A thesis submitted in partial fulfillment of the requirements for the degree of Master of Engineering of the Asian Institute of Technology, Bangkok, Thailand.

Thesis (M.Eng.) - Asian Institute of Technology, 1972

The maximum annual flood flow data of the northern Thailand was proved to follow the extreme value distribution of type I. The parameters of the extreme value distribution of type I, a and b were estimated from the samples. The student's-t distribution was used to estimate the population parameters of the distribution, and corresponding to a and b, respectively. The estimation error is defined by the difference between the estimated return period and the population return period for a specific level of significance. The pseudo-annual floods were generated by the Monte Carlo method with the extreme value distribution of type I. The same method was used with these pseudo-annual floods to obtain the estimation error. This estimation error is insignificantly different from the estimation error from the observed data. Given the life of the structure, the allowable estimation error and the encounter probability, the minimum number of years of flood record can be determined.