Tropical tropopause parameters derived from GPS radio occultation measurements with CHAMP

JOURNAL OF GEOPHYSICAL RESEARCH, VOL. 109,, doi:10.1029/2004jd004566, 2004 Tropical tropopause parameters derived from GPS radio occultation measurements with CHAMP T. Schmidt, J. Wickert, G. Beyerle, and C. Reigber Department 1: Geodesy and Remote Sensing, GeoForschungsZentrum Potsdam, Potsdam, Germany Received 23 January 2004; revised 5 April 2004; accepted 19 April 2004; published 9 July 2004. [1] The temperature structure in the tropical upper troposphere and lower stratosphere (UTLS) region is discussed based on Global Positioning System (GPS) radio occultation (RO) data from the German CHAMP (CHAllenging Minisatellite Payload) satellite mission. Several climatologies for tropopause parameters based on radiosonde data and model analyses have been published in recent years. Both data sources suffer either from low global coverage or poor vertical resolution. The GPS RO technique, on the other hand, is characterized by global coverage, high vertical resolution, all-weather viewing, and long-term stability. CHAMP RO data are available since February 2001. Since May 2001, up to 200 high- resolution temperature profiles per day are available. The temperature bias between CHAMP temperature profiles and radiosonde data as well as ECMWF analyses is less than 0.5 K between 300 30 hpa. On the basis of the May 2001 to November 2003 data set of CHAMP RO data the structure and temporal and spatial variability of the tropical tropopause based on several tropopause definitions (thermal and cold-point tropopause) are discussed. This includes an overview of the global tropopause characteristics, the discussion of the annual cycle and the latitudinallongitudinal structure of the tropical tropopause. In the CHAMP RO temperature data, clear evidence of the stratospheric quasi-biennial oscillation (QBO) was found. The goal of this study is to show the potential of GPS RO for global monitoring of the temperature demonstrated exemplarily for the tropical UTLS region and based on the first 31 months (as of November 2003) of CHAMP RO data. INDEX TERMS: 3309 Meteorology and Atmospheric Dynamics: Climatology (1620); 3300 Meteorology and Atmospheric Dynamics; 3334 Meteorology and Atmospheric Dynamics: Middle atmosphere dynamics (0341, 0342); 3360 Meteorology and Atmospheric Dynamics: Remote sensing; KEYWORDS: CHAMP GPS radio occultation, tropical tropopause, QBO Citation: Schmidt, T., J. Wickert, G. Beyerle, and C. Reigber (2004), Tropical tropopause parameters derived from GPS radio occultation measurements with CHAMP, J. Geophys. Res., 109,, doi:10.1029/2004jd004566. 1. Introduction [2] The tropopause separates the troposphere and stratosphere that have fundamental different characteristics with respect to chemical composition and static stability. In the climate change discussion tropopause parameters have received more attention in recent years since they describe climate variability and change, and are crucial for the understanding of the stratosphere-troposphere exchange [Holton et al., 1995; Sausen and Santer, 2003]. The dynamical, radiative, and chemical connections between the stratosphere and troposphere are a topic of great importance that must be understood for the prediction of global climate change. The exchange of mass, water, and trace gases between the troposphere and stratosphere takes place across the tropopause. In this context the tropical tropopause plays an exceptional role [Holton et al., 1995]. Copyright 2004 by the American Geophysical Union. 0148-0227/04/2004JD004566 [3] The global mean tropopause altitude shows an increase in reanalyses and radiosonde observations during the last 25 years [Randel et al., 2000; Seidel et al., 2001; Santer et al., 2003]. Studies with an atmospheric General Circulation Model (GCM) found the tropopause altitude to be sensitive to surface temperatures, less sensitive to ozone distribution and changes in equator-to-pole temperature gradient, and nearly insensitive to changes in the Earth s rotation [Thuburn and Craig, 1997, 2000]. The tropopause altitude is an excellent parameter for the detection of climate variability. The global mean tropopause altitude has the property to act as a natural filter removing much of the El Niño/Southern Oscillation (ENSO) variability that makes the interpretation of tropospheric and surface temperature changes difficult [Sausen and Santer, 2003]. [4] Thus the continuous identification and monitoring of the tropopause on a global scale is an important goal in atmospheric and climate research. According to the World Meteorological Organization (WMO) the thermal lapse-rate tropopause (LRT) is defined as the lowest level at which the temperature lapse-rate is less than 2 K/km and the lapse-rate 1of13

