Large-scale modeling of rain fields from a rain cell deterministic model
Article first published online: 29 APR 2006
Copyright 2006 by the American Geophysical Union.
Volume 41, Issue 2, April 2006
How to Cite
2006), Large-scale modeling of rain fields from a rain cell deterministic model, Radio Sci., 41, RS2010, doi:10.1029/2005RS003312., , , , , and (
- Issue published online: 29 APR 2006
- Article first published online: 29 APR 2006
- Manuscript Accepted: 18 JAN 2006
- Manuscript Revised: 7 DEC 2005
- Manuscript Received: 8 JUL 2005
- propagation in rain;
- rain modeling;
- radar meteorology
 A methodology to simulate two-dimensional rain rate fields at large scale (1000 × 1000 km2, the scale of a satellite telecommunication beam or a terrestrial fixed broadband wireless access network) is proposed. It relies on a rain rate field cellular decomposition. At small scale (∼20 × 20 km2), the rain field is split up into its macroscopic components, the rain cells, described by the Hybrid Cell (HYCELL) cellular model. At midscale (∼150 × 150 km2), the rain field results from the conglomeration of rain cells modeled by HYCELL. To account for the rain cell spatial distribution at midscale, the latter is modeled by a doubly aggregative isotropic random walk, the optimal parameterization of which is derived from radar observations at midscale. The extension of the simulation area from the midscale to the large scale (1000 × 1000 km2) requires the modeling of the weather frontal area. The latter is first modeled by a Gaussian field with anisotropic covariance function. The Gaussian field is then turned into a binary field, giving the large-scale locations over which it is raining. This transformation requires the definition of the rain occupation rate over large-scale areas. Its probability distribution is determined from observations by the French operational radar network ARAMIS. The coupling with the rain field modeling at midscale is immediate whenever the large-scale field is split up into midscale subareas. The rain field thus generated accounts for the local CDF at each point, defining a structure spatially correlated at small scale, midscale, and large scale. It is then suggested that this approach be used by system designers to evaluate diversity gain, terrestrial path attenuation, or slant path attenuation for different azimuth and elevation angle directions.