Atmospheric Research 116 (2012) 105–115
Contents lists available at SciVerse ScienceDirect
Atmospheric Research journal homepage: www.elsevier.com/locate/atmos
Tropospheric temperature inversion over central China Yunying Li ⁎, Jiping Yan, Xingbin Sui Institute of Meteorology, PLA University of Science and Technology, Nanjing, 211101, China
a r t i c l e
i n f o
Article history: Received 24 July 2011 Received in revised form 10 March 2012 Accepted 12 March 2012 Keywords: Temperature inversion Mechanism Central China
a b s t r a c t A four-tiered pattern of temperature inversion over central China is revealed using radiosonde data at seven stations in China during 1990–2010. The statistical features of climatic distribution, typical temperature structure and seasonal variation of each temperature inversion are obtained. The formation mechanism of this four-tiered pattern is then explored. The results show that there is one temperature inversion layer in the lower troposphere, one in the middle troposphere, and two inversion layers in the upper troposphere. The inversion in the lower troposphere occurs frequently in spring and autumn, and is likely caused by radiation cooling. The inversion in the middle troposphere is a representative of temperature characteristic over central China; it appears in winter and disappears in summer. The inversion in the middle troposphere depends on geographical location and synoptic cold-front system. It is the result of cold advection from northern China at a lower level and warm advections from both western and southern China at a higher level. The two inversions in the upper troposphere are associated with extratropical tropopause and tropical tropopause, respectively. The extratropical tropopause generally occurs in higher latitude and rarely occurs in lower latitude over China, with higher frequencies in winter and spring and a lower frequency in summer. The tropical tropopause varies little with season or latitude, neither does the associated inversion. Only ~35% of the extratropical tropopause samples are accompanied by temperature inversion, while nearly 100% of the tropical tropopause samples are accompanied by inversion. The tropical-tropopause inversion is a temperature inflection from troposphere to stratosphere. The mechanism of the extratropical-tropopause inversion needs further study. © 2012 Elsevier B.V. All rights reserved.
1. Introduction Temperature inversion layers have been frequently observed and identified over China; they are characterized by changes in vertical temperature gradient from negative to positive. Temperature inversion can cut off mass and energy exchanges between the layers above and below the inversion, and therefore influence atmospheric dynamic and thermodynamic processes (Nodzu et al., 2006; Cao et al., 2007). Temperature inversion also plays an important role in the development of potential instability (Fulks, 1951). In addition, when temperature inversion occurs, slowly ascending moist air in a stable environment may be suppressed, resulting in consistent stratiform clouds. For example, Yu et al. (2001, 2004a) pointed ⁎ Corresponding author. E-mail address:
[email protected] (Y. Li). 0169-8095/$ – see front matter © 2012 Elsevier B.V. All rights reserved. doi:10.1016/j.atmosres.2012.03.009
out that the globally unique middle-stratiform clouds (with cloud top height between 680 and 440 hPa) dominated central China (east of 103ºE, around 30ºN) because a divergence layer in the middle troposphere limited clouds' vertical development. Li et al. (2005) confirmed that the existence of a midlevel temperature inversion layer facilitated the formation of mid-level clouds. According to formation mechanisms, temperature inversion can usually be classified as radiation inversion, advection inversion, subsidence inversion, turbulence inversion, frontal inversion, 0 °C level inversion, tropopause inversion, etc. Many kinds of inversion phenomena have been examined all over the world, such as surface-based temperature inversions in Alaska (Bourne et al., 2010; Mölders and Gerhard, 2010), surface temperature inversion in the interior of the Antarctic ice sheet in winter (Connolley, 1996), low-level temperature inversion along the Alaskan Arctic Coast (Kahl, 1990), clear-
106
Y. Li et al. / Atmospheric Research 116 (2012) 105–115
sky temperature inversion over the Arctic Ocean (Devasthale et al., 2010), global radiation inversion and capping inversions at the top of the boundary layer (1-km height) in California, USA, and the west coast of Peru (Rogers et al., 1995), the trade inversion (height of ~2 km) over the tropical and subtropical oceans (von Engeln et al., 2005; Cao et al., 2007), inversion near the 0 °C level (height of ~4 km) over the tropics (Johnson et al., 1996), and global tropopause inversion (height of ~10 km; Birner et al., 2006; Randel et al., 2007; Bian and Chen, 2008; Randel and Wu, 2010). These studies have improved our understanding of global temperature distribution; however, up to now very little is known about the temperature inversion over China (Yuan et al., 2003). Temperature inversions are observed frequently over central China, an area around 30ºN that is a unique climate zone. From west to east along the 30ºN parallel in China, the landform varies from the Tibetan Plateau (TP; the roof of the world with mean altitude > 4500 