Description is given of a technique for determining optimal storm (reference frame) motion based upon application of a dynamic-retrieval method to velocity datasets derived from multiple-Doppler radar observations. The method depends upon the necessary consistency between the steady-state assumption and assumed reference-frame (storm) motion and uses the quantity E, from dynamic-retrieval calculations as a measure by which to judge when this consistency is best achieved. Application of the technique is demonstrated in case examples including an Oklahoma squall line, a Montana hailstorm, and an Oklahoma tornadic storm.
In the squall-line case the question of the dependence of optimal reference-frame motion upon analysis domain (e.g., convective versus stratiform regions of the system) is explored. Similar optimal frame motions for different regions of the system are found. Optimal frame motion corresponds more closely to cell motion than to line motion. In the Montana case the dependence upon analysis domain is again explored, and significant differences between a large domain and subdomain are found. Retrieved pressure compares favorably with independent below-cloud-base measurement of perturbation pressure by aircraft. It is shown that agreement between retrieved and observed pressure patterns is best when optimal reference-frame motion is assumed. In the tornadic-storm case, optimal frame motion is very similar to storm motions derived from reflectivity-core tracking and from numerical simulation of this storm. Investigation of the question of height variation of optimal reference-frame motion is investigated and found to be influenced by, but not completely dependent upon, both the environmental winds and mean in-storm air motion. The notion of using optimal reference-frame motions as a basis for adjustment of nonsimultaneous Doppler radar observations to a common reference time is discussed.