• Bell, G. D., and M. S. Halpert, 1998: Climate assessment for 1997. Bull. Amer. Meteor. Soc.,79, S1–S50.

  • Bojkov, R. D., and V. E. Fioletov, 1997: Changes of the lower stratospheric ozone over Europe and Canada. J. Geophys. Res.,102, 1337–1347.

  • Elliott, W. P., and D. J. Gaffen, 1991: On the utility of radiosonde humidity archives for climate studies. Bull. Amer. Meteor. Soc.,72, 1507–1520.

  • Gibson, R., P. Kållberg, S. Uppala, A. Hernandez, A Nomura, and E. Serrano, 1997: ERA description. ECMWF-ReAnalysis Project Rep. Series, Vol. 1, 72 pp.

  • Halpert, M. S., and G. D. Bell, 1997: Climate assessment for 1996. Bull. Amer. Meteor. Soc.,78, S1–S49.

  • Hoerling, M. P., T. D. Schaack, and A. J. Lenzen, 1991: Global objective tropopause analysis. Mon. Wea. Rev.,119, 1816–1831.

  • Hoinka, K. P., 1998a: Mean global surface pressure series evaluated from ECMWF reanalysis data. Quart. J. Roy. Meteor. Soc.,124, 2291–2297.

  • ——, 1998b: Statistics of the global tropopause pressure. Mon. Wea. Rev.,126, 3303–3325.

  • ——, H. Claude, and U. Köhler, 1996: On the correlation between tropopause pressure and ozone above central Europe. Geophys. Res. Lett.,23, 1753–1756.

  • Hyson, P., 1983: Stratospheric water vapor over Australia. Quart. J. Roy. Meteor. Soc.,109, 285–294.

  • Jackman, C. H., E. L. Fleming, S. Chandra, D. B. Considine, and J. E. Rosenfield, 1996: Past, present, and future modeled ozone trends with comparisons to observed trends. J. Geophys. Res.,101, 28 753–28 767.

  • Jackson, D. R., S. J. Driscoll, E. J. Highwood, J. E. Harries, and J. M. Russell III, 1998: Troposphere to stratosphere transport at low latitudes as studied using HALOE observations of water vapor 1992–1997. Quart. J. Roy. Meteor. Soc.,124, 169–192.

  • Jenkinson, A. F., 1955: Average vector wind distribution of the upper air in temperate and tropical latitudes. Meteor. Mag.,84, 140–147.

  • Kållberg, P., 1998: Aspects of the re-analysed climate. ERA description. ECMWF-ReAnalysis Project Report Series, Vol. 2, 89 pp.

  • Kelly, K. K., A. F. Tuck, and T. Davies, 1991: Wintertime asymmetry of upper tropospheric water between the Northern and Southern Hemispheres. Nature,353, 244–247.

  • ——, M. H. Proffitt, K. R. Chan, M. Loewenstein, J. R. Podolske, S. E. Strahan, J. C. Wilson, and D. Kley, 1993: Water vapor and cloud water measurements over Darwin during the STEP 1987 tropical mission. J. Geophys. Res.,98, 8713–8723.

  • Kritz, M. A., S. W. Rosner, K. K. Kelly, M. Loewenstein, and K. R. Chan, 1993: Radon measurements in the lower tropical stratosphere: Evidence for rapid vertical transport and dehydration of tropospheric air. J. Geophys. Res.,98, 8725–8736.

  • Mote, P. W., K. H. Rosenlof, J. R. Holton, R. S. Harwood, and J. W. Waters, 1995: Seasonal variations of water vapor in the tropical lower stratosphere. Geophys. Res. Lett.,22, 1093–1096.

  • Murphy, D. M., K. K. Kelly, A. F. Tuck, and M. H. Proffitt, 1990: Ice saturation at the tropopause observed from the ER-2 aircraft. Geophys. Res. Lett.,17, 353–356.

  • Nagurny, A. P., 1998: Climatic characteristics of the tropopause over the Arctic Basin. Ann. Geophys.,16, 110–115.

  • Namias, J., and P. F. Clapp, 1951: Observational studies of general circulation patterns. Compendium of Meteorology, T. F. Malone, Ed., Amer. Meteor. Soc., 551–567.

  • Newell, R. E., and S. Gould-Stewart, 1981: A stratospheric fountain? J. Atmos. Sci.,38, 2789–2796.

  • ——, J. W. Kidson, and G. J. Boer, 1969: Annual and biennial modulations in the tropical Hadley circulation. Nature,222, 76–78.

  • ——, ——, D. G. Vincent, and G. J. Boer, 1972: The General Circulation of the Tropical Atmosphere. Vol. I. The MIT Press, 258 pp.

  • Oltmans, S. J., and D. J. Hofmann, 1995: Increase in lower-stratospheric water vapour at a mid-latitude Northern Hemisphere site from 1981 to 1994. Nature,374, 146–149.

  • Oort, A. H., 1983: Global Atmospheric Circulation Statistics 1958–1973. NOAA Prof. Paper 14, 180 pp + 47 microfiches. [Available from U.S. Government Printing Office, Washington, DC 20402.].

  • Peixoto, J. P., and A. H. Oort, 1992: Physics of Climate. American Institute of Physics, 520 pp.

  • ——, and ——, 1996: The climatology of relative humidity in the atmosphere. J. Climate,9, 3445–3463.

  • Phillips, P. D., H. Richner, J. Joss, and A. Ohmura, 1980: ASOND-78: An intercomparison of Väisälä, VIZ and Swiss radiosondes. Pure Appl. Geophys.,119, 259–277.

  • Pielke, R. A., J. Eastman, T. N. Chase, J. Knaff, and T. G. F. Knittel, 1998: 1973–1996: Trends in depth-averaged tropospheric temperature. J. Geophys. Res.,103, 28 909–28 911.

  • Pruvost, P., J. Ovarlez, J. Lenoble, and W. P. Chu, 1993: Comparison of Stratospheric Aerosol and Gas Experiment II and balloon-borne stratospheric water vapor measurements. J. Geophys. Res.,98, 4889–4896.

  • Read, W. G., J. W. Waters, D. A. Flower, L. Froidevaux, R. F. Jarnot, D. L. Hartmann, R. S. Harwood, and R. B. Rood, 1995: Upper-tropospheric water vapor from UARS MLS. Bull. Amer. Meteor. Soc.,76, 2381–2389.

  • Reid, G. C., and K. S. Gage, 1981: On the annual variation in height of the tropical tropopause. J. Atmos. Sci.,38, 1928–1938.

  • Reiter, E. R., 1961: Meteorologie der Strahlströme. Springer Verlag, 473 pp.

  • Rind, D., E.-W. Chiou, W. Chu, S. Oltmans, J. Lerner, J. Larsen, M. P. McCormick, and L. McMaster, 1993: Overview of the Stratospheric Aerosol and Gas Experiment II water vapor observations:Method, validation and data characteristics. J. Geophys. Res.,98, 4835–4856.

  • Salathé, E. P., and R. B. Smith, 1992: In situ observations of temperature microstructure above and below the tropopause. J. Atmos. Sci.,49, 2032–2036.

  • Schumann, U., 1994: On the effect of emmissions from aircraft engines on the state of the atmosphere. Ann. Geophys.,12, 365–384.

  • Sneyers, R., 1990: On the statistical analysis of series of observations. WMO Publ. 415, Tech. Note 143, 192 pp.

  • Soden, B. J., and F. P. Bretherton, 1993: Upper tropospheric relative humidity from the GOES 6.7 μm channel: Method and climatology for July 1987. J. Geophys. Res.,98, 16 669–16 688.

  • ——, and ——, 1996: Interpretation of TOVS water vapor radiances in terms of layer-average relative humidities: Method and climatology for the upper, middle, and lower troposphere. J. Geophys. Res.,101, 9333–9343.

  • Spencer, R. W., and J. R. Christy, 1992: Precision and radiosonde validation of satellite gridpoint temperature and anomalies. Part II: A tropospheric retrieval and trends during 1979–90. J. Climate, 5, 858–866.

  • Stendel, M., J. R. Christy, and L. Bengtsson, 1998: How representative are recent temperature trends? Rep. 264, Max-Planck-Institut für Meteorologie, 30 pp. [Available from Max-Planck-Institut für Meteorologie, Bundesstrasse 55, D-20146 Hamburg, Germany.].

  • WMO, 1957: Definition of the tropopause. WMO Bull.,6, 136.

