Strike probability map for Typhoon Matsa

Strike probability map for Typhoon Matsa, starting from the ECMWF analysis of 4 August 2005 00UTC and incorporating C-band ERS-2 scatterometer data.
ECMWF operates a so-called 'Ensemble Plotting System' (EPS) in which, in addition to the operational high-resolution forecast (OPER), 51 other alternatives forecasts are made at a lower resolution. One of these forecasts, the control (CTRL) is run from the same starting point but at lower resolution; the other 50 members of EPS are each run from slightly perturbed starting points with perturbed physics models - between them representing uncertainty in the initial state and errors built in to ECMWF's non-perfect physics models. Both contribute to incorrect forecasts and so are incorporated into the probability forecast.
The idea is that all these members together provide an error bar for the OPER run - the larger the spread, the less certain the result.
The probabilities range from 0.05 (one out of 20) to 1 (one out of one), so the numbers in the legend should be divided by ten or 100.