average between this level and the next 2 km does not exceed 2 K/km [WMO, 1957]. Alternatively, in the tropics the minimum temperature in a vertical temperature profile (cold-point) and the potential temperature is used to locate the tropical tropopause. The tropical tropopause corresponds to an isentropic surface with a potential temperature of on average about 380 K. In the extra-tropics the LRT corresponds to a surface of constant potential vorticity [Holton et al., 1995]. [5] In the past several studies have been performed leading to tropopause climatologies from different data sources [Hoinka, 1998; Highwood and Hoskins, 1998; Randel et al., 2000; Seidel et al., 2001; Zhou et al., 2001a, 2001b; Gettelman and de Forster, 2002; Santer et al., 2003]. The most important data source for the determination of tropopause parameters are radiosonde data, whereas model analyses, as, e.g., ECMWF or NCEP, suffer from lower vertical resolution and attendant biases [Randel et al., 2000]. Despite good vertical resolution of radiosonde measurements, a global coverage is impossible. On the other hand, investigations of long-term radiosonde data have to take into account inhomogeneous time series due to sensor changes. This lack can be removed by the GPS radio occultation (RO) technique with its (1) high vertical resolution (less than 1 km), (2) long-term stability, (3) all-weather capability, and (4) global coverage [Melbourne et al., 1994; Kursinski et al., 1997; Anthes et al., 2000]. [6] The availability of GPS radio signals has introduced a new promising remote sensing technique for the Earth s atmosphere. GPS-based RO exploit signals received onboard a Low Earth Orbiting (LEO) satellite for atmospheric limb sounding. The GPS signals are influenced by the atmospheric refractivity field resulting in a time delay and bending of the signal. The atmospheric excess phase is the basic observable that is measured with millimetric accuracy [Wickert et al., 2001b]. This is the basis for high vertical resolution and precise refractivity and temperature profiles. The GPS RO technique requires no calibration, is not affected by clouds, aerosols or precipitation, and the occultations are almost uniformly distributed over the globe. These properties makes the RO technique suitable for long-term monitoring of atmospheric temperature change [Steiner et al., 2001]. There are about 2.5 years of CHAMP RO data available (as of November 2003). Thus a long-term analysis with respect to trends is not yet possible. [7] The proof-of-concept GPS RO experiment GPS/MET (GPS/Meteorology) performed between 1995 and 1997 has demonstrated for the first time the potential of GPS-based limb sounding from LEO satellites for deriving atmospheric temperature and tropospheric water vapor profiles [Ware et al., 1996; Kursinski et al., 1997; Rocken et al., 1997]. Further missions with radio occultation experiments followed (Øersted, SAC-C, and CHAMP), with the German geoscience satellite CHAMP measuring continuously in an operational manner since mid-2001 [Wickert et al., 2004; Hajj et al., 2004]. [8] First investigations of the thermal structure and variability in the tropical UTLS region based on GPS RO measurements were performed by Nishida et al. [2000] and Randel et al. [2003] on the basis of the GPS/MET data from Figure 1. GPS radio occultation principle [after Wickert et al., 2004]. 1995 1997 with focus to the so-called prime-times (June July 1995, December February 1996/1997) [Rocken et al., 1997]. First applications of CHAMP RO data to that region with a 1-year data set were shown by Schmidt et al. [2004]. [9] In this paper a thermal tropopause statistics based on GPS RO measurements aboard CHAMP for the May 2001 to November 2003 period is presented. This study is an extension of previous papers (with GPS/MET data [Nishida et al., 2000; Randel et al., 2003], and with a much wider database [Schmidt et al., 2004]). After a brief description of the CHAMP RO experiment, the database and data quality of CHAMP RO measurements are discussed (section 2). In section 3 a short overview of the zonal behavior of the global tropopause is given. Statistics of the tropical tropopause are presented in section 4, whereas section 5 contains a discussion of the lower stratospheric annual temperature cycle and temperature anomalies over the equator region. 2. CHAMP GPS Radio Occultation Experiment [10] CHAMP was launched on July 15, 2000 from Plesetsk (Russia) into an almost circular (eccentricity = 0.004) and near polar (inclination = 87.2) orbit with an initial altitude of 454 km. The GPS radio occultation experiment was successfully started on Feb. 11, 2001 [Wickert et al., 2001a]. Since then, up to 200 high-resolution temperature profiles per day were observed. 2.1. Data Processing [11] A detailed description for deriving vertical atmospheric profiles from CHAMP occultation measurements is presented in Wickert et al. [2001a, 2001b, 2004]. A Black- Jack GPS receiver (provided by NASA/JPL) onboard CHAMP records phase and amplitude variations with high temporal resolution (50 Hz) during an occultation event (Figure 1). By using high precision orbit information from CHAMP and occulting GPS satellites the atmospheric excess phase can be extracted which is related to a bending angle profile. A double differencing method is applied to remove clock errors. For this purpose a GPS ground station network is operated jointly by JPL and GFZ. Ionospheric effects are corrected by a linear combination of the bending angles derived from the two GPS frequencies. Assuming spherical symmetry refractivity profiles are calculated by inverting the bending angle profiles using an Abel inversion. The atmospheric refractivity N = 10 6 (n 1) with refractive index n, is the basic meteorological observable derived with the GPS RO 2of13

