Stochostic Analysis of Spring Meteorological Data in the Upper Midwest

Abstract

The objective of this study was the development of models by which air temperature and precipitation data can be generated for spring periods for use in mathematical runoff models, to predict the range and probability of snowmelt floods. In the upper midwestern United States, spring floods due to a combination of snowmelt and rainfall inflict large amounts of damage. To predict snowmelt floods, information is needed concerning the water equivalent of snow on the ground, soil conditions and hydrometeorological data. The water equivalent of snow and soil conditions can usually be obtained at the time of the forecast. However, it is currently not possible to predict temperature, precipitation and other meteorological factors over a three or four-week interval critical to the spring floods. Simulation of such data by stochastic models should provide a basis for determination of flood probabilities and the range of possible flood magnitudes for current conditions.