Vertical wind observation in the tropical upper troposphere by VHF wind profiler: A case study

Transcription

1 RADIO SCIENCE, VOL. 42,, doi: /2006RS003538, 2007 Vertical wind observation in the tropical upper troposphere by VHF wind profiler: A case study Masayuki K. Yamamoto, 1 Noriyuki Nishi, 2 Takeshi Horinouchi, 1 Masanori Niwano, 3 and Shoichiro Fukao 1 Received 24 June 2006; revised 16 October 2006; accepted 1 November 2006; published 18 May [1] Features of upper tropospheric vertical wind (W) over Sumatra, Indonesia, are presented using data observed by a VHF wind profiler installed at West Sumatra (0.2 S, E). During 5 9 May 2004, W from the middle to upper troposphere (8 14 km) changed in accordance with the cumulus activity over Sumatra. During 5 6 May, 3-hourly averaged W continuously showed upward motions up to 0.09 m s 1. The upward motions were observed in the vicinity of deep convective events, which were continuously seen over Sumatra within a synoptic-scale convectively active envelope. After 7 May, when cumulus activity was suppressed over Sumatra, 3-hourly averaged upward motions of greater than 0.05 m s 1 almost disappeared. During 5 6 May, downward motions up to 0.11 m s 1 were observed above 14 km, while upward motions were observed below 14 km. Estimation of W by the European Centre for Medium-Range Weather Forecasts operational analysis have revealed that a major part of observed downward motions above 14 km is explained by the leeward (southwestward) wind and leeward downward tilt of isentropes that existed over western Sumatra. The observed downward motions above 14 km during 5 6 May suggest that downward motions caused by leeward downward tilt of isentropes can be produced in the vicinity of the convectively active region, and leeward downward tilt of isentropes can suppress an upward transport of air mass into the tropical tropopause layer (TTL) by producing downward motions in the TTL. Citation: Yamamoto, M. K., N. Nishi, T. Horinouchi, M. Niwano, and S. Fukao (2007), Vertical wind observation in the tropical upper troposphere by VHF wind profiler: A case study, Radio Sci., 42,, doi: /2006rs Introduction [2] Dynamical processes in the tropical upper troposphere play an important role in the energy and air mass transport between the troposphere and the stratosphere [Holton et al., 1995]. Vertical wind (or vertical motion; hereafter W) is one of important factors that control the energy and air mass transport in the upper troposphere, and is determined by phenomena which range from 1 Research Institute for Sustainable Humanosphere, Kyoto University, Uji, Japan. 2 Division of Earth and Planetary Sciences, Graduate School of Science, Kyoto University, Kyoto, Japan. 3 Frontier Research Center for Global Change, Japan Agency for Marine-Earth Science and Technology, Yokohama, Japan. Copyright 2007 by the American Geophysical Union /07/2006RS synoptic scale [e.g., Fujiwara and Takahashi, 2001; Wheeler et al., 2000] to mesoscale [e.g., Lane et al., 2001; Piani and Durran, 2001]. [3] VHF wind profiler can directly observe vertical and horizontal winds both in clear and precipitation conditions by receiving echoes from refractivity fluctuations caused by turbulence [e.g., Röttger, 1980; Gage, 1990]. In the tropics, VHF wind profilers have been used to investigate tropospheric W profiles in the northern Australia [e.g., Cifelli and Rutledge, 1994], in the Pacific Ocean [e.g., Gage et al., 1991], and in India [e.g., Jagannadha Rao et al., 2003; Dhaka et al., 2002]. [4] The Indonesian Maritime Continent is one of the regions where deep convection which occurs over it influences global-scale circulation [Ramage, 1968]. Sumatra is one of the largest islands in the Indonesian Maritime Continent. A VHF wind profiler named the Equatorial Atmosphere Radar (EAR) was installed at 1of14