  • ——, 1986: Atmospheric ozone 1985: Global ozone research and monitoring report. Rep. 16, World Meteorological Organization, 392 pp.

  • View in gallery

    Mean height z (km) of the global tropopause (1979–93; 1200 UTC): annual mean [(a), all], northern summer [(b), JJA], and northern winter [(c), DJF]. The height increment is 500 m.

  • View in gallery

    Mean temperature T (K) at the global tropopause (1979–93; 1200 UTC): annual mean [(a), all], northern summer [(b), JJA], and northern winter [(c), DJF]. The increment is 2 K.

  • View in gallery

    Mean potential temperature θ (K) at the global tropopause (1979–93; 1200 UTC): annual mean [(a), all], northern summer [(b), JJA], and northern winter [(c), DJF]. The increment is 4 K.

  • View in gallery

    Mean annual zonal wind u in m s−1 (a), meridional wind v in m s−1 (b), and vertical wind w in 10−1 mm s−1 (c) at the global tropopause (1979–93; 1200 UTC). The increments are 5 m s−1 (u), 2 m s−1 (υ), and 0.2 mm s−1 (w). Areas with negative velocities are shaded.

  • View in gallery

    Mean water vapor mixing ratio m (ppmv) at the global tropopause (1979–93; 1200 UTC): annual mean [(a), all], northern summer [(b), JJA], and northern winter [(c), DJF]. The increment is 2.5 ppmv. Shading in the mixing ratio field indicates 0–5 (none), 5–10 (weak), 10–15 (moderate), 15–20 (strong), greater than 20 ppmv (very strong).

  • View in gallery

    Continued

  • View in gallery

    Vertical profiles of mixing ratio of water vapor of ERA data (1200 UTC), of aircraft data taken during the STEP experiment (Kritz et al. 1993), and of satellite data taken during SAGE II (Rind et al. 1993): (a) STEP flight 22 Jan 1987; (b) STEP flight 23 Jan 1987; (c) STEP flight 18 Feb 1987; (d) SAGE data, 1613 UTC; ϕ = 16°S, λ = 157°W; radiosonde Samoa (RASO), 1735 UTC, ϕ = 14.2°S, λ = 170.3°W; ERA data, ϕ = 13.5°S, λ = 171°W. The horizontal lines indicate the corresponding ERA tropopause of 3.5 PVU.

  • View in gallery

    Monthly mean upper-tropospheric mean relative humidity (rh) (%) for Jul 1987 derived from satellite data [(a) adapted from Soden and Bretherton (1993)] and from ERA data (b). The mean relative humidity is determined for the layer between 200 and 500 hPa. The increment in relative humidity is 10%.

  • View in gallery

    Mean annual rh (%) at the global tropopause (1979–93; 1200 UTC). The increment is 5%. Shading in the relative humidity field indicates 0%–10% (none), 10%–20% (weak), 20%–30% (moderate), 30%–40% (strong), and greater than 40% (very strong).

  • View in gallery

    Monthly mean global tropopause height (dynamical with 3.5 PVU) anomalies from ERA data (full line) and monthly mean tropospheric temperature derived from MSU channel 2R (broken line). A nine-point running mean filter has been applied to both time series.

  • View in gallery

    Trends (△) of tropopause parameters determined for the period 1979–93: height in m per decade (a), temperature in K per decade (b), and mixing ratio of water vapor in % per decade (c). Increments are 100 m (height), 0.5 K (temperature), and 10% (mixing ratio). Positive trends are shaded.

All Time Past Year Past 30 Days
Abstract Views 0 0 0
Full Text Views 661 320 14
PDF Downloads 589 291 11

Temperature, Humidity, and Wind at the Global Tropopause

View More View Less
  • 1 Institut für Physik der Atmosphäre, DLR, Wessling, Germany
Full access

Abstract

This study presents global statistics of tropopause parameters for a 15-yr period (1979–93). The parameters are height, temperature, potential temperature, mixing ratio of water vapor, and zonal, meridional and vertical wind. The global tropopause is derived from ECMWF reanalysis data by applying the thermal and dynamical definitions of the tropopause. The tropopause climatology evaluated from the ECMWF reanalysis data is compared with those provided by radiosonde and satellite data. The meridional and zonal variations in the mean tropopause height and temperature reflect the global jet stream structure. The seasonal variation of tropical tropopause temperatures is characterized by a minimum during Northern Hemispheric summer and a corresponding winter maximum. The tropopause mixing ratio of water vapor attains a minimum in the Tropics that increases gradually poleward. The maximum of the relative humidity occurs in the Tropics above the Pacific and Indian Oceans. There are no global trends in the height and temperature of the tropopause but a decrease of about 10% per decade is evaluated for the mixing ratio of water vapor. The negative trend is particularly strong above both polar regions.

Corresponding author address: Dr. Klaus P. Hoinka, Institut für Physik der Atmosphäre, DLR, Postfach 1116, D-82230 Wessling, Germany.

Abstract

This study presents global statistics of tropopause parameters for a 15-yr period (1979–93). The parameters are height, temperature, potential temperature, mixing ratio of water vapor, and zonal, meridional and vertical wind. The global tropopause is derived from ECMWF reanalysis data by applying the thermal and dynamical definitions of the tropopause. The tropopause climatology evaluated from the ECMWF reanalysis data is compared with those provided by radiosonde and satellite data. The meridional and zonal variations in the mean tropopause height and temperature reflect the global jet stream structure. The seasonal variation of tropical tropopause temperatures is characterized by a minimum during Northern Hemispheric summer and a corresponding winter maximum. The tropopause mixing ratio of water vapor attains a minimum in the Tropics that increases gradually poleward. The maximum of the relative humidity occurs in the Tropics above the Pacific and Indian Oceans. There are no global trends in the height and temperature of the tropopause but a decrease of about 10% per decade is evaluated for the mixing ratio of water vapor. The negative trend is particularly strong above both polar regions.

Corresponding author address: Dr. Klaus P. Hoinka, Institut für Physik der Atmosphäre, DLR, Postfach 1116, D-82230 Wessling, Germany.

1. Introduction

In recent years the tropopause region has received much attention because there is a growing necessity of knowing the temporal and spatial structure of meteorological parameters within the transition zone between the troposphere and the stratosphere. In order to estimate the stratospheric–tropospheric exchange of mass, water, and chemical constituents it is necessary to have precise knowledge about the meteorological conditions of the tropopause. Most commercial air traffic occurs in the tropopause region, in a layer between 9- and 12-km altitude. The growth of subsonic traffic during the last decades caused concern about the pollutant emissions and raised questions on the impact of air traffic on weather and climate in general (e.g., Schumann 1994). During the last two decades a significant loss in stratospheric ozone has been observed around the globe. There is some evidence that there exists a correlation between the height of the tropopause surface and the column of stratospheric ozone (e.g., Hoinka et al. 1996). Thus trends in tropopause parameters can indicate to what extent the stratospheric ozone has changed due to dynamical reasons.

Of particular interest is the tropical tropopause, since it is widely accepted that most water vapor enters the stratosphere here; hence, tropical tropopause conditions have a strong influence on the stratospheric water vapor distribution. Recently Mote et al. (1995) showed that tropical stratospheric air appears to retain information about tropopause conditions that it encountered during more than a year as it rises through the stratosphere. The lower-stratospheric water vapor is also influenced by the annual cycle of tropical tropopause temperatures, as pointed out by Newell and Gould-Stewart (1981). In general, however, the knowledge of the global upper-tropospheric and lower-stratospheric humidity is limited. First water vapor climatologies of the tropical tropopause region, and of adjacent levels in the lower stratosphere and upper troposphere, have been attempted recently based on data collected during the second Stratospheric Aerosol and Gas Experiment (SAGE II; Rind et al. 1993) and with the Halogene Occultation Experiment (HALOE) instrument that flies on the Upper Atmospheric Research Satellite (Jackson et al. 1998).

At present there are few studies on statistics global tropopause statistics whereas the literature is rich in upper-tropospheric statistics related to fixed pressure levels (e.g., Oort 1983). Therefore, data from the 200-hPa level have often been used as a measure for the midlatitude tropopause and similarly the 100-hPa surface is substituted for the tropical tropopause. However, rawinsonde observations confirm that a constant pressure surface, for example, that of 200 hPa, is a bad substitute for the tropopause because it can lead to serious misinterpretations. It is obvious that this approach can represent the tropopause only within limited geographic areas. Recently Hoinka (1998b) described climatological fea-tures of the pressure and zonal wind field at the global tropopause. He showed that the wind at 200 hPa is a good measure of the tropopause wind. This is particularly valid in the Tropics and in the midlatitudes; however, it is not above the polar regions. Although the zonal wind at the tropopause is similar to that at 200 hPa, it should be kept in mind that this does not mean that other tropopause parameters, for example, temperature, are similar to those at 200 hPa.

