GIS in Wildlife Ecology Special Poster Session

Special Posters
ROOM: Galt House, Grand Ballroom & Exhibit Hall
SESSION NUMBER: 26
 

Reconstructing landscapes of ungulate predation
Paige Van de Vuurst
The “landscape of fear” is an ecological theory that is often used to explain behavior modification and landscape use by prey species to avoid predation. However, recent literature has pointed out discrepancies between landscapes that prey species such as ungulates avoid, and landscapes where predation actually occurs. This study explores a landscape of predation based approach, referring to quantifiable landscapes linked to predation events across space, time, and predator taxa. As a case study we used the moose population (Alces alces) of Grand Portage Minnesota, United States. Using GIS, we quantified how satellite-derived environmental data and telemetry derived moose movement data shaped the mortality of 37 moose calves from 2013 to 2015. The observed calf mortality rate averaged 78%, and the largest source of calf mortality (60.5%) was predation caused by American black bears (Ursus americanus) and grey wolves (Canis lupus). Our analysis revealed that landscape heterogeneity use by moose can help to quantify landscapes of predation in space and time at the local level. Furthermore vegetation phenology of adult moose range, cow parturition, and calf depredation, were all consistent and tractable in multidimensional environmental space suggesting that sites of predation may be predictable. This predation focused approach overcomes some limitations and inconsistencies present in landscape of fear assessments, and highlights the need for a standardized theoretical and analytical framework of predation-prey dynamics that account for landscape configuration.
Exploring Connectivity Between HabitatPatches to Identify Conservation Priority Areas for Sonoran Desert TortoisePopulations
Sean Sutor
The border region between the United States and Mexico has been subject to recent intensive human activities that are creating pressure on sensitive natural resources in the biodiverse Sonoran Desert. Landscape alterations resulting in habitat fragmentation have been increasing in the region, with the continued expansion of the US-Mexico border wall presenting a formidable barrier to the movement of desert organisms that are already faced with navigating through a harsh matrix. Anthropogenic activities may thus constrain connectivity (the permeability of a landscape to biotic flows), leading to a reduction in gene flow, population isolation, movement bottlenecks, or other impacts. Understanding the heterogeneity of the landscape and how it may impede or facilitate movement of organisms can help guide management practices and conservation efforts for species that are sensitive to rapid changes in the landscape.
One of these organisms is the Sonoran Desert Tortoise (Gopherus morafkai), a long-lived species that occupies the mountain foothills and incised washes of the Sonoran Desert. Using graph theory and electrical circuit theory in a GIS environment, I modeled structural and functional connectivity within the Sonoran Desert Tortoise range, examining multiple potential scenarios related to impacts of border activities. With these approaches, I have (1) identified important habitat patches and movement corridors to protect, restore, or enhance connectivity for Sonoran Desert Tortoises, (2) ranked habitat patches and movement corridors by their importance for maintaining connectivity, and (3) provided theoretical models indicating the potential impacts of the increasing hardness of an anthropogenic barrier or removal of at-risk, important patches.
Integration of GIS, GPS, Remote Sensing, Drones and Distance Sampling to Estimate White-tailed Deer Densities
Robert E. Kissell, Jr.; Edward L. Warr
Wildlife agencies are increasingly in need of metrics to assess population responses to management actions and diseases. Chronic wasting disease (CWD), is an ever-growing concern as it is being detected in more states over time; for example, Tennessee detected CWD in December, 2018. Efforts to reduce white-tailed deer (Odocoileus virginianus) populations in response to detecting CWD is one approach used and managers need to know if their efforts are effective; density is the preferred metric to guide those managers. Over the last decade Geographic Information Systems (GIS), Global Positioning Systems (GPS), unmanned aerial vehicles (i.e., drones), and infrared imaging have advanced to the point of being able to integrate such that spatial data needed for distance sampling may be easily obtained to estimate density. We used GIS in pre- and post-flight operations, integrated GPS into drone flight planning, and used infrared imagery collected on a drone platform to locate deer along flight lines such that distances from flight lines to deer could be obtained to use in program DISTANCE. We provide an example of density estimation using GIS, GPS, drones, infrared imagery and distance sampling for white-tailed deer in a population near the core area of CWD in Tennessee. We flew 21 transects across approximately 244 ha covering approximately 32% of the property. We programmed flights to follow the terrain to maintain a consistent scale of imagery. We observed 120 deer along 50.8 km of sampled flight lines. Frames with deer were exported, assigned world files, and used in GIS to calculate distances. Density was estimated as 0.457 deer/ha (+/- 0.099 deer/ha), the highest density recorded in Tennessee. Plans for population reduction are scheduled and a comparison will be made to determine if increased harvest was successful. We provide a low cost alternative to manned aerial data collection.
