Implementing a space-time rainfall model for the Sydney region

Abstract

This paper investigates a Spatial Neyman-Scott Rectangular Pulse (SNSRP) model, which is one of only a few models capable of continuous simulation of rainfall in both space and time. The SNSRP is a spatial extension of the Neyman-Scott Rectangular Pulse model at a single point. The model is highly idealized having six parameters: storm arrival, cell arrival, cell radius, cell lifetime and two cell intensity parameters. A spatial interpolation of the scale parameter is used so that the model can be simulated continuously in space, rather than as a multi-site model. The parameters are calibrated using least-squares fits to statistical moments based on data aggregated to hourly and daily totals. The SNSRP model is calibrated to a very large network of 85 gauges over metropolitan Sydney and shows a good agreement to calibrated statistics. A simulation of 50 replicates over the region compares favourably to several observed temporal statistics, with an example given for one site. A qualitative discussion of the simulated spatial images demonstrates the underlying structure of non-advecting cylindrical cells.