Land-use is transforming habitats across the globe, thereby threatening wildlife. Large mammals are especiall affected because they require large tracts of intact habitat and functioning corridors between core habita areas. Accurate land-cover data is critical to identify core habitat areas and corridors, and medium resolution sensor such as Landsat 8 provide opportunities to map land cover for conservation planning. Here, we used all availabl Landsat 8 imagery from launch through December 2014 to identify large mammal corridors and assess thei quality across the Caucasus Mountains (N700,000 km ). Specifically, we tested the usefulness of seasonal imag composites (spring, summer, fall, and winter) and a range of image metrics (e.g., mean and median reflectanc across all clear observations) to map nine land-cover classes with a Random Forest classifier. Using image composite from all four seasons yielded markedly higher overall accuracy than using single-season composites (8 increase) and the inclusion of image metrics further improved the classification significantly. Our final land-cove map had an overall accuracy of 85%. Using our land-cover map, we parameterized connectivity models for thre generic large mammal groups and identified wildlife corridors and bottlenecks within corridors with cost-distanc modeling and circuit theory. Corridors were numerous (in total, 85, 131, and 132 corridors for our thre mammal groups, respectively), but often had bottlenecks or high average cost along the least-cost path, indicatin limited functioning. Our findings highlight the potential of Landsat 8 composites to support connectivity analyse across large areas, and thus to contribute to conservation planning, and serve as an early warning system fo biodiversity loss in areas where on-the-ground monitoring is challenging, such as in the Caucasus.
File: Bleyhl_etal_2017_RSE.pdf
This is a publication uploaded with a php script
How organisms respond to climate change during the winter depends on snow cover, because the subniviu (the insulated and thermally stable area between snowpack and frozen ground) provides a refuge for plants, animals and microbes. Satellite data characterizing either freeze/thaw cycles or snow cover are both available, bu these two types of data have not yet been combined to map the subnivium. Here, we characterized global pattern of frozen ground with and without snow cover to provide a baseline to assess the effects of future winte climate change on organisms that depend on the subnivium. We analyzed two remote sensing datasets: th MODIS Snow Cover product and the NASA MEaSUREs Global Record of Daily Landscape Freeze/Thaw Statu dataset derived from SSM/I and SSMIS. From these we developed a new 500-m resolution dataset that capture global patterns of the duration of snow-covered ground (Dws) and the duration of snow-free frozen groun (Dwos) from 2000 to 2012. We also quantified how Dws and Dwos vary with latitude. Our results show that bot mean and interannual variation in Dws and Dwos change with latitude and topography. Mean Dws increase with latitude. Counter-intuitively though, Dwos has longest duration at about 33°N, decreasing both northwar and southward, even though the duration of frozen ground (either snow covered or not) was shorter than tha at higher latitudes. This occurs because snow cover in mid-latitudes is low and ephemeral, leaving longer period of frozen, snow-free ground. Interannual variation in Dws increased with latitude, but the slopes of this relationshi differed among North America, Europe, Asia, and the Southern Hemisphere. Overall, our results show that for organisms that rely on the subnivium to survive the winter, mid-latitude areas could be functionally colde than either higher or lower latitudes. Furthermore, because interannual variation in Dwos is greater at high latitudes we would expect organisms there to be adapted to unpredictability in exposure to freezing. Ultimately the effects of climate change on organisms during winter should be considered in the context of the subnivium when warming could make more northerly areas functionally colder in winter, and changes in annual variation i the duration of snow-free but frozen conditions could lead to greater unpredictability in the onset and end o winter.
