Start-up seminar by Emy Alerskans

Project name: Hyper-local forecasting system for agricultural applications

"Technological advances within statistics, physics and bio-resources during the last decade make it possible for the agricultural sector to take the next step into the era of intelligent farming. Combining expertise in meteorology, numerical modelling, bio-resources, machine learning and software development lays a strong foundation for successfully reaching the project’s objectives. The demand for more accurate and detailed weather forecasts is ever-increasing and in this aspect, traditional weather forecasts are not keeping up with the demand.

 Multisensor data from Private Weather Stations (PWSs) and related crowdsourced data has the potential to improve weather forecasts significantly. By using PWS data, numerical weather predictions (NWP) can be enhanced through post-processing and data assimilation techniques for crowdsourced data. A dense network of PWSs would, therefore, serve to increase the forecast skill and give downscaled hyper-local forecasts tailored to the customers' needs. The primary objective of the project is to develop a scale-aware post-processing technique for improving NWP model output. Further, this PhD-project will review the possibilities and potentials for including PWS data in short-term weather forecasting. The results of this PhD project will open up new opportunities within the applications of the atmospheric sciences and contribute to research on techniques for improving the performance and detail of short-term forecasts."

Supervisors: Eigil Kaas (PICE, Copenhagen University, John Smedegaard (FieldSense), Andreas Troelsen (FieldSense), Xiaohua Yang (DMI)