m), to the steep slope, then descends through the Sichuan Basin (SB; with an average altitude of ~500 m), to an extensive plain with an average altitude of ~30 m (Fig. 1). Along this typical climate zone, many meteorological phenomena, such as clouds, wind, and temperature distribution, are distinctive (Yu et al., 2004a; Li and Gu, 2006). The aim of this study is to investigate the characteristics of temperature inversion over central China and its relationship with local topographic effects and large-scale circulation. We will use long-term radiosonde measurements to explore the climatological characteristics of the temperature inversion over central China, including its occurrence frequency, height, intensity, seasonality and temperature structure. We will focus on the mechanisms of various inversions. 2. Data processing The radiosonde data from the World Meteorological Organization (WMO) consists of twice daily data (0000 and 1200 UTC,
which corresponds to 0800 and 2000 Beijing Standard Time, respectively) at 118 stations in China. In this study, the data at seven stations from 1990 to 2010 are used, including five stations near 30ºN and two stations far away from 30ºN. The five stations (from west to east) are Chengdu, Chongqing, Enshi, Wuhan, and Hangzhou. The rest two stations are Shenyang and Fuzhou, representing northern China and southern China, respectively (geographical locations are listed in Table 1 and are marked by stars in Fig. 1). About 5–8 groups of the radiosonde measurements are recorded every 1 min, which leads to a vertical resolution of about 50–80 m given the typical ascent rate of 400 m/min of the radiosonde balloons. In the radiosonde dataset, data with TTAA indicator are at mandatory levels (1000, 925, 850, 700, 500, 400, 300, 250, 200, 150, and 100 hPa), data with TTCC indicator are above 100 hPa (70, 50, 30, 20, and 10 hPa), and data with TTBB indicator are at significant levels. Available meteorological elements include geometrical height, temperature, dew-point depression, wind speed, and wind direction at each mandatory pressure level, and temperature and dew point at each significant pressure level. Before analysis, the original codes are decoded and quality controlled. First, temperature lower than the dew point is treated as erroneous. Second, all data are checked against statistic means to remove bad data. Because the number and occurrence height of the significant levels are different at each time, TTAA, TTBB, and TTCC data are combined and interpolated to constant isobaric surfaces using cubic spline interpolation from 1020 to 20 hPa with an interval of 20-hPa to facilitate statistical analysis because TTBB measurements are only provided on pressure surfaces. However, in the lower troposphere below 1000 m, 20-m geometrical height interval is also used in our study to investigate the base height and the thickness of the inversion when station elevations are considered. We also test the sensitivity of two intervals 20-hPa and 20-m at all altitudes and find that the inversion patterns are similar. The inversion occurrence frequency, height, and intensity are identified only when both TTAA and TTBB
Fig. 1. Topography and locations of stations used in this study.
Y. Li et al. / Atmospheric Research 116 (2012) 105–115
107
Table 1 Information about the stations and their data coverages. Stations
Chengdu
Chongqing
Enshi
Wuhan
Hangzhou
Shenyang
Fuzhou
Number Latitude (ºN) Longitude (ºE) Elevation (m) Data start time Data end time Data availability
56294 30.40 104.01 507.3 1990.01 2004.06 193/198
57516 29.35 106.28 260.4 1996.06 2010.12 172/175
57447 30.17 109.28 458.0 2000.06 2010.12 122/127
57494 30.37 114.08 27.0 1990.01 2010.12 249/252
58457 30.14 120.10 43.2 2001.11 2010.12 107/110
54342 41.44 123.27 45.2 1990.01 2010.12 225/228
58847 26.05 119.17 85.4 1996.07 2010.12 169/174
data are available. The monthly inversion intensity is calculated using days with inversions during 1990–2010. Note that data at the seven stations cover different years (Table 1). Temperature inversion is simply defined as the temperature at upper level greater than or equal to the temperature at an immediate lower level, which means a zero or positive temperature gradient in the vertical. The occurrence frequency is defined as the ratio of inversion samples to the total data samples available. The intensity is defined as the temperature difference between the upper and lower levels at 20-hPa interval at all altitudes and 20-m interval below 1000 m. The daily data are averaged to the monthly data at 0000 and 1200 UTC (or 0800 and 2000 Beijing Standard Time; BST), respectively. To reduce statistical error, the corresponding monthly data are set to missing data when data samples are less than five in one month. Data availability at the seven stations (Table 1) shows the effective number of month and the total number of month during the available data period. The National Centers for Environmental Prediction (NCEP) FNL (Final) operational global analysis data from the global data assimilation system (GDAS) with resolution of 1°× 1° in 2000–2008 are also used in this study. Monthly-mean temperature and wind data at four times (i.e., 0000, 0600, 1200, and 1800 UTC, namely 0800, 1400, 2000, 0200 BST) are used to obtain temperature advection. 