• GPS/MET Experiment
• Determination of potential energy (Ep) from GPS/MET profiles
• Vertical wavenumber spectrum of temperature perturbations
• Summary
• Acknowledgements
• References

FIGURE

• Figure 1

Using the temperature profiles obtained with the GPS occultation technique between April 1995 and February 1997, we have studied meso-scale temperature fluctuations in the stratosphere. In particular, we have analyzed the global distribution of potential energy (Ep) caused by atmospheric gravity waves, and vertical wavenumber spectrum of the normalized temperature perturbations.

1. GPS/MET experiment

The GPS/MET (Global Positioning System/Meteorology) experiment has been initiated by the University Corporation for Atmospheric Research (UCAR), successfully providing the international scientific community with a new global high resolution data set of temperature, pressure and refractivity profiles in the 1-60 km height range. These profiles are obtained from the active limb sounding occultation technique as described by Ware et al. [1996] and Rocken et al. [1997]. In this experiment, a GPS receiver aboard Microlab-1 was launched on April 3, 1995, into a low earth orbit (LEO) to observe occulted radio signals from the GPS satellites. Height profiles of atmospheric refractive index were derived from these observations. By assuming the hydrostatic relation for a dry atmosphere, temperature profiles can further be inferred. Rocken et al. [1997] showed 1 K mean temperature agreement with the best correlative data between 5-40 km. The GPS/MET profiles are found to be very unique in revealing detailed temperature structure, including sharp inversions and step-wise increase of temperature gradient near the tropical tropopause, which has not been achieved by a conventional satellite measurement [Nishida et al., 2000].

In this study we have analyzed meso-scale temperature perturbations caused by gravity waves using all published GPS/MET temperature profiles between April 1995 and February 1997, available at the GPS/MET office of UCAR. Then, we have defined the potential energy Ep, from the temperature variance, and studied the energy distribution of gravity waves in the stratosphere as a function of latitude, longitude, season and altitude. We have also retrieved temperature profiles with a better height resolution (about 200 m) from Level-2 GPS/MET data collected during prime times, and analyzed vertical wavenumber spectrum of the normalized temperature fluctuations.

2. Determination of potential energy (Ep) from GPS/MET profiles

We have extracted fluctuating components of temperature, T’, from a single GPS/MET profile by applying a high-pass filter with a cutoff at 10 km. We also calculated N² by differentiating the temperature profile for adjacent three heights. As the GPS/MET profiles are already low-pass filtered in order to reduce noise, where the cut-off of the filter is tuned to pass phase variations of the GPS signal corresponding to vertical scale of 2 to 3 km in the stratosphere and approximately 200 m in the lower troposphere [Rocken et al., 1997]. Accordingly, T’ in the stratosphere consists of temperature fluctuations having a vertical scale between 2-3 km and 10 km. Then, we have calculated the temperature variance, and estimated the potential energy Ep=1/2(g/N)²(T’/T)² by using the observed N². We have noticed that the enhanced amplitudes of T’ above about 45 km are unrealistically large. Therefore, we have restricted the maximum height range of the analysis below about 45 km.

2.1 Comparison of seasonal variations of Ep and E_k from GPS/MET and ST Radar Data

Using GPS/MET data, we have determined the monthly mean values of Ep in a height range of 15-20 km around the MU radar in Shigaraki, Japan (34.9^oN, 136.0^oE), and the ST radar at the White Sands Missile Range, New Mexico (32.4^oN, 106.4^oW) [Tsuda et al., 2000; Nastrom et al., 2000]. Then, we have compared the seasonal variations of Ep with the climatological behavior of the kinetic energy E_k due to gravity waves obtained from long-term observations with MU and WS in 1985-1989 and 1991-1996, respectively. We have found a good consistency of the seasonal variations between Ep and E_k with an enhancement in winter months.

It is noteworthy that the ratio of Ep to E_k agrees reasonably well with a theoretical prediction assuming linear gravity waves, although the difference in the vertical and horizontal resolutions between GPS/MET and the MU radar measurements should be investigated in more detail. These comparisons imply that the utilization of the GPS/MET profiles in the study of stratospheric gravity wave activity has been verified at these specified radar sites.

