Stochastic convective rain‐field simulation using a high‐resolution synoptically conditioned weather generator (HiReS‐WG)

Abstract

A new stochastic high-resolution synoptically conditioned weather generator (HiReS-WG) appropriate for climate regimes with a substantial proportion of convective rainfall is presented. The simulated rain fields are of high spatial (0.5 3 0.5 km 2 ) and temporal (5 min) resolution and can be used for most hydrological applications. The WG is composed of four modules: the synoptic generator, the motion vector generator, the convective rain cell generator, and the low-intensity rainfall generator. The HiReS-WG was applied to a study region on the northwestern Israeli coastline in the Eastern Mediterranean, for which 12 year weather radar and synoptic data were extensively analyzed to derive probability distributions of convective rain cells and other rainfall properties for different synoptic classifications; these distributions were used as input to the HiReS-WG. Simulated rainfall data for 300 years were evaluated for annual rain depth, season timing, wet-/dry-period durations, rain-intensity distributions, and spatial correlations. In general, the WG well represented the above properties compared to radar and rain-gauge observations from the studied region, with one limitation-an inability to reproduce the most extreme cases. The HiReS-WG is a good tool to study catchments' hydrological responses to variations in rainfall, especially small-size to medium-size catchments, and it can also be linked to climate models to force the prevailing synoptic conditions.

Spatial and temporal variability of rainfall and their effects on hydrological response in urban areas – a review

“…The first were introduced in the 1970s and are widely used to reproduce the spatial and temporal variability (see Schertzer and Lovejoy, 2011 for a review). Autoregressive methods, also nowadays often referred to as "rainfall generator models", are used to generate multidimensional random fields while preserving the rainfall spatial autocorrelation, for natural Peleg and Morin, 2014;Niemi et al, 2016) and urban (Sørup et al, 2016) areas. Point-process models are used when the spatial structure of intense rainfall is defined by convective rainfall cells (see McRobie et al, 2013 for an example).…”

Effect of Temporal Resolution of Rainfall on Simulation of Urban Flood Processes

“…Emmanuel et al [11] summarized the characteristic scales of rainfall of different types. Also, stochastic rainfall generator and statistics downscaling methods were widely used to obtain high temporal resolution rainfall series [12][13][14][15][16].…”

Generating synthetic rainfall with geostatistical simulations

“…In that case, it looks reasonable to restrict the use of geostatistical simulations to what they are really able to reproduce, i.e., distributed rain fields. To model and reproduce the rain field in all its diversity, an interesting approach that should be pursued consists in combining geostatistical simulations with other stochastic methods such as fractals, point processes or multiple‐points statistics in order to build hierarchical models which capitalize on the strengths of each involved methods …”