Abstract. This study evaluates commonly used methods of extracting gravity-wave-induced temperature perturbations from lidar measurements. The spectral response of these methods is characterized with the help of a synthetic data set with known temperature perturbations added to a realistic background temperature profile. The simulations are carried out with the background temperature being either constant or varying in time to evaluate the sensitivity to temperature perturbations not caused by gravity waves. The different methods are applied to lidar measurements over New Zealand, and the performance of the algorithms is evaluated. We find that the Butterworth filter performs best if gravity waves over a wide range of periods are to be extracted from lidar temperature measurements. The running mean method gives good results if only gravity waves with short periods are to be analyzed.

We analyze a large‐amplitude mountain wave event, which was observed by a ground‐based lidar above New Zealand between 31 July and 1 August 2014. Besides the lidar observations, European Centre for Medium‐Range Weather Forecasts (ECMWF) data, satellite observations, and ray tracing simulations are utilized in this study. It is found that the propagation of mountain waves into the middle atmosphere is influenced by two different processes at different stages of the event. At the beginning of the event, instabilities in a weak wind layer cause wave breaking in the lower stratosphere. During the course of the event the mountain waves propagate to higher altitudes and are refracted southward toward the polar night jet due to the strong meridional shear of the zonal wind. As the waves propagate out of the observational volume, the ground‐based lidar observes no mountain waves in the mesosphere. Ray tracing simulations indicate that the mountain waves propagated to mesospheric altitudes south of New Zealand where the polar night jet advected the waves eastward. These results underline the importance of considering horizontal propagation of gravity waves, e.g., when analyzing locally confined observations of gravity waves.

We present atmospheric gravity wave (GW) measurements obtained by a Rayleigh/Raman lidar at Lauder, New Zealand, (45°S, 170°E) during and after the Deep Propagating Gravity Wave Experiment campaign. GW activity and characteristics are derived from 557 h of high‐resolution lidar data recorded between June and November 2014 in an altitude range between 28 and 76 km. In this period, strong GW activity occurred in sporadic intervals lasting a few days. Enhanced stratospheric GW potential energy density is detected during periods with high tropospheric wind speeds perpendicular to New Zealand's Southern Alps. These enhancements are associated with the occurrence of quasi‐stationary GW (mountain waves). Surprisingly, the largest response in the mesosphere is observed for conditions with low to moderate lower tropospheric wind speeds (2–12 m/s). On the other hand, large‐amplitude mountain waves excited by strong tropospheric forcings often do not reach mesospheric altitudes, either due to wave breaking and dissipation in the stratosphere or refraction away from New Zealand.

On 4 July 2014, during the Deep Propagating Gravity Wave Experiment (DEEPWAVE), strong low‐level horizontal winds of up to 35 m s−1 over the Southern Alps, New Zealand, caused the excitation of gravity waves having the largest vertical energy fluxes of the whole campaign (38 W m−2). At the same time, large‐amplitude mesospheric gravity waves were detected by the Temperature Lidar for Middle Atmospheric Research (TELMA) located at Lauder (45.0°S, 169.7°E), New Zealand. The coincidence of these two events leads to the question of whether the mesospheric gravity waves were generated by the strong tropospheric forcing. To answer this, an extensive data set is analyzed, comprising TELMA, in situ aircraft measurements, radiosondes, wind lidar measurements aboard the DLR Falcon as well as Rayleigh lidar and advanced mesospheric temperature mapper measurements aboard the National Science Foundation/National Center for Atmospheric Research Gulfstream V. These measurements are further complemented by limited area simulations using a numerical weather prediction model. This unique data set confirms that strong tropospheric forcing can cause large‐amplitude gravity waves in the mesosphere, and that three essential ingredients are required to achieve this: first, nearly linear propagation across the tropopause; second, leakage through the stratospheric wind minimum; and third, amplification in the polar night jet. Stationary gravity waves were detected in all atmospheric layers up to the mesosphere with horizontal wavelengths between 20 and 100 km. The complete coverage of our data set from troposphere to mesosphere proved to be valuable to identify the processes involved in deep gravity wave propagation.