2.2 Global distribution of Ep at 20-30 km in November-February.

We have analyzed distributions of Ep as functions of latitude, longitude and seasons between 15 and 45 km. We have separated a year into four seasons, that is, March to April, May to August, September to October, and November to February. Then, we have calculated Ep from the individual GPS/MET temperature profiles, collected during April 1995 and February 1997.

First, we have determined the latitude-longitude distribution of Ep in the 20-30 km height region in northern hemisphere winter months (from November to February), when the largest number of GPS/MET data was collected (a total of 4,569 data out of 10,853). (see Plate 1 of Tsuda et al. [2000]; or Fig. 1 on page VIII of SPARC Newsletter No. 13, 1999). Note that only few data points were obtained around tropical Africa, and a larger amount of data was collected at middle and high latitudes due to the configuration of the satellite orbits and the distribution of the ground reference sites. Large Ep values (exceeding 6 J/kg) are mostly detected at low latitudes from 25^oN to 25^oS centered around the equator, and they are particularly enhanced over the regions where active convection is expected, such as over the Indonesian archipelago, the Indian ocean, Eastern Atlantic Ocean, and South America. This result strongly suggests that atmospheric waves are actively generated by tropical convection. The Ep values at midlatitudes (higher than 30^o) are generally larger in the northern (winter) hemisphere. However, the Ep values at low latitudes could partially be contributed by equatorial waves that add a zonally symmetric bias.

We have also analyzed longitudinal distributions of Ep at 20-30 km during November and February. Since we are interested in the differences in Ep between continents and oceans, considering the orographic effects, we have focused on a latitude band from 30^oN to 60^oN where the topography can be classified into continents (65^o-130^oW and 0^o-145^oE) and oceans (0^o-65^oW and 130^oW-145^oE). There is a tendency for Ep to be larger over continents than oceans. In particular, Ep was enhanced over North America, while it was depressed in the central Pacific. These land-sea contrasts in Ep could be attributed to the effects of mountain lee waves caused by an interaction between topography and surface winds [Tsuda et al., 2000].

2.3 Latitude distribution of Ep at 15-45 km

We have selected a height range between 15 km and 45 km, and defined five overlapping height regions with a 10 km thickness starting from 15 km and an increment of 5 km. It is noteworthy that the noise on the L2 band of GPS signals considerably increased when anti-spoofing (A/S) encryption of the GPS signals was turned on [Rocken et al., 1997]. Then, effects of ionospheric correction produce artificial fluctuations in the GPS/MET temperature profiles, which could be recognized down to about 30 km. Therefore, we use here only data during prime times under A/S-off conditions for a study of the height variations of Ep distributions. Fig. 1 shows latitudinal variations of Ep in the five altitude ranges, where individual Ep values are averaged every 10^o in 18 latitude bands [Tsuda et al., 2000]. The latitudinal distribution of the GPS occultation events, as shown in a histogram in Fig. 1, is nearly symmetric relative to the equator, with the maximum around 30^o in each hemisphere. However, it was significantly reduced at latitudes higher than 70^o. Therefore, the analyzed results in the polar regions may not be statistically significant.

In the 20-30 km layer in Fig. 1, we have detected the largest value of Ep in the low latitudes, where the latitude range of the enhanced Ep is wider in November-February than in other seasons. Moreover, the Ep values at midlatitudes are larger in the winter hemisphere, which is more evident in May-August, while in equinox (September-October) Ep in the midlatitudes was nearly the same between the northern and southern hemispheres. In the overlying altitude regions, the enhanced peak of Ep near the equator tends to disappear, but the Ep at higher latitudes becomes larger.

It is remarkable that in September-October the latitudinal distribution of Ep is symmetric between the northern and southern hemispheres in the entire height ranges. However, near solstices the Ep distribution involves a large hemispheric asymmetry at middle and high latitudes, which is more pronounced in May-August. There are also differences in the Ep distribution between the June and December solstices. For example, Ep values at 40^o-70^oS in May-August are significantly larger than those at 40^o-70^oN in November-February. Conversely, Ep values at 40^o-70^oS in November-February are smaller than those at 40^o-70^oN in May-August. These results suggest that the latitudinal distribution of the gravity wave activity in the stratosphere depends on hemisphere as well as on season.

