Big Data Applied to Forecast Short Term Solar Weather

Abstract

In this project, we are trying to improve the accuracy of a certain module that is used in Suncast Model system. It’s called the nowcast system which specializes in short range solar weather forecast. In order for the nowcast system to work, it uses certain key components such as MADcast (Uses mmr to detect cloud density w/o expensive physics models), CIRAcast, Statcast and TSIcast. Our focus of improvement lies in integrating the pollutants available in the atmosphere to improve solar irradiance accuracy. We use the data set into the nowcast module as separate addon, mainly for urban and heavily populated areas using weather stations recorded data and implement physics models for light scatter and absorption. This in theory will let us know how much light is hitting the surface after it has passed the cloud cover. Our focus of improvement lies in integrating the pollutants available in the atmosphere to improve solar irradiance accuracy. We integrate the data set into the nowcast module as separate addon, mainly for urban and heavily populated areas using weather stations recorded data and implement physics models for light scatter and absorption. This in theory will let us know how much light is hitting the surface after it has passed the cloud cover.