Gravity-Wave-Driven Seasonal Variability of Temperature Differences between ECMWF IFS and Rayleigh Lidar Measurements in the Lee of the Southern Andes

Abstract

Even now with a growing understanding of stratospheric processes, highly developed numerical models, and increasing computational resources, middle atmosphere temperature (re)analyses have a larger uncertainty than their tropospheric counterparts. Improving the representation of the past (reanalysis), current (analysis), and future (forecast) state of the middle atmosphere in general circulation models (GCMs) is important for the validation and forecasting of tropospheric weather and future climate. It is known that the circulation in the middle and upper atmosphere is strongly influenced by internal gravity waves (GWs) triggered for example, by flow over mountains (Fritts & Alexander, 2003). At the same time, processes in the stratosphere such as anomalies in the winter-and spring-time stratospheric polar vortex impact the tropospheric circulation (Baldwin & Dunkerton, 2001;Byrne & Shepherd, 2018;Garfinkel & Hartmann, 2011).One issue when modeling the middle atmosphere is that there is a limited amount of observations to constrain the current model state (e.g., Eckermann et al., 2018). Above 10 hPa, most of the observations assimilated into the Integrated Forecasting System (IFS) of the European Center for Medium-Range Weather Forecasts (ECMWF) are from satellites and have limited spatial and temporal resolutions. They mainly provide temperature-related data (e.g., Global Navigation Satellite System Radio Occultation (GNSS-RO), Atmospheric Infrared Sounder (AIRS), Advanced Microwave Sounding Unit (AMSU-A)) and the topmost radiances assimilated peak at approximately 1-2 hPa. The range of sensitivity of the satellite observations to certain horizontal and vertical scales of GWs depends on the instrument and viewing geometry (observational filter, see Alexander, 1998) as can be seen in for example, Figure 9 of Preusse et al. (2008). To produce the most accurate representation of the atmospheric state, all the observations irregularly distributed in time and space and each having their limitations and uncertainties are combined with the numerical weather prediction model on a global grid. For the (re)analyses at ECMWF, this is achieved by 4-dimensional variational data assimilation (4D-Var).

Comparing gravity waves in a kilometre-scale run of the IFS to AIRS satellite observations and ERA5

“…The results of this study showed that the spatial structure of these temperature variances was similar for the model and observations in the mid-to high latitudes of the Southern Hemisphere. GW potential energy was found to be underestimated in the middle atmosphere in three IFS versions run at ∼9 km resolution, compared to data from the Compact Rayleigh Autonomous Lidar (CORAL) at Río Grande in the lee of the Southern Andes (Gisinger et al, 2022). GW momentum fluxes in the ECMWF operational analysis, produced using the IFS and 4D variational data assimilation, at a grid spacing of approximately 16 km (T1279 resolution) with 91 model levels, were found to be a factor of 5 lower than in Concordiasi balloon observations (Jewtoukoff et al, 2015).…”