Gravity wave climatology and trends in the mesosphere/lower thermosphere region deduced from low-frequency drift measurements 1984–2003 (52.1°N, 13.2°E)

Abstract

Fluctuations of lower ionospheric drifts in the virtual height range 85–110 km at 52.1°N, 13.2°E during 1984–2003 are presented. These fluctuations may be partly owing to neutral atmosphere gravity waves in the period window 0.7–3 h. The results show maximum wave activity in the mesosphere in summer, with a shift to equinoxes at higher altitudes. Maximum gravity wave amplitudes are found near the regions of strongest vertical mean wind shear. The propagation direction is generally close to E–W, however, during winter at lower heights a more South-Easterly direction is preferred, while at greater heights during summer a North-Easterly direction is visible. Time series of seasonal (3-monthly) mean zonal drift variances show maximum amplitudes around years 1989–1991 and 2000–2002, when maxima of solar activity within the 11-year solar cycle occurred. We interpret the results in terms of gravity wave activity although additional influences like ionospheric perturbations also may lead to possible wind fluctuations correlated with solar activity. Therefore, the conclusions are partly qualitative and require further experimental verification.

Introduction

Internal gravity waves (GW) are the important process that controls the dynamics and temperature field of the mesosphere and lower thermosphere (MLT). GW are filtered by the mean circulation in the stratosphere and mesosphere, and accelerate the mean flow in the mesosphere, which leads to the lower thermospheric zonal wind reversal, meridional circulation and low summer MLT temperatures that cannot be explained without taking into account the effect of breaking GW. Successful modelling of MLT dynamics therefore requires knowledge of GW fluxes and their seasonal and vertical distribution. While some first global climatologies of stratospheric GW activity are available from satellite measurements (Tsuda et al., 2000; McLandress et al., 2000; Venkat Ratnam et al., 2004), analyses of GW activity in the MLT essentially relies on ground-based measurements so that basically local climatologies are available. Fewer studies involve regional or global variability (Manson et al., 2002, Manson et al., 2004). The relative sparseness of multi-instrument studies is, among others, due to peculiarities of the different measuring systems (radars of different wavelength, optical) that allow the detection of GW in different spectral ranges or in different parameters.

Nevertheless, comparatively many publications have dealt with the seasonal variations of GW as obtained from radar measurements (Gavrilov et al., 1995, Gavrilov et al., 2001, Gavrilov et al., 2002; Manson et al., 1997, Manson et al., 1999, Manson et al., 2003; Gavrilov and Jacobi, 2004), and also with the comparison of GW activity in the course of the year over different stations. These works may provide an average climatology of GW although the spectral ranges considered depend on the measuring principle used. For example, medium frequency (MF) radars are able to measure winds with a time resolution of few minutes, while the low-frequency (LF) method only provides lower ionospheric E region drift variances in a period window between 0.7 and 3 h (Gavrilov et al., 2001). Therefore, if a global picture is constructed from the individual radar measurements, the result may be somewhat qualitative, which is also the case with the interpretation of interannual variability. At midlatitudes, in the mesosphere a semiannual oscillation (SAO) of GW activity with maxima in winter and summer is usually found, with the summer maximum being more pronounced for longer GW periods (Gavrilov et al., 2002). The SAO may be explained by the middle atmosphere jets that lead to large intrinsic phase speeds and therefore larger wind variances (Manson et al., 1997). One of the most powerful methods of detecting GW in the mesosphere is using MF radar measurements, owing to their high temporal resolution and long-term reliability with only a few data gaps. However, MF radars usually measure up to altitudes around 95 km, so that above that level measurements are sparse. Analyses of LF drifts have shown that near 100 km the SAO phase shifts with GW amplitude maxima near solstices (Gavrilov et al., 2001).

GW measured at MLT heights are on the one hand dependant on their sources, which are mainly situated in the troposphere (Nastrom and Fritts, 1992a, Nastrom and Fritts, 1992b), although, e.g. Gavrilov and Jacobi (2004) have shown that middle atmosphere GW sources may also play an important role in establishing the observed seasonal distribution. On the other hand in particular GW with small phase speeds are sensitive to wind filtering in the middle atmosphere (Taylor et al., 1993). This is especially true for middle and high latitudes, where the stratospheric and mesospheric zonal winds are strong, while the low-latitude MLT GW activity seems to be more strongly controlled by the GW sources in the lower atmosphere (Manson et al., 1999). Nevertheless, even for midlatitudes especially considering long-period GW their MLT activity and (spatial and temporal) variability is resulting from a mixture of source distribution and middle atmosphere filtering.

One method of revealing the relative influence of GW sources and mean wind filtering on the MLT wind variance distribution is analysing long-term variability of GW and comparison with atmospheric mean circulation or characteristic circulation patterns at different heights. Long-term analyses of GW activity in the MLT, however, are sparse. Gavrilov et al. (1995) analysed 15 years of midlatitude MF radar data. Gavrilov et al. (1999) analysed 12 years of data measured over Japan, while low-latitude measurements over Hawaii (11 years) have been presented by Gavrilov et al. (2004). Major focus of these studies is the presentation of the seasonal and vertical variation of GW over the respective stations. Only few studies explicitly included the analysis of interannual variability on time scales of few years, in particular since the time series available are short, so that conclusions are preliminary. A possible 11-year solar cycle effect was mentioned by Gavrilov et al. (1995), however, comparison of three midlatitude stations by Gavrilov et al. (2002) showed that there are substantial differences between longitudes.

Notwithstanding the great number of publications showing GW climatologies and comparisons between stations, one of the major results is the strong variability of GW activity and parameters as propagation direction also across relatively short distances, which is assumed to be connected either with different wind filtering or local sources (Manson et al., 2003, Manson et al., 2004). Therefore, further long-term measurements are required. Here we present GW analyses using a 20-year database measured at Collm, Germany, using the LF method. The dataset represents an update of that one presented by Gavrilov et al. (2001) that was also used in the comparison study by Gavrilov et al. (2002) and that has been compared to model results by Gavrilov and Jacobi (2004). Including roughly another half of a solar cycle enables us to draw more substantial conclusions on interannual variability of GW over Central Europe.