US-IALE 2018

AUTHORS: Noé U. de la Sancha, Department of Biological Science, Chicago State University; Nancy E. McIntyre, Department of Biological Sciences, Texas Tech University; Sarah A. Boyle, Department of Biology, Rhodes College

ABSTRACT: Loss of habitat has been identified as one of the main contributors to biodiversity loss and extinction. This is especially true for tropical and subtropical rainforests throughout the world, where forest fragmentation is compromising connectivity in formerly continuous habitats. The Atlantic Forest of eastern Paraguay has experienced extreme deforestation in the past 60 years, resulting in only 8-12% of its original forest cover. In this study, we present a network analysis to identify important forest remnant patches in eastern Paraguay. We quantified structural connectivity for forest remnants at five nested scales, with networks ranging from those remnants at least 5 ha (n = 27,804 remnants) up to remnants at least 25 ha (n = 7579 remnants). We used eight graph theory metrics to assess aspects of network complexity, dispersal-route efficiency, and individual remnant importance in supporting connectivity. Patterns of connectivity varied based on the network, but all analyses revealed that the Atlantic Forest remnants of eastern Paraguay comprise a complex network with high path redundancy. Furthermore, we found that connectivity was constrained for organisms incapable of traveling at least 12 km (farthest nearest-neighbor distance). We were able to identify a few remnants that were consistently recovered as valuable connectivity stepping-stones. We conclude that structural connectivity metrics can be examined at the regional scale to identify important forest remnants for conservation.

AUTHORS: Xavier Haro-Carrión*, Bette Loiselle, Jane Southworth – University of Florida

ABSTRACT: We aimed to investigate how land-cover affects tree diversity in a fragmented landscape ( 19,000 ha) of tropical dry forest in coastal Ecuador. Our objectives were to understand the extent of land-cover types and to evaluate the impact of land-cover and forest fragmentation on tree species diversity. Landsat 2013 (dry season) and 2015 (wet season) images were used to classify the area into old-growth forest, plantation forest (mostly teak), secondary forest, pastureland, water, and built areas. We used a Random Forest Classifier (RF) with 94 predictor variables that included Landsat bands, band ratios, vegetation indices, and image transformations for each season. The proportion of correctly classified RF training points was 95%, and Kappa index was 94%, indicating high accuracy. Dry season variables were predominantly used to classify pastureland and forests. About equal numbers of dry and wet season variables were used to classify forestry plantations. We used the generated map and tree species inventories of old growth forest (N=8) and secondary forest (N=13) to evaluate the impact of land-cover and forest fragmentation on tree `diversity. First, we use RF to identify key patch, landscape composition, and plot variables that contribute to tree diversity as indicated by Hills Numbers. Results indicate that land management variables (e.g. proximity to plantations) are important indicators of species diversity numbers in old growth and secondary forests. We also tested the effect of patch, landscape, plot and species traits on endemic tree species occurrence using a generalized linear mixed model (GLMM). The GLMM used wood density as the only significant predictor. However, distance to road and land-cover successfully helped predict endemic tree species occurrence contingent to species, indicating a differential response of species to human disturbance.

AUTHORS: Żaneta Kaszta*, Wildlife Conservation Research Unit (WildCRU), Department of Zoology, University of Oxford; Samuel A. Cushman, USDA Forest Service, Rocky Mountain Research Station; Andy Hearn, Dawn Burnham and David Macdonald - Wildlife Conservation Research Unit (WildCRU), Department of Zoology, University of Oxford

