mm/d for (b) and (d), and ℃ for (f). subtropical high belt, which controls the most of SEC. Over the southeast flank of TP no strong southwesterly wind occurs, but strong southeasterly wind, somewhat like the southeasterly wind in southern North America, as in Figure 6(f). And over the southwestern flank of TP no longer strong northwesterly wind occurs, but wide strong easterly belt. The difference of wind vector fields at 850 hPa (CTL minus NTP) is shown in Figure 10(e). Due to the presence of TP, the westerly wind velocity decreases in the west, turns to the south, and strong southwesterly wind prevails over the southeastern flank, which intensifies the transport of warm and dump air to SEC. Around TP, the differential flow forms a unique pattern with two complete circles, which might be named two-circled differential circulation as well. Besides the effect of the mechanical forcing of TP, it is also possibly related to the effect of the thermal forcing of TP, which will be discussed in details in the following. Figure 10(b) shows the precipitation field in CTL. There is obviously a rainy belt over SEC, with center intensity about 7－8 mm/d, roughly consistent with the WAN RiJin et al. Sci China Ser D-Earth Sci | January 2007 | vol. 50 | no. 1 | 130-144 139 observation (Figure 1). On the contrary, in NTP (Figure 10(d)), the rainfall over most SEC decreases to no more than 2 mm/d, and the SPR rain belt disappears, which strongly suggests that the SPR would not exist if there were no TP. The differential temperature field at 850 hPa of CTL minus NTP is shown in Figure 10(f). The presence of TP increases the temperature of surrounding air, especially in the south by 6℃, because it blocks the southward cold air at the northwest side. In addition, it decreases the air temperature over the northeast Asia by 4℃, and increases the air temperature around the Alaska Mountains by 2℃. The differential temperature pattern of “+ +” possibly relates to the propagation of Rossby waves[27]. TP makes the cold air of the North Hemisphere relocate and affect the temperature and circulation of north hemisphere. The warm and dump southwesterly in low level increases the temperature of SEC and decreases that of northeast China. The zero line happens to lie in the middle of SEC. Therefore, the cold front is intensified through cooling in northeast and warming in southwest, which is favorable to the formation of the polar front and SPR belt. Thus, the presence of TP not only makes the SPR appear, but also affects the circulation and temperature of north hemisphere by the propagation of Rossby waves. It is clearly shown in Figure 11(c) that the south differential circle is stronger than the north one, and at the near north of TP appears strong differential easterly. Thus around the main body of TP, a complete differential cyclonic circulation circle forms significantly. The reason will be seen through the second set of numerical experiments in the following. 5.3 Results of LFTP cases The similar lifting TP experiments have been done by Qian and Qian[25]. But they focused on the circulation in summer season (not in spring). In this LFTP case, the time of middle spring was chosen, solar elevation angel was fixed as that on 10th April and integrated 36 months so as to eliminate the effects of Madden-Julian Oscillation (MJO) and inter-annual variation perturbation. In order to get rid of the spin-up error in early months, the latter 30 months data were chosen to be averaged and the long-term mean fields of every experiment mentioned above were obtained. Figure 11 shows the wind vector at 850 hPa (left pan- 140 els) and rain fields (right panels) of different TP elevations in LFTP case. Here presents the representative elevations of 0 km, 2 km, 4 km and 6 km. The other levels are omitted for convenience. Clearly, when TP does not exist, the westerly jet has no split in Eurasia, and subtropical high circulation dominates mid-low latitudes (Figure 11(a)). In East Asia, the rainfall is very little and no SPR rain belt exists (Figure 11(e)). When TP is at 1 km elevation, the subtropical high in mid-low latitudes splits at the South Asia sub-continent. The southwesterly wind over the southeastern flank of TP increases obviously. The rainfall in SEC increases