Using Weather Records with a Forge Production Model to Forecast Range Forage Production

Abstract

This paper describes a method for calculating site specific forecast yields and their associated probabilities of occurrence. A physically based range forage model, which utilizes beginning soil water content and daily precipitation, mean air temperature, and solar radiation as inputs, calculates the ratio of actual transpiration (T) to potential transpiration (Tp) as a yield index. Annual yield is calculated by the relationship: yield = potential site yield (yield when water is nonlimiting) × T/Tp. By using the current year's beginning soil water content and weather data for a number of years, a population of yields is generated (one yield for each year of weather data). From the population of yields, a mean and various confidence intervals around the mean can be calculated as the forcast yield and its associated confidence intervals. The forecast procedure was tested using 55 years (1917-1971) of weather records and 12 years (1967-1978) of actual yield and soil water data for an upland range site in eastern Montana. An expected two thirds of the field measured yields were within a standard deviation of the forecasted yields for the April, May, and June forecasts.