Figure 2. (a) Meridional and (b) zonal distribution of CHAMP temperature profiles for May 2001 to November 2003. technique. To convert the refractivity profiles into pressure and temperature profiles, the assumption of dry air has to be made because of the ambiguity between the dry and wet part in the resulting refractivity profile. The hydrostatic equation is used to calculate pressure and temperature profiles by downward integration of the refractivity profile. For the initialization of the hydrostatic equation (at 43 km), ECMWF data are used. By applying additional temperature information from operational analyses, water vapor can be calculated in the troposphere. The vertical resolution of the profiles ranges from 0.1 1 km. The resolution along the ray path is around a few hundred km. A more detailed description of the retrieval algorithms used for the CHAMP RO experiment can be found in Wickert et al. [2004]. 2.2. Database [12] Figure 2 shows the number of CHAMP temperature profiles per 5 in latitude or longitude, respectively. The zonal distribution of CHAMP radio occultations shows a nearly symmetric behavior between the northern and southern hemispheres with local maxima in the mid-latitudes and around 20 S and 20 N (Figure 2a). For the tropical region, which the following investigations are focussed on, about 2000 4000 occultations per 5 bin are present (as of November 2003). Because of the CHAMP orbit geometry, the longitudinal occultation distribution (Figure 2b) is nearly constant (1700 2000) in the 5 bins from 180 W 180 E. [13] For the complete time period from May 2001 to November 2003 a total number of 124,491 high resolution temperature profiles are available. The latitude interval between 30 S 30 N is covered by 37,619 profiles whereas between 10 S 10 N 10,167 temperature profiles are the basis for investigations in the deep tropics. For the equator region (4 S 4 N) a data set of 3934 profiles was used (section 5). 2.3. Validation With ECMWF and Radiosondes [14] Figure 3 shows meridional variations of the zonal mean differences between CHAMP dry temperatures and Figure 3. Comparison of CHAMP temperature data with ECMWF analyses for the period May 2001 to November 2003. The contours represent the bias in K (contour interval is ±0.5 K). ECMWF analyses in the upper troposphere and lower stratosphere for the period May 2001 to November 2003. The analysis data were obtained by interpolating 6-hourly ECMWF data sets to the time and location of the occultation for the 60 ECMWF pressure levels. The zonal mean statistics was calculated for 5 steps. The mean bias in the upper troposphere and lower stratosphere (300 30 hpa) is less than 1 K with root-mean-square (RMS) deviations between 1 2 K (not shown here). The cold bias in the middle and lower troposphere (below 300 hpa) is caused by the difference between temperature and dry temperature for nonzero humidity [Kursinski et al., 1997; Marquardt et al., 2003]. The bias above 30 hpa is related to the initialization of the upper boundary for the downward integration of the refractivity profile (see section 2.1). [15] A validation of CHAMP temperatures with ECMWF is not completely independent because of the use of ECMWF data for the initialization of the integration of the hydrostatic equation. We therefore validate the CHAMP temperature data with radiosonde observations as well. Figure 4 shows a comparison of CHAMP RO Figure 4. Comparison of CHAMP temperature data with nearby radiosondes (distance Dd 300 km and time delay Dt 3 hours) for the period May 2001 to October 2003. (a) The dashed lines denote the standard deviation, whereas the shaded area shows the ±1 K interval. (b) Number of corresponding data points. 3of13

Figure 5. Monthly biases of the annual cycle of CHAMP and radiosonde LRT (a) pressure and (b) temperature for the period May 2001 to October 2003 and 10 S 10 N. temperatures with nearby radiosondes. More than 10,000 RO meet the condition that the radiosonde was launched within a distance of less than 300 km and with a time delay less than 3 hours from the CHAMP measurement. The comparisons show a temperature bias less than 0.5 K (RMS deviation 1 2 K) between 250 20 hpa (Figure 4a) confirming CHAMP s excellent data quality especially in the UTLS region. [16] For further validation, tropopause pressure and temperature as determined from radiosonde soundings in the deep tropics were included. Therefore first the annual cycle of LRT pressure and temperature from all available radiosonde stations between 10 S 10 N and May 2001 to October 2003 was calculated. In a second step this annual cycle was subtracted from the annual cycle derived with the CHAMP data (see section 4.1). Figure 5 presents the final result. The bias between both data sets shows a good agreement with overall deviations of (0.6 ± 1.4) hpa for pressure and (0.6 ± 0.7) K for temperature respectively. 