Determining optimal parameters for gravity wave drag schemes using data assimilation techniques: Twin experiments

PULIDO M., S. POLAVARAPU, T. SHEPHERD AND J. THUBURN

Bologna (Italy)

Conferencia; SPARC 4th General Assembly; 2008

SPARC, WCRP

Because most gravity wave drag schemes assume instantaneous verticalpropagation of gravity waves in a column, observations in a singlecolumn can be used to obtain information about input parameters to agravity wave drag scheme.  Here, a 1D variational assimilation techniqueis developed to estimate parameters from gravity wave schemes. The costfunction measures the differences between the zonal and meridionalcomponents of the ´observed´ gravity wave drag field and the gravitywave drag calculated with a scheme. The gradient of the cost function iscalculated with the adjoint of Scinocca´s gravity wave scheme. In thiswork  the sensivity of some key parameters to changes in the gravitywave drag and horizontal wind and temperature profiles is calculatedwith the adjoint model and the performance of the technique is evaluatedthrough twin experiments.