3. Temperature inversions and their formation mechanism 3.1. Climatological distribution The seasonal variations of inversion frequencies over central China are shown in Fig. 2. Station Chengdu, located at the SB leeward of the TP, exhibits a typical temperature inversion distribution over central China, with four distinct inversion layers in the troposphere. The occurrence frequency, height, and seasonal variation of these four inversions are very different. The first inversion layer occurs below 950 hPa, with higher frequency in autumn and winter; the frequency can be up to 50% in November. The second inversion layer appears around 650 hPa and is robust in winter. The occurrence frequency reaches 35% in January, and drops back to less than 5% in July. The third inversion layer occurs ~250 hPa, with its highest frequency of 25% in March, becomes nearly zero in summer, and has a discernible base-height increase from winter to spring. The fourth layer appears above 100 hPa and has almost no seasonal variation; its frequency is nearly 100%. With four remarkable temperature inversion layers, Chengdu becomes a representative station of central China,while many other stations far away from central China have very weak mid-level inversion. The inversion
patterns of the other four stations over central China are very similar to that at Station Chengdu (Fig. 2b–e), though the frequency and central height of each inversion layer vary somewhat from station to station. 3.2. Formation of the inversion layer in the lower troposphere The lowest inversion layer over central China appears below 1000 m, mainly occurs in autumn, winter and spring, and always touches the ground (Fig. 3) when station elevations are considered. Is this inversion a radiation inversion? Radiation inversion usually forms at dusk, strengthens in the early morning on the following day, and diminishes after sunrise when the Earth's surface and lower atmosphere are heated. In general, it starts to develop at 2000 BST and is very shallow, while at 0800 BST the following day, after having developing for several hours, it is stronger than which at 1200 UTC. So the frequencies and intensities at 0800 BST in the morning and 2000 BST in the afternoon should be quite different. A comparison of Fig. 3a and b indicates that the mean inversion frequency at the five stations (Chengdu, Chongqing, Enshi, Wuhan and Hangzhou) appears dominantly at 0800 BST (Fig. 3a) than at 2000 BST (Fig. 3b). At 0800 BST, the frequency can be up to 40% in November and March, and half of the inversion layers can develop to a height of 700 m in March and 500 m in November. At 2000 BST, the frequency only reaches 35%, and half of the inversion layer can develop to 200 m in November; the maximum in March disappears. When comparing the inversion strength at 0800 BST (Fig. 4a) and 2000 BST (Fig. 4b), it is found the maximum is near the ground and the magnitude can be up to 0.09 K per 20 m in November at 0800 BST, while it is only 0.06 K per 20 m in November at 2000 BST; and the maximum in March again disappears. The comparison at 0800 and 2000 BST shows that from dusk to the dawn of the next day the temperature inversion gradually develops and its thickness and strength gradually increase. It forms earlier and lasts longer in autumn than in spring. The characteristics of the surface inversion and its diurnal variation suggest that the inversion in the lower troposphere may be attributed to radiation inversion. When comparing the monthly-mean structures of the temperature and dew-point depression among the five stations (Fig. 5), we find they can be classified into two categories. The first category includes stations Chengdu, Chongqing and Enshi, with a negative or weak positive temperature gradient (Fig. 5a) and smaller dew-point depression (Fig. 5b). The second category includes stations Wuhan and Hangzhou, with a positive temperature gradient and rapid increase of dew-point depression with height. These characteristics may
108
Y. Li et al. / Atmospheric Research 116 (2012) 105–115
a
b
c
d
e
Fig. 2. Seasonal variation of inversion frequency (%) over stations Chengdu (a), Chongqing (b), Enshi (c), Wuhan (d) and Hangzhou (e).
be relevant to their geographical locations. Stations Chengdu, Chongqing and Enshi are located at the SB and have smaller dew-point depression (b4 K); they are in a moist and cloudy environment unfavorable to radiation cooling. So their temperature inversion intensity is relatively weak (not shown), and their inversion layer is washed out when taking monthly average. On the other hand, stations Wuhan and Hangzhou are located in open plain and have larger dew-point depression; they are in a relatively dry and clear-sky environment favorable
to radiation cooling, leading to higher inversion frequency and positive temperature gradient. Fig. 5b also illustrates that within the inversion layer the monthly-mean dew-point depression is less than 6 K, and above the inversion layer the dew-point depression increases or the relative humidity decreases rapidly with height. On average, the mean top height or the mean thickness of the radiation inversion is 300 m. Note that the time of radiosonde detection does not reflect the exact time when an inversion occurs, and the fixed times
Y. Li et al. / Atmospheric Research 116 (2012) 105–115
a
109
b
Fig. 3. Five-station-averaged seasonal variation of inversion frequency (%) at 0800 BST (a) and 2000 BST (b).