2 Table 1. Principal Observation Parameters of the EAR From 5 to 9 May 2004 a Item Standard Mode Vertical Wind Mode Vertical resolution, m Beam direction (0, 0 ), (0, 10 ), (0, 0 ), (0, 0 ), (Az, Ze) (90, 10 ), (180, 10 ), (0, 0 ) (270, 10 ) N coh N FFT N icoh 5 1 Observation time, s Spectral resolution, Hz a N coh,n FFT, and N icoh denote a number of coherent integrations (time domain averaging), FFT points, and incoherent integrations (spectral averaging), respectively. In both observation modes, radar beams were steered on a pulse-to-pulse basis. Kototabang (0.20 S, E, 865 m above sea level; hereafter KT), West Sumatra, Indonesia, in Using EAR wind data, convective features over Sumatra have been revealed [e.g., Renggono et al., 2006, Shibagaki et al., 2006a, Seto et al., 2004, 2006]. Further, EAR observations have revealed phenomena around the tropopause such as an enhancement of turbulence by Kelvin wave breaking [Fujiwara et al., 2003] and an existence of continuous shear instability [Yamamoto et al., 2003]. However, W in the upper troposphere is not always observed by the EAR with a high data rate because of weak Bragg scattering and the limitation of power aperture product. [5] To investigate upper tropospheric W motions by the EAR, the Coupling Processes of the Equatorial Atmosphere (CPEA) project has conducted radiosonde and EAR observations at KT during 5 9 May 2004 [Fukao, 2006]. During 5 9 May 2004, the EAR was operated with an additional observation mode to steer radar beams only to the vertical direction, which contributes the improvement of the data rate of W in the upper troposphere. In this study, two features of upper tropospheric W are presented using W data obtained with the observation mode to intensively observe W. First feature is W changes from the middle to upper troposphere. During 5 6 May, 3-hourly averaged W from the middle to upper troposphere continuously showed upward motions up to 0.09 m s 1. The upward motions were observed in the vicinity of deep convective events which were continuously seen over Sumatra within a synoptic-scale convectively active envelope. After 7 May, when cumulus activity was suppressed over Sumatra, 3-hourly averaged upward motions of greater than 0.05 m s 1 almost disappeared. Second feature is prominent downward motions in the uppermost part of the troposphere in the vicinity of enhanced cumulus activity over Sumatra. To investigate effects of adiabatic processes on the observed downward motions, W caused by adiabatic processes is estimated using European Centre for Medium-Range Weather Forecasts (ECMWF) operational analysis (hereafter ECMWF analysis) data. [6] The paper is organized as follows. In section 2, the data set used for this study is described. In section 3, W variations during 5 9 May 2004 are presented using W data observed by the EAR. W observed by the EAR is compared with that estimated by the ECMWF analysis data. Discussion and conclusions are given in sections 4 and 5, respectively. 2. Data 2.1. Equatorial Atmosphere Radar (EAR) [7] The EAR is a VHF wind profiler operated with a center frequency of 47 MHz and with a peak output power of 100 kw. For the system description of the EAR, see Fukao et al. [2003]. From 5 to 9 May 2004, the EAR was operated by two observation modes; a standard mode to observe vertical and horizontal winds and an additional observation mode to steer radar beams only to the vertical direction, which contributes the improvement of the data rate of W in the upper troposphere (hereafter vertical wind mode). Table 1 shows the observation parameters of the EAR. [8] EAR observations were carried out by alternately changing the two observation modes. In the standard observation mode, the observation time for the vertical direction is s while the total observation time for all the five beam directions is s. In the vertical wind mode, all of radar beams are pointed vertically in the observation time of s to integrate more pulses than the standard observation mode; it contributes a signal-to-noise ratio (SNR) improvement of 7 dbfor Doppler spectra obtained with the vertically pointing radar beam. Similar observation mode to intensively observe vertical direction has been used to investigate vertical wind motions in mesoscale convective systems in the northern Australia [Cifelli and Rutledge, 1994]. W was computed from Doppler spectra observed by the vertical wind mode. In the off-line signal processing to compute W, three Doppler spectra obtained in the same observation time ( s) were averaged to apply a least squares fitting to the echo from atmospheric turbulence [Yamamoto et al., 1988]. Using equation (13) of Yamamoto et al. [1988], an estimation error of W is estimated to be m s 1 in the case of typical spectral width of m s 1. The estimation error is further reduced by averaging W every hour. Using the vertical wind mode, the data rate of hourly averaged W at km is improved to be 85% on average with the minimum of 76% at 14.8 km and the maximum of 2of14

