A global climatology of multiple tropopauses (MT) is discussed based on Global Positioning System (GPS) radio occultation (RO) data from the German CHAMP (CHAllenging Minisatellite Payload) and the US-Argentinian SAC-C (Satelite de Aplicaciones Cientificas-C) satellite mission for the period May 2001–April 2005. In this study we present first detailed investigations about the geographical and temporal distribution of MT during different seasons. The thickness of the layer between the lowest (first) and highest (last) tropopause has a strong annual cycle. In the vicinity of the subtropical jet (STJ) stream region values vary between 4–5 km during winter and 2–3 km during summer, respectively, whereas higher differences were found on the northern hemisphere. It is shown that the occurrence distribution of MT is in good agreement with the mean climatological location of the STJ (30°–40°) on both hemispheres, in particular during winter time.
If you can't find a tool you're looking for, please click the link at the top of the page to "Go to old article view". Alternatively, view our Knowledge Base articles for additional help. Your feedback is important to us, so please let us know if you have comments or ideas for improvement.
 The tropopause region is one of the key regions of the atmosphere. It separates the troposphere and stratosphere that have fundamental different characteristics with respect to chemical composition and static stability. For the determination of the tropopause different definitions and concepts exist [e.g. Pan et al., 2004]. By using a temperature profile the World Meteorological Organization (WMO) defines the (first) thermal tropopause “as the lowest level at which the lapse rate decreases to 2°C/km or less, provided also the averaged lapse rate between this level and all higher levels within 2 km does not exceed 2°C/km.” For any tropopause above that altitude the following condition must be fulfilled: “If above the first tropopause the average lapse rate between any level and all higher levels within 1 km exceeds 3°C/km, then a second tropopause is defined by the same criterion as above. This tropopause may be either within or above the 1 km layer.” [WMO, 1957]. These tropopause definition is used in our study.
 In the climate change discussion tropopause parameters have received more attention in recent years since they describe climate variability and change. The global mean tropopause altitude shows an increase in re-analyses and radiosonde observations during the last decades [Sausen and Santer, 2003]. On the other hand the tropopause region plays an important role in stratosphere-troposphere exchange (STE) [review in Holton et al., 1995 and Stohl et al., 2003]. In the context of STE multiple tropopause (MT) or tropopause break (TB) regions become important because they are linked to, e.g., tropopause fold events as one main mechanism for transport of stratospheric air into the troposphere in the extra-tropics [e.g., Elbern et al., 1998; Baray et al., 2000; Sprenger et al., 2003]. These tropopause folds appear as multiple stable layers (i.e., MT's) in vertical temperature profiles [Stohl et al., 2003] and mostly occur in the vicinity of the polar and subtropical jet (STJ) streams. Thus, the knowledge of the geographical and temporal distribution of MT's (TB's) and the extent of the TB's, i.e. the difference between the last and first thermal tropopause, could give a reference to the possible occurrence and, at least, depending on the extent of the TB a qualitative estimation of the strength of mixing between tropospheric and stratospheric air. This hypothesis must be quantified by experiments. To what extent the MT occurrence distribution can be used for the localisation of jet streams will be discussed in Section 2.
 The most important data source for the determination of tropopause parameters are radiosonde data whereas model analyses, as e.g., ECMWF (European Centre for Medium-Range Weather Forecasts), suffer from lower vertical resolution. Despite of good vertical resolution of radiosonde measurements a global coverage is impossible. The availability of Global Positioning System (GPS) radio signals has introduced a new remote sensing technique for the Earth's atmosphere [Melbourne et al., 1994]. GPS radio occultation (RO) enables precise refractivity and temperature profiles with high vertical resolution (less than 1 km in the tropopause region). 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 (150–200 soundings per day from CHAMP). The CHAMP RO experiment was successfully started on Feb. 11, 2001 and provides RO data continuously in an operational manner since mid-2001 [Wickert et al., 2005]. The amount of data from SAC-C has varied over the last years [Hajj et al., 2004]. Thus, SAC-C data are only available for this study from August–October 2001 and March–November 2002.
 The derivation of vertical temperature profiles from CHAMP/SAC-C occultation measurements is described in detail by Wickert et al. . For the CHAMP/SAC-C data quality in general we refer to several comparison studies [e.g. Wickert, 2004]. Comparisons of CHAMP/SAC-C tropopause parameters with ECMWF and radiosonde data can be found in Schmidt et al. [2004, 2005], the latter also contains the description of the data base and details for the determination of tropopause parameters from CHAMP/SAC-C and ECMWF data used in this study.