the north to the south. Between 800 and 500 hPa, the southerly wind is dominant, and the monthly-mean maximum wind speed is ~1 m s− 1 at 700 hPa. Fig. 6 shows that in winter cold advection is southward in the lower troposphere, while warm advection is northward in the middle troposphere. When cold air is under warm air, temperature inversion appears. On the contrary, in July the northerly wind always dominates below 370 hPa, and the same wind direction is unfavorable to the formation of mid-level inversion. Where the winter inversion mainly arises from the meridional cold advection in the lower troposphere, a certain link between the temperature gradient in the vertical and the strength of the winter monsoon may exist. When examining this relationship at Station Chengdu, we find it to be relatively weak due to Chengdu's unique geographical location: the SB is blocked by Qinling and Daba mountain ranges in the north. However, when we verify this relationship at Station Wuhan, a plain site where cold air blows unhindered from the north, a significant correlation is discovered. Fig. 7 shows time series of meridional wind-speed difference between 700 and 925 hPa and vertical-averaged temperature gradient between 925 and 700 hPa over Station Wuhan in January. Two lines show a high correlation coefficient of 0.84, significant at the 99% confidence
of 0800 and 2000 BST may represent different local times. Thus, the actual frequency of the radiation inversion may be higher than is calculated here. 3.3. Formation of the inversion layer in the middle troposphere The temperature inversion in the middle troposphere over central China has a higher frequency in winter than in summer. For example, over Station Chengdu, the top height of the inversion layer is ~550 hPa, slightly higher than the average altitude of the TP surface (600 hPa), and the bottom height is ~750 hPa (see Fig. 2a), with an inversion intensity of 0.8 K per 20 hPa. To which type does this temperature inversion belong? Certainly, it cannot be a 0 °C-level inversion because the temperature is below 0 °C. Is it an advection inversion, then? To examine the inversion features, temperature advections in the zonal and meridional directions are analyzed, respectively. In the meridional direction, the five-station-mean wind vectors are different in winter and in summer below 800 hPa (Fig. 6). On average, the dominant wind direction below 800 hPa is southward in January, though with a smaller average wind speed of about −1 m s− 1. However, conveyed by the prevailing winter monsoon, cold air is transported from
a
b
Fig. 4. Five-station-averaged seasonal variation of inversion strength (K per 20 m) at 0800 BST (a) and 2000 BST (b).
110
Y. Li et al. / Atmospheric Research 116 (2012) 105–115
a
b
Fig. 5. Monthly-mean temperature (K) profiles (a) and dew-point depression (K) profiles (b) at 0800 BST in November at the five stations.
level. In general, southerly wind dominates 700 hPa and northerly wind dominates 925 hPa over Wuhan in January. When the northerly wind is under the southerly wind, meaning that
Fig. 6. Five-station-averaged meridional wind-speed (m s− 1) profiles in January (solid line) and July (dashed line).
Fig. 7. Meridional wind speed difference between 700 and 925 hPa (solid line) and average temperature gradient from 925 to 700 hPa (dashed line) in January over Wuhan. Left coordinate is for the difference of the wind speed; units: m s− 1, right coordinate is for temperature gradient; units: K/20 hPa.
warm advection is above cold advection, an inversion layer forms. The greater the wind shear is, the stronger the inversion strength has. When we further choose three representative months (January in 1993, 2000 and 2008) of larger wind speed difference and two representative months (January in 1992 and 2002) of smaller wind speed difference to analyze the relationship between these two factors, as shown in Fig. 8, the relation is more remarkable. In months of smaller wind-speed difference, the meridional wind vector is all negative (or southward) at every level below 630 hPa (dashed line in Fig. 8a). The wind speed difference between the upper and lower levels is smaller, and the monthly-mean temperature gradient (dashed line in Fig. 8b) gently increases above 970 hPa. While in months of larger wind speed difference, northerly wind prevails below 830 hPa and southerly wind prevails above 830 hPa (solid line in Fig. 8a), namely cold advection exists at lower levels and warm advection exists at middle levels, which results in conditions favorable to temperature inversion. The strongest inversion occurs at 800 hPa, with a relatively rapidly increasing temperature gradient that is close to positive gradient (solid line in Fig. 8b). When cold advection appears below warm advection and inducing positive gradient, temperature inversion forms. However, which advection is dominant and what causes these advections? When examining the synoptic system, we find in most circumstances temperature inversion occurs accompanied by cold front system. The cold air accompanies the front passage from the north, leading to cold advection in the south–north direction and positive temperature gradient in the vertical. For example, during 8–11 January 2008, a robust anticyclonic circulation system revealed by the NCEP data moved from the north to southern China, with strong northerly wind (Fig. 9a). At 0200 BST on 6 January 2008 before the cold air invaded southern China, the meridional wind vector was positive at all vertical levels, indicating dominant southerly wind (solid line in Fig. 9b). Controlled by weak pressure, no mid-level temperature inversion was seen except for a near-surface radiation inversion (solid line in Fig. 9c). However, when cold air passed through central China at 0200 BST on 11 January 2008, northerly wind prevailed below 850 hPa (dashed line in Fig. 9b), while southerly wind prevailed above 850 hPa, forming a mid-level inversion between 850 and 750 hPa (dashed line in Fig. 9c) over central China, say at Station Hangzhou. There was no notable diurnal
Y. Li et al. / Atmospheric Research 116 (2012) 105–115
a
a
b b
c Fig. 8. Meridional wind speed (m s− 1) profiles (a) and temperature gradient (K per 20 hPa) profiles (b) at Station Wuhan. The solid lines are for the January mean in 1993, 2000, and 2008; the dashed lines are for the January mean in 1992 and 2002.