Figure 1 Latitudinal variations of the zonally averaged Ep and standard deviation at 15-45 km in three seasons. Number of GPS/MET profiles is shown in a histogram, whose total is 1608, 1430 and 1836, respectively. Results in March-April are not shown here, because number of the GPS/MET profiles under A/S-off conditions was insufficient (430).

2.4 Height variations of Ep

We have determined height variations of Ep in the equatorial region (10^oS-10^oN) and at midlatitudes (40^o-60^o). The Ep values decrease between 20 km and 25-30 km altitude, which is more evident at low latitudes, and show a gradual and rapid increase at 25-35 km and above 35 km, respectively. A strong annual variation with maximum in winter is detected in both northern and southern hemispheres, although large hemispheric differences are recognized. On the other hand, seasonal variations are not evident in the equatorial region.

3. Vertical wavenumber spectrum of temperature perturbations

Temperature profiles are re-analyzed by using a shorter data smoothing range (about 200 m) with an improved retrieval algorithm [Hocke, 1997]. Using 83 GPS/MET profiles collected on Feruary 2-10, 1997 in a latitude band between 20o N and 20oS, we have calculated vertical wavenumber spectrum of the normalized temperature fluctuations, and compared them with the radiosonde result determined from 78 profiles over Pontianak (0o, 109oE), Indonesia, in January-February, 1997. The observed spectra are also compared with a model spectrum assuming a linear saturation of gravity waves, N4/(10g2m3), where m is the vertical wavenumber. We have found that the GPS/MET profiles seem to resolve gravity waves with a vertical scale as small as about 400 m at 20-30 km altitude. However, above 30 km, the spectral density from GPS/MET profiles could be suppressed because the data retrieval algorithm artificially reduces large temperature deviation from a standard model atmosphere. In particular, the observed spectral density at 40-50 km obviously exceeds a theoretical model probably due to spurious temperature fluctuations caused by ionosphereic correction and receiver noise [Hocke, 1997].

We have analyzed the spectra at 20-30 km and 30-40 km during three prime times under A/S-off condition in June/July 1995, October 1995 and February 1997. At 20-30 km the spectral slope for short wavelengths (>2 km) is about ?3 as predicted by a saturated gravity wave model. But, the spectral density is sometimes smaller than the model. At 30-40 km both spectral slope and density agree well with the model, especially for wavelengths shorter than about 1.5 km.

We have calculated the normalized temperature variance, (T’/T)2, integrating the spectra for long (10-2.5 km) and short (2-0.4 km) wavelength ranges. The seasonal and latitudinal variations of (T’/T)2 are evident. In particular, (T’/T)2 is largely enhanced near the equator for both long and short wavelength ranges at 20-30 km.

4. Summary

Profiles of atmospheric temperature in the upper troposphere and stratosphere have been obtained by means of radio occultation observations of GPS (Global Positioning System) signals by the GPS/MET (GPS Meteorology) experiment from April 1995 to February 1997.