3. Global Tropopause [17] For the tropopause several definitions were introduced. As discussed in Highwood and Hoskins [1998], the LRT [WMO, 1957] is an arbitrary definition for operational use that has limited physical relevance especially in the tropics. They argued that there is little direct connection between convective processes and the LRT definition. The temperature minimum or cold-point tropopause (CPT) becomes more important for describing tropical tropopause characteristics because these correlate better with convective processes that play an important role in the stratosphere-troposphere exchange. The 100-hPa level that has sometimes also been used for the identification of the tropical tropopause can only serve as a proxy. It will be shown that the 100-hPa altitude has nearly no temporal variability in contrast to the LRT and CPT. [18] Although several tropopause definitions exist, the tropical tropopause has not a sharp boundary. As described in Highwood and Hoskins [1998], Holton and Gettelman [2001], or Gettelman and de Forster [2002], a tropical transition layer (TTL) was introduced in which the interaction of convection, the stratosphere wave driven circulation and horizontal transport processes determine the stratosphere-troposphere exchange. Gettelman and de Forster [2002] suggest that the TTL extends from the level of the lapse-rate minimum at 10 12 km to the CPT. In spite of the limitations of the different tropopause definitions the LRT, CPT and 100-hPa level provide a consistent framework for the current analysis and relate the results to previous studies. [19] Before starting the detailed discussion of tropical tropopause features, a short overview of the global zonal-mean tropopause is given on the basis of the CHAMP data set. Averaged tropopause statistics were computed from individual high-resolution CHAMP temperature profiles that are the basis for all further studies. Figure 6 shows the CHAMP LRT altitude and temperature for the deep tropics for the complete CHAMP RO data set. Starting with the individual measurements zonal means were determined for 5 latitude bins. Figure 7 presents climatological zonal mean values of tropopause altitude, pressure, temperature, and potential temperature. The data points represent the center of a 5 latitude interval. [20] The tropopause has a strong meridional structure. In the tropics (30 S 30 N) the tropopause altitude and the associated pressure are nearly constant. In the deep tropics (10 S 10 N) the LRT altitude reaches values of about 16.5 km (100.3 99.4 hpa), at the margin of the tropics the mean altitude decreases to about 15.8 km (116 117 hpa). The strongest gradients in tropopause altitude and pressure occur between 30 60 on both hemispheres with mean altitudes decreasing to 8.5 km (306.6 hpa) at northern polar latitudes, whereas in the southern polar region the tropopause altitude is nearly constant at about 10 km (260 240 hpa) between 60 S 90 S (Figures 7a and 7b). [21] The monthly mean zonal temperature pattern represents the zonal tropopause altitude and pressure features Figure 6. CHAMP tropical LRT (a) altitudes and (b) temperatures for the period May 2001 to November 2003. The years denote the beginning of the year. 4of13

Figure 7. Zonal means (5 ) of CHAMP LRT parameters for the period May 2001 to November 2003. (Figure 7c). The lowest mean LRT temperature are found in the deep tropics reaching 81.1 C (2.5 N 7.5 N). From the equator region to the north pole the mean temperature is increasing nearly continuously to 54.8 C, whereas in the southern hemisphere the temperature has a local maximum at 57.8 C (50 S 55 S). From there the monthly mean temperature decreases to 67.4 C in the south pole region. [22] The potential temperature (Figure 7d) defined as Q = T(1000hPa/p) k (T, p: air temperature and pressure, respectively; k = 0.286) is highest in the tropics reaching zonal mean values exceeding 376 K and decreases to the poles with lower values in the northern polar region (306 K) compared to southern latitudes (316 K). [23] The asymmetric behavior between the northern and southern hemisphere (from 45 to both poles) in all of the tropopause parameter features are caused by the different distribution of the land-see masses on both hemispheres resulting in different mean circulation characteristics. [24] This general climatology mean picture of the tropopause (Figure 7) is in good agreement with global tropopause parameters published by Hoinka [1998] and Santer et al. [2003]. 4. Tropical Tropopause [25] The focus here is to study the time-dependent and spatial structure of the tropical tropopause based on the CHAMP RO data set. In addition to the LRT parameters shown in section 3 the CPT characteristics and the 100-hPa level features are also included here. 4.1. Time-Averaged Structure [26] On the basis of the individual CHAMP RO measurements, monthly means of the tropopause parameters were calculated. Figures 8a 8d show the individual monthly means of tropopause parameters for the equatorial zone (10 S 10 N). Since the CHAMP GPS receiver software update in March 2002 about 300 400 temperature profiles per month are available in the equatorial region (Figure 8e). These monthly means are the basis for the annual-mean climatological tropopause parameters and allows the discussion of the annual cycle of the tropical tropopause (Figure 9). The tropopause is highest and coldest during the northern hemisphere winter months, and lowest and warmest during the northern summer. The LRT altitude (Figure 9a) reaches its maximum at 17.0 km in December and January and the minimum in August and September (16.2 km). The CPT altitude is on average 400 m higher than the LRT with values varying between 300 m in July and 500 m in September. The 100-hPa level is 16.6 km throughout the time period. In the northern hemisphere winter the 100-hPa level is in the