An Evaluation of SARP (Search and Rescue Planning System) Drift Predictions Using Satellite-Tracked Drift-Buoys.

During the period of 1979 through 1981 the USCG Research and Development Center conducted several experiments in which the drift predictions of the computerized Search and Rescue Planning System (SARP) were compared with the movement of satellite-tracked drift-buoys released in the Gulf Stream east of Florida. Buoy positions at specific times were used as the last known position and incident date-time-group of a simulated search and rescue (SAR) incident; subsequent buoy positions were used to test the accuracy of 24, 48, 72, 96, and 120 hour drift predictions. Of the 680 drift predictions evaluated, in only 41 cases (6.0%) was the interpolated buoy position within the SARP-predicted search area. The accuracy of the drift predictions degraded with increasing prediction time; There was no discernible difference between the predicted (forecast) and analysis (observed) wind results; If the effect of wind is not included in the drift prediction the result is a larger relative drift error; The sea current files used in the SARP drift predictions contain some systematic errors which adversely affect system performances; and The drift error factor presently in use seriously underestimates the total drift error.