For many years forecasting has been done deterministically. Ed Lorenz (Lorenz 1963) was one of the first scientist to suggest that deterministic forecasts could only be good for a period of time due to the non-linear nature of the atmosphere. The non-linearity of the forecast parameter depended upon the scales in which the forecast was made. For large scale modeling, the breakdown of the forecast would not happen for a week or so. For smaller scales, this time would be less. In order to account for this non-linearity, people began to suggest the use of ensembles. This approach was not used operationally until the 1990s (Tracton and Kalnay 1993). The first experiments dealt with the larger scales. Recently, studies have focused on the shorter time scales. This has sparked a debate among many meteorologists. Is a one time lower resolution model a better predictor of the atmosphere than an ensemble of higher resolution model runs? One could argue either way. Assuming that there is a perfect model, there is no debate. Ensemble forecasting would be the way to proceed in forecasting. The debate continues because the model is not perfect. Some modelers believe the atmosphere will be better modeled at lower resolutions. At some point, though, it is argued that smaller scales will be unresolvable and everything will have to modeled explicitly, nothing will be parameterized. This is where short range ensemble forecasting could step in and help in the forecast process. At lower resolutions, ensembles could be run in order to give the forecaster a number of solutions or grouping of possible solutions.

This study is aimed at answering some of the questions regarding short range ensembles versus lower resolution model runs; and at what point do we start using ensembles in the short range. In order to accomplish this task, the goal is to run ensembles using two separate methods. Numerous techniques (regional breeding, different assimilation schemes, initial condition perturbations, and lyaponov vectors) can be used to generate initial conditions. Ensembles can also be generated via different model physical and convective parameterization package combinations. Further studies with ensembles could explore different PBL schemes as well as land-use differences. For this study, it has been decided to use a method of perturbing the initial conditions, first developed by Errico and Baumhefner (1987). The method of varying the convective and model physic packages provided with the MM5V2 has also been employed.

REFERENCES

Cook, L. K., 1998: Western Region Model Diagnostics. WR-Technical Attachment 98-39.

Errico, R. and D. Baumhefner, 1987: Predictability Experiments Using a High-Resolution Limited-Area Model. Mon. Wea. Rev., 115, 488-504.

Lorenz, E. N., 1963: Deterministic nonperiodic flow. J. Atmos. Sci., 20, 130-141.

Mullen, S. L. and D. P. Baumhefner, 1988: The impact of initial condition uncertainty on numerical simulations of large-scale explosive cyclogenesis. Mon. Wea. Rev., 117, 2289-2329.

Tracton, M. S., and E. Kalnay, 1993: Operational ensemble prediction at the National Meteorological Center: Practical aspects. Wea. Forecasting, 8, 379-398.