slightly (figure omitted). When at 2 km, the westerly jet splits into two belts, and South Asia sub-continent is dominated by wave-like westerly flow. The deflected southwesterly wind over the southeastern flank of TP is obvious and the southwesterly wind over SEC increases remarkably (Figure 11(b)). The SPR rain belt begins to take shape (Figure 11(f)). When at 3, 4 and 5 km, the westerly belts over both the north and south of TP strengthen into two westerly jets, and the southwesterly wind over the southeastern flank of TP strengthens to form a SWVC (Figure 11(c)). Then the SPR rain belt center forms (Figure 11(g)). After 5 km, the basic circulation pattern changes little, but the two westerly jets and the southwesterly wind are stronger. The cyclonic bend intensifies at the southeastern flank of TP and the anti-cyclonic vortex develops over SEC (Figure 11(d)). As a result, the rainfall in SEC decreases and the East Asia rain belt moves northward to the Yangtze River and the north of it. At the same time, the rainfall belt center is attracted to southeastern TP (Figure 11(h)). Therefore, it can be seen that the presence of SWVC over the southeastern flank of TP is crucial to the formation of SPR. The relationship between the southwesterly wind velocity and the heating of TP when rising will be probed into in the following section. Figure 12 shows the southwesterly wind velocity at 850 hPa (Vsw) over the southeastern flank of TP (I region) corresponding to the diabatic heating of the main part of TP (K region) when the elevation of TP rises. In this figure, Q is total heating which can be expressed as follows: Q = S+Lp+R, where S is surface sensible heating, Lp is latent heating, and R is total radiation heating. It can be seen that S in- WAN RiJin et al. Sci China Ser D-Earth Sci | January 2007 | vol. 50 | no. 1 | 130-144 Figure 11 Fields of wind vector (left, the shaded areas represent the regions where the velocity exceeds 4 m/s) at 850 hPa and rain (right, unit: mm/d) of LFTP cases. The bold dashed line and black shaded areas are the main part of TP. creases slowly from about 80 W/m2 at the beginning of the rise of TP. The reason can be that the atmospheric optic thickness turns shorter and short wave scatters de- crease when TP rises. Thus the radiation reaching ground increases and the ground temperature rises. At the same time, the surface air temperature drops because WAN RiJin et al. Sci China Ser D-Earth Sci | January 2007 | vol. 50 | no. 1 | 130-144 141 Figure 12 The southwesterly wind velocity VSW at 850 hPa over I region at the southeastern flank of TP corresponds to the diabatic heating of the main part of TP (K region) when the elevation of TP rises. of the higher elevation. So the temperature difference between ground and surface air increases and at last the heat exchange intensifies. When TP rises to 3 km, S reaches 100 W/m2 and then keeps unchanged afterwards possibly because of the balance between shorter atmospheric optic thickness and more cloud overcast. After higher than 7 km, S begins to drop because the effect of increasing overcast has exceeded the effect of the decreasing shorter atmospheric optic thickness. The total radiation heating R changes little in almost the whole process except increasing slowly after higher than 5 km, for the long wave radiation output decreases while the cloud overcast increases. The latent heating almost does not change with a low value about 40 W/m2 till 3 km, then it increases rapidly after higher than 3 km, and it is comparative to sensible heating S with a value of about 100 W/m2 at 6 km. After higher than 6 km, it tends to increase slowly. When Eurasia is leveled, the total diabatic heating Q of TP is 40 W/m2 and already becomes the heat source to air. As TP rises, Q increases too. The increase before it rises to 3 km comes mainly from S and it is quite slow. Q increases rapidly between 3 km and 6 km and then tends to slow down after higher than 6 km. The increase comes mainly from Lp. As a whole, the increasing trend of Q is consistent with Lp very much. This indicates that the increase of the total heating mainly comes from the increase of latent heating. So even in spring, the heating 142 source effect of the rise of TP is mostly represented by the rapid