One characteristic of ensemble forecasts has been their under-dispersive nature. The purpose of this experiment is to explore how to generate a perturbed soil moisture initial condition and determine its impact on short range ensemble forecasts using the NCAR/PSU MM5 version 3. Several Great Plains warm-season short-range forecasts will be carried out three ways: with a perturbed atmosphere, with a perturbed soil moisture, and with both a perturbed atmosphere and perturbed soil moisture. The results will help determine the how much missing spread can be explained by perturbing the soil moisture.