3 96% at 16.5 km, while the data rate of W is 20 25% in the standard observation mode [Alexander et al., 2006] Other Data [9] During 5 9 May 2004, 6-hourly radiosonde soundings were carried out at KT. Launching time is 0000, 0600, 1200, 1800 universal time coordinated (UTC). Note that local standard time (LST) is 7 hours earlier than UTC. RS92-SGP sensor produced by Vaisala was used. Horizontal wind and temperature data derived from radiosondes are averaged with an interval of 150 m. [10] The ECMWF analysis data with 6-hour intervals were used to investigate spatial distributions of temperature and wind around Sumatra. To investigate a detailed isentropic structure around Sumatra, we used the ECMWF analysis data with 60 vertical levels and N256 Gaussian grids; they correspond to a horizontal resolution of 0.35 and vertical resolution of m at km. [11] Hourly blackbody brightness temperature (hereafter T BB ) data derived from the Geostationary Operational Environmental Satellite 9 (GOES 9) IR1 ( mm) sensor were used to examine a horizontal distribution of cumulus activity. Cloud-top altitude was inferred by T BB and vertical profile of temperature. Vertical profile of temperature was computed by averaging temperature data derived from radiosondes from 1 to 9 May Cloud-top altitude was defined as the altitude at which temperature equaled to T BB. However, it is noted that cloud-top temperature inferred from T BB is generally higher than the actual temperature at cloud top because clouds are not regarded as perfect black bodies [Sherwood et al., 2004]; it means that cloud-top altitude inferred from T BB is generally lower than real cloud top, and indicates the lowest altitude where cloud top can exist. 3. Results 3.1. Synoptic-Scale Cumulus Activity [12] Synoptic-scale cumulus activity shifted from the Indian Ocean to the western Pacific Ocean in the beginning of May Figure 1 shows a horizontal distribution of daily averaged T BB. During 3 4 May, in the equatorial band (10 S 10 N), a convectively active region is observed in the west of 100 E (Figures 1a and 1b). On 5 May, some portion of the convectively active region moves eastward and covers Sumatra (Figure 1c). The center of convectively active region which covers Sumatra on 5 May moves northeastward; it reaches to the northeast of Sumatra on 7 May, then to the western Pacific Ocean (east of 120 E) on 9 May (Figures 1d 1g). Convectively active region in the Indian Ocean gradually becomes obscure after 6 May, then almost disappears on 9 May. Using T BB data, Shibagaki et al. [2006b] have shown that the SCC, which passed over Sumatra during 5 6 May, developed during the active phase of intraseasonal variation (ISV; see Zhang [2005] for details of ISV). [13] Figures 2b 2d show a time-altitude plot of the zonal wind (hereafter U), meridional wind (hereafter V) and temperature (hereafter T) at KT, respectively. An intensification of the easterly wind at km and a change from the southerly wind to the northerly wind at km are observed around 0000 UTC 7 May. These wind changes occur as the synoptic-scale convectively active region moves northeastward of Sumatra (see Figures 1c 1e). Other than wind changes seen in the upper troposphere, downward phase propagation of U and T are observed above the tropopause altitude. Because of the downward propagation of warmer T, the tropopause altitude displaced from 16.5 km to 18.0 km on 9 May. These signals in U and T are probably caused by tropopause-level Kelvin wave, as shown in previous studies [e.g., Fujiwara et al., 2003] W From the Middle to Upper Troposphere During 5 9 May [14] W from the middle to upper troposphere observed by the EAR changed as synoptic-scale convectively active region shifted from the Indian Ocean to the western Pacific Ocean. Figure 2a shows a time-altitude plot of W and Figure 3 shows time series of T BB at KT, 3-hourly W averaged over km, and a 3-hourly ratio of upward motion at km. The ratio of upward motion is defined as the fraction of bins counted both in time (3 hour) and altitude (8 14 km). W is averaged every 3 hours to focus on W changes with a timescale of 1 day or longer, and averaged vertically to reduce vertical fluctuations. The lower boundary of W averaging is set to 8 km to reduce effects from local-scale clouds over KT. The upper boundary of W averaging is set to 14 km, because 14 km is the altitude where the air starts to have stratospheric features. Folkins et al. [1999] have shown that the start of increase of ozone and potential temperature occurred at 14 km at Samoa (14.2 S, W), and defined this altitude as the lower boundary of tropical tropopause layer (TTL), where the air gradually changes tropospheric one to stratospheric one. In our case study, a significant contrast of W above 14 km and one below 14 km is found during 5 6 May, when deep convective events were observed over Sumatra; this contrast is presented in section 3.3. On 5 and 6 May, upward motions are prominent at 8 14 km (Figure 2a). The 3-hourly averaged upward motion shows a maximum of 0.18 m s 1 around 1200 UTC 5 May, when very low T BB of less than 200 K, which indicates that cloud top locates at higher than 15.1 km, is observed (Figure 3b). Except for the prominent 3of14

4 Figure 1. Horizontal distribution of daily averaged TBB from 3 to 9 May TBB are averaged over the region. 4 of 14

5 Figure 2. Time-altitude plots of (a) hourly averaged W observed by the EAR, (b) zonal wind, (c) meridional wind, and (d) temperature anomaly observed by radiosondes. The diamonds indicate the locations of cold point tropopause computed from radiosonde data. Temperature anomaly is the deviation from the average during 1 9 May at each altitude range. 5of14

6 Figure 3. (a) Time series of T BB.T BB is averaged over the area centered on KT. (b) Time series of W averaged over km. W is observed by the EAR and is averaged every 3 hours. (c) Time series of ratio of upward motion at km observed by the EAR. The ratio is computed with 3-hour intervals. upward motion around 1200 UTC 5 May, upward motions up to 0.09 m s 1 are continuously observed when W is averaged over 3 hours in time and 8 14 km in altitude. The ratio of upward motions at 8 14 km frequently exceeds 70%, and is greater than 80% around UTC 5 May and UTC 6 May (Figure 3c). The averaged W and the ratio of upward motions at 8 14 km during 5 6 May are 0.05 m s 1 and 70%, respectively. [15] The upward motions during 5 6 May were observed in the vicinity of deep convective events over Sumatra. Figure 4 shows a horizontal distribution of T BB around KT from 0600 UTC 5 May to 0000 UTC 7 May. T BB data are plotted with 6-hour intervals and not averaged in time. T BB of less than 230 K, which locates at 11.6 km or higher altitudes, is continuously observed around KT; this result indicates that deep convective events existed around KT within a synoptic-scale convectively active envelope (see Figures 1c and 1d). In addition to continuous deep convective events around KT, convective events with a diurnal variability were embedded around KT during 5 6 May; lowest T BB was seen around 1200 UTC (19 LST) on 5 and 6 May. Renggono et al. [2006] have shown that shallowconvective precipitating events and stratiform precipitating events are observed at KT in the early afternoon ( UTC or LST) and in the nighttime ( UTC or LST) on 5 and 6 May, respectively. KT is located at the mountainous region of Sumatra (see Figure 5), and many 6of14