File: Zhu_etal_2017_RSE.pdf
This is a publication uploaded with a php script
Green-leaf phenology describes the development of vegetation throughout a growing season and greatl affects the interaction between climate and the biosphere. Remote sensing is a valuable tool to characteriz phenology over large areas but doing at fine- to medium resolution (e.g., with Landsat data) is difficul because of low numbers of cloud-free images in a single year. One way to overcome data availability limitation is to merge multi-year imagery into one time series, but this requires accounting for phenologica differences among years. Here we present a new approach that employed a time series of a MODIS vegetatio index data to quantify interannual differences in phenology, and Dynamic Time Warping (DTW to re-align multi-year Landsat images to a common phenology that eliminates year-to-year phenologica differences. This allowed us to estimate annual phenology curves from Landsat between 2002 and 201 from which we extracted key phenological dates in a Monte-Carlo simulation design, including green-u (GU), start-of-season (SoS), maturity (Mat), senescence (Sen), end-of-season (EoS) and dormancy (Dorm) We tested our approach in eight locations across the United States that represented forests of differen types and without signs of recent forest disturbance. We compared Landsat-based phenological transitio dates to those derived from MODIS and ground-based camera data from the PhenoCam-network The Landsat and MODIS comparison showed strong agreement. Dates of green-up, start-of-season an maturity were highly correlated (r 0.86-0.95), as were senescence and end-of-season dates (r > 0.85) an dormancy (r > 0.75). Agreement between the Landsat and PhenoCam was generally lower, but correlatio coefficients still exceeded 0.8 for all dates. In addition, because of the high data density in the ne Landsat time series, the confidence intervals of the estimated keydates were substantially lower than i case of MODIS and PhenoCam. Our study thus suggests that by exploiting multi-year Landsat imager and calibrating it with MODIS data it is possible to describe green-leaf phenology at much finer spatia resolution than previously possible, highlighting the potential for fine scale phenology maps using th rich Landsat data archive over large areas.
File: Baumann_etal_2017_IJAEOG.pdf
This is a publication uploaded with a php script
Wetland loss is a global concern because wetlands are highly diverse ecosystems that provide important good and services, thus threatening both biodiversity and human well-being. The Paraná River Delta is one of the larges and most important wetland ecosystems of South America, undergoing expanding cattle and forestry activitie with widespread water control practices. To understand the patterns and drivers of land cover change in th Lower Paraná River Delta, we quantified land cover changes and modeled associated factors. We developed lan cover maps using Landsat images from 1999 and 2013 and identified main land cover changes. We quantified th influence of different socioeconomic (distance to roads, population centers and human activity centers), lan management (area within polders, cattle density and years since last fire), biophysical variables (landscap unit, elevation, soil productivity, distance to rivers) and variables related to extreme system dynamics (floodin and fires) on freshwater marsh conversion with Boosted Regression Trees. We found that one third of the freshwate marshes of the Lower Delta (163,000 ha) were replaced by pastures (70%) and forestry (18%) in only 14 years. Ranching practices (represented by cattle density, area within polders and distance to roads) were th most important factors responsible for freshwater marsh conversion to pasture. These rapid and widesprea losses of freshwater marshes have potentially large negative consequences for biodiversity and ecosystem services A strategy for sustainable wetland management will benefit from careful analysis of dominant land use and related management practices, to develop an urgently needed land use policy for the Lower Delta
File: Sica_etal_2016_Science of the Total Environment.pdf
This is a publication uploaded with a php script
File: Alix-Garcia_etal_2016_LandUsePolicy.pdf
This is a publication uploaded with a php script
Weather and climate affect many ecological processes, making spatiall continuous yet fine-resolution weather data desirable for ecological research and predictions Numerous downscaled weather data sets exist, but little attempt has been made t evaluate them systematically. Here we address this shortcoming by focusing on four majo questions: (1) How accurate are downscaled, gridded climate data sets in terms of temperatur and precipitation estimates? (2) Are there significant regional differences in accuracy amon data sets? (3) How accurate are their mean values compared with extremes? (4) Does thei accuracy depend on spatial resolution? We compared eight widely used downscaled dat sets that provide gridded daily weather data for recent decades across the United States We found considerable differences among data sets and between downscaled and weathe station data. Temperature is represented more accurately than precipitation, and climat averages are more accurate than weather extremes. The data set exhibiting the best agreemen with station data varies among ecoregions. Surprisingly, the accuracy of the dat sets does not depend on spatial resolution. Although some inherent differences amon data sets and weather station data are to be expected, our findings highlight how muc different interpolation methods affect downscaled weather data, even for local comparison with nearby weather stations located inside a grid cell. More broadly, our results highligh the need for careful consideration among different available data sets in terms of whic variables they describe best, where they perform best, and their resolution, when selectin a downscaled weather data set for a given ecological application.
