Analysis of a long record of annual maximum rainfall in Athens, Greece, and design rainfall inferences

An annual series of maximum daily rainfall extending through 1860-1995, i.e., 136 years, was extracted from the archives of a meteorological station in Athens. This is the longest rainfall record available in Greece and its analysis is required for the prediction of intense rainfall in Athens, where currently major flood protection works are under way. Moreover, the statistical analysis of this long record can be useful for investigating more generalised issues regarding the adequacy of extreme value distributions for extreme rainfall analysis and the effect of sample size on design rainfall inferences. Statistical exploration and tests based on this long record indicate no statistically significant climatic changes in extreme rainfall during the last 136 years. Furthermore, statistical analysis shows that the conventionally employed Extreme Value Type I(EV1 or Gumbel) distribution is inappropriate for the examined record (especially in its upper tail), whereas this distribution would seem as an appropriate model if fewer years of measurements were available (i.e., part of this sample were used). On the contrary, the General Extreme Value (GEV) distribution appears to be suitable for the examined series and its predictions for large return periods agree with the probable maximum precipitation estimated by the statistical (Hershfield's) method, when the latter is considered from a probabilistic point of view. Thus, the results of the analysis of this record agree with a recently land internationally) expressed scepticism about the EV1 distribution which tends to underestimate the largest extreme rainfall amounts. It is demonstrated that the underestimation is quite substantial (e.g., 1:2 for large return periods and this fact must be considered as a warning against the widespread use of the EV1 distribution for rainfall extremes.