troposphere, and during the southern hemisphere winter months the 100-hPa level is located above the LRT, but not above the CPT. The latter is different to a radiosonde climatology (1961 1990) from Seidel et al. [2001], who have found the 100-hPa level altitude above the CPT and thus into the stratosphere during the southern hemisphere winter. Figure 9b shows the annual cycle of LRT and CPT pressure supporting the discussion for LRT and CPT altitude: minimum of LRT pressure in December January (93.7 92.8 hpa) and maximum in August September (107 hpa). The CPT pressure is about 5 10 hpa lower than the LRT pressure (Figure 9b). [27] Figure 9c shows the annual cycle of tropopause temperature from CHAMP RO data. In the equatorial zone the tropopause temperature is lowest in the northern hemisphere winter months and highest during northern summer, whereas the CPT is colder than LRT, and the LRT is colder than the 100-hPa level temperature. CPT monthly mean temperatures reaching 84.4 C in January and increasing to 79.3 C in August. LRT values are about 1 K higher than CPT temperatures. 5of13

Figure 8. Monthly means of CHAMP tropical tropopause parameters for the period May 2001 to November 2003 (LRT, lapse-rate tropopause; CPT, cold-point tropopause; 100, 100-hPa-level). [28] Typical values for monthly mean potential temperatures in the equatorial zone found in the CHAMP RO measurements are between 368.8 K (August) and 375.1 K (December) for LRT and 374.5 K (August) and 381.2 K (January) for the CPT with lower values during northern hemispheric summer. As depicted in Figure 9d, the 100-hPa level potential temperature is not in phase with LRT and CPT potential temperature. [29] The explanation of the annual cycle in tropical tropopause altitude (pressure) and temperature which is linked to the discussion of the annual temperature cycle in the tropical lower stratosphere (section 5) has changed during the last decades. The classical concept [e.g., Reed and Vlcek, 1969] explains the annual tropical tropopause cycle due to a stronger Hadley circulation during the northern hemispheric winter compared to the northern hemispheric summer months. This is caused by a cooling of the landmasses, and therefore the temperature gradient between the subtropical and tropical region is enhanced. This effect is stronger in the northern hemispheric winter because of the different distribution of the land-see masses between both hemispheres. Finally, the upwelling in the rising Hadley branch leads to adiabatic cooling in the tropopause region and lower stratosphere (review in Reid and Gage [1996]). In the mid-1990s the concept of extratropical wave activity driving the tropical lower stratosphere was introduced [Yulaeva et al., 1994]. This approaches the large-scale stratospheric circulation in the center of interest which is driven by the radiative heating cycle and the dissipation of Rossby and gravity waves [Yulaeva et al., 1994; Holton et al., 1995; Reid and Gage, 1996]. Because of the generally strongest wave activities during the wintertime, especially during northern hemispheric winter months, the annual cycle in tropo- 6of13

Figure 9. Annual cycle of CHAMP tropical tropopause parameters based on the period May 2001 to November 2003 (LRT, lapse-rate tropopause; CPT, cold-point tropopause; 100, 100-hPa level). pause temperature and altitude as well as the annual temperature cycle in the lower stratosphere (section 5) can be explained. [30] The longitude-dependent behavior of the tropical tropopause is shown in Figure 10. Here each data point represents an area-mean of ±10 longitude for the complete time interval of available CHAMP RO data. On average each bin contains 550 occultation measurements. The longitudinal variations of the tropopause altitude are relatively small. The minimum LRT altitude is found between 70 E 90 E with 16.5 km (102.9 hpa) and the maximum between 150 W 170 W with a height of 16.8 km (97.2 hpa). The CPT shows an even smaller variation. The CPT altitude is nearly constant at 17.0 km (91 94 hpa). The longitude-dependent temperature and potential temperature distribution is more structured. One can clearly recognize the tropopause temperature minimum between 110E 150W reaching values less than 83 for both the LRT and CPT. The western part of this region is the so-called warm pool region of the western Pacific with deep convection as a main feature leading to low tropopause temperatures. 4.2. Tropopause Spatial Structure [31] The spatial, i.e., latitudinal-longitudinal structure of the tropical tropopause derived from CHAMP RO measurements for the northern hemisphere winter and summer months is shown in Figures 11 13. The basis for the contour plots are mean tropopause parameters representing an area of ±5 in latitude and longitude. The averages for the winter and summer plots are based on 2- or 3-year means, respectively. The 6 latitudes (25 S, 15 S, 5 S, 5 N, 15 N, and 25 N) and 35 longitudes (175 W, 165 W, 155 W,..., 155 E, 165 E, and 175 E) are the centers of the bins. Each of these areas contain on average about 40 CHAMP RO measurements. Thus 7of13