To the author’s knowledge global fields of tropopause parameters and their trends have not yet been published. Therefore, the global statistics of following tropopause parameters are presented here: height, temperature, potential temperature, mixing ratio of water vapor, relative humidity, and zonal, meridional and vertical wind. Decadal trends of the global field of tropopause parameters are also presented. From a statistical point of view it is also discussed whether the data at a fixed pressure level are a useful substitute for the tropopause data. The presented climatology is based on the ECMWF reanalysis (ERA) dataset, which has been created at the European Centre for Medium-Range Weather Forecasts by rerunning a selected model version with its analysis cycle without changes for the whole period (Gibson et al. 1997). This study is considered as an extension of our previous study of tropopause parameters (Hoinka 1998b).

In the present paper emphasis is put on describing the climatology of the tropopause water vapor because it is widely recognized to be of fundamental importance in determining the present climate and its sensitivity to increasing greenhouse gases. Much of the uncertainty related to upper-tropospheric and lower-stratospheric water vapor stems from problems in accurately measuring its distribution and climatological variations. Multiannual global datasets such as the ERA data provide an excellent base to decribe consistently the climatology of tropospheric humidity (Kållberg 1998). Notwithstanding, one has to keep in mind that climatological statistics derived from data series of reanalyses provided by ECMWF or the National Centers for Environmental Prediction–National Center for Atmospheric Research (NCEP–NCAR) and of satellites represent only an approximative description of the real atmosphere.

The paper is structured as follows: Section 2 describes the data and method; in section 3 statistics of tropopause height, temperature, and wind are given; section 4 presents statistics of the tropopause humidity; and finally section 5 discusses trends.

2. Data and method

The present study considers a data sample of 15 yr that starts in January 1979 and ends on December 1993. The data used were taken from the ERA data, which were provided by the ECMWF within the reanalysis Project (Gibson et al. 1997). This time series of data with global coverage was produced by applying the analysis scheme introduced in 1993 to the entire data series. The second important point is that all available observational data were considered in the analyses. Because the operational analyses have to be done within a limited period of time, they consider only those observational data that are available before the analysis starts. Data that arrived after this so-called cutoff time are not considered in the operational analysis. Thus the observation density was improved for the ERA analyses. Model-level data are used because they contain about twice the number of levels in the tropopause region compared to the standard pressure level data. The grid resolution is 1.125° in latitude and longitude. The analyses are restricted to one time per day, 1200 UTC.

Following Hoinka (1998b), the location of the tropopause is evaluated by applying the dynamical (potential vorticity) and the thermal (lapse rate) definitions of the tropopause (WMO 1957, 1986). The dynamical tropopause is useful only in the extratropics, since the potential vorticity surfaces tend to be vertical near the equator. A global quasi-horizontal surface has been constructed by merging the extratropical dynamical tropopause with the tropical thermal tropopause. The two tropopauses are combined in the following way: for latitudes greater than 36° the dynamical tropopause is applied, whereas for latitudes smaller than 19° the thermal tropopause is used. Within the transition zone a weighted average of dynamically and thermally derived tropopause pressures is taken that ensures a continuous and smooth variation of the tropopause pressure between tropical and extratropical latitudes.

The WMO (1986) defines the dynamical tropopause by the value 1.6 PVU (potential vorticity unit) where 1 PVU is equal to 1.0 × 10−6 K m2 kg−1 s−1. Other studies suggest values between 1 and 3 PVU. Currently, there has been no conclusive evidence for a particular value of potential vorticity to divide the stratosphere and the troposphere. In view of the uncertainty of applying a particular value of potential vorticity, Hoerling et al. (1991) examined tropopause analyses based on the application of several potential vorticity thresholds. They showed that with the value suggested by the World Meteorological Organization (WMO) the tropopause pressures are much too high compared to the thermal tropopause. Analyses with threshold values ranging from two to five units showed that the tropopause pressure was systematically overestimated for values less than three and systematically underestimated for values greater than four units. In the present study the dynamical tropopause is assigned to the vertical position where the potential vorticity first exceeds the threshold value of ±3.5 PVU. The positive values are used for the Northern Hemisphere because the potential vorticity is positive above this hemisphere. In the Southern Hemisphere the potential vorticity is less than zero, so there the negative values are applied. For further details on deriving the tropopause and its parameters from analysis data, see Hoinka (1998b).

3. Tropopause height, temperature, and wind

a. Height

Because the jet stream is correlated with a steep horizontal gradient in the tropopause height, strong concentration of isohypses or isobars can be interpreted as jet stream position. Zonal variations in this steepness can also be seen as an indicator for the occurrence of jet maxima and minima. Figure 1 depicts the height of the tropopause for the entire 15-yr period (Fig. 1a, all), summer [Fig. 1b, June–August (JJA)], and winter [Fig. 1c, December–February (DJF)]. It shows the same features as those observed for the pressure that are discussed in Hoinka (1998b). The strongest meridional steepness of the tropopause surface occurs in the western Pacific jet region during DJF where it reaches up to 4500 m per 1000 km meridional distance. The steepness is very strong during winter when the jet in this region strengthens to about 60 m s−1 (Namias and Clapp 1951). In summer the jet maximum has moved to the mid-Pacific by weakening down to 20 m s−1. Above northern America to the east of the orographically induced lee trough there is a region of high jet stream intensity that is very prominent during winter (Reiter 1961). The climatology indicates a jet strengthening of up to 45 m s−1, which is about twice the magnitude that is observed during summer. Also above the northern part of the Asian continent appears a region with great steepness in the meridional gradients indicating a wintertime strengthening of the jet stream. Another wintertime jet maximum is observed above the Arabian peninsula, which is also evident in the steep tropopause surface in that region shown in Fig. 1 and also in a wind maximum of about 40 m s−1 at the tropopause (not shown). During wintertime the jet axis is located directly above the Tibetian Plateau whereas during summer it is shifted toward the north and the easterly wind regime dominates the southern part of Asia (see, e.g., Reiter 1961).

In the Southern Hemisphere during summer there is a jet stream of up to 35 m s−1 above Australia (Jenkinson 1955). He pointed out that during JJA there are jet maxima also observed above South America at about 30°S and at the southern edge of Africa. Both jet maxima are not obvious from the steepness of the tropopause surface given in Fig. 1. Nevertheles, in general the meridional and zonal variations in the tropopause height reflect to a great deal the global jet stream structure.

The averaged heights of the tropopause range between about 9 km in both polar regions and 16 km in the Tropics. From a series of radiosonde data, Nagurny (1998) evaluated the thermal tropopause above the Arctic and showed that its mean annual height amounts to about 9-km altitude. Figure 1c exhibits that in winter the tropopause above the western Pacific and above northern South America is higher than 16 km whereas in summer (Fig. 1b) the tropopause height is greater than 16 km only over the Indian Ocean and above the Tibetian Plateau. Based on radiosonde data, Reid and Gage (1981) showed that, over large areas of the Tropics, tropopause heights are indeed lowest in northern summer and highest in northern winter.

The difference in the thermal structure of the troposphere between the equator and the poles results in different heights of constant pressure surfaces. This difference is about 2000 m at 200 hPa and 1500 m at 100 hPa (Peixoto and Oort 1992). The meridional variation of the tropopause height of about 7000 m is much larger than the corresponding height variations of the 200- or 100-hPa surface. Obviously neither pressure level is a good substitute for the global tropopause. Nevertheless, pressure levels can be used over limited areas. In the Tropics the 100-hPa level is a good proxy for the tropopause, whereas in the midlatitudes 200 hPa is better, and the 300-hPa surface is appropriate in the polar regions.

b. Temperature

The global structure of the temperature, given in Fig. 2, corresponds well with those of pressure and height. The location of the region where the meridional gradients in temperature are strongest extends more to the south above the Northern Hemisphere than those of height and pressure. This is because the jet stream axis is on the southern side of the steep gradients in tropopause height whereas the strongest temperature gradients occur above the jet stream axis. This is correspondingly valid for the Southern Hemisphere.

