EXTREME HURRICANE WIND SPEEDS - ESTIMATION, EXTRAPOLATION AND SPATIALSMOOTHING

Estimation of the extremal behaviour of hurricane wind speeds is compl icated by the lack of accurate extreme data at any one site. As a resu lt, wind engineers have developed climatological/physical models from which hurricane events can be simulated and used as a basis for infere nce. We present an application of non-linear spatial regression techni ques to hurricane wind speed data simulated at locations on the Gulf a nd Atlantic coasts of the United States. Modelling the spatial variati on in extremal behaviour provides a means of pooling data, thus increa sing the extent of information available for inference at each site. E stimates of distributional models for extremal behaviour and, conseque ntly, return levels are more precise than those obtained by standard m ethods applied to individual site data. We also adapt these spatial te chniques to estimate the distribution of directions of the r-largest h urricane wind speeds at each site. The techniques we describe are prel iminary steps to spatially modelling the joint behaviour of wind direc tion and speeds. We anticipate that this method of estimating return l evels for winds in a given direction will yield estimates with greater accuracy than current techniques enable. (C) 1998 Elsevier Science Lt d. All rights reserved.