A probabilistic forecasting model of tropical cyclone-generated storm surge for a straight coastline

Typescript (photocopy).A dynamic and statistical algorithm is constructed for a probabilistic forecasting of storm surge. The three essential constituents of this algorithm are: (1) an appropriate tropical cyclone model, (2) a justified and easily handled storm surge model, and (3) the correct interpretation of the errors of the official tropical cyclone forecast. The tropical cyclone model described by empirical equations is flexible in structure by allowing the coefficients in equations to be adjusted according to observed values of atmospheric variables. Thus, it provides more realistic atmospheric forcing for the storm surge model than the forcing described by the standard project hurricane. A bathystrophic storm surge simulation model which also considers the two-dimensional atmospheric pressure gradient is reformulated. The simple format of this simulation model requires much shorter computing time than other currently available models (e.g., SPLASH, SSURGE) while its accuracy is competitive with other models. The concept of uncertainty in the forecasts of the track and intensity is applied to the storm surge forecasting. The probability distribution of the errors in the official tropical cyclone forecast is interpreted as a bivariate normal distribution. The two variates are the speed and the direction of the motion of the tropical cyclone. A modified polar coordinate system with one ray aligned to the (curving) official forecast track is used to design the probability integration scheme. The data input to this forecasting model contains the observed values and the forecast track and intensity of the tropical cyclone. The output gives time series of the sample mean and standard deviation envelope of the storm surge. The probabilities of given surge levels being reached can be derived. Examples of this probabilistic storm surge forecast prove useful and suitable for use in preparedness activities for coastal areas during tropical cyclone threats