7 Figure 4 7 of 14

8 Figure 5. Topography of Sumatra. Horizontal and vertical dash-dotted lines indicate the longitude and latitude of KT, respectively. studies have shown that local circulation plays a role in diurnal variability of cumulus convection in the mountainous region of Sumatra [e.g., Mori et al., 2004; Sasaki et al., 2004]. During 5 6 May, larger-scale circulation, as seen in the synoptic-scale eastward moving cloud clusters, caused the continuous development of convective events over Sumatra in addition to the convective events with diurnal variability at the mountainous region of Sumatra. Both of local cloud systems around KT and larger-scale cloud clusters existed over Sumatra probably produced the continuous upward motions at 8 14 km observed at KT. For details of W motions in the lower and middle troposphere at KT from 5 to 6 May, see Renggono et al. [2006]. [16] After 7 May, both upward and downward motions are observed at km (Figure 2a). Upward motions of greater than 0.05 m s 1 are almost absent when W is averaged over 3 hours in time and 8 14 km in altitude (Figure 3b), and the ratio of upward motions is almost always less than 70%. The averaged W and the ratio of upward motions at 8 14 km during 7 9 May are 0.01 m s 1 and 57%, respectively. After 7 May, TBB at KT is almost always greater than 250 K (temperature of 250 K locates at 9.0 km) and gradually increases as the synoptic-scale convectively active region moves northeastward of Sumatra (see Figures 1e 1g and 3a). The results of TBB indicate that upward motions disappear as cumulus activity is suppressed over Sumatra. For details of W motions in the lower and middle troposphere at KT from 7 to 9 May, see Seto et al. [2006]. They also have shown that cumulus activity at KT was suppressed after 7 May because of the dryness of the lower troposphere, and the lower tropospheric dry air was transported by westerly wind existed in the west of synoptic-scale convectively active region Downward Motions Above 14 km [17] During 5 6 May, W above 14 km shows a different feature from that below 14 km. After 1800 UTC 5 May, downward motions become prominent from 14 km to the tropopause altitude, while upward motions are dominant at 8 14 km (Figure 2a). This Figure 4. Horizontal distribution of TBB from 0600 UTC 5 May to 0000 UTC 7 May TBB is averaged over the region and is plotted every 6 hours. Horizontal and vertical dashed lines in each plot indicate the longitude and latitude of KT, respectively. 8 of 14

9 the measurement might become erroneous when and where shear instability occurs. A horizontal tilt of isentropes produced by shear instability can cause a contamination of horizontal wind components to the Doppler velocity measured by the vertically pointing beam [Yamamoto et al., 2003]. Using temperature and horizontal wind derived from radiosondes, Richardson number (hereafter Ri) above 14 km are computed to show effects of shear instability on vertical wind measurement are negligible in our case. A spurious component of vertical wind (Dw) by shear instability is given by Dw ¼ DQ u ; Figure 6. Altitude profile of W derived from the EAR. W are averaged from 1800 UTC 5 May to 0000 UTC 7 May. contrast of W continues until around 0000 UTC 7 May, except the prominent updraft even above 14 km during UTC 6 May. This prominent updraft perhaps occurs in the vicinity of convective part of cloud systems, as seen in the small T BB which reaches to 210 K around 1200 UTC 6 May. The cloud-top altitude inferred from T BB of 210 K is 13.9 km. However, the actual cloud top probably locates at higher altitude, because clouds are not regarded as perfect black bodies [Sherwood et al., 2004]. On 7 May, the contrast of W disappears; downward motions are prominent above 10 km. On that day, the center of synoptic-scale convectively active region existed in the northeast (around 0 10 N, E) and northwest (around 0 10 N, E) of Sumatra, and cumulus convection over Sumatra was relatively inactive (see Figure 1e). The downward motions above 10 km on 7 May were probably associated with the subsidence existed around the synoptic-scale convectively active region. [18] Figure 6 shows an altitude profile of W averaged from 1800 UTC 5 May to 0000 UTC 7 May. Below 14 km, the averaged W showed upward motions of m s 1. On the other hand, downward motions are prominent above 14.0 km. W at km ranges from 0.11 to 0.04 m s 1, and the averaged W at km is 0.03 m s 1. When vertical wind is measured from Doppler velocity in the vertically pointing beam, where DQ denotes mean effective off-vertical beampointing angle, u projected component of horizontal wind velocity into the radar vertical wind shear [Muschinski, 1996]. Because DQ is defined to have a positive value, the sign determines the direction of vertical motion (upward or downward). From 1800 UTC 5 May to 0000 UTC 7 May, cases that can cause spurious downward motions are found only at 14.4 and 14.6 km at 1800 UTC on 5 May, when Ri of less than 0.25, westward vertical wind is negative) and easterly wind (u is negative) are observed. [19] As a generation mechanism for the downward motions above 14 km, both adiabatic and diabatic processes are candidates. Johnson et al. [1990] have shown a similar W profile as observed by the EAR from 1800 UTC 5 May to 0000 UTC 7 May. They used radiosonde data to show that upward motions from the middle to upper troposphere and downward motions near the tropopause exist over a several hundred km in stratiform cloud regions of a midlatitude squall line. As a generation mechanism of the downward motions, they have pointed out two factors; the first factor is downward sloping isentropes to the rear of the convective line, and the second is a radiative cooling at the stratiform cloud top. Using a VHF wind profiler at Pohnpei (7 N, 157 E), Balsley et al. [1988] also have shown an existence of upward motions from the middle to upper troposphere and downward motions of m s 1 above 15 km during the stratiform rainfall events in the tropics W Estimation by the ECMWF Operational Analysis [20] In this subsection, the ECMWF analysis data are used to investigate adiabatic processes to produce the upper tropospheric downward motions observed by the EAR. Figure 7 shows time-altitude plots of horizontal wind at 0.18 S, E (closest grid from KT) derived from the ECMWF analysis. Horizontal wind derived from the ECMWF analysis shows a similar tendency observed by radiosondes such as an intensification of 9of14

