Directional connectivity in hydrology and ecology

Abstract

Quantifying hydrologic and ecological connectivity has contributed to understanding transport and dispersal processes and assessing ecosystem degradation or restoration potential. However, there has been little synthesis across disciplines. The growing field of ecohydrology and recent recognition that loss of hydrologic connectivity is leading to a global decline in biodiversity underscore the need for a unified connectivity concept. One outstanding need is a way to quantify directional connectivity that is consistent, robust to variations in sampling, and transferable across scales or environmental settings. Understanding connectivity in a particular direction (e.g., streamwise, along or across gradient, between sources and sinks, along cardinal directions) provides critical information for predicting contaminant transport, planning conservation corridor design, and understanding how landscapes or hydroscapes respond to directional forces like wind or water flow. Here we synthesize progress on quantifying connectivity and develop a new strategy for evaluating directional connectivity that benefits from use of graph theory in ecology and percolation theory in hydrology. The directional connectivity index (DCI) is a graph-theory based, multiscale metric that is generalizable to a range of different structural and functional connectivity applications. It exhibits minimal sensitivity to image rotation or resolution within a given range and responds intuitively to progressive, unidirectional change. Further, it is linearly related to the integral connectivity scale length--a metric common in hydrology that correlates well with actual fluxes--but is less computationally challenging and more readily comparable across different landscapes. Connectivity-orientation curves (i.e., directional connectivity computed over a range of headings) provide a quantitative, information-dense representation of environmental structure that can be used for comparison or detection of subtle differences in the physical-biological feedbacks driving pattern formation. Case-study application of the DCI to the Everglades in south Florida revealed that loss of directional hydrologic connectivity occurs more rapidly and is a more sensitive indicator of declining ecosystem function than other metrics (e.g., habitat area) used previously. Here and elsewhere, directional connectivity can provide insight into landscape drivers and processes, act as an early-warning indicator of environmental degradation, and serve as a planning tool or performance measure for conservation and restoration efforts.

Reviews and syntheses: on the roles trees play in building and plumbing the critical zone

“…Hydrologic connectivity can be quantified in multiple ways (Larsen et al, 2012;Spence and Phillips, 2015). However, metrics of connectivity that work well in some settings are not always transferable to different locales (James and Roulet, 2007).…”

Linking metrics of landscape pattern to hydrological process in a lotic wetland

“…It measures the strength and directionality of connections between nodes (i.e., vegetation patches). It appears to be diagnostic of degrading ecosystem pattern (Larsen et al 2012), and is more sensitive to changing conditions in the ridgeslough mosaic than the suite of existing landscape pattern indices. While DCI explicitly considers longitudinal connectivity, the link to hydrologic conveyance has yet to be tested.…”

“…Multiple proposed mechanisms for creation and maintenance of the ridge-slough landscape generate anisotropic directional patterning Larsen et al 2007;Larsen and Harvey 2010;Cheng et al 2011;Cohen et al 2011;Heffernan et al 2013). Until recently (Larsen et al 2012), however, none of the landscape pattern metrics used to assess spatial and temporal variation in landscape condition explicitly accounted for longitudinal hydrologic connectivity. Specifically, the explicit link between pattern and hydrology shown in Kaplan et al (2012) is not clearly enumerated by any existing metrics despite the obvious relevance of that link, and despite important progress towards measuring landscape condition using a small number of simple spatial or statistical metrics (Wu et al 2006;Watts et al 2010;Nungesser 2011).…”

“…The second uses a promising approach proposed by Larsen et al (2012) called the directional connectivity index (DCI) which enumerates landscape connectivity based on graph-theory. It measures the strength and directionality of connections between nodes (i.e., vegetation patches).…”

A Random Set Approach for Modeling Integrated Uncertainties of Traffic Islands Derived from Airborne Laser Scanning Points

“…These networks were identifi ed and mapped using bands 3 and 5 of a Landsat ETM + image path 225/row 65 from the year 2000 and digitized on a scale of 1:50 000 (Brandão Jr. and Souza Jr., 2006). The software code is conceptually based on mathematical graph theory, of growing interest in landscape ecology (Larsen et al, 2012). Our application implements graph theory to refl ect the spatial decisions that loggers make in building road networks connecting sites for wood extraction; these spatial decisions, in turn, are modeled by adaptations of Prim's algorithm, a classical optimization model in operations research (Prim, 1957;Miller, 2000).…”