ABSTRACT: Changes in land use/cover are the main drivers of global biodiversity loss, thus tools to evaluate effects of landscape change on biodiversity are crucial. However, landscape-level approaches are not widely implemented by the planning agencies, mainly due to lack of extensive species-specific data, complexity of ecological guidelines and lack of integrated protocols which would increase their usability. In this study we integrated spatial statistics, landscape ecology and landscape genetics into a GIS-based toolkit, to evaluate impact of development and forest restorations strategies on landscape connectivity and population dynamics of clouded leopard in Malaysian state Sabah. The toolkit was based on an existing clouded leopard habitat suitability model and a development plan for Sabah. Using the cumulative resistant kernel approach and individual-based spatial population genetics models, we evaluated the impacts of 59 scenarios (new segments of highways and railroads, upgraded roads and new forest restoration areas) on landscape connectivity and clouded leopard population dynamics. For most disruptive scenarios we proposed and evaluated alternatives. We found that applying all planned development strategies in Sabah without forest restoration will decrease landscape connectivity by 23% and significantly reduce population size. We therefore strongly advise that development of roads and railways should be paired with forest restorations in strategic places to mitigate negative impacts. Our study was based on clouded leopard in Sabah, however, the toolkit we developed can be applied to any landscapes, ecosystems and species/set of species.

AUTHORS: Maxwell C. Wilson*, School of Life Sciences, Arizona State University; Guang Hu, Department of Landscape Architecture, School of Civil Engineering and Architecture, Zhejiang Sci-Tech University; Lin Jiang, School of Biological Sciences, Georgia Institute of Technology; Mingjian Yu, College of Life Sciences, Zhejiang University; Jianguo Wu, School of Life Sciences and School of Sustainability, Arizona State University, Center for Human-Environment System Sustainability, State Key Laboratory of Earth Surface Processes and Resource Ecology, Beijing Normal University

ABSTRACT: Habitat fragmentation, the loss of habitat in the landscape and spatial isolation of habitat patches, has long been considered a serious threat to biodiversity on global and regional scales. However, the study of fragmentation has been haunted by definitional and conceptual challenges. In particular, few studies have implemented multi-scale approaches or studied the impacts of habitat fragmentation beyond the species level. Here we perform such an analysis, studying the impacts of habitat fragmentation on both species diversity and functional composition of woody plants at local, patch, and regional scales using the Hierarchical Patch Dynamics Paradigm. We find that fragmentation significantly affects plant communities at and between all scales, specifically altering alpha diversity at the patch and landscape scales while driving variation in functional composition across all scales. Further, we find that many of these effects are scale-dependent and that the scale-dependence of fragmentation-biodiversity relationships necessitates the use of multi-scale approaches. We conclude by calling for the increased understanding of scale-dependence in conservation planning and for the continued study of habitat fragmentation beyond the species level.

AUTHORS: Matthew E. Fagan*, University of Maryland, Baltimore County; Douglas C. Morton, Bruce D. Cook, Jeffrey Masek – NASA Goddard Space Flight Center; Feng Zhao, Xi'an University of Science and Technology; Ross F. Nelson, NASA Goddard Space Flight Center; Chengquan Huang, University of Maryland, College Park

ABSTRACT: Commercial pine plantation monocultures in the southeast U.S. annually produce the greatest volume of industrial roundwood in the world. The area of industrial forest has expanded dramatically in recent years, but the potential impact of pine monocultures on the remaining natural forest habitat has not been widely examined. Pine plantations are regularly harvested every 10 to 15 years and may not function as habitat for many species of interest. Here, we integrate a novel map of plantation forest extent in 2011 with historical maps of forest disturbance to assess the fragmentation of natural habitat in the southeastern U.S. If pine plantations are not considered to be forest habitat, forest area decreases by 28%, core forest area decreases by a third, and the area within 30 m of a forest edge increases by 50%. Because of the rapid harvesting and regrowth of pine monocultures, we further characterized the age and intermittency of current forest-plantation and forest-nonforest edges across the region. We found that repeated disturbance and regrowth of regional forests has caused intermittent edges to be the dominant edge type, and that the majority of intermittent edges are associated with extant pine plantations. We conclude that edge effects in regions dominated by industrial forests are potentially widespread, but their magnitude is likely to be contingent on complex historical dynamics of disturbance and regrowth.

