The use of dynamical constraints in the analysis of mesoscale rawinsonde data

Abstract

This study seeks to generate dynamically consistent analyses of mesoscale rawinsonde data by coupling analyses of winds and geopotentials on pressure surfaces. The approach used here follows in a general way the one taken by Lewis and Bloom (1978) in their analysis of surface data; the observations are first objectively analyzed onto a regular grid in space and time, and the gridded fields are then adjusted in a least-squares sense using a set of forecast equations of horizonatal momentum as dynamical constraints. However, the scheme described in this paper has a more consistent treatment of the adjustments to the fields in time. The adjustment process is tested on objectively analyzed rawinsonde soundings from two cases from the data archives of the National Severe Storms Laboratory (NSSL); 8 June 1966 and 22 May 1966. The principal results of this study are: (1) the initial winds and geopotentials produce unphysical residuals when they are put into the momentum constraints; (2) the adjusted fields have greatly reduced residuals, (3) the adjustments to the wind and geopotential fields have root-mean-square values of 1—2 m s^-1 and 30—50 m² s^-2 respectively, and (4) the vertical velocity fields computed from the adjusted winds are more consistent with the weather events that occurred in the two data cases.