Parameter estimation of Neyman-Scott processes for temporal point rainfall simulation

The role of the data aggregation scale on parameters estimation of the cluster-based Neyman-Scott point processes applied to rainfall simulation is investigated. Extensive calculations showed that in estimating the parameters by the method of moments the choice of the aggregation scale of the data significantly affects the estimates of the continuous process parameters. Besides, when equations at different aggregation scales are employed, the optimisation procedure may fail if the selected scales are too close. An alternative estimation procedure is suggested, based on the scale of fluctuation of the observed process. The estimates obtained with the proposed procedure are at least as good as the better estimates obtained with the procedure using different scales, both in terms of reproduction of the second order statistics and extreme values for different aggregation scales, as shown by Monte Carlo simulations. On the whole, the use of the procedure based on fluctuation scale appears to be a valuable alternative in the estimation of the parameters of the Neyman-Scott processes. (C) 1999 Elsevier Science B.V. All rights reserved.