Brett G. Dickson

USING CIRCUIT THEORY TO MODEL CONNECTIVITY IN ECOLOGY, EVOLUTION, AND CONSERVATION

Connectivity among populations and habitats is important for a wide range of ecological processes. Understanding, preserving, and restoring connectivity in complex landscapes requires connectivity models and metrics that are reliable, efficient, and process based. We introduce a new class of ecological connectivity models based in electrical circuit theory. Although they have been applied in other disciplines, circuit-theoretic connectivity models are new to ecology. They offer distinct advantages over common analytic connectivity models, including a theoretical basis in random walk theory and an ability to evaluate contributions of multiple dispersal pathways. Resistance, current, and voltage calculated across graphs or raster grids can be related to ecological processes (such as individual movement and gene flow) that occur across large population networks or landscapes. Efficient algorithms can quickly solve networks with millions of nodes, or landscapes with millions of raster cells. Here we review basic circuit theory, discuss relationships between circuit and random walk theories, and describe applications in ecology, evolution, and conservation. We provide examples of how circuit models can be used to predict movement patterns and fates of random walkers in complex landscapes and to identify important habitat patches and movement corridors for conservation planning.

INFLUENCE OF VEGETATION, TOPOGRAPHY, AND ROADS ON COUGAR MOVEMENT IN SOUTHERN CALIFORNIA

Abstract Models of individual movement can help conserve wide-ranging carnivores on increasingly human-altered landscapes, and cannot be constructed solely by analyzing the daytime resting locations typically collected in carnivore studies. We examined the movements of 10 female and 7 male cougars (Puma concolor) at 15-min intervals during 44 nocturnal or diel periods of hunting or traveling in the Santa Ana Mountain Range of southern California, USA, between 1988 and 1992. Cougars tended to move in a meandering path (mean turning angle ∼54°), and distance moved (mean and mode ∼300 m) was not correlated with turning angle. Cougars used a broader range of habitats for nocturnal or diel movements than for previously described daybed locations for this same population. Riparian vegetation ranked highest in a compositional analysis of vegetation types selected during movement; grassland, woodland and urbanized sites were least selected. During periods of stasis (we presume many of these were stalking locations), patterns of selection were less marked. Cougars spent a disproportionate amount of time in highly ranked vegetation types, and traveled slowest through riparian habitats and fastest through human-dominated areas. Our results suggest that travel speed may provide an efficient index of habitat selection in concert with other types of analysis. Hunting or traveling individuals consistently used travel paths that were less rugged than their general surroundings. Traveling cougars avoided 2-lane paved roads, but dirt roads may have facilitated movement. Maintenance and restoration of corridors between large wildlands is essential to conserving cougars in southern California. Our results indicate that riparian vegetation, and other vegetation types that provide horizontal cover, are desirable features in such corridors, that dirt roads should not impede cougar use of corridors, and that corridors should lie along routes with relatively gentle topography. Our results suggest that cougars do not key in on highway-crossing structures in a way that creates a prey trap. Our empirical frequency distributions of distances and turning angles, along with cougar responses to vegetation, topography, and roads can help parameterize an individually-based movement model for cougars in human-altered landscapes.

Home-Range and Habitat Selection by Adult Cougars in Southern California

Understanding the impact of habitat fragmentation, roads, and other anthropogenic influences on cougars (Puma concolor) requires quantitative assessment of habitat selection at multiple scales. We calculated annual and multiyear home ranges using a fixed-kernel (FK) estimator of home range for 13 adult female and 2 adult male radiotagged cougars that were monitored October 1986 through December 1992 in the Santa Ana Mountain Range of southern California, USA. Using compositional analysis, we assessed diurnal use of vegetation types and areas near roads at 2 orders of selection (second- and third-order; Johnson 1980). Mean annual and multiyear 85% FK home ranges for males were larger than those reported by previous studies in California. Mean wet-season 85% FK home ranges were significantly larger than those of the dry season. At both scales of selection and across seasons, cougars preferred riparian habitats and avoided human-dominated habitats. Grasslands were the most avoided natural vegetation type at both scales of selection. Although cougar home ranges tended to be located away from high- and low-speed 2-lane paved roads (second-order avoidance), cougars did not avoid roads within their. home range, especially when roads were in preferred riparian areas. Protection of habitat mosaics that include unroaded riparian areas is critical to the conservation of this cougar population.

