Author(s): Boosik Kang

Linked Author(s):

Keywords: Downscaling; Self-similarity; Stochastic model; Quantile mapping

Abstract: As the preprocessing for incorporating the numerical weather/climate model with the hydrologic rainfall-runoff model, a simple bias correction and statistical downscaling schemes are introduced. To remove the systematic biases embedded in the numerical weather/climate model, the Quantile Mapping was applied first, through which the bias corrected numerical weather/climate output will have the equivalent probability distribution with the observed variables. For statistical downscaling, the Stochastic Spatio-Temporal Random Cascade Model (SST-RCM) was applied. The SST-RCM can reproduce the stochastic spatial and temporal dependency and the spatial intermittency showing the self-similarity characteristics. The convective storm event in Aug. 6, 2002 was utilized for scaling analysis and parameter estimation. In that event, the rainfall showed the relatively even distribution over the whole country and the rain field was showing mixed feature of mono- and multifractals. To simplify the problem, simple scaling was used for estimating the self-similarity parameter, beta (β) from the MKP (Mandelbrot-Kahane-Peyriere) function. The beta has strong dependency on the rainfall amount in the mother grid (background rainfall). The methodology for the multiple scaling for the multifractal images will be devised in the future.

DOI:

Year: 2007