Abstract
Omega dropwindsonde and other in situ (INS) data collected during the NOAA/Hurricane Research Division's (HRD) Hurricane Field Program are used as a ground truth dataset for the evaluation of VISSR Atmospheric Sounder (VAS) soundings over the subtropical Atlantic. The experiments were coordinated with the Cooperative Institute for Meteorological Satellite Services at the University of Wisconsin. The focus of this study is to determine whether soundings derived from VAS radiances are an improvement over the first-guess data used as a starting point in the sounding retrieval process. First guess inputs for this study are provided by NMCs Regional Analysis and Forecast System (RAFS) nested grid model (NGM).
In a case study, an objective algorithm is used to analyze the INS, VAS, and first-guess data at and below 500 mb from an HRD experiment on 1–2 September 1988. The case study is supplemented by a statistical investigation of data composited from other HRD experiments. In particular, we examine VAS estimates of horizontal temperature and moisture gradients to see if they represent improvements over the first guess.
The temperature and moisture descriptions in the vicinity of a 500 mb cold low were improved by the VAS in the case study; however, VAS temperature gradients were found to be generally less accurate than those of the first guess. Temperature gradients from the VAS were also consistently stronger than INS or first-guess gradients. The composite study found that large-scale VAS moisture gradients were better than those of the first guess. Other results indicate a preferred mode for VAS modifications to the guess: the primary impact of the VAS radiances on the first guess was to improve the description of the phasing of horizontal features. The VAS representation of the amplitude of features, however, was not consistently an improvement. This suggests that in tropical applications, VAS data may be most suitable for subjective forecasting uses; if VAS data are to be used in numerical weather prediction, strongest weight should be given to the representation of the location of weather features (troughs, ridges, etc.), and relatively weak weight should be given to the representation of the strength of these features.
