Uncertainty Quantification and Attribution in Models of the Earth System

The recent State of Global Water Resources report by the World Meteorological Organization concluded that the hydrological cycle is spinning out of balance due to climate change and human activities. Floods, droughts, landslides and wildfires have already been the most damaging natural disasters in the first two decades of this century. Simulation models enable us to estimate past, current and potential future states of the hydrological cycle, including its extremes, so that we can guide policy making and support adaptation planning in search for a resilient society. But models are of limited use if we do not understand their robustness and if we cannot estimate the uncertainties they carry in their predictions. Additionally, attributing model uncertainty to its causes is a key approach to identify needs for scientific advancement. We develop and apply methods for uncertainty quantification and attribution in Earth system models to identify knowledge gaps and to provide robust policy support. I will provide examples from our work on advancing global sensitivity analysis algorithms and software for improved uncertainty quantification/attribution in Earth system science and beyond.