The temperature above the north polar region is close to 217 K (−56°C) whereas the temperature above the south polar region is about 8 K lower. For the thermal tropopause a mean annual temperature of 217 K is determined from a series of radiosonde data taken between 85° and 90°N (Nagurny 1998), which corroborates the tropopause temperature above the Arctic. A comparable difference of about 10 K between both polar regions is evaluated by Oort (1983) for the 100- and 200-hPa levels. At the 200-hPa level the temperatures approach −50°C above the north polar region, −55°C above the Tropics, and −60°C above the south polar region. The 100-hPa level values are −53°C, −78°C, and −62°C, respectively. This comparison suggests that for all latitudes the temperature at the 100-hPa level provides a good estimate of the tropopause temperature.

In the Tropics the tropopause is very high at about 16 km. Therefore the temperature is about 192 K (−81°C), which is obtained in particular above the western Pacific. Newell et al. (1972) found that the largest region of lowest 100-hPa temperatures (colder than −82.5°C in the seasonal mean) occurs over the western Pacific Ocean in winter. Secondary regions of limited extent appear over northern South America and Africa. Newell and Gould-Stewart (1981) presumed a significant water vapor transport from the troposphere into the stratosphere in these regions. They characterized this feature by the term “stratospheric fountain.” They suggested that the “fountain” is most active in the west Pacific in winter and moves to the monsoon region for the Northern Hemisphere summer.

For convenience the potential temperature (θ) is given in Fig. 3. It is about 308 K above the northern polar region whereas it is close to 304 K above the Antarctic. The midlatitude transition zone between the Tropics and the polar regions shows a belt of strong meridional gradients that extends less toward the south (north) in the Northern (Southern) Hemisphere than the corresponding strong gradient belts in the pressure and temperature fields. This is because there is a meridional shift between the regions of steep gradient in tropopause temperature, pressure, and height as mentioned above. The meridional gradient in θ is again very strong above eastern Asia. For the other regions of the transition zone the gradients are around 15 K per 10° latitude.

c. Wind

The mean zonal, meridional, and vertical wind at the tropopause are displayed in Fig. 4. The annual and seasonal structures of the zonal component were discussed in Hoinka (1998b). Mean fields of the meridional component at the 100- and 200-hPa surface (Oort 1983) show the same global structure as that in Fig. 4b: strong northerly flow above North America, eastern Asia, and the southern Indian Ocean. There is also a regional maximum of northerly flow above the western coast of South America. Generally, there is only a slight difference in the magnitudes of the wind amplitudes between the values at 100 or 200 hPa and those at the tropopause where they are largest. Thus, like the zonal wind component, the meridional wind at the 200-hPa level is a good measure of the tropopause wind. This is particularly true for the Tropics and in the midlatitudes because the zonal wind shows the same coherent structures within the layer between 500 and 50 hPa. Above the polar regions the 200-hPa level is not a useful proxy for the tropopause.

In the present study globally gridded analysis data were compared with measurements. The globally gridded analysis data are derived from a mixture of observational data and model forecast data. The observed data suffer from the fact that they contain errors related to the measurement system and caused by the spatially inhomogeneous distribution of radiosondes. Peixoto and Oort (1992) estimated the uncertainties due to spatial gaps in the observed data based on a 10-yr climatology. They indicated that at 200 hPa above both hemispheres, but particularly above the southern one, the meridional wind fields must be considered with care.

The vertical component of the wind (w) is given in Fig. 4c. It shows various regions with a distinct vertical component. At the eastern edge of Asia downward motion of about −0.8 mm s−1 is observed. An area with maximum upward motion of similar magnitude is apparent over central Africa. Above Australia and west of it there is downward motion of up to −0.8 mm s−1. Unfortunately vertical wind statistics of the upper troposphere and lower stratosphere are not available as yet. Peixoto and Oort (1992) showed that in the upper branch of the Hadley cell (≈200 hPa) the temporal average of the mean zonal vertical velocity in the Tropics is around ±1.5 mm s−1, whereas in the extratropical regions the vertical velocity vanishes at this pressure level. Zonally averaged time-mean upwelling velocities at tropopause level in the Tropics have been estimated to be about 0.2 mm s−1 (Newell et al. 1969). In situ measurements of the vertical velocity by aircraft indicate that there is no difference in behavior between the upper-tropospheric and lower-stratospheric vertical wind (e.g., Salathé and Smith 1992). They reported that the tropopause is not marked by significant microstructure features. This suggests that the magnitudes mentioned above represent a reasonable estimate of the observed vertical velocities in the tropopause region. Nevertheless, a satisfactory explanation of the global field of the vertical velocity at the tropopause cannot be given here.

4. Humidity

Figure 5 shows the water vapor mixing ratio at the tropopause. It is desirable to compare these data with observed regional statistics or global data. Humidity data are often considered to be the least reliable of all radiosonde data (see, e.g., Elliott and Gaffen 1991). Comparing various radiosonde types shows that the measured relative humidity at 200 hPa varies by up to 10% (Phillips et al. 1980). Oort (1983) did not attempt an analysis of the humidity fields above the 300-hPa level due to the vagueness of the upper-tropospheric humidity data. The recently published long-term climatology of relative humidity in the atmosphere by Peixoto and Oort (1996) also considers only the humidity up to the 300-hPa level. The main hope in obtaining global data on the atmospheric moisture content lies in using satellite data. At present, there is no long-term statistical study about the satellite-derived upper-tropospheric humidity field. But there are statistics describing mean structures of short periods, seasons, or months that can be used, however, only in a limited sense, for comparison with the field of mixing ratio of water vapor shown in Fig. 5.

In the following, aircraft profile and satellite-derived data are used for a comparison. In situ measurements performed with the ER-2 aircraft (Murphy et al. 1990) will be compared with the ERA data. Then airborne profiles taken during the Sratosphere–Troposphere Exchange Program (STEP) experiment in 1987 (Kritz et al. 1993) will be shown to verify the humidity structures obtained from daily profiles of ERA data. The first publication of a global study of water vapor in the region of the tropical tropopause was that of Rind et al. (1993) using data of SAGE II. Recently, Jackson et al. (1998) presented observations of water vapor in the equatorial region made by the HALOE instrument. A further dataset is obtained from the Microwave Limb Sounder (MLS; Read 1995). Finally, humidity data derived from data collected by the Geostationary Operational Environmental Satellite (GOES) will also be used (Soden and Bretherton 1993).

a. Comparison with in situ measurements

Figure 5 shows that the mixing ratio of water vapor at the tropopause has a minimum in the Tropics that increases gradually poleward; this is evident during the entire year and for the seasons JJA and DJF. A similar structure of water vapor mixing ratio is to be expected for the lower stratosphere and upper troposphere.

Above the Tropics lowest mixing ratios occur during DJF (Fig. 5c), particularly in January, when a minimum in tropopause temperature is observed (Fig. 2c). Very low tropopause mixing ratios of water vapor of less than 3.5 ppmv are found during DJF in the Tropics above the western Pacific Ocean, above the Indian Ocean, and northern South America, approximately the “fountain” regions of Newell and Gould-Stewart (1981). But during northern summer the situation is completely different. The mixing ratios for the same region are between 6 and 7 ppmv. Airborne in situ measurements during the STEP experiment indicate that above Central America the mixing ratios are stronger than above northern Australia (Kelly et al. 1993). This is roughly corroborated by the mean annual situation in the ERA data as depicted in Fig. 5. Above Australia there is a prominent maximum of up to about 18 ppmv during the Southern Hemispheric winter (JJA) whereas during summer (DJF) magnitudes around 5 ppmv occur. Profile data taken by aircraft sporadically above Australia during Southern Hemisphere summer and winter suggest vaguely that the tropopause is more moist during summertime (Hyson 1983).

Above both polar regions the tropopause is relatively humid during summer and dry during winter. The seasonal variation above the Arctic ranges between 10 and 30 ppmv and above the Antarctic between 10 and 15 ppmv. In the polar regions the tropopause statistics show values between 10 ppmv (all; Antarctic) and 20 ppmv (all; Arctic). Similar magnitudes were taken by airborne measurements reported by Murphy et al. (1990). They have shown that above Stavanger (59°N) the measured mixing ratios within 1 km of the tropopause range between 5 and 40 ppmv, and above Punta Arenas (53°S) the mixing ratios range between 4 and 10 ppmv. The corresponding median mixing ratios are 25 ppmv (Stavanger) and 8 ppmv (Punta Arenas). This asymmetry of upper-tropospheric water vapor exists not only between both polar regions but is well known to exist in general between the Northern and Southern Hemispheres. Like the mixing ratio field at the tropopause and the in situ aircraft measurements, satellite water vapor measurements also indicate that, taken annually, the Southern Hemispheric upper troposphere and lower stratosphere is drier than its northern counterpart (e.g., Kelly et al. 1991). The hemispheric asymmetry in humidity is also well known for the entire atmosphere; using ERA data, for example, Hoinka (1998a) showed that the hemispherically averaged water vapor surface pressure is 2.45 (2.28) hPa for the Northern (Southern) Hemisphere.

