Assessment of middle atmosphere climatology using lidar for aerospace applications

• Evaluation of ERA-5 and NRLMSISE 2.0 biases in the middle atmosphere ; • Creation of a Steady-State Climatology in multiple sites removing extreme atmospheric events ; • Evaluation of the impact of extreme events on the density profile. This study evaluates temperature and density variations in the upper stratosphere and in the mesosphere using ERA-5 and NRLMSISE-2.0 models, by comparing them with our forty-years long temperature lidar observations. NRLMSISE-2.0, while reliable in terms of seasonal trends, is less accurate daily: it shows high biases with the lidar at high altitudes (globally less than 4 K below 50 km, up to 14 K at 80km). The European Climate and Weather Forecast reanalysis model ERA-5 shows a bias to the lidar of -5 K the upper stratosphere but shows a larger difference reaching +20 K in the mesosphere. Lidar observations can reveal non-Gaussian distributions because of extreme events, such as Mesospheric Inversion Layers (MILs), minor and Major Sudden Stratospheric Warmings (m- and M-SSWs) or Double Stratopauses (DSs). These events are atmospheric particularities which should not be mixed with steady-state profiles to compute a mean climatology if the aim is to create a reference. To address this issue, a new methodology was developed to remove those events, resulting in what we called ”Steady-State” climatology. Our method based on lidar observation in four sites (Observatory of Haute-Provence, Reunion Island, Table Mountain and Mauna Loa) allowed us to give statistical information about these extreme phenomena. Furthermore, our findings show that when m-SSWs and MILs occur, the density profile deviates strongly from the NRLMISE-2.0 density, which highlights the need for refined atmospheric models for aerospace applications. Corresponding density differences between NRLMSISE and lidar can reach -15% and +25% at most for MILs and m-SSWs, respectively.