Remote Sensing and Wildlife Management: Using Object Based Image Analysis for the Detection of Nesting Laughing Gulls
Benjamin F. Martini; Douglas A. Miller
Remote sensing has long been used to study wildlife; however, manual methods of detecting wildlife in aerial imagery are often time consuming and prone to human error. Object based image analysis (OBIA) is a promising new technology that addresses both these issues through automation; however, it has not yet been extensively applied to wildlife surveys. We are using the OBIA software eCognition to detect the nests of a breeding colony of Laughing gulls (Leucophaeus atricilla) in Joco Marsh as part of an ongoing monitoring effort at the John F Kennedy International Airport conducted by the USDA. Our technique uses a combination of high resolution 4-band aerial imagery, LiDAR point cloud data, and land cover data as part of a feature extraction ruleset that classifies nest objects using the site, tone, shape, size, and association elements of image interpretation. Preliminary testing on a subset of the imagery revealed that our ruleset was successful at extracting 89% of the nests; however, it also had a high rate of errors of commission due to similarities between the nests and surrounding dead vegetation. After scaling up to the full set of imagery and fine tuning the ruleset using sublevel objects, we achieved a nest extraction rate of 97% while simultaneously reducing errors of commission from 50% to 34%. Although ruleset refinement remains ongoing, our results thus far demonstrate the value of this technology for detecting the nests of colonial nesting shorebirds like the laughing gull, as well as the value of OBIA techniques to wildlife sciences. This approach removes the need to manually search entire sets of imagery for nests and changes the user’s task to verifying output results and eliminating known errors of commission that the software is unable to differentiate from actual nests, a much more efficient and less error prone methodology.
AssessingWatershed Condition and Brook Trout (Salvelinus fontinalis) Distribution Using Variable-width Riparian Buffersand Remote Sensing Data in the Arrowhead Region of Minnesota
Patrick Landisch; Lisa H. Elliott; William Severud; Mark Nelson; Jody Vogeler; Joseph Knight
GIS and spatial analysis are integral tools for understanding interactions between spatiotemporal patterns of landscape characteristics and the distributions of fish and wildlife. Landscape characteristics (e.g., land use, land cover, and disturbance) within riparian areas may be especially important drivers of the distributions of aquatic species. While the National Riparian Areas Base Map (NRABM) provides riparian buffer estimates across broad geographic areas, these variable-width buffers can be refined using the Riparian Buffer Delineation Model (RBDM). This model uses a Lidar-derived 10-m DEM and hydric soils layers to create buffers that more accurately delinate riparian areas. To demonstrate the utility of this spatial approach, we created variable-width riparian buffers for watersheds within Minnesota’s Lake Superior basin, assessed landscape characteristics within and outside NRABM riparian buffers at multiple hydrologic unit code (HUC) scales, and applied these data to a random forest analysis of brook trout (Salvelinus fontinalis) occurrence. Our metrics of watershed-level landscape characteristics included attributes from the 2016 National Land Cover Database, Landsat time series-based forest canopy disturbance data (1974-2018), and delineations of land ownership and protection status. Land cover in the study area’s 12-digit hydrologic unit code (HUC12) watersheds is 82% forested, with 15% disturbed at least once (1974-2018). Variable-width riparian buffers—as quantified from the NRABM—comprise ~17% of the total area and are 89.9% forested, with 9.7% disturbance. Brook trout were recorded in 44% of 174 surveyed HUC12s. Preliminary results from random forest analysis show that percentages of developed and forested cover within NRABM riparian buffers are more important for distinguishing occupied watersheds than are those cover types across the watershed as a whole. We will also compare composition of RBDM riparian areas, and relative importance for brook trout occupancy. This application of riparian and watershed analysis offers a strategic approach for incorporating spatial characteristics into management activities.