2. Multiple Tropopause Climatology
2.1. Meridional Structure of the Tropopause
 By applying the WMO definition on the individual CHAMP/SAC-C temperature profiles all lapse rate tropopauses (LRT) were determined between 500–70 hPa. Figure 1 shows climatological zonal mean values of tropopause altitude, pressure, temperature, and potential temperature (Θ = T(1000 hPa/p)κ; T, p: air temperature and pressure, respectively; κ = 0.286) for the first or lowest (LRTfirst) and last or highest LRT (LRTlast). Each data point represents the center of a ±5° latitude interval. In the tropics (30°S–30°N) in addition to the LRT parameters the cold-point tropopause (CPT) is plotted. Figure 1 shows the well known meridional tropopause structures as already discussed in a former study [Schmidt et al., 2004]. Obviously, the altitude difference Δ(LRTlast − LRTfirst) or the extent of the TB is largest between 30°–50° on both hemispheres with mean maximum differences of up to 3 km (Figure 1a) or 85 hPa (Figure 1b). The associated temperature differences reach maximum values of ∼7 K (Figure 1c) and ∼36 K for the potential temperature, respectively (Figure 1d). In the following we focus on the discussion of the temporal and spatial distribution of MT's and the differences Δ(LRTlast − LRTfirst).
2.2. Time-latitude Structure
Figure 2a shows the time-latitude section of the occurrence distribution of MT's (at least two LRT's) found in the GPS RO temperature profiles. Generally, MT's mostly occur in the latitude bands between 30°–50° on both hemispheres with on average higher values on the northern hemisphere (NH) compared with the southern hemisphere (SH). Between 30°–50° a clear annual cycle for the occurrence of MT's is evident with largest MT probability (>80%) during the winter months. This increased MT occurrence in that latitude band follows nearly the climatological location of the STJ, which is around 30°N–35°N during winter and moves to 40°N–45°N during summer, respectively [Holton, 2004]. The annual zonal variations of the MT probability on the SH are smaller compared with the NH. This characteristic is also found in the annual zonal variation of the STJ due to the more homogenous distribution of the land-sea masses [Holton, 2004]. At higher latitudes on both hemispheres also an annual cycle can be seen, but less clearly than in the STJ region: an increased MT probability is observed during late autumn continuing until late winter caused by intensifying of the polar jet [Holton, 2004]. In the tropics the tropopause is usually well defined throughout the year [Schmidt et al., 2005] resulting in fever MT's. During the NH summer however Figure 2a shows a slightly increased probability of MT's around the equator. This feature can be seen more clearly in Figure 3a and seems to be in agreement with the occurrence of the easterly equatorial jet stream.
 The time-latitude section of Δ(LRTlast − LRTfirst), i.e. the TB, is shown in Figure 2b. The difference is strongly correlated to the occurrence distribution of MT's. The thickness of the layer Δ(LRTlast − LRTfirst) reaching maximum values >5 km in the vicinity of the NH STJ during winter is in good agreement with results from Pan et al. . The authors found regarding the chemical composition the extra-tropical tropopause should be considered as a transition layer with (statistically) the LRT as the center. The thickness of the layer “appears to be ∼2–3 km for locations away from the STJ region, however, it expands into a thicker layer in the vicinity of the STJ due to enhanced mixing activity near tropopause break”. Considering LRTfirst and LRTlast as the lower and upper boundary of this transition layer Figure 2b (and also Figure 4) confirms the results from Pan et al.  with respect to the thickness of this layer in the STJ stream region.
2.3. Latitude-longitude Structure
 To study the geographical distribution of MT's and to answer the question which longitude region the “signals” in Figure 2 come from, we have divided the Earth's surface in a latitude-longitude grid with a resolution of 10° × 15°. For the data availability within the single grid cells see Schmidt et al. . The averaged occurrence distribution of MT's and the difference Δ(LRTlast − LRTfirst) were calculated for each latitude-longitude bin for different seasons as shown in Figure 3 and 4.
 For the discussion of the relationship between the MT occurrence and the mean wind field, Figure 3 includes the monthly mean zonal wind speed (m/s) at the 200 hPa level from ECMWF analyses (same period as for the MT analysis). The maximum zonal wind speed at 200 hPa serves as a roughly indicator for the mean location of the jet streams [Holton, 2004].