Figure 10. Area means of CHAMP tropical tropopause parameters for the period May 2001 to November 2003 (LRT, lapse-rate tropopause; CPT, cold-point tropopause; 100, 100-hPa level). already this single satellite constellation leads to a global coverage of the tropical region for seasonal plots of climate parameters where no interpolation is necessary. With increasing numbers of RO experiments in the near future the spatial information density (not only in the UTLS region) of atmospheric climate parameters will increase rapidly. [32] The results shown in Figures 11 13 are consistent with climatologies based on radiosonde measurements and meteorological analyses [Highwood and Hoskins, 1998; Hoinka, 1998; Randel et al., 2000]. The highest LRT altitudes during December February reaching values of >17.0 km are located in the tropical western Pacific region and over the northern part of South America. A smaller maximum altitude (16.8 km) is located over Central Africa and the western Indian Ocean (Figure 11a). The associated pressure patterns are shown in Figure 11b: in the areas of highest LRT the pressure drops to values of less than 90 hpa. The coldest LRT temperatures (Figure 11c) are less than 82 C correlated with the maximum LRT altitudes. In the western Pacific region, tropopause temperatures of less than 84 C are reached. The potential temperature shows little spatial variation with values up to 375 K in December February (Figure 11d). [33] The CPT pressure pattern is different compared to the LRT (Figure 13a). Tropical CPT pressure with values between 80 88 hpa are not correlated with CPT temperatures as found for the LRT. The mean winter CPT altitude is low over the western Pacific where the CPT temperature is also low (Figure 13c). This is in agreement with Gettelman and de Forster [2002]. They noted that the cold point is even lower over some convectively active regions than over convectively nonactive regions. The CPT temperature for the northern hemisphere winter months, however, shows a picture similar to the LRT 8of13

Figure 11. LRT (a) altitude, (b) pressure, (c) temperature, and (d) potential temperature for the northern hemisphere winter months (December February). Contour intervals are (a) 0.2 km, values above 16.8 km with solid lines; (b) 5 hpa, values below 95 hpa with solid lines; (c) 2 K, values below 82 C with solid lines; (d) 5 K. with respect to the location of the minimum temperatures, but with about 1 2 K lower temperatures reaching 86 C between 165 E 165 W in the western Pacific (Figure 13c). As shown by Zhang [1993] and Gettelman et al. [2002], there is a spatial-temporal relationship between coldest tropopause temperatures and deep convection, whereas the relationship is stronger during northern hemisphere winter than northern hemisphere summer months. [34] During the northern hemisphere summer the picture for the LRT is different compared to the winter months (Figure 12): the area of highest tropopause (>17 km) is found over the south-asian monsoon region (Figure 12a). The temperature in that region is less than 80 C (Figure 12c) as also in the western Pacific area but here with LRT altitudes reaching only 16.2 km. An averaged temperature of less than 82 C were not observed at any location in June August. The potential temperature is decreasing to less than 370 K (western Pacific and Indian ocean) compared with the December February time period (Figure 12d). [35] The CPT pressure distribution during northern summer is comparable to the winter months with little variation, but with about 10 20 hpa higher pressure values (Figure 13b). As in Figure 12b, also the CPT is high over the south-asian monsoon region. The CPT June August temperature shows values below 80 C in the western Pacific region and over Southeast Asia, that are in agreement with the LRT summer months (Figure 13d). 5. Annual Temperature Cycle and the QBO [36] A further aspect of this paper is to discuss the temperature variability in the UTLS over the equator region based on CHAMP RO measurements. Figure 14 shows all individual CHAMP temperature measurements (3934) at 18 km altitude over the equator (4 S 4 N) from May 2001 to November 2003. This is the well-known temperature cycle just above the tropical tropopause similar to the annual tropopause temperature cycle (Figure 9c) with minima during northern hemisphere winter [Reed and Vlcek, 1969; Yulaeva et al., 1994; Reid and Gage, 1996]. The 9of13

Figure 12. Same as Figure 11, but for the northern hemisphere summer months (June August). 10 of 13

Figure 13. CPT (a and b) pressure and (c and d) temperature for the northern hemisphere winter (December February) and summer (June August) months. Contour intervals are 2 hpa (Figures 13a and 13b) and 2 K (Figures 13c and 13d). determination of the amplitude of the annual temperature cycle by harmonic analysis between 10 30 km (in steps of 0.5 km) for the tropical region (30 S 30 N) leads to Figure 15 showing the largest temperature amplitudes in a layer between the tropopause and 20 km. The maximum amplitude found in the CHAMP RO data is 8.5 K at 18 km which is in almost exact agreement with Randel et al. [2003] for the GPS/MET data from 1995 1997 but with Figure 14. Individual CHAMP temperature measurements at the altitude of 18 km over the equator region (4 S 4 N). The years denote the beginning of the year. 11 of 13