Adapting Habitat Suitability in a Dynamic Seasonal Environment
Nicolas Jaffe; William F. Porter
Understanding species densities and distributions is an integral part of wildlife management. Habitat suitability models are commonly applied to address this need. Where coarse-scale species distribution modeling is unsuitable, expert or empirically based habitat assessments may be utilized effectively. However, model utility can be limited by environmental heterogeneity, even where habitat quality is otherwise predictive. This is particularly acute at the edge of a specie’s range. In the Upper Great Lakes Region of North America, white-tailed deer (Odocoileus virginianus) near the northern edge of their range, marked by increased frequency of severe winters. To reduce mortality, white-tailed deer in this region migrate to cedar swamps until spring. However, migration and associated mortality are conditional on multiple aspects of winter severity, which vary over space and time. This presents challenges for estimating annual densities of white-tailed deer. To address these challenges, our study integrates an existing habitat suitability model of white-tailed deer, previously applied in the specie’s central range, with a spatial analysis of winter severity data. We specifically examine seasonal triggers of migration, cumulative winter severity, and timing of spring thaw. Our objectives are to evaluate how spatial structure and interactions between measures of winter severity influence annual variation in local white-tailed deer densities. The model’s spatial and temporal accuracy is validated using annual population estimates from local deer management units. This study demonstrates how a broadly applicable habitat model can be adapted to account for environmental heterogeneity at the edge of a specie’s climatic range, potentially informing future management decisions.
The spatial ecology of dynamic interactions between White-Tailed Deer and hunters in Love County, Oklahoma
Rhiannon Kirton
White-tailed deer (Odocoileus virginianus) are the most widely distributed game species in North America and hunting is key to managing populations. However, hunters also indirectly influence deer behavior, which may have implications on management success. Using a novel dataset containing simultaneous high-resolution GPS tracking data of 83 hunters and 37 white tailed deer during the 2008 and 2009 hunting seasons, I identify encounters between hunters and deer to show how these encounters influence deer behaviour, flight response, and resource selection. The results highlight how deer alter movement in response to interactions with anthropogenic hunters and their subsequent resource use. This work utilizes new methods to gain greater insight into questions that can have important impacts for the knowledge of animal movement, physiology, spatial ecology and wildlife management.
Remote habitat suitability analysis for the mitigation of gophertortoise and solar development conflict in Florida
Alexia J. Goodman
The gopher tortoise (Gopherus polyphemus) is an endemic keystone species to the longleaf pine-wiregrass ecosystem of the Southeastern United States. Over 300 vertebrate and invertebrate species use gopher tortoise burrows to meet life requisites. Gopher tortoises are listed as endangered in their range west of Mobile Bay, Alabama and are being considered for listing in the eastern portion of their range. Current conservation practices for the gopher tortoise are proving successful, but the species has been introduced to a new threat, solar farm development. Solar farm development is increasing in the Southeast because private landowners are willing to lease their land to developers for a fraction of the cost in other areas of the country. Solar farm development requires large areas of open land and reduces the habitat quality for native wildlife. Approximately 80 percent of the gopher tortoise range is owned by private landowners, making it difficult for consistent habitat conservation to occur across their range. This habitat suitability analysis will use various environmental factors and gopher tortoise observation density calculations to determine what private land to prioritize for future conservation efforts in the state of Florida. All habitat suitability criteria were based on published habitat suitability indices used by biologists in the field (Inkley 1986, NRCS 2017). Through the analysis of iNaturalist gopher tortoise observations and data obtained from the USDA Gridded Soil Survey Geographic Database, areas of highly suitable gopher tortoise habitat were identified using ESRI ArcMap 10.7. These areas of highly suitable gopher habitat are recommended to be considered as a priority for exclusion from the future development of solar farms. Continuing federal and state conservation programs, including NRCS conservation easements, and promoting the use of published solar development best management practices will allow the gopher tortoise to coexist with the future of renewable energy.