Finally, single vertical profiles of tropical humidity data at particular days are considered in order to show that the ERA data contain the basic features of the humidity structures in the tropopause region. This is done by comparing humidity data collected by aircraft, radiosondes, and satellite during the experiments STEP and SAGE with ERA data of the same day and from the same region. Figure 6 gives the vertical profiles of four days. The ERA profiles that are compared with STEP aircraft profiles (Figs. 6a–c) are obtained by averaging the data profiles of the four surrounding grid points. Figure 6 shows that the coarse structures of the stratospheric profiles of mixing ratio of all data sources compare very well. It is clear that the finescale structures encountered by the aircraft cannot be resolved by the ERA data. Also the comparison of satellite, radiosonde, and ERA data exhibit only weak differences (Fig. 6d), which are within the error range of the measurement system used during SAGE II (Pruvost et al. 1993).

b. Comparison with satellite data

Rind et al. (1993) derived monthly mean values of tropical water vapor mixing ratios at the 100-hPa level for the period 1985–89. In January, values between 3 and 6 ppmv occur above the Tropics. Embedded are various regions with lower mixing ratios. The values are about 2 ppm above the region between the western Pacific (north of Australia) and the eastern coast of Africa. Low mean values occur also above equatorial Africa. A similar distribution of mixing ratios is determined by the ERA data for the DJF period (Fig. 5c). Another region with relatively low mixing ratios is located above northern South America; this is apparent in the SAGE as well as in the ERA data. However, the comparison of magnitudes between the ERA and SAGE data shows that the SAGE data are much drier. Rind et al. (1993) pointed out that a comparison between the upper-tropospheric radiosonde climatology of Oort (1983) and SAGE data found SAGE to be much drier, by approximately 50%, for most SAGE measurement periods.

The July mean of SAGE data shows relatively small mixing ratios around 2 ppmv above the Arabian Sea west of India; a further minimum is observed above Indonesia and above the equatorial eastern Pacific. Above Australia there is a weak maximum in mixing ratio. The ERA data exhibit roughly the same distribution, with the exception that there is a maximum above the area west of India. This difference might be due to the different length of the considered periods: July 1985–89 (SAGE) and JJA 1979–93 (ERA). This seems to be reasonable because for a different period (JJA 1992–96) there appears a maximum above that region in the mixing ratio derived from HALOE data (Jackson et al. 1998).

For the period 1992–96, Jackson et al. (1998) determined the mean mixing ratio of water vapor at the 100-hPa pressure level, which is in the Tropics close to the tropopause. Above the Tropics they obtained for DJF magnitudes around 3 ppmv. This compares reasonably well with the value of around 5 ppmv obtained from the ERA data. However, various limited areas in the Tropics differ from this picture. Above South America, the southern Indian Ocean, and Australia the mixing ratio increases to more than 3.6 ppmv. On the contrary, the ERA data (Fig. 5c) show a decrease in mixing ratio above these regions: less than 3.5 ppmv above Australia and less than 4.0 ppmv above South America. The satellite data show a minimum less than 2.4 ppmv above southeast Asia; the ERA data show a similar but more extended area with a maximum of up to 15% farther to the northeast. The strong increase above Australia during JJA (Fig. 5b) is not reproduced by the satellite-derived mixing ratios. But in both data sources an increase occurs above the Indian Ocean to the west of India, which is related to the occurrence of the summer monsoon in the region.

Read et al. (1995) showed that upper-tropospheric water vapor for a 100-hPa-thick layer at 215 hPa based on MLS data has the basic structure expected from climate model simulations. However, there is an apparent difference between the tropopause data field (Fig. 5) and the MLS data fields (Read et al. 1995). At the tropopause the polar regions are more humid than the tropical regions whereas the opposite holds for the MLS data field. Because the latter are at a fixed 215-hPa level, the data show whether the air is in the stratosphere and hence should be drier, or in the troposphere where it should be moister. Thus, the polar data field indicates lower-stratospheric characteristics and the tropical one upper-tropospheric features. Correspondingly, the mixing ratios derived from satellite data are significantly higher in the Tropics compared to the tropopause field, which shows only 5–8 ppmv. The global maximum values in specific humidity are observed well beneath the tropical tropopause at the 300-hPa surface because of the strong tropical convection (Oort 1983). This is in accord with the statistics based on satellite data (Read et al. 1995).

In the midlatitudes, the tropopause and the 215-hPa surface are approximately at the same height, which suggests their data could be compared. Significant meridional deformation of the isolines in the mixing ratio is observed at the tropopause above the northern midlatitudes. This is particularly notable above the North Atlantic storm region and above the North American continent. Very large values of up to 30 ppmv are obtained above northern Asia as well as above the eastern edge of this continent. Above the Southern Hemisphere such features do not appear. The satellite data also show a moist region above the western part of the Atlantic storm track region and west of the North American continent. In the satellite-derived humidity this is particularly evident during Northern Hemispheric winter whereas during summer very high values are prominent above the eastern edge of Asia. The MLS data show moisture streams extending far north from the Tropics to high northern latitudes along the eastern U.S. coast and eastern Japan. In the same areas the tropopause is lower in comparison to the neighboring region in the zonal direction. This results in an increase in humidity as is obvious from the mixing ratios at the tropopause shown in Fig. 5.

A comparison of the magnitudes in mixing ratio at the tropopause and those derived from satellite data shows that the latter are significantly larger. In the midlatitudes values around 10 ppmv appear at the tropopause whereas those derived from satellite data are between 50 and 100 ppmv. The reason for this is twofold. One reason is that the satellite data represent an integral quantity over a vertical layer of 100 hPa. A second reason is related to the sensitivity of the measurement system. Water vapor values can be deduced from the MLS measured signal over the vertical range where this signal is strong enough to be distinguished from other effects. This range corresponds to water vapor abundances between approximately 100 and 300 ppmv. Consequently, when the atmosphere has less than 100 ppmv humidity, more than 50% of the signal comes from the dry air continuum emission. Read et al. (1995) pointed out that at 215 hPa the rms error in humidity is roughly 25 ppmv, which is a large rms error when the concentrations are low.

Soden and Bretherton (1993, 1996) derived upper-tropospheric relative humidity from satellite data by considering vertically averaged values over the layer between 200 and 500 hPa. Figure 7 compares these data with mean upper-tropospheric relative humidities that are evaluated from the ERA data by averaging the profile data over the same layer. The considered period is July 1987. The region is limited to North and South America. Note that the mean level derived from the ERA data is about 10%–15% (in relative humidity units) lower than that derived from the satellite data. A band of high upper-tropospheric moisture coincides with the ITCZ and is particularly evident over the convective active region of Central America. A moist tongue over the midlatitude storm track region in the Southern Hemisphere and a moist band over the Atlantic storm track region in the Northern Hemisphere are obvious. The latter extends in a southwest–northeast direction off the east coast of North America. East of the Atlantic storm track region low relative humidity values extend from Europe toward the southwest. Between the equator and 20°S east of South America the values of relative humidity diminish to less than 10% (ERA data) and 20% (satellite data). Above the southern end of South America the satellite data show another moist region with values up to 60% whereas the ERA data exhibit a relative minimum. In general, the regional structure of both humidity datasets compares very well, which suggests that the ERA data describe the humidity even in the upper troposphere in a reasonable manner. This is not too surprising because the ERA analyses consider also satellite radiance data as input for the analysis that are related to temperature and humidity structures in the atmosphere (Gibson et al. 1997). However, the quality of satellite retrievals varies considerably over the years.

The long-term average of the relative humidity at the tropopause is given in Fig. 8. The ERA data show that at the tropical tropopause the maximum relative humidity occurs. Above the region north of Australia and southeastern Asia very high values of more than 50% were obtained from the ERA data. Also west of northern South America there is an extended region with relative humidities of more than 50%. The ERA data also show that the relative humidity is around 15% above the Arctic whereas above the Antarctic it is more than 25%. The seasonal variations are weak (not shown); the structure during northerly summer is similar to the annual structure. During DJF the wet tropical belt above the Indian Ocean extends farther to the west and covers all of central Africa. Also above northern South America the relative humidity increases.