AUTHORS: Ho Yi Wan*, Northern Arizona University; Samuel A. Cushman, U.S. Forest Service

ABSTRACT: Globally, human-induced environmental change is degrading habitats and driving biodiversity loss and species extinctions across nearly all taxonomic groups. Threatened and endangered species are particularly vulnerable to habitat loss and degradation. The 2015 IUCN red list identifies habitat loss and degradation as the main threat to 85% of all 79,837 species being assessed. Habitat loss and fragmentation increase spatial isolation of populations, reduce population size, and disrupt dispersal behavior and population connectivity. Protecting habitat that enhances broad-scale connectivity for landscape-level ecological processes such as dispersal and gene flow is a critical task for conserving rare species. Meanwhile, distribution and genetic data on rare species are often obscure because of small population size and low detection probability, and existing data are usually restricted to small or local region. As a result, identifying areas important for conservation over large landscapes can be challenging. In this study, we combined multi-scale optimization habitat selection model, resistance-based connectivity network (UNICOR), and landscape genetics simulation (CDPOP) into a single, spatially explicit framework for identifying broad-scale core habitats, long-distance species dispersal and gene flow patterns. We used the federally listed threatened Mexican spotted owl as an example to illustrate this framework. We identified potential suitable habitats and areas of high connectivity across a large portion of its range in Arizona, and tested how differences in species’ maximum dispersal distance affect prediction in connectivity corridor network and genetic diversity. Our analysis showed a strong positive relationship between landscape connectivity and genetic diversity, suggesting the importance of maintaining corridors and linkages between remnant habitats. We developed maps of suitable habitat, connectivity, and genetic diversity that inform broad-scale management and conservation planning. This framework provides a feasible and powerful solution for assessing effects of human and natural disturbances on species habitats and connectivity at the landscape level.

AUTHORS: Damilola Eyitayo*, Brian C. McCarthy – Ohio University

ABSTRACT: Habitat modification remains one of the greatest threats to biodiversity and achievement of conservation goals require an understanding of how pervasive forms of landscape modification such as utility right-of-way impact plant communities. While most studies on habitat modification have focused on animal species, particularly insects and birds, fewer studies have investigated this topic with respect to plants. In this study, I investigated the impact of utility right-of-way on plant community composition and how changes in environmental variables account for differences in species composition. By sampling vegetation along a 14-km electricity transmission right-of-way in Zaleski State Forest and in an old-field, data on plant species composition and abiotic variables such as soil pH, litter depth, organic matter content and topographic position were collected based on stratified sampling techniques. The utility right-of-way corridor had greater diversity of plant species than adjacent forest and edge habitat areas, while open areas in the old-field were dominated by a few herbaceous species. Differences in species richness between powerline corridors and open old-field areas may be due to disturbance frequency and type of management regime of these areas. In addition, site variables such as aspect and soil organic matter were significant predictors of species richness. These results support a growing sense that powerline corridors can be considered novel habitats, that if properly managed could help achieve biodiversity goals.

AUTHORS: Andrea Santos Garcia*, Center of Nuclear Energy in Agriculture, University of São Paulo, and Global Landscape Initiative, Institute on the Environment, University of Minnesota; Ângela Lúcia Bagnatori Sartori, Federal University of Mato Grosso do Sul; Maria Victoria Ramos Ballester, Center of Nuclear Energy in Agriculture, University of São Paulo

ABSTRACT: Land use change impacts ecological processes due to both the modification of the biophysical component of the landscape and the interference on ecological interactions. Such impacts are well studied in temperate and tropical rain forests, but still poorly addressed in Neotropical Savannas -South America's second largest biome. In this study we aimed to estimate the effects of landscape changes on the reproductive activity of trees. We monitored 52 trees of the Hymenaea stignocarpa species in an area of 2000 ha for one year to analyse phenology, flowering and fruiting activity. This species presents a post-zygotic self-incompatibility and is highly associated with local fauna; thus, fruiting is not a direct product of flowering. The analysed population presented a marked reproductive seasonality. Flowering in edge areas was concentrated in January while in core areas it was in April. But no differences in seasonality between trees located in the fragment core and edge were observed for fruiting. We observed a significantly larger number of trees flowering outside core areas compared to those located inside it. However, overall fruiting activity rates also did not significantly differ between core and edge areas. Reproductive activity was related to landscape characteristics such as the distance to a fragment edge or canopy density. Larger trees located in areas with lower vegetation density were most likely to flower. On the other hand, trees surrounded by higher number of co-specifics and near forest edges were most likely to bear fruit. We argue that flowering and fruiting activity are directly or indirectly related to landscape characteristics and thus, varied spatially. Reproduction activities are dependent on the physiology of each plant. However, flowering is also modelled by availability of light while fruiting is modelled by disperser behaviour. Finally, we emphasize that attention should be paid to such patterns when managing multifunctional landscapes due to the importance of reproduction patterns in biodiversity maintenance.

