Premium
A coupled stochastic space‐time intermittent random cascade model for rainfall downscaling
Author(s) -
Kang Boosik,
Ramírez Jorge A.
Publication year - 2010
Publication title -
water resources research
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 1.863
H-Index - 217
eISSN - 1944-7973
pISSN - 0043-1397
DOI - 10.1029/2008wr007692
Subject(s) - downscaling , intermittency , scale (ratio) , spatial ecology , environmental science , climatology , cascade , meteorology , similarity (geometry) , stochastic modelling , temporal scales , spatial dependence , mathematics , geology , statistics , geography , computer science , precipitation , cartography , turbulence , engineering , chemical engineering , ecology , artificial intelligence , image (mathematics) , biology
Analysis of Next Generation Weather Radar rainfall data indicates that for the central United States, rainfall exhibits a composite behavior with respect to its spatial and temporal scaling characteristics. Our data analysis shows that rainfall fluctuations at spatial scales smaller than a reference scale exhibit self‐similarity and that at scales larger than the reference scale, rainfall fluctuations are scale dependent. Accordingly, we present a new methodology for downscaling large‐scale rainfall consistent with this composite character of rainfall variability. The new downscaling model is a composite of a stochastic space‐time submodel that preserves the spatial and temporal dependency characteristics at scales larger than the reference scale and an intermittent random cascade submodel that preserves the statistical self‐similarity and spatial intermittency at scales smaller than the reference scale. The new model is applied to downscale summer daily rainfall for the central United States from a scale of 256 km to a scale of 2 km. We show that the new model reproduces quite well the intermittency and self‐similarity features and the interscale and across‐scale correlation structures of observed rainfall with a relatively low computational burden.