Improvement of COAMPS-TC Quantitative Precipitation Forecasts for Varied Tracks Over and Near the Island of Taiwan

Previous studies of COAMPS-TC QPF's have focused on a single iteration cycle for a particular storm. This has resulted in a small sample size with little variation in model storm tracks and less opportunity to discover errors from multiple sources. This study utilizes simulations at multiple iteration times for storms representing four track categories for a total of sixteen runs. Each simulation is examined in detail to identify the sources of error most relevant to the problem of surface flooding.

The recent Typhoon Soulik (2013) is used as a northern landfall case. The historic Typhoon Morakot (2009) moved across the center of the island. The looping track of Typhoon Tembin (2012) allows it to be both a southern landfall and an island parallel, offshore storm. The COAMPS-TC accumulated precipitation is compared to both specific rainfall totals from the dense gauge network and to flooding threshold values to determine which QPF errors would have the largest impact to emergency management. Key sources of error to investigate are cross-track and along-track biases in the storm position, intensity during the initial spinup, and impacts of the microphysical parameterization. Identifying the most important sources of error will be key in developing idealized simulations to isolate and improve these processes in the model.