• From each GPS/MET temperature profile we have extracted mesoscale temperature perturbations with a vertical scale ranging from 2 to 10 km, as well as the background N². Then, we have defined the potential energy, Ep, from the temperature variance.
• The seasonal variation of Ep determined with the GPS/MET data in the lower stratosphere is consistent with the climatological behavior of the kinetic energy of gravity waves, E_k, observed with ST radars in Japan and New Mexico.
• We have analyzed the latitude-longitude distribution of Ep at 20-30 km from November to February. Ep values are particularly enhanced over convectively active regions around the equator, which strongly suggests that gravity waves are generated by tropical convection.
• Using GPS/MET data without anti-spoofing, latitudinal variations of Ep were determined between 15 and 45 km altitude. The results indicate that large Ep values are concentrated near the equator at 20-30 km. But, Ep becomes larger at midlatitudes at 30-40 km and higher altitude regions. At midlatitudes, Ep was larger in winter months in each hemisphere, although differences also existed between northern and southern hemispheres. During equinoxes, a remarkable latitudinal symmetry of Ep relative to the equator was found.
• The longitudinal distribution of Ep at latitudes between 30^oN and 60^oN during November and February indicates an enhanced gravity wave activity over continents, possibly suggesting orographic generation of mountain lee waves.
• We have also determined height variations of Ep at 15-45 km. The Ep values decrease between 20 km and 25-30 km altitude, and show a gradual and rapid increase at 25-35 km and above 35 km, respectively.
• We have employed an improved retrieval software so as to obtain temperature profiles with a better height resolution (about 200 m) from Level 2 data of the GPS/MET experiment. We have analyzed vertical wavenumber spectrum of the normalized temperature fluctuations at 20-30 km, and found that the observed spectra agree with a saturated gravity wave model with a vertical scale as small as about 400 m.
• Latitude variations of the spectra are investigated at 20-30 km and 30-40 km by using GPS/MET data during a prime time in June-July, October and February. The spectral slope is nearly ?3, consistent with themodel. However, the spectral density at 20-30 km is slightly smaller than the model value.
• Temperature variance caluculated by integrating the spectra in wavelength ranges of 10-2.5 km and 2-0.4 km is largely enhanced in the equatorial region.

This study has clarified that the GPS occultation technique provides important and unique data for studies of the global distribution of atmospheric gravity waves.

5. Acknowledgements

The GPS/MET program is sponsored primarily by NSF, and temperature profiles derived from the GPS occultation measurements were provided by the UCAR GPS/MET office. This study is supported by the Japanese GPS-Meteorology project of Science and Technology Agency. We deeply thank D. Hunt and C. Rocken for providing us Level 2 data of the GPS/MET experiment. We wish to acknowledge K. Hocke, M. Nishida and M. Iwata for their help in analyzing the vertical wavenumber spectra.

6. References

Hocke, K., Inversion of GPS meteorology data, Annales Geophys., 15, 443-450, 1997.

Nastrom, G. D., A. R. Hansen, T. Tsuda, M. Nishida, and R. Ware, A comparison of gravity wave energy observed by VHF radar and GPS/MET over central North America, J. Geophys. Res., 105, 4685-4687, 2000.

Nishida, M., T. Tsuda, C. Rocken, and R. H. Ware, Seasonal and longitudinal variations in the tropical tropopause observed with the GPS occultation technique (GPS/MET), J. Meteorol. Soc. Japan, in print, 2000.

Rocken, C., R. Anthes, M. Exner, D. Hunt, S. Sokolovskiy, R. Ware, M. Gorbunov, W. Schreiner, D. Feng, B. Herman, Y.-H. Kuo, X. Zou, Analysis and validation of GPS/MET data in the neutral atmosphere, J. Geophys. Res., 102, 29849-29866, 1997.

Smith, S. A., D. C. Fritts, and T. E. VanZandt, Evidence of a saturation spectrum of atmospheric waves, J. Atmos. Sci., 44, 1,404-1,410, 1987.

Tsuda, T., M. Nishida, C. Rocken and R. H. Ware, A global morphology of gravity wave activity in the stratosphere revealed by the GPS occultation data (GPS/MET), J. Geophys. Res., 105, 7257-7273, 2000.

Tsuda, T., M. Nishida, C. Rocken and R. H. Ware, A global distribution of gravity wave activity in the stratosphere, SPARC Newsletter, No. 13, 28-30, 1999.

Ware, R., M. Exner, D. Feng, M. Gorbunov, K. Hardy, B. Herman, Y. Kuo, T. Meehan, W. Melbourne, C. Rocken, W. Schreiner, S. Sokolovskiy, F. Solheim, X. Zou, R. Anthes, S. Businger, and K. Trenberth, GPS sounding of the atmosphere From Low Earth Orbit: Preliminary Results, Bull. Am. Meteorol. Soc., 77, 19-40, 1996.