Seasonal weather prediction for Canadian farmers

点击翻译按钮获取中文摘要

Canadian farmers base almost all crop production decisions by reflecting on short term, medium term and long range weather factors. There are many services available in Canada that provide useful information that farmers can use for short term weather based decisions. Farmers rely on the science of climatology to generate indexes of "climate normal" to satisfy the need for full season and multi-season business planning purposes. However, there is a significant gap in the availability of reliable, medium range weather forecasts, sometimes referred to as trend forecasts, for most important farming regions in Canada. The purpose of this research is to develop and improve regional and site-specific medium range weather forecasts and bridge the gap between short term forecasts, which usually provide weather information for the next ten to fourteen days, and climatology which provides the long term historical averages that are currently used for season long crop model delivery. For a specific example, the Weather INnovations (WIN) company produces phenology models for many crops. These provide a forecast of the timing of crop life cycle stages, using a range of inputs including the location, the planting date and the cultivar used in the field, for which the model is to be produced. Each day in the growing season the crop model is updated using three types of weather files. First, the observed weather from in-field weather stations or interpolated by a network of nearby stations provides that crop's actual weather experience since its planting date. Second, the short term ten to fourteen day forecast is used to update the model for near-term crop development expectations. Lastly, climatology is used for the climate normal data for the balance of the growing season. In the past, expecting the weather to be "average" in the longer term was the best that could be provided to analytical crop models. However, the advances in meteorology can now provide sufficient accuracy in medium range forecast to improve models. This research will test some of those medium range models and statistical processes so crop phenology forecasting can be more accurate in the period between the short term forecast and when crop maturity or harvest occurs.