Date of this Version
Forecasting storm total snow accumulation is one of the most difficult aspects of meteorological forecasting. The forecaster has to interpret three main variables in order to forecast snowfall accurately. These forecasting variables are the duration of the snowfall, the amount of liquid water the storm will produce, and the snow density or snow ratio. With the advancement of computer models in recent history, the need for a quick and easy interpretation of these variables has grown, and to improve on previous forecasting techniques’ disadvantages with including the three snow forecasting variables. The Cobb Method snowfall forecasting algorithm utilizes model data and interprets all variables of snowfall forecasting and quickly produces snowfall amounts for storm events. By using past model and observational data, model forecast errors can be eliminated, and a better interpretation of the Cobb Method’s accuracy can be determined compared to observations. The results indicate that the Cobb Method is 77.7% accurate to observations without considering errors in observational data. Dividing the study data into three groups of snowfall totals and three groups of snow ratios, the Cobb Method is still shown to have accuracy between 70% and 80%. In an attempt to improve the Cobb Method, two simple linear modifications were made. The two modifications show similar results; underforecasted amounts become more accurate to observations while near exact and overforecasted amounts become more overforecasted. This study shows that the Cobb Method is a step in the right direction towards more accurate snowfall forecasting, and with increased research on the variables affecting snowfall and more accurate model data, the difficulties of forecasting snowfall amounts will become much easier.
Adviser: Mark Anderson