10 Figure 7. analysis. Time-altitude plots of (a) U and (b) V at 0.18 S, E derived from the ECMWF easterly wind at km around 0000 UTC 7 May, a change from southerly wind to northerly wind at km around 0000 UTC 7 May, and an intensification of northerly wind at km during 5 6 May. This results indicates that the ECMWF analysis reproduces the wind field around KT. [21] W caused by adiabatic processes (hereafter W est ) is expressed by using the first law of thermodynamics and by neglecting a radiative heating rate: W @t ; ð2þ where q denotes potential temperature, t time, x zonal direction, y meridional direction, z vertical direction. W est is estimated by using q, U, and V derived from the ECMWF reanalysis. Distributions of U, V, and q from which W est are computed are presented. Figure 8b shows a horizontal distribution of q at 15.0 km. In the western coastal region of Sumatra, a large gradient of q is found especially at 3 S 3 N. q is less than K in the land region of Sumatra, while q is greater than K in the adjacent region of western Sumatra. [22] Figures 8c and 8d show a vertical structure of U and V, respectively, with q along the thick solid line shown in Figure 8b. The thick solid line is parallel to the direction of q gradient (from the northeast to the southwest) and passes over at 0.18 S, E (closest grid of KT). Around 100 E, a southwestward downward tilt of q is seen above km (Figure 8c). Above km, a northeasterly wind is found (Figures 8c and 8d). [23] Figure 9 shows an altitude profile of W est at 0.18 S, E (closest grid of KT). The downward motions reaches to 0.06 m s 1 at 14 km, and has an average of 0.04 m s 1 at km. The W estimation by the ECMWF analysis indicates that a major part of the observed downward motions above 14 km is explained by the leeward (southwestward) wind and the leeward downward tilt of isentropes. The fact that a downward tilt of isentropes is a cause for producing downward motions is similar to the case of midlatitude squall line shown by Johnson et al. [1990]. Above 13 km, downward motions are evident in the ECMWF analysis, while downward motions are observed above 14 km by the EAR (see Figure 6). The fact that vertical motions by diabatic processes is not taken into consideration may explain the discrepancy of the W motion at km. 4. Discussion [24] Hereafter possible mechanisms that produce the contrast of upper tropospheric q between the land and adjacent region of Sumatra are discussed. Figure 8a shows a distribution of divergence at 14 km. At 3 S 3 N, a contrast of divergence between the land region and the adjacent sea region of western Sumatra is found; relatively large divergence of greater than s 1 is seen only in the land region of Sumatra. Further, a contrast of convergence between the land region and the adjacent sea region of western Sumatra is found especially at 6 8 km altitude; relatively large convergence of s 1 is seen only in the land 10 of 14

