• Vejleder/eksaminator:
  Christoffer Karoff, Geoscience, AU

• Censor:
  Thorkild Maack Rasmussen, Luleå University of Technology, Sverige

Abstract:

Reliable simulations and prediction of past, present and future climate is important to understand the natural variability and the impact of anthropogenic forcing on the postindustrial climate.

At the moment the global climate is changing due to anthropogenic forcing and higher than average temperatures on both a seasonal and monthly scale are frequently observed. These changes are mostly predicted and simulated by numerical climate models such as large global climate model, where the goal is to understand and describe the internal processes.

A alternative method to these slow and at times costly climate model are machine learning models. State of the art machine learning techniques can be valuable for a large variety of prediction problems and other branches of climate change research.