A space-time Neyman-Scott rainfall model with defined storm extent

Abstract

In previous formulations, a model based on a Neyman-Scott point process coupled with a spatial Poisson process of cell centers has been used to generate raincells in space and time. For large regions, this specification yields unrealistic cross-correlations between distant sites. This observation motivates further development of the model, where a storm boundary is used to modulate the spatial extent of cell occurrences with respect to the storm center. Multi-site first- and second-order properties of this model are derived and used to fit rainfall statistics from 24 pluviograph (6 minute intensity) gauges and 52 daily gauges across the Sydney metropolitan region. Copyright 2008 by the American Geophysical Union.