Grant: Engage Grants Program

Success in the harvest of crops and their delivery to market depends on knowing when the produce will beready to be picked. Cantaloupe melons are particularly challenging, as the harvest period is short and therequired human and transportation resources demanding. This requires careful planning based on the bestpossible prediction of crop yield. Scotia Weather Services (SWS) wishes to extend their abilities in weatherforecasting to provide forecasts of crop yield to the agri-food industry. The proposed work will use machinelearning technology to develop forecast models that predict melon harvest (number of melons) per day up to atleast two weeks in advance. A standard Data Mining approach will be used to develop and test predictivemodels. The predictive models will forecast daily harvest given the current day of the year, observed andforecasted weather data and field information as inputs. Melon harvest data going back to 2007, provided byVermuelen Farms will be used to build the models. An NSCC weather station will provide observed weatherdata proximal to Vermeulen Farms over the same period and Scotia Weather Services will provide high qualitydaily weather forecasts up to two weeks in advance. The major challenge will be extending accurate harvestpredictions out to two weeks and beyond so as to provide the most distant possible horizon for farm and marketplanning. The majority of the funding will be used to pay two Research Associates. Remaining funds will beused to purchase automated weather monitoring and photography equipment that will be used this year tocapture more detailed site specific data in the field. The long-term goal is to work toward a cloud-basedharvest forecasting system for a variety of crops.