Cell identification, verification, and classification using shape analysis techniques

Abstract

[ACCESS RESTRICTED TO THE UNIVERSITY OF MISSOURI AT AUTHOR'S REQUEST.] The use of object-oriented approaches in both the verification of meteorological variables (especially precipitation) and the short-term forecasting of storms (nowcasting) has recently expanded. This two-part dissertation addresses the issue of verification and the possible nowcasting application of using shape analysis. Part one deals with a newly developed object-oriented verification tool using Procrustes shape analysis methodology. Using the Procrustes verification tool, the examination of the errors of matched truth and forecast objects for the different nowcasts can be assessed via a penalty function. This penalty function is based on errors in translation, rotation, dilation, shape, and intensity of a forecast object and, as these penalties can be assessed separately, the mode of the error in the forecast can be determined. Part two deals with the potential to classify convective storm cells based on shape characteristics combined with radar-derived products and near-storm environmental (NSE) data. Obtaining radar-derived products and NSE data from the Warning Decision Support System-Integrated Information (WDSS-II) system allows for fields of data to be easily overlaid for an identifiable cell in a domain. The ability to use this data overlay with classification trees may result in the ability to add physical information to a nowcast system for growth, decay, and morphology as well as provide grounds for correctly matching storm cells in an object-oriented verification technique.