Predictability of the 24–25 January 2000 Snow Storm with respect to model grid resolution and initial state

On 24-25 January 2000, an unexpectedly intense winter storm caused by an explosive cyclogenesis off the southeastern coast of the United States brought heavy snowfalls from the Carolinas, through the Washington D.C. area, and into New England, with at least 5 deaths reported. Record snowfall amounts fell across North Carolina, with Raleigh-Durham area reporting a snowfall total from the storm of over 20 inches breaking not only the snowfall record of a single storm, but also the highest total in one month. The intensity and position of the storm was ill-handled by the operational models, posing a forecasting challenge for forecasters in the affected region. The short-range models (Eta and NGM) tracked the storm and precipitation too far east In this study, mesoscale model MM5 is used to investigate the predictability of the 24-25 January 2000 snowstorm. The triple-nested MM5 initialized from 00Z 24 January 2000 with the highest grid resolution of 3.3km simulated well the explosive coastal cyclogenesis in terms of the cyclone strength and location as well as the heavy onshore precipitation band. The model was running as operational. The success of the high-resolution control simulation shows that the storm could have been well forecasted with conventional data in real-time. Various sensitivity experiments suggest that the difficulty of the real-time operational numerical forecast come from insufficient model grid resolution, errors in the model initial state, and the strong nonlinearity inherent in the dynamic system. Other factors particularly the physical parametrization may also have contributed to the operational forecast failure.