Comparison between non‐orographic gravity‐wave parameterizations used in QBOi models and Strateole 2 constant‐level balloons

Gravity‐wave (GW) parameterizations from 12 general circulation models (GCMs) participating in the Quasi‐Biennial Oscillation initiative (QBOi) are compared with Strateole 2 balloon observations made in the tropical lower stratosphere from November 2019–February 2020 (phase 1) and from October 2021–...

Full description

Bibliographic Details
Main Authors: Lott, F, Rani, R, McLandress, C, Podglajen, A, Bushell, A, Bramberger, M, Lee, H, Alexander, J, Anstey, J, Chun, H, Hertzog, A, Butchart, N, Kim, Y, Kawatani, Y, Legras, B, Manzini, E, Naoe, H, Osprey, S, Plougonven, R, Pohlmann, H, Richter, JH, Scinocca, J, García‐Serrano, J, Serva, F
Format: Journal article
Language:English
Published: Wiley 2024
Description
Summary:Gravity‐wave (GW) parameterizations from 12 general circulation models (GCMs) participating in the Quasi‐Biennial Oscillation initiative (QBOi) are compared with Strateole 2 balloon observations made in the tropical lower stratosphere from November 2019–February 2020 (phase 1) and from October 2021–January 2022 (phase 2). The parameterizations employ the three standard techniques used in GCMs to represent subgrid‐scale non‐orographic GWs, namely the two globally spectral techniques developed by Warner and McIntyre (1999) and Hines (1997), as well as the “multiwaves” approaches following the work of Lindzen (1981). The input meteorological fields necessary to run the parameterizations offline are extracted from the ERA5 reanalysis and correspond to the meteorological conditions found underneath the balloons. In general, there is fair agreement between amplitudes derived from measurements for waves with periods less than 1 $$ 1 $$ h and parameterizations. The correlation between the daily observations and the corresponding results of the parameterization can be around 0.4, which is 99 % $$ 99\% $$ significant, since 1200 days of observations are used. Given that the parameterizations have only been tuned to produce a quasi‐biennial oscillation (QBO) in the models, the 0.4 correlation coefficient of the GW momentum fluxes is surprisingly good. These correlations nevertheless vary between schemes and depend little on their formulation (globally spectral versus multiwaves for instance). We therefore attribute these correlations to dynamical filtering, which all schemes take into account, whereas only a few relate the gravity waves to their sources. Statistically significant correlations are mostly found for eastward‐propagating waves, which may be due to the fact that during both Strateole 2 phases the QBO is easterly at the altitude of the balloon flights. We also found that the probability density functions (pdfs) of the momentum fluxes are represented better in spectral schemes with constant sources than in schemes (“spectral” or “multiwaves”) that relate GWs only to their convective sources.