Data assimilation of stratospheric ozone using a high-resolution transport model

Abstract

We describe a method for assimilating sequentially tracer measurements in isentropic chemistry-transport models (CTMs) of the stratosphere. The parametrisation of the forecast error covariance and its evolution is largely based on simplifications described in [2000] and [2000]. The model used here is a high resolution isentropic advection model which is driven by ECMWF (European Center for Medium range Weather Forcast) meteorological analyses. The assimilation on isentropic surfaces allow us to exploit the well-established correlation between tracer mixing ratio and potential vorticity in the formulation of the forecast error covariance. Multiple 20-day sequential assimilations of MLS (Microwave Limb Sounder onboard UARS satellite) ozone data during an ozone depletion event are performed. χ2 (chi-square) and OmF (observation minus forecast) statistics are used to optimise the assimilation system by adjusting parameters of the error covariance. The quality of the analysis is found to be significantly improved when the strong correlation between ozone and potential vorticity is taken into account.