increase of latent heating. The mechanism of it consists with that in summer[28]. But the effect of sensible heating before it rises to 3 km is not neglectable, and together with the effect of the mechanical deflected flow of TP, it makes SPR take shape preliminarily. Finally, what needs to scrutinize is Vsw. From 0 km to 1 km, Vsw increases rapidly from 0.9 m/s to 4.1 m/s, then it increases slowly to 5.2 m/s at 3 km. From 3 km to 6 km, it increases rapidly again with value 9 m/s at 6 km, then it slows down again. As a whole, except for the initial increase of mechanical deflected flow of TP, the Vsw is consistent with Q almost linearly. This strongly suggests that the southwesterly Vsw is essentially related to the thermal effect of TP. On the other hand, if it is reckoned according to the ratio of 16.6 W/m2 increase of Q corresponding to 1 m/s increase of Vsw averaged between 3 km to 6 km, there is only 0.8 m/s contribution of Q to Vsw before 1 km while the contribution of deflected flow is 2.4 m/s. And from 1 km to 2 km, they are 0.3 m/s and 0.6 m/s, respectively. From this point, the increase of the southwesterly is mostly attributed to the effect of mechanical deflected flow of TP before it rises to 2 km, especially before 1 km. This result agrees with the theoretical calculation of Zhang and Qian[36] that the critical height between climbing and deflecting is below 1500 m for TP. In brief, in spring, in the process of the rise of TP, Q WAN RiJin et al. Sci China Ser D-Earth Sci | January 2007 | vol. 50 | no. 1 | 130-144 increases continuously and so does its effect of heating to air. The increase of the total heating of TP relies on the slow increase of sensible heating before 3 km and the rapid increase of latent heating after 3 km. The southwesterly wind Vsw is almost linearly related to total heating after the mechanical deflected flow of TP makes it increase rapidly at the early rise. These strongly suggest that the rise of TP not only splits and deflects the westerly wind belt, but also causes the rapid increase of total diabatic heating in spring. And this strong heating of TP produces cyclonic vortex source and thus strengthens cyclonic circulation in lower levels. And the cyclonic circulation strongly intensifies the southeastern deflected flow of TP. Consequently, SWVC appears and SPR forms. 6 Conclusions and discussion This study presents the large-scale features of the atmospheric circulation associated with SPR based on the climatic means. Through the comparison analysis of the circulation and precipitation between East Asia and North America, a mechanism for the rainy season over SEC climatologically was proposed. Two sets of sensitive numerical experiments were employed to verify the proposed mechanism. The main conclusions of this study are as follows: (1) In spring, the presence of SWVC over the southeastern flank of TP is directly responsible for the climatic formation of SPR. It causes strong velocity convergence at its rear and is crucial for rich moisture to be transported to and converge over southeastern China. (2) In spring, the emergence of SWVC over the southeastern flank of TP is attributed to the mechanical and thermal effects of TP. The southwesterly wind is related to the deflected flow round TP and the sensible heating over southeastern TP. The rise of TP not only splits and deflects the westerly wind belt, but also causes the rapid increase of diabatic heating over TP. And the heating produces cyclonic vortex source, strengthens deflected southwesterly wind of southeastern TP and leads to the emergence of SWVC. (3) The mechanical and thermal effects of the giant TP are the essential climatic cause of SPR formation. These effects play a primary role on the formation of SPR and contribute significantly to the unique East Asia atmospheric circulation. As for all kinds of synoptic or climatic events, some are results of the effect of sea-land distribution only, such as sea-land winds. Some are not distinct without topography, such as the Asia monsoon. It is known that monsoon is caused by the changes of the thermal contrast of sea-land. It relates to not only the total area and geophysical position, but also the topography of the continent. The existence of TP is the key to the distinct