File: Behnke_etal_2016_EcologicalApplications.pdf
This is a publication uploaded with a php script
Animal movement patterns in space and time are a central aspect of animal ecology. Remotely-sensed environmental indices can play a key role in understanding movement patterns by providing contiguous, relatively fine-scale data that link animal movements to their environment. Still, implementation of newly available remotely-sensed data is often delayed in studies of animal movement, calling for a better flow of information to researchers less familiar with remotely-sensed data applications. Here, we reviewed the application of remotely-sensed environmental indices to infer movement patterns of animals in terrestrial systems in studies published between 2002 and 2013. Next, we introduced newly available remotely-sensed products, and discussed their opportunities for animal movement studies. Studies of coarse-scale movement mostly relied on satellite data representing plant phenology or climate and weather. Studies of small-scale movement frequently used land cover data based on Landsat imagery or aerial photographs. Greater documentation of the type and resolution of remotely-sensed products in ecological movement studies would enhance their usefulness. Recent advancements in remote sensing technology improve assessments of temporal dynamics of landscapes and the three-dimensional structures of habitats, enabling near real-time environmental assessment. Online movement databases that now integrate remotely-sensed data facilitate access to remotely-sensed products for movement ecologists. We recommend that animal movement studies incorporate remotely-sensed products that provide time series of environmental response variables. This would facilitate wildlife management and conservation efforts, as well as the predictive ability of movement analyses. Closer collaboration between ecologists and remote sensing experts could considerably alleviate the implementation gap. Ecologists should not expect that indices derived from remotely-sensed data will be directly analogous to field-collected data and need to critically consider which remotely-sensed product is best suited for a given analysis.
File: Neumann_etal_MovementEcology_2015.pdf
This is a publication uploaded with a php script
In the former Eastern Bloc countries, there have been dramatic changes in forest disturbance and forest recovery rates since the collapse of the Soviet Union, due to the transition to open-market economies, and the recent economic crisis. Unfortunately though, Eastern European countries collected their forest statistics inconsistently, and their boundaries have changed, making it difficult to analyze forest dynamics over time. Our goal here was to consistently quantify forest cover change across Eastern Europe since the 1980s based on the Landsat image archive. We developed an algorithm to simultaneously process data from different Landsat platforms and sensors (TM and ETM+) to map annual forest cover loss and decadal forest cover gain. We processed 59,539 Landsat images for 527 footprints across Eastern Europe and European Russia. Our results were highly accurate, with gross forest loss producer's and user's accuracy of N88% and N89%, respectively, and gross forest gain producer's and user's accuracy of N75% and N91%, based on a sample of probability-based validation points.We found substantial changes in the forest cover of Eastern Europe. Net forest cover increased from 1985 to 2012 by 4.7% across the region, but decreased in Estonia and Latvia. Average annual gross forest cover loss was 0.41% of total forest cover area, with a statistically significant increase from 1985 to 2012. Timber harvesting was the main cause of forest loss, accompanied by some insect defoliation and forest conversion, while only 7.4% of the total forest cover loss was due to large-scale wildfires and windstorms. Overall, the countries of Eastern Europe experienced constant levels or declines in forest loss after the collapse of socialism in the late 1980s, but a pronounced increase in loss in the early 2000s. By the late 2000s, however, the global economic crisis coincided with reduced timber harvesting in most countries, except Poland, Czech Republic, Slovakia, and the Baltic states. Most forest disturbance did not result in a permanent forest loss during our study period. Indeed, forest generally recovered fast and only 12% of the areas of forest loss prior to 1995 had not yet recovered by 2012. Our results allow national and sub-national level analysis and are available on-line (http://glad.geog.umd.edu/europe/) to serve as a baseline for further analyses of forest dynamics and its drivers.
