SHORT-RECORD-BASED EXTREME WIND SIMULATION

In order to utilize limited historical wind records for estimating ext reme wind speeds for natural hazards damage mitigation, a Markov chain model for generating long-term annual extreme winds, on the basis of short-term records, is investigated. Basically, this simulation model consists of three components. They are state of wind speeds, wind spee d distribution functions, and transition probability matrices. The bas ic strategy of our simulation model is to generate the time series of hourly wind speeds in parts: for those wind associated with well-behav ed climates and those with extreme winds. Applications of this model t o generate long-term extreme winds, on the basis of short records at H ouston Intercontinental Airport of Texas, are demonstrated.