HFRAD Hindcast: An OTH-B Radar and Model Comparison Study

Hindcast analysis of OTH radar data differs from purely predictive or forecast analysis in that it is made after the fact of radar measurement, permitting the removal of a substantial part of the random variability inherent in forecasting radar performance. In this paper, hindcast analysis is used to verify the performance of the MITRE OTH radar performance model, HFRAD. Data was collected during November and December of 1990 with the ANFPS-118 East Coast Radar System ECRS, and during January and February of 1991 with the West Coast Radar System WCRS. This data was analyzed and compared to hindcast estimates of performance using a new version V507g of HFRAD that incorporates a new and improved, clutter model. Performance was analyzed in terms of target signal-to- noise ratio SNR, both for the aggregate data set and for the data as a function of azimuth because of its location near the auroral oval, the ECRS is known to exhibit a strong dependence of performance on azimuth. In addition, these internal model components were analyzed in the same way subclutter visibility SCV, surface clutter peak, background noise i.e., noise measured with transmitter operating, noise measured with transmitter off ambient noise , and the ratio of background noise to ambient noise. As anticipated, the comparison is generally in good agreement, although notable discrepancies still exist, particularly near the auroral oval. Suggestions for further improvements to HFRAD are presented.