File: Potapov_etal_RSE_2015.pdf
This is a publication uploaded with a php script
In an effort to increase conservation effectiveness through the use of Earth observation technologies, a group of remote sensing scientists affiliated with government and academic institutions and conservation organizations identified 10 questions in conservation for which the potential to be answered would be greatly increased by use of remotely sensed data and analyses of those data. Our goals were to increase conservation practitioners' use of remote sensing to support their work, increase collaboration between the conservation science and remote sensing communities, identify and develop new and innovative uses of remote sensing for advancing conservation science, provide guidance to space agencies on how future satellite missions can support conservation science, and generate support from the public and private sector in the use of remote sensing data to address the 10 conservation questions. We identified a broad initial list of questions on the basis of an email chain-referral survey. We then used a workshop-based iterative and collaborative approach to whittle the list down to these final questions (which represent 10 major themes in conservation): How can global Earth observation data be used to model species distributions and abundances? How can remote sensing improve the understanding of animal movements? How can remotely sensed ecosystem variables be used to understand, monitor, and predict ecosystem response and resilience to multiple stressors? How can remote sensing be used to monitor the effects of climate on ecosystems? How can near real-time ecosystem monitoring catalyze threat reduction, governance and regulation compliance, and resource management decisions? How can remote sensing inform configuration of protected area networks at spatial extents relevant to populations of target species and ecosystem services? How can remote sensing-derived products be used to value and monitor changes in ecosystem services? How can remote sensing be used to monitor and evaluate the effectiveness of conservation efforts? How does the expansion and intensification of agriculture and aquaculture alter ecosystems and the services they provide? How can remote sensing be used to determine the degree to which ecosystems are being disturbed or degraded and the effects of these changes on species and ecosystem functions?
File: Rose_etal_ConsBio_2015_0.pdf
This is a publication uploaded with a php script
Himalayan forests are undergoing rapid changes due to population growth and economic development and their associated bird communities are among the most threatened and least-studied on earth. In the Chinese Himalaya, traditionally managed Tibetan sacred forests are keystone structures for forest bird conservation. Yet, it remains unclear which fine-scale habitat characteristics of the sacred forests are best associated with Himalayan forest bird species. Our goal here was to quantify the relationship between forest habitat characteristics and bird communities in Tibetan sacred forests to understand habitat associations of common forest birds in the Chinese Himalaya. In 2010 and 2011, we conducted bird point counts and habitat surveys at 62, 50-m radius, sample points distributed within and adjacent to six Tibetan sacred forests in northwest Yunnan, China. From this data, we constructed habitat-occupancy relationship models for 35 bird species and documented tree-use patterns of 14 common arboreal foraging bird species. Our modeling results revealed that large diameter trees and heterogeneity in vertical vegetation structure were the most important habitat characteristics, and were positively associated with occupancy of 63 % of the study bird species. Furthermore, we found that occupancy of eight bird species of conservation concern was related to specific thresholds of forest integrity characteristics. For example, predicted occupancy of three of eight species was high in forested habitats with[15 % bamboo cover and was greatly reduced when bare ground cover exceeded 5 %. We found that bird species foraged on pine (Pinus densata, 58 % more than it was available) and poplar (Populus davidiana, 41 %) in higher proportion to their availability, but that foraging success was highest on fir (Abies spp.), oak (Quercus spp.), willow (Salix spp.) and Chinese Larch (Larix potaninii). Our findings suggest that, although conservation is not a primary management goal of Tibetan sacred forests, these lands harbor critical habitat features for forest breeding birds of the Chinese Himalaya.
File: Wood_etal_2015_Birds_TibetanSacredForests_0.pdf
This is a publication uploaded with a php script