This paper describes the results of a comprehensive analysis of the atmospheric conditions during the Deep Propagating Gravity Wave Experiment (DEEPWAVE) campaign in austral winter 2014. Different datasets and diagnostics are combined to characterize the background atmosphere from the troposphere to the upper mesosphere. How weather regimes and the atmospheric state compare to climatological conditions is reported upon and how they relate to the airborne and ground-based gravity wave observations is also explored. Key results of this study are the dominance of tropospheric blocking situations and low-level southwesterly flows over New Zealand during June–August 2014. A varying tropopause inversion layer was found to be connected to varying vertical energy fluxes and is, therefore, an important feature with respect to wave reflection. The subtropical jet was frequently diverted south from its climatological position at 30°S and was most often involved in strong forcing events of mountain waves at the Southern Alps. The polar front jet was typically responsible for moderate and weak tropospheric forcing of mountain waves. The stratospheric planetary wave activity amplified in July leading to a displacement of the Antarctic polar vortex. This reduced the stratospheric wind minimum by about 10 m s−1 above New Zealand making breaking of large-amplitude gravity waves more likely. Satellite observations in the upper stratosphere revealed that orographic gravity wave variances for 2014 were largest in May–July (i.e., the period of the DEEPWAVE field phase).

Middle atmospheric lidar temperature observations conducted above Sodankylä, Finland (67.4°N, 26.6°E), during December 2015 are compared with two estimates of the atmospheric state computed by the integrated forecast system (IFS) of the European Centre for Medium‐Range Weather Forecasts (ECMWF). The first set corresponds to an hourly sampling of the middle atmosphere by high‐resolution analyses and very short‐range forecasts produced by the operational IFS cycle 41r1 at a horizontal resolution of 16 km. The second set is retrieved from the upgraded IFS cycle 41r2 (horizontal resolution 9 km), which was running in parallel with cycle 41r1 during the validation before it became operational. A remarkable agreement between both IFS datasets and the lidar temperature observations above Sodankylä is found below 45 km altitude. Above 45 km altitude, within the sponge layer of the IFS, both IFS datasets depict lower temperatures than the observations, with the 9 km runs showing the coldest temperatures. Various sensitivity experiments conducted with the IFS are analyzed and compared with the lidar observations to investigate the impact of the different changes implemented in IFS cycle 41r2. It is found that both the scientific changes and the horizontal resolution upgrade contribute to the colder mesosphere above Sodankylä. The data assimilation seems to amplify this effect even further.

Two Rayleigh lidars were employed at a southern-hemisphere mid-latitude site in New Zealand (45 • S) and a northern-hemisphere high-latitude site in Finland (67 • N) in order to observe gravity waves between 30 and 85 km altitude under wintertime conditions. Two-dimensional wavelet analysis is used to analyze temperature perturbations caused by gravity waves and to determine their vertical wavelengths and phase progression. In both datasets, upward phase progression waves occur frequently between 30 and 85 km altitude. Six cases of large-amplitude wave packets are selected which exhibit upward phase progression in the stratosphere and/or mesosphere. We argue that these wave packets propagate downward and we discuss possible wave generation mechanisms. Spectral analysis reveals that superpositions of two or three wave packets are common. Furthermore, their characteristics often match those of upward-propagating waves which are observed at the same time or earlier. In the dataset means, the contribution of upward Preprint submitted to JASTP March 8, 2017 phase progression waves to the potential energy density E p is largest in the lower stratosphere above Finland. There, E p of upward and downward phase progression waves is comparable. At 85 km one third of the potential energy carried by propagating waves is attributed to upward phase progression waves. In some cases E p of upward phase progression waves far exceeds E p of downward phase progression waves. The downward-propagating waves might be generated in situ in the middle atmosphere or arise from reflection of upward-propagating waves.

Two mountain wave events, which occurred over northern Scandinavia in December 2013 are analysed by means of airborne observations and global and mesoscale numerical simulations with horizontal mesh sizes of 16, 7.2, 2.4 and 0.8 km. During both events westerly cross-mountain flow induced upward-propagating mountain waves with different wave characteristics due to differing atmospheric background conditions. While wave breaking occurred at altitudes between 25 and 30 km during the first event due to weak stratospheric winds, waves propagated to altitudes above 30 km and interfacial waves formed in the troposphere at a stratospheric intrusion layer during the second event. Global and mesoscale simulations with 16 and 7.2 km grid sizes were not able to simulate the amplitudes and wavelengths of the mountain waves correctly due to unresolved mountain peaks. In simulations with 2.4 and 0.8 km horizontal resolution, mountain waves with horizontal wavelengths larger than 15 km were resolved, but exhibited too small amplitudes and too high energy and momentum fluxes. Simulated fluxes could be reduced by either increasing the vertical model grid resolution or by enhancing turbulent diffusion in the model, which is comparable to an improved representation of small-scale nonlinear wave effects.