much better temporal and spatial data density. This strong annual cycle in the tropical lower stratosphere is caused by the large-scale stratospheric circulation which is driven by seasonal changes in the radiative heating and forcing due to dissipation of vertically propagating Rossby and gravity waves [Reed and Vlcek, 1969; Yulaeva et al., 1994; Holton et al., 1995; Randel et al., 2002]. [37] Considering time series as in Figure 14 monthly means and the annual cycle based on 2- or 3-year values, respectively, were calculated but separately for each altitude level between 10 35 km in 0.5 km intervals over the equator (4 S 4 N). Subtracting the annual temperature cycle from the individual monthly means leads to Figure 16 showing the monthly temperature anomalies in the UTLS over the equator region based on the CHAMP RO measurements. The downward propagating patterns in the lower stratosphere have a period of about 2 years with maximum deviations of ±4.5 K between 22 and 31 km. At the tropopause level the anomalies are decreased to about ±0.5 K. This temperature anomaly patterns are caused by the stratospheric QBO [Randel et al., 1999; Baldwin et al., 2001; Huesmann and Hitchman, 2001; Ribera et al., 2003]. Figure 16 is also in agreement with QBO patterns found in the GPS/MET data published by Randel et al. [2003]. 6. Summary and Outlook [38] Global and tropical tropopause parameters on the basis of CHAMP GPS RO measurements for the period May 2001 to November 2003 have been discussed. In this study a total number of 124,491 high-resolution and highquality temperature profiles were used, whereas 37,619 temperature profiles for the tropical region (30 S 30 N) were exploited. It could be shown that the temperature bias between CHAMP temperature profiles and radiosonde data as well as ECMWF analyses is less than 0.5 K between 300 30 hpa. [39] As already discussed by Highwood and Hoskins [1998], the tropopause definitions have their limitations, especially in the tropics. However, to provide a consistent framework for this study and to compare the results with Figure 16. Temperature anomalies over the equator region (4 S 4 N) from CHAMP measurements for the period May 2001 to November 2003. Contours are ±0.5 K. The heavy dashed line shows the monthly mean CPT altitude. previous investigations tropical LRT, CPT, and also 100-hPa level statistics have been calculated. The results, as, e.g., monthly and seasonal means of global and tropical tropopause features, are in good agreement with other climatologies based on radiosonde and operational analyses data published in the past [Reid and Gage, 1996; Hoinka, 1998; Highwood and Hoskins, 1998; Randel et al., 2000; Seidel et al., 2001; Zhou et al., 2001a, 2001b; Gettelman and de Forster, 2002; Santer et al., 2003]. Noteworthy is the fact that for the discussion of the seasonal properties of the tropical tropopause characteristics no interpolation between the different bins was necessary (Figures 11 13). This is due to the global coverage with CHAMP temperature data. [40] Beside the discussion of the tropopause features in the 31 months of CHAMP temperature data a clear evidence of the lower stratospheric QBO was found. The temperature anomalies reach ±4.5 K between 22 31 km. At the CPT level the anomalies decrease to values of about ±0.5 K. [41] Because of the accuracy, high vertical resolution, and globally distributed temperature data in the tropopause region, the relatively new RO technique is suitable for global monitoring of the UTLS as an important part of the atmosphere. The CHAMP RO experiment generates the first long-term RO data set. Other satellite missions will follow (GRACE, TerraSAR-X, COSMIC, METOP) generating some thousand profiles of refractivity and temperature daily. Thus the RO technique will be established for global monitoring of temperature in the UTLS region. [42] Acknowledgments. The authors would like to thank William Randel for his helpful comments and suggestions. We thank Katrin Schöllhammer and the Institute for Meteorology at the Free University Berlin for delivering radiosonde data and the ECMWF for supplying global weather analyses. Figure 15. Annual cycle of temperature amplitude for the UTLS region based on CHAMP temperature measurements (May 2001 to November 2003). Amplitudes >5 K are plotted in solid lines. The dashed line shows the CPT altitude. References Anthes, R. A., C. Rocken, and Y. H. Kuo (2000), Applications of COSMIC to meteorology and climate, Terr. Atmos. Ocean Sci., 11, 115 156. Baldwin, M. P., et al. (2001), The quasi-biennial oscillation, Rev. Geophys., 39, 179 229. Gettelman, A., and P. M. de Forster (2002), A climatology of the tropical tropopause layer, J. Meteorol. Soc. Jpn., 80, 911 924. 12 of 13

Gettelman, A., M. L. Salby, and F. Sassi (2002), Distribution and influence of convection in the tropical tropopause region, J. Geophys. Res., 107(D10), 4080, doi:10.1029/2001jd001048. Hajj, G. A., C. O. Ao, B. A. Iijima, D. Kuang, E. R. Kursinski, A. J. Mannucci, T. K. Meehan, L. J. Romans, M. de la Torre Juarez, and T. P. Yunck (2004), CHAMP and SAC-C atmospheric occultation results and intercomparisons, J. Geophys. Res., 109, D06109, doi:10.1029/ 2003JD003909. Highwood, E. J., and B. J. Hoskins (1998), The tropical tropopause, Q. J. R. Meteorol. Soc., 124, 1579 1604. Hoinka, P. H. (1998), Statistics of the global tropopause pressure, Mon. Weather Rev., 126, 3303 3325. Holton, J. R., and A. Gettelman (2001), Horizontal transport and the dehydration of the stratosphere, Geophys. Res. Lett., 28, 2799 2802. 