A novel methodology of stochastic short term forecasting of cloud boundaries

Following the chaos theory proposed by Lorenz, probabilistic approaches have been widely used in numerical weather prediction. This paper introduces a convection diffusion equation to quantify the dynamic growth of uncertainty for short term forecasts of cloud boundaries. The equation is inserted into a numerical weather prediction model, weather research and forecast. A two parameter model based on wind velocity dispersion and surface evaporation rates parameterizes the stochastically motivated, but deterministic, equation. Prediction verification tests in comparison to observed data show good predictive capability for an hour, with a gradual loss of predictive power continuing for predictions up to three hours shown qualitatively. The methodology can be applied to a variety of topics in numerical weather prediction research.