Models of Regional Habitat Quality and Connectivity for Pumas (Puma concolor) in the Southwestern United States

The impact of landscape changes on the quality and connectivity of habitats for multiple wildlife species is of global conservation concern. In the southwestern United States, pumas (Puma concolor) are a well distributed and wide-ranging large carnivore that are sensitive to loss of habitat and to the disruption of pathways that connect their populations. We used an expert-based approach to define and derive variables hypothesized to influence the quality, location, and permeability of habitat for pumas within an area encompassing the entire states of Arizona and New Mexico. Survey results indicated that the presence of woodland and forest cover types, rugged terrain, and canyon bottom and ridgeline topography were expected to be important predictors of both high quality habitat and heightened permeability. As road density, distance to water, or human population density increased, the quality and permeability of habitats were predicted to decline. Using these results, we identified 67 high quality patches across the study area, and applied concepts from electronic circuit theory to estimate regional patterns of connectivity among these patches. Maps of current flow among individual pairs of patches highlighted possible pinch points along two major interstate highways. Current flow summed across all pairs of patches highlighted areas important for keeping the entire network connected, regardless of patch size. Cumulative current flow was highest in Arizona north of the Colorado River and around Grand Canyon National Park, and in the Sky Islands region owing to the many small habitat patches present. Our outputs present a first approximation of habitat quality and connectivity for dispersing pumas in the southwestern United States. Map results can be used to help target finer-scaled analyses in support of planning efforts concerned with the maintenance of puma metapopulation structure, as well as the protection of landscape features that facilitate the dispersal process.

Fuel-Reduction Treatment Effects on Avian Community Structure and Diversity

Abstract We assessed responses of the breeding bird community to mechanical thinning and prescribed surface fire, alone and in combination, between 2000 and 2006 in ponderosa pine (Pinus ponderosa) forests in northern Arizona, USA. Fuel-reduction treatments did not affect species richness or evenness, and effects on density of 5 commonly detected species varied among species. Populations of some species, such as the western bluebird (Sialia mexicana), increased following burning treatments, whereas others, such as the mountain chickadee (Poecile gambeli), decreased in response to thinning treatments. Our results also identified a temporal response component, where avian community composition and structure changed synchronously on all treatments over time. Given the modest effects these small-scale fuel-reduction treatments had on avian composition and the specific density responses of particular species, our results suggest that land managers should consider implementing prescribed surface fire after thinning projects, where appropriate.

Community occupancy responses of small mammals to restoration treatments in ponderosa pine forests, northern Arizona, USA.

In western North American conifer forests, wildfires are increasing in frequency and severity due to heavy fuel loads that have accumulated after a century of fire suppression. Forest restoration treatments (e.g., thinning and/or burning) are being designed and implemented at large spatial and temporal scales in an effort to reduce fire risk and restore forest structure and function. In ponderosa pine (Pinus ponderosa) forests, predominantly open forest structure and a frequent, low-severity fire regime constituted the evolutionary environment for wildlife that persisted for thousands of years. Small mammals are important in forest ecosystems as prey and in affecting primary production and decomposition. During 2006-2009, we trapped eight species of small mammals at 294 sites in northern Arizona and used occupancy modeling to determine community responses to thinning and habitat features. The most important covariates in predicting small mammal occupancy were understory vegetation cover, large snags, and treatment. Our analysis identified two generalist species found at relatively high occupancy rates across all sites, four open-forest species that responded positively to treatment, and two dense-forest species that responded negatively to treatment unless specific habitat features were retained. Our results indicate that all eight small mammal species can benefit from restoration treatments, particularly if aspects of their evolutionary environment (e.g., large trees, snags, woody debris) are restored. The occupancy modeling approach we used resulted in precise species-level estimates of occupancy in response to habitat attributes for a greater number of small mammal species than in other comparable studies. We recommend our approach for other studies faced with high variability and broad spatial and temporal scales in assessing impacts of treatments or habitat alteration on wildlife species. Moreover, since forest planning efforts are increasingly focusing on progressively larger treatment implementation, better and more efficiently obtained ecological information is needed to inform these efforts.

Multispecies Conservation Planning on U.S. Federal Lands

This chapter reviews past approaches and some significant recent advances in multispecies conservation planning. Most comprehensive conservation strategies are similar in that they invoke a set of key conservation planning principles, for example, the selection of reserve sites is based on characteristics such as their representation, resilience, and redundancy, or complementarity, irreplaceability, and vulnerability. A conservation strategy has representation and complementarity if it provides habitat for each species at one or more locations on the landscape. The design of most comprehensive multispecies conservation planning efforts invokes some form of a “coarse filter” and/or “fine filter” approach. The coarse filter is usually considered to function at broad spatial scales and to reflect underlying ecological processes that are operative over long temporal scales. The fine filter is most often used in reference to individual species or groups of functionally related species. The concept of the mesofilter is introduced to bridge the gap between more traditional coarse and fine filter approaches. The mesofilter concentrates on habitat elements that are too small to be the focus of reserve design strategies, but that often may be limiting the populations of some species. Sample elements retained by the mesofilter include large logs and snags, riparian zones, seeps and springs, and rock outcrops.

