A methodology was developed to quantify the uncertainty associated wit
h maximum likelihood parameter estimates used in the wet day amount co
mponent of stochastic rainfall models. Traditional point parameter est
imates were combined with interval estimates to construct a probabilis
tic region of parametric uncertainty. Regions of model uncertainty wer
e then constructed from the region of parametric uncertainty. The wet
day amount model was implemented by repeatedly selecting parameters fr
om a bivariate normal distribution. The results of this test suggested
that the amount model was sensitive to small changes in parameters. T
he model was least sensitive to parametric uncertainty for small wet d
ay amounts and increased in sensitivity toward large amounts.