A Procedural Approach to Forecasting Nocturnal Air Temperature for Frost Prediction

Accurate forecasting of nocturnal air temperatures for frost prediction can lead to the development of effective strategies for damage reduction to the crops due to seasonal frost formation. A Real-time forecasting system must be able to run on low power computing machines and must be able to be deployed in the field where sight specific historic meteorological observations may not be available. This paper presents a procedural approach to forecast nocturnal air temperature for frost prediction using fuzzy matching to extrapolate regression function structural parameters from limited historical data. The paper also presents a case study of frost formation in Shepparton, Australia.