Title: Developing diagnostic tools for the next generation of Earth System Models and data sets
Host: Prof. Erik Thomson
Place: Room Vinden (8301 Medicinaregatan 7b) University of Gothenburg, and Zoom https://gu-se.zoom.us/j/67557529829?pwd=GvjdSFB3qKCj7Y74eUboDxASFU7Wz7.1 (Meeting ID: 675 5752 9829; Passcode: 593110)
This seminar is part of the Faculty of Science thematic area seminar A Healthy Living Planet.
Abstract
Earth System Models (ESMs) are the tools that project climate change and offer alternative futures. Typical climate projections are based on scenarios that prescribe forcing agents and are useful for distinguishing analyzing how reducing carbon dioxide emissions can reduce future temperatures. Over the past 50 years, the development of ESMs requires to modify the previous model components, which change the future response to the anthropogenic forcing agents. A critical component of this paradigm is that we can only know how to test these models, which are always using an incomplete model. In the 1980s, model developers were building atmosphere-ocean models and testing them. In the 1990s, the sulfate aerosols were prescribed in the atmospheric models until sulfate emissions were included and this required algorithms to generate sulfate aerosols. In the 2000s, the carbon-cycle and short-lived forcing agents were introduced. In the 2010s, computing power expanded to develop high-resolution ESMs (10-25km grids). Today, we're in the mid-2020s and more model refinements and new model components are being incorporated, while we now have a longer data records that are being archived to test the model components.
This raises some critical questions:
- How do we test these Earth System Models?
- What data are available for testing the models?
- What new diagnostic tools are needed to test if ESMs are working?
In this talk, I will discuss how to test Earth System Models and how to understand whether the results can be useful. In short, the complexity of testing ESMs needs to be assessed along with testing the models.
Welcome!
