GCM parameter estimation based on LETKF

RUIZ J. J.; M. PULIDO

Viena

Conferencia; European Geophysical Union. General assembly 2011.; 2011

EGU

The Local Ensemble Transform Kalman Filter is used for the simultaneous estimation of model state and parame-ters associated to the convective parametrization of a simplfied general circulation model. The impact of parameterestimation upon initial conditions and short range forecasts are estimated using twin experiments. This approachallows the evaluation of the technique under the assumption of perfect knowledge of the true system evolution andits associated parameters.Two approaches are tested for the inclusion of parameter estimation within a data assimilation cycle, one whichuses an ensemble with perturbations in the initial conditions and in the parameters. The other approach uses twoseparate ensembles, one ensemble based on the perturbations of the initial conditions and the other ensemble basedonly on the perturbations of the parameters. The second approach produces a less noiser estimation of the param-eter, although the reduction of the analysis error is similar to that obtained with the first approach which is alsocomputationaly cheaper.The sensitivity of the analysis error and estimated parameters to the spread of the parameter ensemble and to theparameter initial value is explored.The Ensemble Transform Kalman Filter successfully estimates the parameters associated with the convectivescheme. The method also produces a good synchronization when the true model parameters are time dependent.