Role of spatial anisotropy in design storm generation: Experiment and interpretation

Research output: Contribution to journalArticleScientificpeer-review

Researchers

Research units

  • Hydrology and Flood Warning
  • Bureau of Meteorology Australia

Abstract

Rainfall accumulation depths over a given area are strongly dependent on the shape of the storm together with its direction of advection. A method to produce design storms exhibiting anisotropic spatial scaling is presented by combining a state-of-the-art stochastic rainfall generator STEPS with the linear generalized scale invariance (GSI) notation. The enhanced model is used to create ensembles of design storms based on an extreme storm with a distinct rainband shape observed in Melbourne, Australia. Design storms are generated both with and without accounting for anisotropy. Effect of anisotropy on precipitation characteristics is studied using the entire region covered by the radar (radar scale) and at a significantly smaller catchment scale. A rainfall-runoff model is applied to route the rainfall through the catchment into streamflow. Accounting for anisotropy allows for a more realistic description of precipitation features at the radar scale. At the catchment scale, anisotropy increases the probability of high rainfall accumulations, which translates into greater flood volumes. No discernible difference was observed in streamflow characteristics after controlling for the accumulation over the catchment. This could be explained by a lower importance of anisotropy relative to other factors affecting streamflow generation, and by the difficulties in creating representative rainfall temporal properties at the catchment scale when the radar scale is used for model calibration. The proposed method provides a tool to create ensembles of design storms when the anisotropic shape of the fields is of importance.

Details

Original languageEnglish
Pages (from-to)69-89
Number of pages21
JournalWater Resources Research
Volume52
Issue number1
Publication statusPublished - 8 Jan 2016
MoE publication typeA1 Journal article-refereed

    Research areas

  • 1805 Computational hydrology, 1847 Modeling, 1854 Precipitation, 1869 Stochastic hydrology, anisotropy, design storm, ensemble, precipitation simulation, rainband, scaling

Download statistics

No data available

ID: 1638371