AUTHORS: Samuel A. Cushman, US Forest Service; David W. Macdonald, University of Oxford; Andrew J. Loveridge, University of Oxford; Laila Bahaa-el-din, University of Kwazulu Natal; Dominic Bauer, University of Oxford; Kristen Kesch, University of Oxford; Helen Bothwell, Northern Arizona University; Nic Elliot, Kenya Wildlife Trust

ABSTRACT: Habitat loss and human-wildlife conflict are major drivers of population decline and extinction, particularly in large carnivores, such as the African lion. Landscape connectivity is critical for maintaining gene flow and reducing demographic stochasticity. While the status of National Parks is relatively secure, areas adjacent to them that link to other protected areas face the threat of conversion to land-uses that exclude wildlife. We used an empirically optimized resistance surface to calculate resistant kernel and factorial least cost path predictions of population connectivity across the Kavango-Zambezi Transfrontier Conservation Area and surrounding landscape. We mapped and ranked the relative importance of (1) dispersal areas outside National Parks, (2) corridors between the key areas, and (3) areas of highest human-lion conflict risk. To increase overall connectivity of the population and reduce human-lion conflict, efforts should be made to minimize anthropogenic disturbance in the identified dispersal areas and corridors, and wildlife-related activities should be encouraged to prevent land-use conversion.

AUTHORS: Grant Paton*, The University of North Carolina at Charlotte; Alexandra Shoffner, Michigan State University; Andrew Wilson, Gettysburg College, Sara Gagné, The University of North Carolina at Charlotte

ABSTRACT: As humans continue moving to urban areas, there is a growing need to understand the effects of urban intensification on native wildlife populations. Forest species in remnant habitat are particularly vulnerable to urban intensification, but the mechanisms behind these effects are poorly understood. The objective of our study was to identify the traits that best explained variation in the responses of forest bird species to urbanization intensity. To do so, we used occurrence data for 58 forest species derived from 16,541 forested point counts from the Second Pennsylvania Breeding Bird Atlas. For each species, we estimated the effect of urbanization intensity on occurrence at each of ten landscape scales, controlling for other measures of landscape heterogeneity, local habitat attributes, and species detectability, using generalized linear models and an information theoretic approach. From these analyses we extracted the largest effect size for urbanization intensity across scales for each species. Effect sizes ranged from -1.49 to 0.90, with a mean ± SE of -0.36 ± 0.49. We then collated data on multiple traits for each species from field guides and published literature. We also collected song traits from species recordings. Trait data will be used as predictors of variation in urbanization intensity effect sizes among species in phylogenetic regression models. Preliminary results indicate that use of nest cavities, clutch size, clutches per year, fledgling success, sociality, frugivory, granivory, lifespan, nesting height, offspring’s duration in nest, omnivory, biparental nest construction, sedentarism, mean song frequency, and song length are positively associated with urbanization intensity effect sizes, whereas foraging height, song frequency range, and territory size are negatively associated. Our results will identify species traits that facilitate or hinder a species’ ability to utilize urban environments. Species with inhibiting traits can then be prioritized in urban conservation programs.

AUTHORS: Brandon MacDougall*, Heather Sander – University of Iowa

ABSTRACT: Urban habitats, however small, fragmented, or degraded they may be, are becoming increasingly important areas for species conservation. In recent years, patterns of urban biodiversity have been identified and described for many groups of organisms; however, these studies have generally been conducted in large urban conglomerates. Much less is known about these patterns of diversity for small cities, as well as for cities located where expansive, non-urban habitat degradation has occurred, such as the vast agricultural region of the American Midwest. This lack of knowledge is an impediment to species conservation in these cities, where a higher proportion of tree cover and a greater mix of habitat diversity (when compared to the surrounding landscape) may provide unique opportunities for native species that remain. This study strives to address the impacts of urbanization on meso-predator species presence in an urban area located within an agricultural matrix. We monitor the presence and distribution of these species using motion-detecting cameras, with sampling sites covering a wide spectrum of variability in landscape structure along an urban-rural gradient in Iowa City, Iowa. Meso-predator species in this area vary in their response to urbanization; however, overall species diversity is positively correlated with site variables such as patch size and distance to riparian corridors, and less related to attributes corresponding to urbanization intensity. These findings can support efforts to retain ecological balance in human-dominated landscapes through the preservation of a wide range of meso-predator species.