 In general, increased values of MT occurrence distribution are observed between ∼25°–50° on both hemispheres with maximum values between 30°–40° during winter time. This latitude band agrees well with the mean climatological location of the STJ as already pointed out in Section 2.2. In detail, in some regions MT occurrence maxima coincide with the 200 hPa wind speed maxima whereas in other regions the MT maxima are slightly shifted poleward compared with the location of the wind maxima. There is an excellent agreement between the wind speed and the MT occurrence and TB (>5 km) maxima over the Tibetian plateau (part of the monsoon circulation [Holton, 2004]) during NH summer (Figure 3a and Figure 4a). The strong zonal wind maxima (jet) near 35°N over Japan and the western Pacific, and the eastern part of the North American continent during NH winter (Figure 3b) coincide also with an enhanced MT occurrence (>80%). These locations exhibit the largest TB's with values between 4–5 km (Figure 4). The other locations of increased MT occurrence during winter time along ∼35° on both hemispheres are slightly shifted poleward in relation to the wind speed maxima (over the Indian Ocean, South Australia, and the western Pacific (Figure 3a), and over the Mediterranean area and the Middle East (Figure 3b)). This shift seems to be plausible if the TB's occur poleward of the STJ core which is below the tropical, and therefore higher, side of the break. Such an example is illustrated, e.g., in Pan et al. .
 There is also a region with enhanced MT occurrence (60–70%) and TB (2–3 km) east of South America (35°S) during both the summer and winter season. This is an active region of mountain waves generated due to a continuously high zonal wind speed (on average about 30 m/s at 200 hPa, see Figure 3) over the Andes [de la Torre et al., 2005] leading to permanent irregularities in the upper troposphere and lower stratosphere temperature field.
 It is also seen in Figure 3 and Figure 4 that the MT occurrence and the extent of the TB's are smaller during summer compared with the winter season (not for the Tibetian plateau). This statement is also valid for the mean zonal wind speed at 200 hPa in the jet stream areas showing the linkage between both, the MT occurrence and the wind field. In the Arctic and Antarctic region nearly no MT's are observed during summer. In the tropics a peak over the Indian ocean is noticeable (Figure 3a) which is in good agreement with the expected location of the easterly equatorial jet already mentioned in Section 2.2.
 At higher latitudes on the SH during winter the influence of the polar vortex leads to increased values of MT's at the edge of the Antarctic continent (Figure 3a). Figure 3b and Figure 4b show an increased MT occurrence and TB's at high northern latitudes in the longitude range between 60°W–60°E. This is linked to the intensifying of the polar jet stream and their associated storm tracks in the north Atlantic region and Europe during winter [Holton, 2004]. The correlation of this enhanced MT occurrence with the polar jet is not as clear as with the STJ because the polar jet is much more variable in space and time. Therefore the polar jet is not to be seen so clear in the mean zonal wind.
 The regions of enhanced MT occurrence and large TB's shown in Figure 3 and Figure 4 are in good agreement with geographical tropopause fold distributions presented by Sprenger et al.  confirming the correlation between MT's, TB's, and tropopause folds.
 The results deduced from the GPS RO data were compared with independent 6-hourly operational weather analyses from the ECMWF showing qualitatively a good agreement in the STJ stream region (not shown here), but quantitatively GPS RO data exhibit about twice the MT occurrence and the difference Δ(LRTlast − LRTfirst). In the tropics (30°N–30°S) ECMWF shows nearly no MT occurrence during the time considered here. At higher latitudes on both hemispheres ECMWF reproduces the RO MT occurrence results also only qualitatively and partly over the edge of the Antarctic continent and Europe. The absolute values are 2–4 times smaller. All the differences between GPS RO and ECMWF data are supposed to result from the poorer vertical resolution of the latter data set (60 levels from the Earth's surface to 0.1 hPa).
3. Summary and Outlook
 A first global climatology of MT's has been presented and discussed. The basis for this study are CHAMP GPS RO measurements for the period May 2001–April 2005 and SAC-C data for several periods in 2001 and 2002 resulting in a total number of 214,697 high-resolution and high-quality temperature profiles.
 MT's mostly occur in the vicinity of the jet streams. The linkage between enhanced MT's and TB's and the maximum mean zonal wind speed at the 200 hPa level (jets) could be shown, whereas the MT maxima are mostly slightly shifted poleward in relation to the wind maxima. The probability of MT's is higher during winter than during summer on both hemispheres, but generally the MT occurrence is larger on the NH. The difference Δ(LRTlast − LRTfirst) is correlated with the occurrence distribution of MT's. Largest values (>5 km) are observed in the STJ stream region.
 Weather analyses as ECMWF can reproduce our findings only qualitatively in the STJ area. At higher latitudes and in the tropics only weak agreement could be achieved.
 The CHAMP RO experiment generates the first long-term RO data set. Other satellite missions will follow (GRACE, TerraSAR-X, COSMIC, Metop) generating together with CHAMP some thousand meteorological profiles daily.
 The authors would like to thank Jet Propulsion Laboratory (JPL) for supplying SAC-C data, the German Weather Service (DWD) and ECMWF for supplying global weather analyses and the two anonymous referees for their helpful comments and suggestions. This work is partially funded through DFG project GW-CODE (ER 474/1-1), DFG priority program CAWSES SPP 1176.