appearance of the Asia monsoon. Sometimes, the topography is a decisive factor for some climatic phenomena, such as SPR. In the view of the results of the model experiments with or without TP and with TP rising, it is clear that the existence of TP not only yields SPR, but also amplifies the rain belt of the polar front near East Asia’s coast remarkably. It is quite uncertain to deduce that SPR is not related to the effect of topography only based on the concurrent rapid rain increasing over the region from SEC to the south of Japan in early spring[14]. So the relative importance of sea-land distribution and topography must be analyzed concretely case by case. Several problems are worth further discussing. First, the temporal span and spatial position of SPR remain unclear. It is easy to locate the time of the expiry of SPR when considering the abrupt changes of the atmospheric circulation from winter to summer as the symbols of the end of SPR. But it is difficult to find an abrupt change in circulation corresponding to the onset of SPR because of the continuity of the winter circulation around SPR onset. As to the spatial distribution of SPR, it is accepted in tradition that SPR lies between the middle and lower reaches of the Yangtze River and the Nanling Mountains[12,14]. In fact, the SPR rain belt coincides with the axes of the Nanling Mountains and the Wuyi Mountains, as shown in Figure 1. Its southern border may move southward and reach 24°N at least. Second, it is interesting to note that the SPR rain belt does not move with seasonal evolution except for the summer monsoon season, as shown in Figure 2, which might be related to those mountains. But how do the mountains affect the SPR rain belt? Third, besides the alleged time-lag effect, what is the other factors affecting the rainfall of SPR? The final also the most important question is what’s the relationship between SPR and the East Asia monsoon? All these questions deserve more attention and further studies. We would like to thank NCAR for access to Xie and Arkin’s rainfall data archives and for providing the NCEP/NCAR reanalysis data. WAN RiJin et al. Sci China Ser D-Earth Sci | January 2007 | vol. 50 | no. 1 | 130-144 143 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 Li C Y, Qu X. Large scale atmospheric circulation evolution corresponding to the onset of the South China Sea summer monsoon. Chin J Atmos Sci (in Chinese), 2000, 24(1): 1―14 Wu G X, Zhang Y S. The effect of thermal and mechanical forcing of Tibet Plateau and the onset of the Asia summer monsoon, Part I: The place of onset. Chin J Atmos Sci (in Chinese), 1998, 22(6): 825 －838 Jin Z H, Tao S Y. The features of the phases of the onset, activity and intervals of the South China Sea summer monsoon. Clim Environ Res (in Chinese), 2002, 7(3): 267－278 Yeh T C, Luo S W, Zhu B Z. The structure of stream fields over the Tibetan Plateau and surrounding areas and the balance of quantity of heat in the troposphere. Acta Meteorol Sin (in Chinese), 1957, 28(2): 108－121 Zhao P, Chen L X. Climatic features of atmospheric heat source/sink over the Qinghai-Xizang Plateau in 35 years and its relation to rainfall in China. Sci China Ser D-Earth Sci, 2001, 44(9): 858－864 Yanai M, Song Z. Seasonal heating of the Tibetan Plateau and its effects on the evolution of the Asian summer monsoon. J Meteor Soc Jpn, 1992, 70: 189－221 Yanai M, Li C. Mechanism of heating and the boundary layer over the Tibetan Plateau. Mon Wea Rev, 1994, 122(2): 305－323[DOI] Wu G X, Li W P. The sensible heat air pump of Tibet Plateau and the Asia Summer Monsoon. In: Yeh, T C, ed. Zhao Jiuzhang Corpus (in Chinese). Beijing: Science Press, 1997. 116－126 Wu G X, Zhang Y S. The effect of thermal and mechanical forcing of Tibet Plateau and the onset of the Asia summer monsoon, Part 2: The date of onset. Chin J Atmos Sci (in Chinese), 1999, 23(1): 51－ 61 Li M C, Pan J F, Tian S C, et al. The Forecast Method of Spring Persistent Low-temperature and Rains (in Chinese). Beijing: Science Press, 1977. 