change and inversion always maintained when cold advection was present (not shown). The case analysis indicates that perhaps the cold advection resulted from cold front system contributes more than the warm advection to the midlevel inversion. In the zonal direction, on the other hand, the westerlies prevail in the middle troposphere in winter over central China. When comparing the surface air temperature of the TP with its downstream regions at the same latitude and altitude in February, it is found that during the local daytime, the surface air Fig. 9. Geopotential height (gpm) and wind vector (m s− 1) at 925 hPa (a), meridional wind speed (m/s) profiles (b), and temperature (K) profiles (c) at 0200 BST on 06 January 2008 (solid line) and 0200 BST on 11 January 2008 (dashed line) at Station Hangzhou. Pentacle in a stands for the location of Station Hangzhou.
111
112
Y. Li et al. / Atmospheric Research 116 (2012) 105–115
a
b
a
b
Fig. 10. Latitude–pressure cross section of air temperature (K) and wind vector (m s− 1) at 1400 BST (a) and 0200 BST (b) in February. Shaded area indicates topography.
Fig. 11. Diurnal variation of the temperature difference (K; left coordinate; dashed line) and temperature advection (K m s− 1; right coordinate; solid line) between location (30°N, 90°) and (30°N, 105°E) in February (a) and July (b).
temperature over the TP is higher than that in its downstream regions, but the opposite circumstance occurs during the local nighttime (Fig. 10a and b). Carried by the westerlies, the warm air moves from the TP to its downstream regions during the day, while at night warm advection is very weak or cold air moves downstream. Let us define T1 as the temperature at 30°N, 90°E at 550 hPa (to represent the air temperature over the TP), T2 as the temperature at 30°N, 105°E at 550 hPa (air temperature over the downstream plain region), and U as the mean zonal wind speed between these two points. We use U ×(T1 −T2) to describe the strength of the temperature advection (Fig. 11). In the daytime of 1400 and 2000 BST, (T1 −T2) derived from the NCEP data is positive, with a temperature difference of 4 K, and U× (T1− T2) exhibits a strong warm advection, with an intensity of 30 K m s− 1. At the nighttime of 0200 and 0800 BST, (T1 −T2) is negative, with a temperature difference of −4 K, and U× (T1− T2) exhibits a strong cold advection, with an intensity of −30 K m s− 1 (Fig. 11a). The TP is thought by some as a source of cooling in winter, but the data show that winter TP surface air temperature in the daytime is relatively higher than that of the downstream regions at the
same latitude and altitude. The warm advection in the middle troposphere may provide a condition favorable to form inversion in downstream regions. Using a sensitivity test with a high vertical resolution regional model, Advanced Regional Eta-coordinate Model (AREM; Yu, et al., 2004b), Li et al. (2005) showed that the daytime temperature inversion to be much stronger than that at night, and confirmed the contribution of the warm advection to the inversion formation. But to what extent does the zonal warm advection in the middle troposphere contribute to the inversion formation? Is it a determining factor for inversion formation? When calculating the thickness of the warm advection layer, we find it to be very thin, and with the daytime occurrence frequency of 50%, such inversion can hardly be found in the climatological temperature profile when all data are averaged. The area influenced by the TP's thermal effect is also limited to within the SB. On the other hand, as a heat source in summer, the TP should produce stronger warm advection and should lead to much stronger downstream inversion if warm advection is very important in forming inversion. However, the inversion frequency in the middle troposphere in summer over central China is very low (see Fig. 2). The maximum temperature difference between
Y. Li et al. / Atmospheric Research 116 (2012) 105–115
the TP and the plains is up to 6.3 K during the day, and the minimum reaches 4.4 K at night (Fig. 11b). The maximum advection strength only reaches 16 K m s− 1, less than that in winter because of the relatively weaker westerlies. The relatively weaker warm advection and lower inversion frequency imply that zonal warm advection may not be the determining factor in generating the inversion, but it does provide a condition favorable to the inversion formation in winter. The five stations have different central height of the midlevel inversion (see Fig. 2). Of all five stations, Chengdu is the closest station to the TP and has higher inversion frequency and strongest inversion intensity because of the thermal effect of the TP. With increasing distance away from the TP (see Fig. 1), the inversion occurrence height falls from 650 hPa (Station Chengdu) to 850 hPa (Station Hangzhou). The midlevel inversion center is at 650 hPa (~3754 m in January) over Chengdu, at 700 hPa (~3059 m in January) over Chongqing, at 750 hPa (~2623 m in January) over Enshi, at 820 hPa (~1857 m in January) over Wuhan, and at 850 hPa (~1617 m in January) over Hangzhou. Further analysis reveals a link between the occurrence height and the station's