The precipitation output of a mesoscale atmospheric numerical model is usually interpreted as the average rainfall intensity over the grid cell of the model (typically 30¿30 km to 60¿60 km). However, rainfall exhibits considerable heterogeneity over subgrid scales (i.e., scales smaller than the grid cell), so it is necessary for hydrologic applications to recreate or simulate the small-scale rainfall variability given its large-scale average. Rainfall disaggregation is usually done statistically. In this paper, a new subgrid scale rainfall disaggregation model is developed. It has the ability to statistically reproduce the rainfall variability at scales unresolved by mesoscale models while being conditioned on large-scale rainfall averages and physical properties of the prestorm environment. The model is based on two extensively tested hypotheses for midlatitude mesoscale convective systems <Perica and Foufoula-Georgiou, 1996>: (1) standardized rainfall fluctuations (defined via a wavelet transform) exhibit simple scaling over the mesoscale, and (2) statistical scaling parameters of rainfall fluctuations relate to the convective available potential energy (CAPE), a measure of the convective instability of the prestorm environment. Preliminary evaluation of the model showed that the model is capable of reconstructing the small-scale statistical variability of rainfall as well as the fraction of area covered with rain at all analyzed subgrid scales. The performance evaluation was based on comparison of summary statistics and spatial pattern measures of simulated fields with those of known fields observed during the Oklahoma-Kansas Preliminary Regional Experiment for Storm-Central (PRE-STORM). ¿ American Geophysical Union 1996