Gridding of design rainfall parameters for the IFD revision project for Australia

As part of the revision of Australian Rainfall and Runoff, the Bureau of Meteorology has revised the Intensity-Frequency-Duration (IFD) design rainfall estimates. Daily point rainfall data are available for over 8000 sites across Australia. However, as IFD estimates are required Australia wide, the point data need to be gridded to provide estimates for areas where point data are not available. The grids serve as the basis for deriving IFD estimates for Annual Exceedance Probabilities (AEP) of 50% to 1%. Different gridding approaches are available to interpolate spatial data with thin plate smoothing splines being one widely used technique. This paper examines the gridding of the Generalised Extreme Value (GEV) parameters using a thin plate smoothing spline to estimate design rainfalls across Australia. When gridding rainfall data, the interaction between topography and precipitation is important for the accurate interpolation of rainfall data. Large rainfalls tend to occur in areas of high elevation where the density of data is sparse as rainfall gauges are often located in more accessible lowland areas. As a consequence, at high elevations, rainfall is likely to be underestimated when it is spatially interpolated without reference to elevation. Thin plate smoothing splines allow for the incorporation of the topographic dependence of rainfall, which this study confirms as important for the modelling of Annual Maximum Series rainfall data. This investigation also explores the implications of Digital Elevation Model (DEM) resolution and the number of knot points when gridding precipitation using the smoothing spline software ANUSPLIN.