geographical location (including station elevation and its surrounding topography). For example, Chengdu and Chongqing are located in the SB surrounded by mountains, especially the Qinling and Daba mountain ranges in the north (see Fig. 1). The Qinling Mountains extend from east to west, with an average elevation of 1000 m and several peaks higher than 1500 m. The Daba Mountains are located south of the Qinling Mountains and extend from northwest to southeast, with an average elevation of 1000 m. These two mountain ranges have strong blocking effects on low-level cold air. On the other hand, the height of warm advection from the TP is generally ~4000 m (~620 hPa). Influenced by these two factors, the central height of inversion over Chengdu and Chongqing is above 700 hPa. Compared with Chongqing, Chengdu is closer to the TP and has a higher inversion center height and stronger inversion intensity or higher inversion frequency, i.e., it is affected by much more powerful thermal effects of the TP. Enshi is located at the east bridge of the SB, and cold air is not blocked by mountains. In addition, Enshi is farther away from the TP than Chengdu and Chongqing, and is affected much less by the TP, leading to a relatively lower inversion height of 750 hPa. Wuhan and Hangzhou are located in the plains. With very low elevations, they are not subject to mountain blocking or mid-level warm advection impact. Cold air at lower levels can invade easily from the north, resulting in very low height of temperature inversion center. Overall, the presence of the middle inversion layer is a representative feature over central China, a typical monsoon region around 30ºN. The winter inversion in the middle troposphere may result from the combined effect of west–east oriented warm advection and/or south–north oriented warm advection in the middle troposphere, and north–south oriented cold advection in the lower troposphere. The cold advection carried by the cold front system in the lower troposphere may have a greater contribution to the temperature inversion than that of the two warm advections.
113
level inversion appears at ~240 hPa, with an ascending bottom height from winter to spring. The higher-level inversion appears at ~90 hPa, with persistent maximum frequency of 100%, exhibiting a feature of tropopause. Many studies (e.g., Santer et al., 2003; Birner et al., 2006; Li et al., 2009) interpreted there are two types of tropopause: one is called extratropical tropopause (known as the first tropopause) at a relatively lower height, and the other is called tropical tropopause (known as the second tropopause) at a relatively higher height, usually above 100 hPa. In this study, we calculate the frequency of the tropopause according to the 1957 definition by the WMO, namely, the lowest level at which the lapse rate is 2 °C/km or less, provided that the average lapse rate between this level and all higher levels within 2 km does not exceed 2 °C/km. We confirm that sometimes these two tropopauses coexist in the mid latitude. On average, the tropical tropopause pressure is ~90 hPa and nearly has no seasonal variation (dashed line in Fig. 12), and its occurrence frequency is about 60% (dash-dotted line in Fig. 12; note that radiosonde often could not reach above 100 hPa and therefore no tropical tropopause is reported). The extratropical tropopause pressure is ~270 hPa in January and rises to 200 hPa from January to May (solid line in Fig. 12), and its frequency is up to 60% in March but zero in summer (dotted line in Fig. 12). The occurrence frequency of the extratropical tropopause varies with latitude. It has a higher occurrence frequency at higher latitude and in winter and spring (e.g., at Station Shenyang, a representative station of northern China, it is 100% in winter and 30% in summer), while has a much lower frequency in lower latitude and in summer (e.g., at Station Fuzhou, a representative station of southern China, it is 16% in winter and zero in summer). On the other hand, the tropical tropopause has a higher frequency in lower latitudes and a lower frequency in higher latitudes, with little seasonal variation (e.g., annual-mean occurrence frequency is 70% at Station Fuzhou and is 50% at Station Shenyang). However, the tropopause is not always accompanied by temperature inversion. Comparing the frequency of the trapopause in Fig. 12 with the frequency of the inversion in Fig. 2, It is found only ~35% of the extratropical tropopause samples are associated with temperature inversion, and nearly 100% of the tropical tropopause samples are associated with temperature inversion. In general, the tropical tropopause is a temperature inflection between troposphere and stratosphere, namely, a temperature inversion. However, why is sometimes the extrotropical tropopause also associated with temperature inversion?
3.4. Formation of the upper troposphere inversion There are two inversion layers over central China in the upper troposphere above 400 hPa (see Fig. 3). The lower-
Fig.12. Five-station-averaged pressure (solid line) and frequency (dotted line) of extrotropical tropopause, and pressure (dashed line) and frequency (dash-dotted line) of tropical tropopause.