3－6 Wu B J, Peng Z B. The advance of the study on the spring persistent rain over southeastern China. Sci Tech Bull (in Chinese), 1996, 12(2): 65－70 Chen S D, Wang Q Q, Qian Y F. A pilot study on the basic climatic features of flood season rainfall over southeastern China and the relationship to abnormal sea surface temperature. J Tropical Meteorol (in Chinese), 2003,19(3): 260－268 Wang Q Q, Chen S D. The SVD analyses of the relationship between the flood season rainfall over southeastern China and tropical sea temperature. Drought Meteorol (in Chinese), 2004, 22(3): 11－ 16 Tian S F, Yasunari T. Climatological aspects and mechanism of spring persistent rains over central China. J Meteorol Soc Jpn, 1998, 76(1): 57－71 Xie P P, Arkin P A. Global precipitation: a 17-year monthly analysis based on gauge observations, satellite estimates and numerical model outputs. BAMS, 1997, 78(11): 2539－2558[DOI] Yang D S, Liu Y B, Liu S G. Mechanical Meteorology (in Chinese). Beijing: Meteorological Press, 1983. 183－185 Wu G X, Liu Y M, Liu P. The effect of the spatial inhomogeneous heating on the formation and variation of subtropical high belt (I): scale analyses. Acta Meteorol Sin (in Chinese), 1999, 57(3): 257－ 263 Liu Y M, Wu G X, Liu H, et al. The effect of the spatial inhomogeneous heating on the formation and variation of subtropical high belt (III): condensation latent heating and the South Asia High and 144 View publication stats 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 the West Pacific Subtropical High. Acta Meteorol Sin (in Chinese), 1999, 57(5): 525－537 Flohn H. Contributions to a Meteorology of the Tibetan Highlands. In: Atmospheric Science Paper, 130. Fort Collins: Colorado State University Press, 1968. 1－122 Manabe S, Terpstra T B. The effects of mountains on the general circulation of the atmosphere as identified by numerical experiments. J Atmos Sci, 1974, 31: 3－42 Hohn D G, Manabe S. The role of mountains in the South Asian monsoon circulation. J Atmos Sci, 1975, 32: 1515－1541 Kuo H L, Qian Y F. Numerical simulation of the development of mean monsoon circulation in July. Mon Wea Rev, 1982, 110(2): 1879－1897[DOI] Wu A M, Li Y Q. Numerical experiment about the effect of Tibetan Plateau on the mean monsoon circulation over Asia. Plateau Meteorol (in Chinese), 1997, 16(2): 153－164 Qian Z A, Wu T W, Lu S H, et al. The numerical simulation of the climatic cause of the drought climate over northwest in summer―the influences of plateau topography and circulation fields. Chin J Atmos Sci (in Chinese), 1998, 22(5): 753－726 Qian Y, Qian Y F. The numerical sensitive experiments concerning the effect of the rise of Tibetan Plateau on the atmospheric circulation in summer. Acta Meteorol Sin (in Chinese), 1996, 54(4): 474－ 483 Wu G X. The currently progress on the study of climatic mechanicals of Tibetan Plateau in China. The Quaternary Period Study (in Chinese), 2004, 24(1): 1－9 Wu G X, Liu Y M, Liu X, et al. How does the heating of Tibetan Plateau affect the climatic pattern over Asia in summer. Chin J Atmos Sci (in Chinese), 2005, 29(1): 47－56 Wu G X, Zhang X H, Liu H. The study on global sea-air-land model (GOALS/LASG) and its simulations. J Applied Meteorol (in Chinese), 1997, 8(suppl): 15－28 Shi G Y. An accurate calculation and representation of the infrared transmission function of the atmospheric constitutes. Dissertation for the Doctoral Degree. Tohoku University of Japan, 1981. 191 Slingo J M. The development and verification of a cloud prediction scheme for the ECMWF model. Quart J Roy Meteor Soc, 1987, 113: 899－927[DOI] Shao H, Qian Y F, Wang Q Q. The effects of the diurnal variation of solar radiation on climate modeling of R15L9. Plateau Meteor (in Chinese), 1998, 17: 158–169 Xue Y K, Sellers P J, Kinter J J, et al. A simplified biosphere model for global climate studies. J Clim, 1991, 4: 345－364[DOI] Liu H, Wu G X. Impacts of land surface on climate of July and onset of summer monsoon: A study with an AGCM plus SSiB. Adv Atmos Sci, 1997, 14: 289－308 Wu T W, Liu P, Wang Z Z, et al. The performance of atmospheric component model R42L9 of GOALS/LASG. Adv Atmos Sci, 2003, 20(5): 726－742 Wang Z Z, Wu G X, Wu T W, et al. Simulation of Asian monsoon seasonal variations with climate model R42L9/LASG. Adv Atmos Sci, 2004, 21(6): 879－889 Zhang Y C, Qian Y F. Simulation of the rise of Tibetan Plateau on the atmospheric critical height. Acta Meteorol Sin (in Chinese), 1999, 57(2):157－167 WAN RiJin et al. Sci China Ser D-Earth Sci | January 2007 | vol. 50 | no. 1 | 130-144