Environmental impacts to responsive space launch at the Eastern and Western Ranges

Space launch is, and will continue to be, sensitive to weather. In order to develop a truly responsive launch capability significant research needs to be conducted in specification and prediction of the atmosphere below 50,000 ft. Present research shows that weather is the leading cause of cancelled space launches (51% at Eastern Range and 58% at Western Range). The ability to forecast weather in support of current requirements was examined. Almost four years (2000 – 2004) of metric data was obtained from the Air Force Space Command (AFSPC). Metrics include weather warnings, weather advisories (watches) and forecasts of Launch Commit Criteria (LCC). Data was grouped by season based on the LCC statistics for selected sites. These sites include Cape Canaveral Air Force Station (CCAFS), Patrick AFB, Melbourne Stations, Kennedy Space Center (KSC) and Vandenberg AFB. The criteria were chosen based on the meteorological conditions found in the LCC. The metric data obtained are number of forecast and observed occurrences of each event grouped into eight mutually exclusive categories. These eight categories include required, issued, met desired lead time (DLT), did not meet DLT, met DLT 50-99%, met DLT 1-49%, negative or zero LT, and false alarms.

Calculations for the statistical measures of accuracy were developed. The Hit Rate (HR) represents that fraction of observed “yes” events were correctly forecast. For this study, this is defined as the probability of issued warnings that met the desired lead-time but ignores false alarms. The False Alarm Ratio (FAR) is the fraction of predicted “yes” events that did not actually occur. The last statistical formula is the Success Score (SS). The SS measures the fraction of events that were correctly predicted. It can be thought of as the accuracy of correctly forecasted events. Other examinations include analyzing lead times for various meteorological events and the ability to issue a forecast within the desired lead time.

Results demonstrate current shortfalls in forecasting across several key environmental parameters which include lightning, convective and non-convective winds, precipitation and temperature. Both ranges show a large number of false alarms (forecasted but did not verify) for some of the environmental parameters. Even more significant are the low success scores or the probability of issued warnings meeting the desired lead time based on LCC. Ongoing research is focused on improvements in weather prediction which will lead to significant increases in operational responsiveness and decreased cost. Further research is required to improve weather forecasting so that responsive space launch will be realized.