Greater Sage-grouse Least-cost Paths (LCPs) in the SageCon study area, southeastern and central Oregon

This dataset is a component of habitat connectivity analyses for Greater Sage-grouse (GSG) in southeastern and central Oregon conducted in 2014 in accord with the Sage Grouse Conservation Partnership (SageCon) by The Nature Conservancy in Oregon (TNC) – see Jones (2015). Spanning the SageCon Assessment Area within Oregon plus a 10-mile buffer, the data comprise least-cost paths (LCPs) relating to the structural connectivity, or ‘continuity’, of sage-grouse habitat amid the network of habitat patches defined by the ‘lek kernels’ dataset (the nodes of the analysis network). Each LCP identifies the single-cell wide route of least cumulative resistance for a sage grouse moving between a given pair of adjacent lek kernels.

Least-cost paths were identified and mapped with the Linkage Mapper toolset (McRae and Kavanagh 2011) using for inputs the lek kernels data and the cost-weighted distance (CWD) surface. Linkage statistics yielded by Linkage Mapper were then used to characterize the relative ‘quality’ and ‘robustness’ of each LCP.

In context of the broader analysis, LCPs serve both as discrete representations of linkages between adjacent pairs in the lek kernel network and as a conceptual basis for least-cost corridors. LCPs and corridors derive from the spatial configuration and continuity of habitat in the study landscape. While they depict modeled routes of least cumulative resistance, neither necessarily correlate with known routes of sage-grouse migration nor describe the likelihood of particular routes attempted by individuals.

The least-cost approach to modeling connectivity serves to complement the study’s circuit theoretic component in several respects. First, delineation of LCPs provides an intuitive and distinct visualization of the full analysis network. Second, metrics of linkage ‘quality’ and ‘robustness’ (combined in this study into a single metric; see ‘Linkage Statistics’) enable distinct comparison between linkages as represented by the LCPs. Third, corridors demarcate broad belts of land with relatively greater habitat continuity; such ‘linkage zones’ are useful for framing potential conservation actions and for constraining models based in circuit theory.

METHODS

Using Linkage Mapper, a preliminary network of lek kernels (n = 362) was constructed in which linked kernel pairs were defined by adjacency in either Euclidean or cost-weighted space. Minimum accumulated CWDs were calculated between each lek kernel pair, and an LCP mapped for each linkage with the exclusion of those that would intersect an intermediate lek kernel (n = 964).

Network Refinement

The initial analysis network of lek kernels and associated LCPs resulted in a large set of linkage corridors with extensive overlap, making it difficult to proceed with analyses and interpretation. Consequently, a decision was made to further refine the network prior to continuing with the analyses.

With guidance from the Advisory Team, a ruleset was devised to hone the lek kernel network to remove potential linkages of relatively low importance and facilitate interpretation of the remaining individual linkage zones.

Network refinement began with the application of an appropriate threshold CWD value over which LCPs would be removed; this threshold was determined through iterative modification of maximum CWD values with visual review of the resulting networks. Based on the Advisory Team’s recommendations, all LCPs of > 120 cw-km were removed with the exception of three required to maintain a minimum of two linkages for every lek kernel (n = 54); this latter ‘path redundancy rule’was adopted so as to support analysis of at least one alternative movement route with the effective loss of any linkage to fire or other disturbance event. Second, a few linkages (n = 3) were reinstated into the analysis set.

Next, network constellations determined to be presently well-connected were removed from the full analysis network. These linkages (n = 647) and associated lek kernels (n = 149) were first classified as either ‘internal’or ‘external’. Internal linkages (n = 426) were defined as those connecting two lek kernels within the same core area or BBD area and the associated lek kernels (n = 122) defined as those connected only by internal linkages. External linkages (n = 221) were defined as those < 90 cw-km and < 11.3 (Euclidean) km in length, the latter equal to the mean plus one standard deviation of straightline distances (km) measured edge-to-edge between all lek kernel pairs within any single core area. The associated external lek kernels (n = 27) were defined as those connected only by ‘internal’linkages and/or linkages < 90 cw-km and < 11.3 km.

Linkage Statistics

To better inform comparisons between linkages in the refined analysis network (263 LCPs and 213 lek kernels) for conservation planning, statistics for each linkage were used to derive two linkage metrics and, in turn, a single composite linkage index.

The first metric, a measure of linkage ‘quality,’was based inversely on the ‘CWD to Path Length Ratio’, the total cumulative cost along an LCP divided by the Euclidean distance along the same path. This statistic, independent of LCP length, is a measure of the average resistance encountered along an LCP.

The second metric, interpreted as a measure of linkage ‘robustness’, stems from the ‘CWD to Effective Resistance Ratio’. The ‘effective resistance’statistic, calculated using Circuitscape within defined ‘linkage zones’, serves as a measure of the relative isolation of lek kernels that accounts for the availability of multiple movement routes. The ‘CWD to Effective Resistance Ratio,’in turn, can be understood as a measure of average corridor width, the availability of multiple, low-resistance routes within a corridor, and –by extension –the ‘robustness’of the linkage to being severed.

Raw statistic values from LCPs were standardized from 0 –1 to constitute each linkage metric and the metrics then multiplied to produce the ‘linkage index’.

The result served as an integrated measure of both linkage quality and average corridor width; higher values indicate stronger support for more focused consideration of protective conservation action in the linkage at finer scales. Conversely, lower values suggest poor quality and/or tenuous connections where restoration actions may be warranted when considered with the priorities of decision makers.

While the linkage index derives both from a statistic defined at scale the of a single-cell width (LCPs ) and a statistics defined at the scale of the linkage zones, the index was mapped to cells (LCPs) for the sake of greater visual clarity when superimposed over raster model outputs.