Leipzig University (DE)
As the global water cycle intensifies, the Earth’s surface will experience more extreme weather and climate events. Increasingly intense and frequent hydrological extremes, such as heavy precipitation events (HPEs), will result in unprecedented alteration of terrestrial ecosystem processes. Prior research has succeeded to track and catalogue several weather phenomena that act as drivers of hydrological extremes, such as atmospheric rivers (ARs). ARs are narrow filaments of extensive water vapour transport in the lower troposphere. While mild-intensity ARs provide vital supply of freshwater, high intensity ARs can cause detrimental impacts along their tracks. Recent advancements in the global catalogisation of ARs offer great predictive potential for hydrological extremes and their impacts on the Earth’s land surface. The AR research community has raised the need for a better understanding of current and future AR-related impacts. Specifically, there is a lack of understanding how the controls of hydrological extremes propagate to changes in land-surface dynamics. The main objective of this project is to gain a better understanding of the interactions between AR-driven hydrological extremes and the the vast suit of terrestrial surface processes. ARNETLAB proposes to understand the complex interplay between atmospheric drivers, extreme weather phenomena and ecosystem impacts using a multi-layer approach. This ambition requires a powerful methodological framework that is, firstly, capable of effectively encoding high numbers of nonlinear interactions and, secondly, merging them together into a joint representation. To this end, we will introduce multilayer networks to the analysis of land-surface dynamics. First, to represent the layer of AR transport, we will develop a novel transport network formalism that characterizes the spatial transport of ARs along their trajectories. This analysis will reveal the ‘global infrastructure network of ARs’, including prominent pathways, basins and regional clusters of AR dynamics. Here, our currently developed novel AR catalogue (0.5°×0.5°, 1940-2022, 6h) will serve as a robust basis. Second, we represent the layer of hydrological extremes via spatiotemporal synchronization patterns, considering seasonal timing and interannual recurrence times. While prior studies have shown that ARs can trigger HPEs along coastlines when they landfall, the degree of their inland penetration still have to be examined more in-depth. The quantification of these phenomena allows us to integrate the third layer, i.e., the land surface. We will systematically introduce complex networks to the study of the Earth’s land ecosystem variables, exploring the full range of remote-sensing derived suite of ESA land data as they are curated in the Earth System Data Lab. To study potential land-atmosphere interactions, we will, for instance, explore ESA soil moisture products, energy fluxes, and vegetation responses. Nonlinear time series analysis measures enable us to link the three layers and obtain a multi-layer representation of the interactions between the Earth’s atmosphere and land-surface processes. Finally, we will examine whether the obtained network representation can be used to design a data-driven adaptive network model that can aid at scenario-based analyses of weather extremes on land-surface dynamics.