Detecting Caves with Thermal Imagery and Topographic Analysis Techniques
Jeff S. Jenness; J. Judson Wynne; Derek L. Sonderegger; Timothy N. Titus; Murzy D. Jhabvala
Caves often support sensitive animal populations ranging from bats to subterranean-adapted invertebrates, salamanders, and fishes. Developing a framework to detect caves using GIS and remote sensing could serve as an invaluable conservation tool. To this end, we examined whether cave entrances could be detected in thermal imagery employing methods used for analyzing surface topography. Under ideal conditions, cave entrance temperatures are resolvable in the thermal infrared – appearing as cooler features in daytime imagery and warmer features at night. These thermal anomalies were predicted to appear as distinct “topographic” features on a 3-D representation of a thermal image. We tested this idea using aircraft-borne thermal imagery acquired during two flights (at “predawn” and “midday”) of Pisgah Lava Beds, California using: (1) Topographic Position Index (TPI), (2) thermal gradient (or slope), and (3) curvature. TPI is the relative value of a specific pixel compared to the average thermal value of a neighborhood of pixels, where relatively warm regions present themselves appear as thermal hills, while cooler areas are rendered as thermal valleys. Thermal slope represents the rate of change of thermal values over space, indicating where temperature values change quickly. Curvature represents the rate of change in slope on the “thermal surface” depicted as regions of relatively cool concavity (dips or valleys) and relatively warm convexity (hills and peaks). We determined that these terrain-based analytical methods, which described the “shape” of the thermal landscape, hold significant promise for cave detection. We also compared thermal profiles along transects radiating out from cave entrances to similar profiles conducted at random points on the landscape. Similarly, this analysis revealed distinct signatures of cave entrances. As these findings may be useful in the conservation of sensitive cave-roosting populations, we provide recommendations for future research directions in terrestrial cave detection using thermal imagery.
Effects of Rapid ʻŌhiʻa Death on Root-Obligate Cave Arthropod Communities, Hawaiʻi Island, USA
J. Judson Wynne; Jeff S. Jenness; Asa Aue
On the Hawaiian islands, native animal populations are limited to birds, fishes, one bat species, and invertebrates. As such, endemic arthropod species have received considerable attention from researchers and conservation biologists. To date, perhaps the most prominent elision is an examination of the potential impacts of Rapid ʻŌhiʻa Death (ROD) on root-obligate cave arthropod communities. ʻŌhiʻa (Metrosideros polymorpha), a keystone canopy species of Hawaiian native forests, may also be a keystone species for endemic cave arthropod communities. Unfortunately, over the past several years, a highly aggressive pathogen, Ceratocystis lukuohia (the fungus that causes ROD) has killed hundreds of thousands of ʻōhiʻa trees on Hawaiʻi Island alone. ʻŌhiʻa roots penetrate into the deep zone areas of caves, where these roots can represent the center of the food web for endemic subterranean-adapted arthropod communities. Subsequently, the loss of ʻōhiʻa trees overlying cave systems is expected to have a negative cascading impact on endemic cave arthropod communities. Using several geospatial techniques, we examined the relationship between the distribution of native forests with an ʻōhiʻa tree component, the current distribution of ROD, and caves supporting root-obligate subterranean-adapted endemic arthropod communities. When recording cave locations, researchers record the coordinate data of the cave entrance. However, root-obligate endemic arthropod communities often occur hundreds of meters to kilometers from the entrance; thus the entrance location is not representative of where the communities of interest occur. In this study, we estimated spatial relationships between the underground extent of caves and the above-ground extent of ROD-affected areas using a 3-dimensional analysis of mapped caves to obtain average linear cave extent and average depth from surface. These metrics enabled us to derive average values of linear extent and depth, which we applied to unmapped caves. This approach improved our ability to predict cave communities likely to be affected by ROD.

 
Organizers: Jeff Jenness, Chair, Spatial Ecology and Telemetry Working Group
 
Supported by: Spatial Ecology and Telemetry Working Group

Special Posters
Location: Galt House Date: September 28, 2020 Time: 5:00 pm - 6:00 pm
  • We are closely monitoring the situation regarding COVID-19 and its potential impacts on our conference. We are preparing for all possible scenarios, but at this time we plan to proceed with the conference as scheduled.
toggle