5. Trends

During the last two decades a significant year-round loss in ozone has been observed around the globe; above the midlatitudes and Tropics this decrease is between 2% and 4% (1975–95; Jackman et al. 1996). Various studies indicate that there is a correlation between the tropopause height and the column of total stratospheric ozone. The idea behind this is that with increasing height of the tropopause the depth of the stratospheric ozone column decreases and vice versa. Hoinka et al. (1996) have shown that for midlatitudes a correlation of about 0.5 exists. This would be associated with a change in ozone per change in tropopause pressure by about 0.5 Dobson Units (DU) hPa−1, which is equivalent to 18 DU km−1.

Therefore, in this section trends of tropopause parameters are discussed in order to show their local, regional, and global structures. The considered 15-yr period is not very long but sufficient as a basis for a trend calculation. An prerequisite of the used data series is that it does not contain unphysical trends. Stendel et al. (1998) have discussed the reliability of temperature trends determined from surface measurements, satellite observations, and reanalyses. They pointed out that there exists a considerable degree of uncertainty. Good agreement was found in data-rich regions whereas for regions without conventional observations from radiosondes, trends cannot reliably be estimated from reanalyses. On the other hand, Hoinka (1998a) showed that the time series of ERA data does not exhibit a trend in integral quantities such as mean global total surface pressure, water vapor pressure, and dry-air pressure. This suggests that the time series is reliable and derived trends can be considered as realistic, at least for the integral quantities. In spite of the uncertainty, trends in tropopause parameters are presented here in order to show the tropopause characteristics derived from ECMWF reanalyses.

The trends are determined from the series of monthly mean values by removal of the zonal annual cycle from the original data. Global or regional averages are then obtained by weighting the gridded data value by the corresponding representative grid area. Figure 9 (full line) shows the time series of monthly mean global tropopause heights anomalies for the 15-yr period. Variations in tropopause height in consecutive years of up to 100 m are apparent. It is interesting to note that the time series of the monthly mean tropopause height exhibits a similar temporal behavior as that determined for the monthly mean tropospheric temperatures, which are derived from Microwave Sounding Unit (MSU) data (Bell and Halpert 1998) (Fig. 9, broken line) or from radiosonde data based on a 63-station network (Halpert and Bell 1997) (not shown). An increase in the mean tropospheric temperature is highly correlated with an increase in the height of the mean tropopause and vice versa. The correlation coefficient between the MSU tropospheric temperatures and the ERA tropopause heights amounts to 0.73 ± 0.04. The monthly standard deviation of the mean global tropopause varies between 5 and 10 m; a corresponding standard deviation for the MSU data series is not available.

The regression coefficient derived from the annual mean ERA tropopause heights and MSU temperatures is 0.003 K m−1 with an error of 0.0006 K m−1. An alternative way of obtaining a height–temperature relation is to calculate the height difference between two pressure surfaces, which is referred to as the thickness Δz of the corresponding layer. Considering the troposphere then, the two pressure surfaces are defined by the surface pressure ps and by the tropopause pressure ptr. Thickness is related to the mean temperature T of the layer Δz = ztrzs by
i1520-0493-127-10-2248-eq1
where R is the gas constant and g the gravitational acceleration. Thus a change in Δz over some time period is related to a change in Δ T. For the entire tropospheric layer, approximately 1000–200 hPa, a 0.1 K change in T corresponds to about 5-m thickness change. The ratio 0.02 K m−1 is much larger than the regression coefficient. The reason for this difference is not obvious.

Table 1 gives trends and the trend-to-noise ratios of tropopause parameters. The confidence level of a trend is tested by the trend-to-noise ratio, and by the nonparametric Mann–Kendell trend test (Sneyers 1990). In both cases values of 1.0, 2.0, and 3.0 point to 80%, 95%, and 99% confidence levels, respectively. In general, the global and hemispheric trend values—except the mixing ratios—are weak and have a confidence level of less than 80%. The pressure trends for the globe, NH, and SH are +0.1 hPa dayr−1. In the following the abbreviation dayr (decayears) is used. Hoinka (1998b) determined −0.1, +0.5, and −0.6 hPa dayr−1 for the globe, NH, and SH, respectively. These trends were derived by using the complete time series where the annual cycle was retained.

The global mean tropopause height shows a trend of −7 m dayr−1 (Table 1). Applying the regression coef-ficient derived from the ERA and MSU data series, a decrease results in the mean temperature of about −0.021 K dayr−1. This value is smaller than the trend values derived from MSU temperatures of −0.054 K dayr−1 and from the NCAR–NCEP reanalyses data of −0.075 K dayr−1 (Pielke et al. 1998). Both values are related to the period from 1979 to 1996. For the Northern (Southern) Hemisphere a change in mean height of −1 (−13) m dayr−1 is determined from the ERA data suggesting a change in mean tropospheric temperature of −0.003 (−0.039) K dayr−1. Pielke et al. (1998) showed that the corresponding MSU temperature trends are −0.006 (NH) and −0.114 K dayr−1 (SH). The negative trends in the tropopause height and the related cooling trends in the mean tropospheric temperature are derived within a confidence level of less than 80%. In general, the trends resulting from both datasets compare well. One reason for the remaining differences is that the MSU and ERA data series are based on different periods. For instance, Pielke et al. (1998) showed that the trend values (MSU and NCAR–NCEP reanalyses) for the period 1973–96 differ considerably from those for 1979–96. Their results exhibit a warming trend within a 95% confidence interval for the period 1973–96. Spencer and Christy (1992) showed that the MSU-2R data exhibit a decadal trend of +0.039 K dayr−1 (globe), +0.094 K dayr−1 (NH) and −0.016 K dayr−1 (SH) for the period of 1979–90.

It is well known that for smaller regions or even locally there are much stronger trends in pressure and height. Above Europe, Bojkov and Fioletov (1997) found an increase in height of 100 to 300 m dayr−1 (1973–94). Using radiosonde data, Hoinka et al. (1996) determined a trend in tropopause pressure above Munich of −4.5% dayr−1 (1979–92), which is equivalent to an increase in height of about 350 m dayr−1. A rather strong negative trend is determined for the mixing ratio of water vapor. For both hemispheres and globally the trends are negative around −1.0 ppmv dayr−1, which is equivalent to a reduction of about 10% of the mean tropopause mixing ratio of water vapor during one decade. This is also apparent in the midlatitudes and the Tropics. Above both polar regions, the trend of the mixing ratio has its maximum. The same is true for pressure, height, and temperature trends. The trends above the Antarctic are strongest compared to the other parts of the earth. The mean height increases by 80 m dayr−1 and the mean temperature decreases by 1.2 K dayr−1.

Hoinka (1998b) showed that the sign of the pressure trends could vary depending on the latitude and longitude. The global field of the trends for tropopause height (Fig. 10a) shows the same structure as that determined for the tropopause pressure. One has to keep in mind that the reliability of trends determined from reanalyses data is good in regions with a dense coverage of radiosonde data data whereas in regions with poor coverage of data there is a considerable degree of uncertainty (Stendel et al. 1998). Above the Atlantic storm track region the trend in the tropopause height is more than 300 m dayr−1. Figure 10b gives the global fields of the trends for tropopause temperature. In the midlatitudes and the Tropics the trends are weaker than in the polar regions. There is a negative trend of less than −1 K dayr−1 above the Atlantic storm track region and above the Tibetian Plateau. The Antarctic shows very strong negative trend values of less than −2.5 K dayr−1. These strong negative temperature trends appear in the same areas where strong negative tropopause pressure trends of less than −3 hPa dayr−1 were determined (Hoinka 1998b). Bojkov and Fioletov (1997) found trends in the temperature between −0.2 and +0.5 K dayr−1 above Canada at tropopause level (≈11 km) whereas the trends were around −0.5 K dayr−1 at the same height above Europe. This is confirmed by the regional temperature trends shown in Fig. 10b. Depending on the location the significance of the trends given in Fig. 10 is between 70% and 99%.

