Dynamic process connectivity explains ecohydrologic responses to rainfall pulses and drought

Abstract

Ecohydrologic fluxes within atmosphere, vegetation, and soil systems exhibit a joint variability that arises from forcing and feedback interactions. These interactions cause fluctuations to propagate between variables at many time scales. In an ecosystem, this connectivity dictates responses to climate change, land-cover change, and weather events and must be characterized to understand resilience and sensitivity. We use an information theory-based approach to quantify connectivity in the form of information flow associated with the propagation of fluctuations between variables. We apply this approach to study ecosystems that experience changes in dry-season moisture availability due to rainfall and drought conditions. We use data from two transects with flux towers located along elevation gradients and quantify redundant, synergistic, and unique flow of information between lagged sources and targets to characterize joint asynchronous time dependencies. At the Reynolds Creek Critical Zone Observatory in Idaho, a dry-season rainfall pulse leads to increased connectivity from soil and atmospheric variables to heat and carbon fluxes. At the Southern Sierra Critical Zone Observatory in California, separate sets of dominant drivers characterize two sites at which fluxes exhibit different drought responses. For both cases, our information flow-based connectivity characterizes dominant drivers and joint variability before, during, and after disturbances. This approach to gauge the responsiveness of ecosystem fluxes under multiple sources of variability furthers our understanding of complex ecohydrologic systems.

Debates—Does Information Theory Provide a New Paradigm for Earth Science? Causality, Interaction, and Feedback

“…Since the direct computation of the four unknowns in equation using three equations (Williams & Beer, ) requires an additional condition that is not provided by IT directly, different methods have been developed to compute , , or (see; James et al, , for a review of different methods). Recently, a method for temporal information partitioning has been proposed and used to analyze ecohydrologic interactions based on flux tower and weather station data (Goodwell & Kumar, , ; Goodwell et al, ). This method quantifies redundancy as a function of the mutual dependency between two sources.…”

Entropy and Intermittency of River Bed Elevation Fluctuations

“…Shannon (1948) defined entropy as a measure of variability, uncertainty, and complexity. The concept of entropy has been used in various fields of science and engineering including hydrology and geomorphology, such as basin geomorphology, water distribution systems, surface and subsurface hydrology, and water quality assessment (Fiorentino et al, 1993; Goodwell & Kumar, 2017; Goodwell et al, 2018; Leopold & Langbein, 1962; Pincus, 1991; Singh, 1997; Tejedor et al, 2017). For example, in order to characterize the energy distribution of river networks, Leopold and Langbein (1962) used the entropy concept and suggested that different spatial and temporal scales of landforms contain important information about the energy distributions in river networks.…”

Information transfer from causal history in complex system dynamics

“…For example, a recent study has argued that the spontaneous formation of a self-organized structure is reflected as decrease of joint entropy of the system as well as increase of contemporaneous inter-dependencies among interacting components [2]. However, most of the existing information-theoretic approaches are anchored on characterizing either bivariate information transfer using transfer entropy or momentary information transfer [3][4][5][6][7], or the interactions among a specific set of variables by using methods based on partial information decomposition [8][9][10][11][12], which becomes difficult when more than Last, summary and conclusions are given in Section IV.…”