AUTHORS: Hayley R. Tumas*, Brian Shamblin, Nathan P. Nibbelink – University of Georgia; Mark Woodrey, Grand Bay National Estuarine Research Reserve; Richard Chandler, Campbell J. Nairn – University of Georgia

ABSTRACT: Common species important for ecological restoration and ecosystem processes are understudied in landscape genetics, but stand to lose as much genetic diversity as rare species from habitat fragmentation. Genetic diversity, maintained in fragmented populations through gene flow, is critical for successful restoration and conserving adaptive potential in a changing climate. Salt marshes are ecosystems valuable to humans and wildlife that are in widespread decline nationally and globally. Black needlerush (Juncus roemerianus Scheele) is the dominant foundational plant species in northeastern Gulf of Mexico salt marshes, and a target species for restoration in the area. We used 19 microsatellite markers to measure levels of genetic and genotypic diversity and characterize population structure in this important clonal macrophyte. Least cost transect analysis (LCTA) was used within a model selection framework to delineate potential dispersal pathways and identify landscape factors (wetland, ocean, forest cover, developed land) influencing population connectivity. Genetic and genotypic diversity were greater than expected for a species believed to use predominantly clonal reproduction, and samples structured into two admixed clusters across the study area. The proportion of developed land across coastal transects and Euclidean distance had the greatest influence on population connectivity, and coastal transects were identified as important for dispersal. Results have important implications for restoration and management techniques that aim to preserve adaptive potential in J. roemerianus and sustain ecosystem processes throughout the salt marsh. This study expands upon an understudied taxon and ecosystem, and demonstrates methodology that could be applied to other common, understudied species.

AUTHORS: Stephanie Burgess*, University of Mississippi; Ryan Garrick, University of Misssissippi

ABSTRACT: Knowledge of movement patterns and the manner in which wildlife interact with their landscape are crucial to wildlife management. Conventional methods of observing wildlife movement, including game cameras and drift fences, are costly, labor intensive, and difficult to conduct at a large scale. Landscape genetic analyses can be used to identify landscape features that cause resistance to wildlife movement with less sampling while encompassing large areas in a single analysis. However, there are several knowledge gaps within the field of landscape genetics that must be addressed before broad conservation application is possible. Specifically, little is known about the transferability, or applicability to nearby areas, of landscape genetics models. Using multispectral imaging, we created land cover classifications for two forests, Holly Springs National Forest (HSNF) and William B Bankhead National Forest (BNF), within the species range of Plethodon mississippi. We then used microsatellite genotypes from P. mississippi individuals within each forest to create landscape resistance models. To test the transferability of the models, the parameters of the HSNF model were applied to the BNF landscape and tested for significance against BNF genetic data, and vice versa.Land cover analysis has shown that although both forests are within the species range for P. mississippi, they have different land cover ratios. While the land cover within both forests is primarily pine and hardwood forest, BNF has 9% more wetlands than HSNF, and HSNF has 8% more agricultural land. Preliminary analyses have shown man-made structures impede salamander gene flow in HSNF, and agricultural lands increase gene flow. If, when applied to the BNF landscape, man-made and agricultural properties produce similar effects, it would support the use of amphibian landscape genetics models across the species’ variable range. Conversely, if they do not produce similar effects, it would indicate the models are location-specific.

