Constraining the Large Uncertainties in Earth System Model Projections with a Big Data Approach
Climate scientists around the world develop and use highly complex Earth system models to project future climate change. As the number and level of complexity of these model increase, it is becoming impractical to use traditional tool to analyze the model outputs, and in turn could hinder the discovery of new knowledge that could be critical the society when dealing with future climate change.
The project COLUMBIA is an interdisciplinary project that aims to develop an innovative tool, based on the state-of-the-art machine learning technology, to efficiently analyze large amount of model data to better understand
why some models behave very differently than the others. Combined with our current knowledge of how the climate system works based on the observational evidence, we will constraint the large spread in these model simulations. Also using our novel tool, we hope to filter out those climate models that do not represent well the important dynamics observed in nature. Eventually, optimal future climate projections at global and regional scale
will be achieved using only the best performing models. The project team plans to apply the new tool to determine the tropical Pacific variability and its connection to climate in Europe. The dynamic#of ocean's heat and carbon budgets, and their associated impacts for future climate change will be explored as well.
Activities in the project will strengthen the interactions between climate modeling and observational communities as well as create new interdisciplinary bridge between natural and computational scientists. The proposed work is also well timed as the process for upcoming IPCC-AR6 has just begun and the first batch of new Earth system model simulations will become available in late 2018. New knowledge from COLUMBIA will support the advancement of climate science both nationally and internationally.