Finally, Figure 10c exhibits that most trends in mixing ratio are negative. There are positive trends above several tropical areas and above the North American continent. In various limited regions the trends in mixing ratio are close to −30% dayr−1 of the mean value: west of South America, above the tropical Atlantic Ocean, and above Australia. Above the latter region there is a decadal decrease of 20% in mixing ratio at the tropopause; for the lower stratosphere (15–21 km) a much stronger decrease from 4.8 to 3.7 ppmv dayr−1 was reported between 1970 and 1980 by Hyson (1983). In general Fig. 10c shows positive trends above the North American continent and above small limited areas, all concentrated in the Tropics. A reason for this structure cannot be given here. Oltmans and Hofmann (1995) determined trends of mixing ratio of water vapor above Boulder, Colorado, based on balloon-borne hygrometers that were launched between 1981 and 1994. They reported a mean value of 11.9 ppmv dayr−1 for the layer 12–14 km, which is close to the value obtained from the ERA data (Fig. 5). The decadal trend derived from balloon data is about 5%, which is equivalent to 0.6 ppmv. The ERA data (Fig. 10c) also show positive trends above the North American continent; for the Colorado region results have a trend value of about 5% dayr−1.

6. Summary

This study presents statistics of the height, temperature, wind, and humidity of the global tropopause. For the first time, a comprehensive description of the climatology of the global tropopause is presented based on 15 years of analysis data. The comparison between radiosonde climatologies and ECMWF reanalysis data corroborates the mean annual and seasonal global structure of wind and temperature. The annual mean and seasonal global fields of water vapor mixing ratio and relative humidity derived by the ERA data compare well with the data derived from satellite.

The following statistical features of parameters at the global tropopause are presented in this paper.

  • The meridional and zonal variations in the tropopause height and temperature reflect the global jet stream structure.

  • The seasonal variation of tropical tropopause temperatures is characterized by a minimum during Northern Hemispheric summer and a winter maximum.

  • The global mean tropopause height is positively correlated with the global mean tropospheric temperature.

  • The tropopause mixing ratios of water vapor has a minimum in the Tropics that increases gradually poleward.

  • The relative humidity has its maximum in the Tropics above the Pacific and Indian Oceans.

  • There are no global trends in the height and temperature of the tropopause.

  • There is a global mean decrease of 10% dayr−1 in the water vapor mixing ratio at the tropopause.

  • The negative trends in the mixing ratio of water vapor are very strong above the Arctic and the Antarctic.

The comparison between the data at the tropopause and those at fixed pressure levels suggests that it is possible to use fixed pressure level data as a measure for tropopause data, although only within limited geographic areas. The height of the 100-hPa level is a good estimate for the tropopause height in the Tropics, the 200-hPa level is appropriate in the midlatitudes, and the 300-hPa level is a good estimate for both polar regions. This is equivalent to the tropopause pressure data. The temperature at 100 hPa represents a good approximation to the global tropopause temperature. Like the zonal wind, the meridional wind at 200 hPa is a good measure of the tropopause wind except above the polar regions where this approach is not advisable. For the vertical wind there are neither statistics from the 100-hPa or 200-hPa levels nor observational data. Therefore, a meaningful suggestion cannot be made.

Acknowledgments

The data were kindly provided by the ECMWF within a “special project” under the title “The Climatology of the Global Tropopause.” Fruitful discussions with Manolo Castro (Complutense University, Madrid) on technical issues were much appreciated. I thank John Christy (University of Alabama) for providing the MSU data shown in Fig. 9.

REFERENCES

  • Bell, G. D., and M. S. Halpert, 1998: Climate assessment for 1997. Bull. Amer. Meteor. Soc.,79, S1–S50.

  • Bojkov, R. D., and V. E. Fioletov, 1997: Changes of the lower stratospheric ozone over Europe and Canada. J. Geophys. Res.,102, 1337–1347.

  • Elliott, W. P., and D. J. Gaffen, 1991: On the utility of radiosonde humidity archives for climate studies. Bull. Amer. Meteor. Soc.,72, 1507–1520.

  • Gibson, R., P. Kållberg, S. Uppala, A. Hernandez, A Nomura, and E. Serrano, 1997: ERA description. ECMWF-ReAnalysis Project Rep. Series, Vol. 1, 72 pp.

  • Halpert, M. S., and G. D. Bell, 1997: Climate assessment for 1996. Bull. Amer. Meteor. Soc.,78, S1–S49.

  • Hoerling, M. P., T. D. Schaack, and A. J. Lenzen, 1991: Global objective tropopause analysis. Mon. Wea. Rev.,119, 1816–1831.

  • Hoinka, K. P., 1998a: Mean global surface pressure series evaluated from ECMWF reanalysis data. Quart. J. Roy. Meteor. Soc.,124, 2291–2297.

  • ——, 1998b: Statistics of the global tropopause pressure. Mon. Wea. Rev.,126, 3303–3325.

  • ——, H. Claude, and U. Köhler, 1996: On the correlation between tropopause pressure and ozone above central Europe. Geophys. Res. Lett.,23, 1753–1756.

  • Hyson, P., 1983: Stratospheric water vapor over Australia. Quart. J. Roy. Meteor. Soc.,109, 285–294.

  • Jackman, C. H., E. L. Fleming, S. Chandra, D. B. Considine, and J. E. Rosenfield, 1996: Past, present, and future modeled ozone trends with comparisons to observed trends. J. Geophys. Res.,101, 28 753–28 767.

  • Jackson, D. R., S. J. Driscoll, E. J. Highwood, J. E. Harries, and J. M. Russell III, 1998: Troposphere to stratosphere transport at low latitudes as studied using HALOE observations of water vapor 1992–1997. Quart. J. Roy. Meteor. Soc.,124, 169–192.

  • Jenkinson, A. F., 1955: Average vector wind distribution of the upper air in temperate and tropical latitudes. Meteor. Mag.,84, 140–147.

  • Kållberg, P., 1998: Aspects of the re-analysed climate. ERA description. ECMWF-ReAnalysis Project Report Series, Vol. 2, 89 pp.

  • Kelly, K. K., A. F. Tuck, and T. Davies, 1991: Wintertime asymmetry of upper tropospheric water between the Northern and Southern Hemispheres. Nature,353, 244–247.

  • ——, M. H. Proffitt, K. R. Chan, M. Loewenstein, J. R. Podolske, S. E. Strahan, J. C. Wilson, and D. Kley, 1993: Water vapor and cloud water measurements over Darwin during the STEP 1987 tropical mission. J. Geophys. Res.,98, 8713–8723.

  • Kritz, M. A., S. W. Rosner, K. K. Kelly, M. Loewenstein, and K. R. Chan, 1993: Radon measurements in the lower tropical stratosphere: Evidence for rapid vertical transport and dehydration of tropospheric air. J. Geophys. Res.,98, 8725–8736.

  • Mote, P. W., K. H. Rosenlof, J. R. Holton, R. S. Harwood, and J. W. Waters, 1995: Seasonal variations of water vapor in the tropical lower stratosphere. Geophys. Res. Lett.,22, 1093–1096.

  • Murphy, D. M., K. K. Kelly, A. F. Tuck, and M. H. Proffitt, 1990: Ice saturation at the tropopause observed from the ER-2 aircraft. Geophys. Res. Lett.,17, 353–356.

  • Nagurny, A. P., 1998: Climatic characteristics of the tropopause over the Arctic Basin. Ann. Geophys.,16, 110–115.

  • Namias, J., and P. F. Clapp, 1951: Observational studies of general circulation patterns. Compendium of Meteorology, T. F. Malone, Ed., Amer. Meteor. Soc., 551–567.

  • Newell, R. E., and S. Gould-Stewart, 1981: A stratospheric fountain? J. Atmos. Sci.,38, 2789–2796.

  • ——, J. W. Kidson, and G. J. Boer, 1969: Annual and biennial modulations in the tropical Hadley circulation. Nature,222, 76–78.

  • ——, ——, D. G. Vincent, and G. J. Boer, 1972: The General Circulation of the Tropical Atmosphere. Vol. I. The MIT Press, 258 pp.

  • Oltmans, S. J., and D. J. Hofmann, 1995: Increase in lower-stratospheric water vapour at a mid-latitude Northern Hemisphere site from 1981 to 1994. Nature,374, 146–149.

  • Oort, A. H., 1983: Global Atmospheric Circulation Statistics 1958–1973. NOAA Prof. Paper 14, 180 pp + 47 microfiches. [Available from U.S. Government Printing Office, Washington, DC 20402.].

  • Peixoto, J. P., and A. H. Oort, 1992: Physics of Climate. American Institute of Physics, 520 pp.

  • ——, and ——, 1996: The climatology of relative humidity in the atmosphere. J. Climate,9, 3445–3463.

