TRANS-STAT: statistics for environmental studies, Number 16. Extreme value statistics applied to environmental radionuclide studies: Chapters 7 and 8

The commonly used procedure for estimating the Extreme Value distribution of a sequence of measurements implicitly assumes the samples are from a stationary random process. This ignores the well known systematic and cyclic trends that typically occur. Ignoring these trends tends to overestimate the magnitude of the extremes and their associated statistics. This chapter has presented a procedure for including such trends within the parameter estimation algorithm. The result is a better estimate of extreme values