114
Y. Li et al. / Atmospheric Research 116 (2012) 105–115
So far the physical mechanism underlying the observed extrotropical tropopause inversion layer (TIL) is not yet well understood. Randel et al. (2007) argued that radiative effects associated with water vapor and ozone should play important roles. Kunz et al. (2009) highlighted the importance of radiative cooling from water vapor. Wirth (2003, 2004) emphasized the importance of the asymmetry between upper level cyclones and anticyclones and their effects on the local stratification around the tropopause. Wirth and Szabo (2007) and Erler and Wirth (2011) tested this idea in idealized baroclinic life cycle experiments and found a TIL forms above anticyclones and remains in the mean due to non-linear effects of wave breaking. Other potentially important mechanisms involve gravity waves, turbulence, deep convection, and the Brewer–Dobson circulation (Müller and Wirth, 2009; Birner, 2010). Different processes may be important in different regions and seasons (Miyazaki et al., 2010a, 2010b).
a
4. Summary and discussions 4.1. Summary A four-tiered temperature inversion pattern over central China is revealed using radiosonde data, and inversion mechanisms are explored. Temperature inversion in the lower troposphere can be categorized as radiation inversion, with higher frequency in autumn and spring, and can occur over almost any region in China. Mid-level tropospheric inversion, a typical vertical temperature distribution over central China, mainly appears in winter and disappears in summer, and possibly results from the combined effect of eastward warm advection and northward warm advection in the middle troposphere, and southward cold advection originated from the cold front system in the lower troposphere in winter. The uppertroposphere inversions can be categorized as extratropical tropopause inversion and tropical tropopause inversion though the mechanism of the extratropical tropopause inversion is not yet well understood.
b
4.2. Discussions Temperature inversion in this study is defined as the temperature at an upper level greater than or equal to the temperature at an immediate lower level. As a result, many cases with smaller lapse rates of temperature are not included in our statistics. The sounding data are only available twice daily at 0800 and 2000 BST, not the local representative daytime and nighttime, so the diurnal variation of eastward warm advection in the middle troposphere could not be fully studied. However, the observations indicate that the inversion intensity over Station Chengdu at 2000 BST is slightly stronger than that at 0800 BST (not shown), and Li et al. (2005) showed a stronger inversion during the daytime and a weak or no inversion at night using a numerical model (the AREM), indicating that the eastward warm advection from the TP does contribute to the formation of the mid-level temperature inversion, though it is not the determining factor. On the other hand, subsiding warming seems to help the formation of the mid-level inversion over Chengdu and Chongqing, but it is difficult to separate subsiding warming from advection warming using the data we have access to. Birner (2006) pointed out that because of substantial variability in the height of the extratropical tropopause (linked to
Fig. 13. Temperature (K) profiles in January, April, July, and October at stations Shenyang (a) and Fuzhou (b). Logarithmic coordinate is used in the vertical.
synoptic-scale eddies), the extratropical tropopause inversion structure is washed out when taking ground-based averages, whereas it is a ubiquitous and clear feature in tropopause coordinate. In our studies, the inversion is also found above the extratropical tropopause in January in northern China (instead of in southern China) from the monthly averaged temperature profiles (Fig. 13) based on the ground coordinate, which we intend to study in future.
Acknowledgments We thank two anonymous reviewers for suggestions that helped improve the paper. This research was jointly supported by the National Basic Research Program of China
Y. Li et al. / Atmospheric Research 116 (2012) 105–115
(grant no. 2010CB951904) and the National Natural Science Foundation of China (grant no. 41075034). References Bian, J., Chen, H., 2008. Statistics of tropopause inversion layer over Beijing. Adv. Atmos. Sci. 25 (3), 381–386. Birner, T., Sankey, D., Shepherd, T.G., 2006. The tropopause inversion layer in models and analyses. Geophys. Res. Lett. 33, L14804. doi:10.1029/ 2006GL026549. Birner, T., 2010. Residual circulation and tropopause structure. J. Atmos. Sci. 67, 2582–2600. Birner, T., 2006. Fine-scale structure of the extratropical tropopause region. J. Geophys. Res. 111, D04104. doi:10.1029/2005JD006301. Bourne, S.M., Bhatt, U.S., Thoman, R., Zhang, J., 2010. Surface-based temperature inversions in Alaska from a climate perspective. Atmos. Res. 95, 353–366. Cao, G., Giambelluca, T.W., Stevens, D.E., Schroeder, T.A., 2007. Inversion variability in the Hawaiian trade wind regime. J. Clim. 20, 1145–1160. Connolley, W.M., 1996. The Antarctic temperature inversion. Int. J. Climatol. 16, 1333–1342. Devasthale, A., Will'en, U., Karlsson, K.G., Jones, C.G., 2010. Quantifying the clear-sky temperature inversion frequency and strength over the Arctic Ocean during summer and winter seasons from AIRS profiles. Atmos. Chem. Phys. Discuss. 10, 2835–2858. Erler, A.R., Wirth, V., 2011. The static stability of the tropopause region in adiabatic baroclinic life cycle experiments. J. Atmos. Sci. 68, 1178–1193. Fulks, J.R., 1951. The instability line. Compendium of Meteorology. American Meteorological Society, Boston, pp. 647–652. Johnson, R.H., Ciesielski, P.E., Hart, K.A., 1996. Tropical inversions near the 0 °C level. J. Atmos. Sci. 53, 1838–1885. Kahl, J.D., 1990. Characteristics of the low-level temperature inversion along the Alaskan Arctic Coast. Int. J. Climatol. 10, 537–548. Kunz, A., Konopka, P., Uller, R.M., Pan, L.L., Schiller, C., Rohrer, F., 2009. High static stability in the mixing layer above the extratropical tropopause. J. Geophys. Res. 114, D16305. doi:10.1029/2009JD011840. Li, W., Fan, C., Yi, F., 2009. Characteristics of tropopause over Wuhan and Haikou. Chin. J. Space Sci. 29, 409–416. Li, Y., Gu, H., 2006. Relationship between middle stratiform clouds and large scale circulation over eastern China. Geophys. Res. Lett. 33, L09706. doi:10.1029/2005GL025615. Li, Y., Yu, R., Xu, Y., Zhou, T., 2005. AREM simulations of cloud features over eastern China in February. Adv. Atmos. Sci. 22, 260–270. Miyazaki, K., Watanabe, S., Kawatani, Y., Tomikawa, Y., Takahashi, M., Sato, K., 2010a. Transport and mixing in the extratropical tropopause region in a high-vertical-resolution GCM, part I: potential vorticity and heat budget analysis. J. Atmos. Sci. 67, 1293–1314.
