Measuring the Accuracy of Hyperlocal Forecasting through Onset of Precipitation Predictions

Students from Concord-Carlisle Weather Services performed an independent study of the accuracy of the hyperlocal forecasting application DarkSky and other hyperlocal forecasters by examining predictions for the time of onset of precipitation. Participants collected data by recording the forecasted onset time, and the observed actual onset of precipitation at the location of prediction. The totality of collected data was used to create a comprehensive view of the precision of onset of precipitation predictions relative to other forecasting systems and to offer a perspective on the accuracy and usefulness of this type of application and the emerging field of hyperlocal forecasting that depends on large data sets and computational capabilities instead of a team of meteorologists to deliver a forecast.

This public-private collaboration between DarkSky and Concord-Carlisle High School will benefit both partners. DarkSky will gain valuable feedback on the accuracy of their product. Students engaged in this unique opportunity will be exposed to questions and problems in a new frontier in weather forecasting as it evolves to meet the needs of a twenty-first century world. Students will gain experience in the NGSS science and engineering practices such as planning and carrying out investigation, analyzing and interpreting data, and in sharing their findings at AMS they will be practicing obtaining, evaluating and communicating information. In addition, students will learn how science and technology concepts translate to the business world. As such, this project offers a means of integrating aspects of STEM education, and aligns with the ideals of the NGSS in focusing on technology and the applications of science.