11 Figure 8 11 of 14

12 propagation of lee waves, where zonal wind or meridional wind becomes 0 m s 1, exist 3 km and km in the east of 100 E, respectively (see Figures 8c and 8d). Therefore the effect of lee waves on upper tropospheric q is not at least a direct one. [26] Though a major part of the observed downward motions above 14 km from 1800 UTC 5 May to 0000 UTC 7 May is explained by the southwestward (leeward) wind and the leeward downward tilt of isentropes, a radiative cooling rate necessary for producing the observed downward motions is estimated using W rad 1 ðp 0=pÞ k Q R ; c p Figure 9. Altitude profile of W est (thick black curve) at 0.18 S, E averaged from 1800 UTC 5 May to 0000 UTC 7 May The first (thin black curve), second (thick gray curve), and third (thin gray curve) terms of the right side of equation (2) are also plotted. is computed as an averaged gradient of in the time domain at each vertical grid. region of Sumatra (not shown). This fact suggests that cloud systems over the land region contains larger or stronger convective part of cloud system than the adjacent sea region, and that the larger or stronger convective part over the land region contributes to the stronger cooling of the upper troposphere in the land region than the adjacent sea region. [25] Gravity waves are another candidates that affect q in the upper troposphere. Lee waves generated by the mountains at western Sumatra can produce the contrast of upper tropospheric q, if they reach to the upper troposphere. However, the critical levels of the upward where W rad denotes the vertical motion caused by radiative heating, p pressure, c p the specific heat at constant pressure, Q R the radiative heating rate, p 0 = 1000 hpa, and k = If we assume that the whole part of the downward motions is caused by radiative cooling, the radiative cooling rate necessary for producing the downward motion of 0.03 m s 1 averaged over km is computed to be 5.3 K day 1. [27] During 5 6 May, upward motions terminated around 14 km (see Figure 6). This result indicates that the level of 14 km can be regarded as the lower boundary of the TTL, at which most of cumulus convection terminates. The downward motions above 14 km observed by the EAR and the southwestward (leeward) downward tilt of isentropes found in the ECMWF analysis suggest that downward motions caused by tilts of isentropes can be produced in the vicinity of convectively active region, and that a leeward downward tilt of isentropes in the vicinity of convectively active region can suppress an upward transport of air mass into the TTL by producing downward motions in the TTL. [28] As generation mechanisms for producing continuous upward motions at 8 14 km during 5 6 May, latent heat release produced in convection [e.g., Houze, 1989] is one of candidates. T from the middle to upper troposphere (8 14 km) during 5 6 May was warmer than that after 7 May (Figure 2d). This fact indicates the existence of latent heat release during 5 6 May. However, it is noted that the upward motions were found even when relatively large T BB of K, which locates Figure 8. Longitude-latitude plots of (a) divergence at 14.0 km and (b) potential temperature () anomaly at 15.0 km. In Figure 8a, solid contours are plotted with an interval of s 1. Orange (blue) indicates positive (negative) values. The anomaly is the deviation from the average (355.5 K) in the region shown in Figure 8a. Horizontal and vertical dashed lines in Figures 8a and 8b indicate the longitude and latitude of KT, respectively. Vertical sections of (c) U, (d) V, and (e) T anomaly along the thick solid line indicated in Figure 8b. In Figures 8c 8e, is plotted as solid line contours. T anomaly in Figure 8e is the deviation from the average in the region shown in Figure 8a at each altitude range. Vertical dashed lines in Figures 8c 8e indicate the location of KT. Data are averaged in time from 1800 UTC 5 May to 0000 UTC 7 May of 14