115
Miyazaki, K., Sato, K., Watanabe, S., Tomikawa, Y., Kawatani, Y., Takahashi, M., 2010b. Transport and mixing in the extratropical tropopause region in a high-vertical-resolution GCM, part II: relative importance of largescale and small-scale dynamics. J. Atmos. Sci. 67, 1315–1336. Mölders, N., Gerhard, K., 2010. A case study on wintertime inversions in Interior Alaska with WRF. Atmos. Res. 95, 314–332. Müller, A., Wirth, V., 2009. Resolution dependence of the tropopause inversion layer in an idealized model for upper-tropospheric anticyclones. J. Atmos. Sci. 66, 3491–3497. Nodzu, M.I., Ogino, S.Y., Tachibana, Y., Yamanaka, M.D., 2006. Climatological description of seasonal variations in lower-tropospheric temperature inversion layers over the Indochina Peninsula. J. Clim. 19, 3307–3319. Randel, W.J., Wu, F., Forster, P., 2007. The extratropical tropopause inversion layer: global observations with GPS data, and radiative forcing mechanism. J. Atmos. Sci. 64, 4489–4496. Randel, W.J., Wu, F., 2010. The polar summer tropopause inversion layer. J. Atmos. Sci. 67, 2572–2581. Rogers, D.P., Yang, X., Norris, P.M., Johnson, D.W., Martin, G.M., Friehe, C.A., Berger, B.W., 1995. Diurnal evolution of the cloud-topped marine boundary layer. Part I: nocturnal stratocumulus development. J. Atmos. Sci. 52, 2953–2966. Santer, B.D., Sausen, R., Wigley, T.M.L., Boyle, J.S., AchutaRao, K., Doutriaux, C., Hansen, J.E., Meehl, G.A., Roeckner, E., Ruedy, R., Schmidt, G., Taylor, K.E., 2003. Behavior of tropopause height and atmospheric temperature in models, reanalyses, and observations: decadal changes. J. Geophys. Res. 108, 4002. doi:10.1029/2002JD002258. Von Engeln, A., Teixeira, J., Wickert, J., Buehler, S.A., 2005. Using CHAMP radio occultation data to determine the top altitude of the Planetary Boundary Layer. Geophys. Res. Lett. 32, L06815. doi:10.1029/ 2004GL022168. Wirth, V., 2003. Static stability in the extratropical tropopause region. J. Atmos. Sci. 60, 1395–1409. Wirth, V., 2004. A dynamical mechanism for tropopause sharpening. Meteorol. Z. 13, 477–484. Wirth, V., Szabo, T., 2007. Sharpness of the extratropical tropopause in baroclinic life cycle experiments. Geophys. Res. Lett. 34, L02809. doi:10.1029/ 2006GL028369. Yu, R., Wang, B., Zhou, T., 2004a. Climate effects of the deep continental stratus clouds generated by Tibetan Plateau. J. Clim. 17, 2702–2713. Yu, R., Yu, Y., Zhang, M., 2001. Comparing cloud radiative properties between the eastern China and the Indian monsoon region. Adv. Atmos. Sci. 18, 1090–1102. Yu, R., Xue, J., Xu, Y., 2004b. AREMS Meso-scale Heavy Rainfall Numerical Prediction Model System (in Chinese). China Meteorological Press, Beijing. 232. Yuan, R., Ma, C., Fan, A., 2003. Characteristics of the capping inversion above the atmospheric convective boundary layer. J. Univ. Sci. Technol. Chin 33, 247–252 (in Chinese).