Ecosystem Insights through Extreme Values: A Fresh Look at Meteorological Drivers (Papers Track)

Christian Reimers (Department of Biogeochemical Integration, Max Planck Institute for Biogeochemistry); Claire Robin (Department of Biogeochemical Integration, Max Planck Institute for Biogeochemistry); Alexander Winkler (Department of Biogeochemical Integration, Max Planck Institute for Biogeochemistry)

Paper PDF Slides PDF Poster File Cite
Climate Science & Modeling

Abstract

Understanding the influence of meteorological drivers on the ecosystem is a central problem in Earth’s system science. Deriving these influences directly from observations is crucial. However, natural systems often have complex interactions, where multiple drivers influence each other and the target variable simultaneously. These interactions complicate our understanding of individual variable effects. For instance, the relative importance of soil moisture and temperature on net biome production remains unclear, with in-situ measurements and earth’s system models yielding contradicting results. In this work, we propose a novel approach: training on extreme values only. By focusing only on these values, we can better approximate the influences of meteorological drivers. We demonstrate the potential of this approach through a simple example, validating it analytically and empirically.