AUTHORS: Indranil Mondal*, Pallavi Ghaskadbi, Zehidul Hussain, Gautam Talukdar, Parag Nigam, Bilal Habib – Wildlife Institute of India

ABSTRACT: The Eastern Vidarbha Landscape (EVL) is located in the centre of India. It has a human population of 11 million people, but at the same time, it has a tiger population of 150-200 individuals in forests that cover 45% of its total geographical area. Forests outside protected (PAs) areas are very fragmented and under continuous anthropogenic pressure. EVL has a very prominent human-wildlife interface, and a high number of conflict cases are reported every year, where losses are suffered by both humans and wildlife. Nonetheless, EVL, consisting of 6 protected areas, forms a pivotal role in connecting the Central Indian tiger populations. It harbours a large number of tigers even outside PAs and gives us a unique opportunity to study them in a human-dominated environment.We used data on tiger presence from outside PAs and habitat variables to model corridors (9,370 sq. km.) and 9 pinch-points using Circuit Theory. These corridors were validated (76% accurate) using tracking data from 5 dispersing radio-collared tigers. We used the data from radio-collars to study the time spent by the dispersing tigers using a Linear Time Density Function and identified pockets in the corridors where they were spending a considerable amount of time. Based on the habitat characteristics of these pockets we extrapolated the results to all corridor areas in entire landscape, to identify areas for prioritized conservation action. Several tiger mortalities are reported from EVL due to electrocution of individuals at illegally electrified guard fences around agricultural fields. Therefore we modelled the electrocution potential for the entire landscape and identified 121 and 1184 villages in Very High Risk and High-Risk categories respectively, to help managers tackle the issue. In the end, we propose a model of triage to raise funds from unconventional sources to aid corridor conservation in penny-scarce scenarios.

AUTHORS: Hongbo Yang, Jianguo Liu – Michigan State University

ABSTRACT: Protected areas have long been considered as the cornerstone of ecological conservation and have been expanding rapidly around the world. While designating more land as protected areas is important, there is a growing concern about the degree to which existing protected areas are effective in achieving desirable ecological outcomes and how their effectiveness is affected by socio-ecological factors. However, at large spatial scales, the ecological effectiveness of protected areas and the underlying determinants remain unclear. Here, we evaluated the effectiveness of each of the protected areas in China established before 2000 in reducing deforestation – one of the biggest threats to global ecosystem health – between 2000 and 2015. We found that, at varying degrees, the majority of protected areas were identified as effective in reducing deforestation. Without their establishment, an area of about 160,000 hectares would have been deforested, an amount of forest that can sequestrate about 120 Megaton of carbon annually. Protected areas in regions with high deforestation pressures often contribute more to reducing deforestation. Some other attributes of protected areas, such as protection strictness, are also correlated with their effectiveness in reducing deforestation. These results provide important information to improve the effectiveness of protected areas and fully realize the promise they provide for biodiversity conservation in China and beyond.

AUTHORS: Mason Fidino and Seth Magle - Institution: Urban Wildlife Institute, Lincoln Park Zoo, Chicago

ABSTRACT: Habitat fragmentation via urbanization can significantly alter connectivity for wildlife populations, and therefore the collective ability of a species to persist in urban habitat remnants. To explore this relationship, we used a long-term dataset spanning 15 years to assess trends in the metapopulation dynamics of the black-tailed prairie dog, a potential keystone species, in 384 urban habitat remnants in Denver, Colorado, USA. We developed a Bayesian multi-state model to predict the annual probability that 1) prairie dogs locally colonize an undeveloped habitat remnant, 2) prairie dogs go locally extinct in an undeveloped habitat remnant and 3) a habitat remnant is developed. Both available habitat remnants and number of prairie dog colonies declined precipitously over the course of study, though we also monitored new colonization events, which were most correlated to habitat connectivity. Between 2002 – 2007 an average of 7.74% of sites were developed annually, however, observed trends indicate that this development rate may be slowing. Despite this, the rate of decline of urban prairie dogs estimated by our model closely reflected the observed data, and extinction rates generally exceeded colonization rates for this species across all years. Using data on declines during the study we project a 40% probability that prairie dogs will be extirpated from this area by 2067, though that probability could range as high as 50% or as low as 20% based on differing development scenarios. We suggest that prairie dog fulfill important ecological roles in urban landscapes, and could persist in the Denver area with appropriate management and habitat protections.