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, 403 439. Huesmann, A. S., and M. H. Hitchman (2001), The stratospheric quasibiennial oscillation in the NCEP reanalyses: Climatological structures, J. Geophys. Res., 106(D11), 11,859 11,874. Kursinski, E. R., G. A. Hajj, K. R. Hardy, J. T. Schofield, and R. Linfield (1997), Observing Earth s atmosphere with radio occultation measurements using the Global Positioning System, J. Geophys. Res., 102, 23,429 23,465. Marquardt, C., K. Schoellhammer, G. Beyerle, T. Schmidt, J. Wickert, and C. Reigber (2003), Validation and data quality of CHAMP radio occultation data, in First CHAMP Mission Results for Gravity, Magnetic and Atmospheric Studies, edited by C. Reigber, H. Lühr, and P. Schwintzer, pp. 384 396, Springer-Verlag, New York. Melbourne, W. G., E. S. Davis, G. A. Hajj, K. R. Hardy, E. R. Kursinski, T. K. Meehan, and L. E. Young (1994), The application of spaceborne GPS to atmospheric limb sounding and global change monitoring, JPL Publ., 94 18. Nishida, M., A. Shimizu, T. Tsuda, C. Rocken, and R. H. Ware (2000), Seasonal and longitudinal variations in the tropical tropopause observed with the GPS occultation technique (GPS/MET), J. Meteorol. Soc. Jpn., 78, 691 700. Randel, W. J., F. Wu, R. Swinbank, J. Nash, and A. O Neill (1999), Global QBO circulation derived from UKMO stratospheric analyses, J. Atmos. Sci., 56, 457 474. Randel, W. J., F. Wu, and D. J. Gaffen (2000), Interannual variability of the tropical tropopause derived from radiosonde data and NCEP reanalyses, J. Geophys. Res., 105(D12), 15,509 15,523. Randel, W. J., R. R. Garcia, and F. Wu (2002), Time-dependent upwelling in the tropical lower stratosphere estimated from the zonal-mean momentum budget, J. Atmos. Sci., 59, 2141 2151. Randel, W. J., F. Wu, and W. R. Rios (2003), Thermal variability of the tropical tropopause region derived from GPS/MET observations, J. Geophys. Res., 108(D1), 4024, doi:10.1029/2002jd002595. Reed, R. J., and C. L. Vlcek (1969), The annual temperature variation in the lower tropical stratosphere, J. Atmos. Sci., 26, 163 167. Reid, G. C., and K. S. Gage (1996), The tropical tropopause over the western Pacific: Wave driving, convection, and the annual cycle, J. Geophys. Res., 101(D16), 21,233 21,241. Ribera, P., D. Gallego, C. Pena-Ortiz, L. Gimeno, R. Garcia-Herrera, E. Hernandez, and N. Calvo (2003), The stratospheric QBO signal in the NCEP reanalysis, 1958 2001, Geophys. Res. Lett., 30(13), 1691, doi:10.1029/2003gl017131. Rocken, C., et al. (1997), Analysis and validation of GPS/MET data in the neutral atmosphere, J. Geophys. Res., 102(D25), 29,849 29,866. Santer, B. D., et al. (2003), Behavior of tropopause height and atmospheric temperature in models, reanalyses, and observations: Decadal changes, J. Geophys. Res., 108(D1), 4002, doi:10.1029/2002jd002258. Sausen, R., and B. D. Santer (2003), Use of changes in tropopause height to detect influences on climate, Meteorol. Z., 12(3), 131 136. Schmidt, T., J. Wickert, C. Marquardt, G. Beyerle, C. Reigber, R. Galas, and R. König (2004), GPS radio occultation with CHAMP: An innovative remote sensing method of the atmosphere, Adv. Space Res., 33(7), 1036 1040. Seidel, D. J., R. J. Ross, J. K. Angell, and G. C. Reid (2001), Climatological characteristics of the tropical tropopause as revealed by radiosondes, J. Geophys. Res., 106(D8), 7857 7878. Steiner, A. K., G. Kirchengast, U. Foelsche, L. Kornblueh, E. Manzini, and L. Bengtsson (2001), GNSS occultation sounding for climate monitoring, Phys. Chem. Earth. A, 26, 113 124. Thuburn, J., and G. C. Craig (1997), GCM test of theories for the height of the tropopause, J. Atmos. Sci., 54, 869 882. Thuburn, J., and G. C. Craig (2000), Stratospheric influence on tropospheric height: The radiative constraint, J. Atmos. Sci., 57, 17 28. Ware, R., et al. (1996), GPS soundings of the atmosphere from low earth orbit: Preliminary results, Bull. Am. Meteorol. Soc., 77, 19 40. Wickert, J., et al. (2001a), Atmosphere sounding by GPS radio occultation: First results from CHAMP, Geophys. Res. Lett., 28, 3263 3266. Wickert, J., R. Galas, G. Beyerle, R. König, and C. Reigber (2001b), GPS ground station data for CHAMP radio occultation measurements, Phys. Chem. Earth. A, 26, 503 511. Wickert, J., T. Schmidt, G. Beyerle, R. König, C. Reigber, and N. Jakowski (2004), The radio occultation experiment aboard CHAMP: Operational data analysis and validation of atmospheric profiles, J. Meteorol. Soc. Jpn., 82(1B), 381 395. World Meteorological Organization (WMO) (1957), Definition of the tropopause, WMO Bull. 6, Geneva. Yulaeva, E., J. R. Holton, and J. M. Wallace (1994), On the cause of the annual cycle in tropical lower-stratospheric temperatures, J. Atmos. Sci., 51, 169 174. Zhang, C. (1993), On the annual cycle in highest, coldest clouds in the tropics, J. Clim., 6, 987 990. Zhou, X., M. A. Geller, and M. Zhang (2001a), Cooling trend of the tropical cold point tropopause temperatures and its implications, J. Geophys. Res., 106(D2), 1511 1522. Zhou, X., M. A. Geller, and M. Zhang (2001b), Tropical cold point tropopause characteristics derived from ECMWF reanalyses and soundings, J. Clim., 14, 1823 1838. G. Beyerle, C. Reigber, T. Schmidt, and J. Wickert, Department 1: Geodesy and Remote Sensing, GeoForschungsZentrum Potsdam, Telegrafenberg A17, D-14473 Potsdam, Germany. (tschmidt@gfz-potsdam.de) 13 of 13