1 - Principles of meteorology and numerical weather prediction

Abstract

Numerical weather prediction (NWP) models are important tools in the process of generating forecasts of wind and solar power output from a farm. Before running an NWP model or being able to interpret its output, however, modelers and forecasters ought to develop an understanding of several foundational principles that undergird a successful NWP forecast. These foundational principles include atmospheric motion, observation sources and quality, data assimilation, the need for postprocessing model output, the value of probabilistic predictions, and how to perform validation and verification of the forecast. Additionally, knowledge about how the NWP model is discretized in space and time, the conditions under which the physical parameterizations have been tested and work well, and the quality of the initial and boundary conditions are all essential to producing a useful forecast. All of these principles are discussed, as is an example of tailoring an NWP model (WRF-Solar) specifically for solar power forecasting.