  • Phillips, P. D., H. Richner, J. Joss, and A. Ohmura, 1980: ASOND-78: An intercomparison of Väisälä, VIZ and Swiss radiosondes. Pure Appl. Geophys.,119, 259–277.

  • Pielke, R. A., J. Eastman, T. N. Chase, J. Knaff, and T. G. F. Knittel, 1998: 1973–1996: Trends in depth-averaged tropospheric temperature. J. Geophys. Res.,103, 28 909–28 911.

  • Pruvost, P., J. Ovarlez, J. Lenoble, and W. P. Chu, 1993: Comparison of Stratospheric Aerosol and Gas Experiment II and balloon-borne stratospheric water vapor measurements. J. Geophys. Res.,98, 4889–4896.

  • Read, W. G., J. W. Waters, D. A. Flower, L. Froidevaux, R. F. Jarnot, D. L. Hartmann, R. S. Harwood, and R. B. Rood, 1995: Upper-tropospheric water vapor from UARS MLS. Bull. Amer. Meteor. Soc.,76, 2381–2389.

  • Reid, G. C., and K. S. Gage, 1981: On the annual variation in height of the tropical tropopause. J. Atmos. Sci.,38, 1928–1938.

  • Reiter, E. R., 1961: Meteorologie der Strahlströme. Springer Verlag, 473 pp.

  • Rind, D., E.-W. Chiou, W. Chu, S. Oltmans, J. Lerner, J. Larsen, M. P. McCormick, and L. McMaster, 1993: Overview of the Stratospheric Aerosol and Gas Experiment II water vapor observations:Method, validation and data characteristics. J. Geophys. Res.,98, 4835–4856.

  • Salathé, E. P., and R. B. Smith, 1992: In situ observations of temperature microstructure above and below the tropopause. J. Atmos. Sci.,49, 2032–2036.

  • Schumann, U., 1994: On the effect of emmissions from aircraft engines on the state of the atmosphere. Ann. Geophys.,12, 365–384.

  • Sneyers, R., 1990: On the statistical analysis of series of observations. WMO Publ. 415, Tech. Note 143, 192 pp.

  • Soden, B. J., and F. P. Bretherton, 1993: Upper tropospheric relative humidity from the GOES 6.7 μm channel: Method and climatology for July 1987. J. Geophys. Res.,98, 16 669–16 688.

  • ——, and ——, 1996: Interpretation of TOVS water vapor radiances in terms of layer-average relative humidities: Method and climatology for the upper, middle, and lower troposphere. J. Geophys. Res.,101, 9333–9343.

  • Spencer, R. W., and J. R. Christy, 1992: Precision and radiosonde validation of satellite gridpoint temperature and anomalies. Part II: A tropospheric retrieval and trends during 1979–90. J. Climate, 5, 858–866.

  • Stendel, M., J. R. Christy, and L. Bengtsson, 1998: How representative are recent temperature trends? Rep. 264, Max-Planck-Institut für Meteorologie, 30 pp. [Available from Max-Planck-Institut für Meteorologie, Bundesstrasse 55, D-20146 Hamburg, Germany.].

  • WMO, 1957: Definition of the tropopause. WMO Bull.,6, 136.

  • ——, 1986: Atmospheric ozone 1985: Global ozone research and monitoring report. Rep. 16, World Meteorological Organization, 392 pp.

Fig. 1.
Fig. 1.

Mean height z (km) of the global tropopause (1979–93; 1200 UTC): annual mean [(a), all], northern summer [(b), JJA], and northern winter [(c), DJF]. The height increment is 500 m.

Citation: Monthly Weather Review 127, 10; 10.1175/1520-0493(1999)127<2248:THAWAT>2.0.CO;2

Fig. 2.
Fig. 2.

Mean temperature T (K) at the global tropopause (1979–93; 1200 UTC): annual mean [(a), all], northern summer [(b), JJA], and northern winter [(c), DJF]. The increment is 2 K.

Citation: Monthly Weather Review 127, 10; 10.1175/1520-0493(1999)127<2248:THAWAT>2.0.CO;2

Fig. 3.
Fig. 3.

Mean potential temperature θ (K) at the global tropopause (1979–93; 1200 UTC): annual mean [(a), all], northern summer [(b), JJA], and northern winter [(c), DJF]. The increment is 4 K.

Citation: Monthly Weather Review 127, 10; 10.1175/1520-0493(1999)127<2248:THAWAT>2.0.CO;2

Fig. 4.
Fig. 4.

Mean annual zonal wind u in m s−1 (a), meridional wind v in m s−1 (b), and vertical wind w in 10−1 mm s−1 (c) at the global tropopause (1979–93; 1200 UTC). The increments are 5 m s−1 (u), 2 m s−1 (υ), and 0.2 mm s−1 (w). Areas with negative velocities are shaded.

Citation: Monthly Weather Review 127, 10; 10.1175/1520-0493(1999)127<2248:THAWAT>2.0.CO;2

Fig. 5.
Fig. 5.

Mean water vapor mixing ratio m (ppmv) at the global tropopause (1979–93; 1200 UTC): annual mean [(a), all], northern summer [(b), JJA], and northern winter [(c), DJF]. The increment is 2.5 ppmv. Shading in the mixing ratio field indicates 0–5 (none), 5–10 (weak), 10–15 (moderate), 15–20 (strong), greater than 20 ppmv (very strong).

Citation: Monthly Weather Review 127, 10; 10.1175/1520-0493(1999)127<2248:THAWAT>2.0.CO;2

Fig. 6.
Fig. 6.

Vertical profiles of mixing ratio of water vapor of ERA data (1200 UTC), of aircraft data taken during the STEP experiment (Kritz et al. 1993), and of satellite data taken during SAGE II (Rind et al. 1993): (a) STEP flight 22 Jan 1987; (b) STEP flight 23 Jan 1987; (c) STEP flight 18 Feb 1987; (d) SAGE data, 1613 UTC; ϕ = 16°S, λ = 157°W; radiosonde Samoa (RASO), 1735 UTC, ϕ = 14.2°S, λ = 170.3°W; ERA data, ϕ = 13.5°S, λ = 171°W. The horizontal lines indicate the corresponding ERA tropopause of 3.5 PVU.

Citation: Monthly Weather Review 127, 10; 10.1175/1520-0493(1999)127<2248:THAWAT>2.0.CO;2

Fig. 7.
Fig. 7.

Monthly mean upper-tropospheric mean relative humidity (rh) (%) for Jul 1987 derived from satellite data [(a) adapted from Soden and Bretherton (1993)] and from ERA data (b). The mean relative humidity is determined for the layer between 200 and 500 hPa. The increment in relative humidity is 10%.

Citation: Monthly Weather Review 127, 10; 10.1175/1520-0493(1999)127<2248:THAWAT>2.0.CO;2

Fig. 8.
Fig. 8.

Mean annual rh (%) at the global tropopause (1979–93; 1200 UTC). The increment is 5%. Shading in the relative humidity field indicates 0%–10% (none), 10%–20% (weak), 20%–30% (moderate), 30%–40% (strong), and greater than 40% (very strong).

Citation: Monthly Weather Review 127, 10; 10.1175/1520-0493(1999)127<2248:THAWAT>2.0.CO;2

Fig. 9.
Fig. 9.

Monthly mean global tropopause height (dynamical with 3.5 PVU) anomalies from ERA data (full line) and monthly mean tropospheric temperature derived from MSU channel 2R (broken line). A nine-point running mean filter has been applied to both time series.

Citation: Monthly Weather Review 127, 10; 10.1175/1520-0493(1999)127<2248:THAWAT>2.0.CO;2

Fig. 10.
Fig. 10.

Trends (△) of tropopause parameters determined for the period 1979–93: height in m per decade (a), temperature in K per decade (b), and mixing ratio of water vapor in % per decade (c). Increments are 100 m (height), 0.5 K (temperature), and 10% (mixing ratio). Positive trends are shaded.

Citation: Monthly Weather Review 127, 10; 10.1175/1520-0493(1999)127<2248:THAWAT>2.0.CO;2

Table 1.

Trends in tropopause parameters between 1979 and 1993 and the corresponding trend-to-noise ratios (t/n) derived from a time series of monthly mean values. The ratios of 1.0, 2.0, and 3.0 point to confidence levels of 80%, 95% and 99%, respectively. The midlatitudes and Tropics are limited by 65°S and 65°N, the southern end of the Arctic is a 66.5°N, and the northern end of the Antarctic at 66.5°S. The asterisk indicates that the magnitude of the value is less than 0.1 but not zero.

Table 1.

Save