13 at km, was observed (see Figures 3a and 3b). In such periods, upward motions produced by latent heat release probably become weak or almost absent. Pandya and Durran [1996] have shown that upward motion phase of gravity waves produced by convection can cause continuous upward motions from the middle to upper troposphere. Deep convective events which continuously observed in the vicinity of KT (see Figures 4c 4h) probably produced the upward motions through the generation of gravity waves, even when deep convective events did not exist at KT. 5. Conclusions [29] In this study, two features of upper tropospheric W were presented using data obtained by the EAR and the ECMWF analysis. First feature is W changes from the middle to upper troposphere. During 5 6 May, 3-hourly averaged W at 8 14 km continuously showed upward motions up to 0.09 m s 1. The averaged W during 5 6 May was 0.05 m s 1. The upward motions were observed in the vicinity of deep convective events which were continuously seen over Sumatra within a synopticscale convectively active envelope. After 7 May, when cumulus activity was suppressed over Sumatra, 3-hourly averaged upward motions of greater than 0.05 m s 1 almost disappeared. The averaged W during 7 9 May was 0.01 m s 1. Second feature is the prominent downward motions from 14 km to the tropopause in the vicinity of enhanced cumulus activity over Sumatra. During 5 6 May, downward motions up to 0.11 m s 1 were observed above 14 km. W estimation by the ECMWF analysis has revealed that a major part of the observed downward motions are explained by the leeward (southwestward) wind and leeward downward tilt of isentropes, both of which existed over western Sumatra. As possible mechanisms that produce the southwestward downward tilt of isentropes, a larger fraction of convective part of cloud systems over the land region than the adjacent sea region of western Sumatra and modulations of isentropes by gravity waves were discussed. [30] The downward motions above 14 km observed by the EAR and the leeward downward tilt of isentropes found in the ECMWF analysis suggest that downward motions caused by tilts of isentropes can be produced in the vicinity of convectively active region associated with the cloud cluster, and that a leeward downward tilt of isentropes in the vicinity of cloud cluster can suppress an upward transport of air mass into the TTL by producing downward motions in the TTL. [31] Diabatic processes that can produce upward motions from the middle to upper troposphere and downward motions in the TTL are not evaluated quantitatively in this study. Because our present data set is limited, future observations are necessary to understand interactions between vertical air motions and cloud microphysics. [32] Acknowledgments. The authors thank Fumitaka Tsujino of Research Institute for Sustainable Humanosphere (RISH) of Kyoto University for helping data analysis. GOES 9 IR1 data distributed by Kochi University, Japan were used in this study. EAR and radiosonde data are provided from the joint project between Japan and Indonesia, called Coupling Processes in the Equatorial Atmosphere (CPEA). The former (Japan) side is supported by Grant-in-Aid for Scientific Research on Priority Area-764 funded by the Ministry of Education, Culture, Sports, Science, and Technology (MEXT) of Japan. EAR belongs to RISH, Kyoto University and is operated by RISH and National Institute of Aeronautics and Space (LAPAN), Indonesia. The present study was partially supported by Grant-in-Aid for Scientific Research of the Ministry of Education, Culture, Sports, Science, and Technology (MEXT) of Japan. References Alexander, S., T. Tsuda, J. Furumoto, T. Shimomai, T. Kozu, and M. Kawashima (2006), A statistical overview of convection during the first CPEA campaign, J. Meteorol. Soc. Jpn., 84A, Balsley, B. B., W. L. Ecklund, D. A. Carter, A. C. Riddle, and K. S. Gage (1988), Average vertical motions in the tropical atmosphere observed by a radar wind profiler on Pohnpei (7 N latitude, 157 E longitude), J. Atmos. Sci., 45, Cifelli, R., and S. A. Rutledge (1994), Vertical motion structure in maritime continent mesoscale convective systems: Results from a 50-MHz profiler, J. Atmos. Sci., 51, Dhaka, S. K., R. K. Choudhary, S. Malik, Y. Shibagaki, M. D. Yamanaka, and S. Fukao (2002), Observable signatures of a convectively generated wave field over the tropics using Indian MST radar at Gadanki (13.5 N, 79.2 E), Geophys. Res. Lett., 29(18), 1872, doi: /2002gl Folkins, I., M. Loewenstein, J. Podolske, S. J. Oltmans, and M. Proffitt (1999), A barrier to vertical mixing at 14 km in the tropics: Evidence from ozonsondes and aircraft measurements, J. Geophys. Res., 104, 22,095 22,102. Fujiwara, M., and M. Takahashi (2001), Role of the equatorial Kelvin wave in stratosphere-troposphere exchange in a general circulation model, J. Geophys. Res., 106, 22,763 22,780. Fujiwara, M., M. K. Yamamoto, H. Hashiguchi, T. Horinouchi, and S. Fukao (2003), Turbulence at the tropopause due to breaking Kelvin waves observed by the Equatorial Atmosphere Radar, Geophys. Res. Lett., 30(4), 1171, doi: / 2002GL Fukao, S. (2006), Coupling processes in the equatorial atmosphere (CPEA): A project overview, J. Meteorol. Soc. Jpn., 84A, Fukao, S., H. Hashiguchi, M. Yamamoto, T. Tsuda, T. Nakamura, M. K. Yamamoto, T. Sato, M. Hagio, and Y. Yabugaki (2003), Equatorial Atmosphere Radar (EAR): System 13 of 14

