Investigating vertical gravity wave spectra to calibrate a gravity wave perturbation model: comparison with ERA5 reanalysis products and application to infrasound propagation simulations

Infrasound technology is used to monitor the atmosphere and to verify compliance with the Comprehensive Nuclear Test-Ban-Treaty. Acoustic signals recorded by the International Monitoring System allow to characterize sources of interest.

Fine-scale atmospheric perturbations of the order of a few kilometers to a few hundred meters in vertical wavelength are necessary to explain the duration and amplitude of acoustic signals. Such internal gravity wave (GW) perturbations must be added to atmospheric specifications for propagation simulations. Indeed, operational meteorological products underestimate or miss that part of the GW spectrum. The GW universal spectrum approach provides a convenient framework to quickly derive vertical perturbation profiles of GW using inverse Fourier transform along the atmospheric column.

Using radiosonde and lidar measurements from the Observatoire De Haute-Provence in South of France (43° 55′ 51″ N, 5° 42′ 48″ E) and from La Réunion Island (21° 04′ 47″ S, 55° 22′ 59″ E) across many years, we characterize monthly GW vertical wavenumber spectra in different altitude layers. We fit those spectra using the modified Desaubies analytical model in order to retrieve relevant parameters (namely the maximum amplitude of the spectrum and the characteristic wavenumber m*). We also compare the observed spectra and their related parameters and quantities, notably kinetic and potential energies, to those derived from ERA5 products.

Using the calibrated parameters of the GW spectra, we derive the associated ensemble perturbation profiles in a stochastic approach using bootstrap techniques. The goal is to be representative of the observed vertical distribution of the spectra. The ensembles of perturbation profiles are then used as input to infrasound propagation simulations. We discuss how waveform simulations used in operational monitoring can benefit from these better-constrained atmospheric fine-scale uncertainties.