Estimating the parameters of the extreme value type 1 distribution for low flow series in Ireland

In this study two different models have been developed and tested with data from 55 hydrometric stations in the Shannon River Basin in Ireland for estimating the location and the scale parameters of the EV1 distribution of the Minimum, 3-, 7-, 10-, 15-, and 30-days sustained low flow series at ungauged locations. The first is a simple linear model while the other is a fuzzy clustering model. Both models have been calibrated using the unconstrained and the constrained least squares methods. Moreover five different input scenarios including various combinations of some explanatory variables have been investigated with the two models. The results showed that: (i) the simple linear model calibrated by the constrained least squares method was the best model to estimate the EV1 location parameter using catchment area, mean annual rainfall, mean elevation, mean slope, and soil as explanatory variables, and (ii) the fuzzy clustering model calibrated by the unconstrained least squares method was the best for the EV1 scale parameter using only the first four above mentioned explanatory variables.