14 description and first results, Radio Sci., 38(3), 1053, doi: /2002rs Gage, K. S. (1990), Radar observations of the free atmosphere: Structure and dynamics, in Radar in Meteorology, edited by D. Atlas, pp., Am. Meteorol. Soc., Boston, Mass. Gage, K. S., B. B. Balsley, W. L. Ecklund, D. A. Carter, and J. R. McAfee (1991), Wind profiler-related research in the tropical Pacific, J. Geophys. Res., 96, Holton, J. R., P. H. Haynes, M. E. McIntyre, A. R. Douglass, R. B. Rood, and L. Pfister (1995), Stratosphere-troposphere exchange, Rev. Geophys, 33, Houze, R. A. (1989), Observed structure of mesoscale convective systems and implications for large-scale heating, Q. J. R. Meteorol. Soc., 115, Jagannadha Rao, V. V. M., D. N. Rao, M. V. Ratnam, K. Mohan, and S. V. B. Rao (2003), Mean vertical velocities measured by Indian MST radar and comparison with indirectly computed values, J. Appl. Meteorol., 42, Johnson, R. H., W. A. Gallus, and M. D. Vescio (1990), Neartropopause vertical motion within the trailing stratiform region of a midlatitude squall line, J. Atmos. Sci., 47, Lane, T. P., M. J. Reeder, and T. L. Clark (2001), Numerical modeling of gravity wave generation by deep tropical convection, J. Atmos. Sci., 58, Mori, S., J.-I. Hamada, Y. I. Tauhid, M. D. Yamanaka, N. Okamoto, F. Murata, N. Sakurai, H. Hashiguchi, and T. Sribimawati (2004), Diurnal land-sea rainfall peak migration over Sumatera Island, Indonesian maritime continent observed by TRMM satellite and intensive rawinsonde soundings, Mon. Weather Rev., 132, Muschinski, A. (1996), Possible effect of Kelvin-Helmholtz instability on VHF radar observations of the mean vertical wind, J. Appl. Meteorol., 35, Pandya, R. E., and D. R. Durran (1996), The influence of convectively generated thermal forcing on the mesoscale circulation around squall lines, J. Atmos. Sci., 53, Piani, C., and D. R. Durran (2001), A numerical study of stratospheric gravity waves triggered by squall lines observed during the TOGA COARE and COPT-81 experiments, J. Atmos. Sci., 58, Ramage, C. S. (1968), Role of a tropical maritime continent in the atmospheric circulation, Mon. Weather Rev., 96, Renggono, F., M. K. Yamamoto, H. Hashiguchi, S. Fukao, T. Shimomai, M. Kawashima, and M. Kudsy (2006), Raindrop size distribution observed with the Equatorial Atmosphere Radar (EAR) during the Coupling Processes in the Equatorial Atmosphere (CPEA-I) observation campaign, Radio Sci., 41, RS5002, doi: /2005rs Röttger, J. (1980), Structure and dynamics of the stratosphere and mesosphere revealed by the VHF radar investigations, Pure Appl. Geophys., 118, Sasaki, T., P. Wu, S. Mori, J.-I. Hamada, Y. I. Tauhid, M. D. Yamanaka, T. Sribimawati, T. Yoshikane, and F. Kimura (2004), Vertical moisture transport above the mixed layer around the mountains in western Sumatra, Geophys. Res. Lett., 31, L08106, doi: /2004gl Seto, T. H., M. K. Yamamoto, H. Hashiguchi, and S. Fukao (2004), Convective activities associated with intraseasonal variation over Sumatera, Indonesia observed with the Equatorial Atmosphere Radar, Ann. Geophys., 22, Seto, T. H., M. K. Yamamoto, H. Hashiguchi, S. Fukao, M. Abo, T. Kozu, and M. Kudsy (2006), Observational study on westerly wind burst over Sumatra, Indonesia by the Equatorial Atmosphere Radar A case study during the first CPEA campaign, J. Meteorol. Soc. Jpn., 84A, Sherwood, S. C., P. Minnis, and M. McGill (2004), Deep convective cloud-top heights and their thermodynamic control during CRYSTAL-FACE, J. Geophys. Res., 109, D20119, doi: /2004jd Shibagaki, Y., T. Shimomai, T. Kozu, S. Mori, Y. Fujiyoshi, H. Hashiguchi, M. K. Yamamoto, S. Fukao, and M. D. Yamanaka (2006a), Multiscale aspects of convective systems associated with an intraseasonal oscillation over the Indonesian maritime continent, Mon. Weather Rev., 134, Shibagaki, Y., T. Kozu, T. Shimomai, S. Mori, F. Murata, Y. Fujiyoshi, H. Hashiguchi, and S. Fukao (2006b), Evolution of a super cloud cluster and the associated wind fields observed over the Indonesian maritime continent during the first CPEA campaign, J. Meteorol. Soc. Jpn., 84A, Wheeler, M., G. N. Kiladis, and P. J. Webster (2000), Largescale dynamical fields associated with convectively coupled equatorial waves, J. Atmos. Sci, 57, Yamamoto, M., T. Sato, P. T. May, T. Tsuda, S. Fukao, and S. Kato (1988), Estimation error of spectral parameters of mesosphere-stratosphere-troposphere radars obtained by least squares fitting method and its lower bound, Radio Sci., 23, Yamamoto, M. K., M. Fujiwara, T. Horinouchi, H. Hashiguchi, and S. Fukao (2003), Kelvin-Helmholtz instability around the tropical tropopause observed with the Equatorial Atmosphere Radar, Geophys. Res. Lett., 30(9), 1476, doi: / 2002GL Zhang, C. (2005), Madden-Julian Oscillation, Rev. Geophys., 43, RG2003, doi: /2004rg S. Fukao, T. Horinouchi, and M. K. Yamamoto, Research Institute for Sustainable Humanosphere, Kyoto University, Uji, Kyoto , Japan. (m-yamamo@rish.kyoto-u.ac.jp) N. Nishi, Division of Earth and Planetary Sciences, Graduate School of Science, Kyoto University, Kitashirakawa Oiwake, Sakyo, Kyoto , Japan. M. Niwano, Frontier Research Center for Global Change, Japan Agency for Marine-Earth Science and Technology, Yokohama, Kanagawa , Japan. 14 of 14