A comparison of Landsat TM and MODIS vegetation indices for estimating forage phenology in desert bighorn sheep (Ovis canadensis nelsoni) habitat in the Sonoran Desert, USA

Sonoran Desert bighorn sheep (Ovis canadensis mexicana) occupy rugged upland areas that experience irregular periods of vegetation growth associated with precipitation events. These episodic and often spatially limited events provide important forage and preformed water resources that may be important drivers of animal movement and habitat use. Habitat-use models that incorporate forage phenology would broaden our understanding of desert bighorn ecology and have considerable potential to inform conservation efforts for the species. Field-based methods are of limited utility to characterize vegetation phenology across large areas. Vegetation indices (VI) derived from satellite imagery are a viable alternative, but may be confounded by areas of high relief and shadow effects that can degrade VI values. The varying spatial and temporal resolutions of readily available satellite sensors, such as the Landsat thematic mapper (TM) and moderate-resolution imaging spectrometer (MODIS), present additional challenges. In this study, we sought to minimize degrading effects of terrain on TM- and MODIS-based estimates of vegetation phenology. We compared effects of high topographic relief on time series MODIS- and TM-based VI such as the normalized difference vegetation index (NDVI) and enhanced vegetation index (EVI) using VI departures from average (DA) in shaded and unshaded areas. Sun elevation angle negatively impacted TM-derived NDVI and EVI values in areas of steep terrain. In contrast, MODIS-derived NDVI values were insensitive to sun elevation and terrain effects, whereas MODIS-derived EVI was degraded in areas of steep terrain. Time series MODIS NDVI and EVI DA values differed significantly during months of low sun elevation angle. Average MODIS EVI departure values were ≥20% lower than NDVI under these conditions, confounding time series estimates of plant phenology. Our best results were obtained from MODIS 16-day composited NDVI. These remote-sensing-based VI estimates of seasonal plant phenology and productivity can be used to inform models of habitat use and movements of desert bighorn over large areas.

Modelling and mapping dynamic variability in large fire probability in the lower Sonoran Desert of south-western Arizona.

In the lower Sonoran Desert of south-western Arizona, climate change and non-native plant invasions have the potential to increase the frequency and size of uncommon wildfires. An understanding of where and why ignitions are more likely to become large fires will help mitigate the negative consequences of fire to native ecosystems. We use a generalised linear mixed model and fire occurrence data from 1989 to 2010 to estimate the relative contributions of fuel and other landscape variables to large fire probability, given an ignition. For the 22-year period we examined, a high value for the maximum annual Normalised Difference Vegetation Index was among the strongest predictors of large fire probability, as were low values of road density and elevation. Large fire probability varied markedly between years of moderate and high fine fuel accumulation. Our estimates can be applied to future periods with highly heterogeneous precipitation. Our map-based results can be used by managers to monitor variability in large fire probability, and to implement adaptive fire mitigation at a landscape scale. The approaches we present have global applications to other desert regions that face similar threats from changing climate, altered fuels and potential punctuated changes in fire regimes.

A new model of landscape-scale fire connectivity applied to resource and fire management in the Sonoran Desert, USA

Understanding where and when on the landscape fire is likely to burn (fire likelihood) and the predicted responses of valued resources (fire effects) will lead to more effective management of wildfire risk in multiple ecosystem types. Fire is a contagious and highly unpredictable process, and an analysis of fire connectivity that incorporates stochasticity may help predict fire likelihood across large extents. We developed a model of fire connectivity based on electrical circuit theory, which is a probabilistic approach to modeling ecological flows. We first parameterized our model to reflect the synergistic influences of fuels, landscape properties, and winds on fire spread in the lower Sonoran Desert of southwestern Arizona, and then defined this landscape as an interconnected network through which to model flow (i.e., fire spread). We interpreted the mapped outputs as fire likelihood and used historical burned area data to evaluate our results. Expected fire effects were characterized based on the degree to which future fire exposure might negatively impact native plant community recovery, taking into account the impact of repeated fire and major vegetation associations. We explored fire effects within habitat for the endangered Sonoran pronghorn antelope and designated wilderness. Model results indicated that fire likelihood was higher in lower elevations, and in areas with lower slopes and topographic roughness. Fire likelihood and effects were predicted to be high in 21% of the currently occupied range of the Sonoran pronghorn and 15% of the additional habitat considered suitable. Across 16 designated wilderness areas, highest predicted fire likelihood and effects fell within low elevation wilderness areas that overlapped large fire perimeters that occurred in 2005. As ongoing changes in climate and land cover are poised to alter the fire regime across extensive and ecologically important areas in the lower Sonoran Desert, an analysis of fire likelihood and effects can contribute new and important information to fire and fuels management. Our novel approach to modeling fire connectivity addresses challenges in quantifying and communicating wildfire risk and is applicable to other ecosystems and management issues globally.