Accounting for time changes in atmospheric infrasound observations

A new method computes how infrasound sensitivity depends on atmospheric parameters that vary in space and time.

Long range acoustic waves in the atmosphere are sensitive to temperature and wind variations in the upper atmosphere. Infrasound observations have potential to improve numerical weather models through data assimilation methods.

Acoustic waves propagate slowly. Atmospheric parameters can change during this propagation. Most infrasound simulation tools assume stationary atmosphere models and do not take this change into account.

A method was presented to compute linear sensitivities of infrasound observables to atmospheric parameters varying in space and time. These are called 4D sensitivity kernels or time variable observation operators. Adjoint based methods can easily handle this computation.

Variational methods with data assimilation windows longer than the propagation duration can accommodate these observations. Their ensemble counterparts can also handle them.

Experiments using a Lorenz'96 low-order model with circumnavigating Lamb waves show the benefit of assimilating long range infrasound. The results demonstrate that time variability of atmospheric parameters should be taken into account. This applies even for propagation times limited to 2 or 3 time steps of the atmosphere model.

Key facts

  • Infrasound observations can improve numerical weather models.
  • Acoustic waves propagate slowly enough that atmospheric properties can change during travel.
  • Time variability should be considered even for propagation times of 2 or 3 time steps.

All sections

Edition