2022 ARCHIVES: Poster Presentations
At this year's virtual conference, we are offering poster presenters the opportunity to provide a 2-3 minute recorded lightning presentation about their poster! These presentations will be shown during a poster session each day from 11:45 AM to 12:30 PM (Eastern). Poster presenters will also setup virtual booths where they can post their poster and abstract, related images, links, and their recorded lightning presentation. Attendees can visit the virtual booths, connect with poster presenters directly, or via the text-based chat in their booth.
- DAY ONE - Remote Sensing Data and Applications {Monday, April 11, 2022; 11:45 am-12:30 pm ET}
- DAY TWO - Characterizing and Interpreting Landscapes {Tuesday, April 12, 2022; 11:45 am-12:30 pm ET}
- DAY THREE - Landscape Processes and Biodiveristy {Wednesday, April 13, 2022; 11:45 am-12:30 pm ET}
DAY ONE - Remote Sensing Data and Applications
Monday, April 11, 2022
Quantifying the Landscape Configuration and Composition of Tamaulipan Thornscrub in South Texas using UAVs
Student-Graduate
Authors: Lori D. Massey, Caesar Kleberg Wildlife Research Institute, Texas A&M University-Kingsville, Kingsville, TX, USA; Humberto L. Perotto-Baldivieso, Caesar Kleberg Wildlife Research Institute, Texas A&M University-Kingsville, Kingsville, TX, USA; Evan Tanner, Caesar Kleberg Wildlife Research Institute, Texas A&M University-Kingsville, Kingsville, TX, US; J. Alfonso Ortega S., Caesar Kleberg Wildlife Research Institute, Texas A&M University-Kingsville, Kingsville, TX, USA
Description: Tamaulipan thornscrub is a subtropical semi-arid vegetation community that provides valuable habitat ecologically and economically across south Texas and northern Mexico. The focus of my research is to use unmanned aerial vehicle’s (UAVs) to quantify the landscape configuration and composition of Tamaulipan thornscrub across south Texas. My specific objectives are 1) use spectral signatures to identify different brush species present in the vegetation community 2) calculate the composition of brush species found in Tamaulipan thornscrub 3) quantify the landscape configuration across south Texas for Tamualipan thornscrub. For this research, I established two study sites in the Tamaulipan thornscrub range and will fly a UAV with a multispectral camera at two altitudes 50 meters and 100 meters above ground level (AGL). We will also use collector and a handheld GPS unit to identify thornscrub species and their locations in 5 m x 5 m quadrants throughout the sites. Orthomosaics created will be used to determine spectral signatures for specific vegetation present in the thornscrub community. Digital libraries will be created for Tamaulipan thornscrub vegetation, that can be used to classify imagery. These libraries will also be used with Sentinel-2 imagery to classify thornscrub vegetation at a larger scale. Allowing us to quantify the landscape configuration and distribution of Tamaulipan thornscrub across south Texas, this research will show the potential use of very-high resolution imagery acquired from UAVs in rangeland management studies. This is one of the first studies that will help quantify the composition and spatial distribution of Tamaulipan thornscrub in south Texas, which may be used for selecting locations where habitat management practices may be applied to yield better results.
Student-Graduate
Authors: Lori D. Massey, Caesar Kleberg Wildlife Research Institute, Texas A&M University-Kingsville, Kingsville, TX, USA; Humberto L. Perotto-Baldivieso, Caesar Kleberg Wildlife Research Institute, Texas A&M University-Kingsville, Kingsville, TX, USA; Evan Tanner, Caesar Kleberg Wildlife Research Institute, Texas A&M University-Kingsville, Kingsville, TX, US; J. Alfonso Ortega S., Caesar Kleberg Wildlife Research Institute, Texas A&M University-Kingsville, Kingsville, TX, USA
Description: Tamaulipan thornscrub is a subtropical semi-arid vegetation community that provides valuable habitat ecologically and economically across south Texas and northern Mexico. The focus of my research is to use unmanned aerial vehicle’s (UAVs) to quantify the landscape configuration and composition of Tamaulipan thornscrub across south Texas. My specific objectives are 1) use spectral signatures to identify different brush species present in the vegetation community 2) calculate the composition of brush species found in Tamaulipan thornscrub 3) quantify the landscape configuration across south Texas for Tamualipan thornscrub. For this research, I established two study sites in the Tamaulipan thornscrub range and will fly a UAV with a multispectral camera at two altitudes 50 meters and 100 meters above ground level (AGL). We will also use collector and a handheld GPS unit to identify thornscrub species and their locations in 5 m x 5 m quadrants throughout the sites. Orthomosaics created will be used to determine spectral signatures for specific vegetation present in the thornscrub community. Digital libraries will be created for Tamaulipan thornscrub vegetation, that can be used to classify imagery. These libraries will also be used with Sentinel-2 imagery to classify thornscrub vegetation at a larger scale. Allowing us to quantify the landscape configuration and distribution of Tamaulipan thornscrub across south Texas, this research will show the potential use of very-high resolution imagery acquired from UAVs in rangeland management studies. This is one of the first studies that will help quantify the composition and spatial distribution of Tamaulipan thornscrub in south Texas, which may be used for selecting locations where habitat management practices may be applied to yield better results.
Extending the timeline of land cover change data to refine our understanding of grassland degradation in northern China
Student-Graduate
Authors: Brooke Iacone, George Washington University; Ginger Allington, George Washington University; Ryan Engstrom, George Washington University; Mia Murray, George Washington University
Description: High-resolution CORONA imagery acquired by the United States through spy missions in the 1960’s may provide critical quantitative insight into historic land cover change in prior decades, particularly in regions such as northern China, which have experienced significant grassland degradation and desertification. Since the 1960’s, the region has undergone many policy changes that have led to periods of intensified cultivation and, in recent years, restoration. However, growing demand for resources from China’s grasslands and the impacts of climate change continue to threaten the long-term health of these vital ecosystems. Most large scale studies of land cover change only extend back to the 1980s, due to the availability of remotely sensed data, but this may be after significant changes had already occurred. CORONA imagery presents an opportunity to expand the timeline of available data on land use and land change in northern China. However, CORONA imagery is single-band, so we cannot derive the same information on surface conditions that is available from multi-spectral sources. Here, we evaluate the feasibility and accuracy of using derived texture and spatial contextual features from single band images for classification. Contextual features generate bands from statistical quantification of features such as edge patterns, pixel groups, gaps, and texture. This approach has the potential to increase the amount of quantitative information derived from CORONA imagery, furthering our understanding of historic land cover types.
In this study, four CORONA images were georeferenced using ground control points and Landsat TM imagery. Contextual features were derived at several scales from a mosaic of the georeferenced images and combined with a Landsat MSS NDVI to create a multi-band image. We used an unsupervised clustering algorithm, paired with cluster statistics to derive a final set of classes. Preliminary results suggest that there are several distinct classes which were accurately classified by the contextual features, particularly urban and agricultural areas. These classes, along with visual interpretation of the mosaic and modern imagery, reveal that urban areas, agricultural areas and exposed dunes have increased significantly since the 1960’s, while bodies of water and vegetated dunes have decreased. The georeferenced mosaic and extracted land cover classes from this study can also inform future change detection analyses of Naiman Banner.
Student-Graduate
Authors: Brooke Iacone, George Washington University; Ginger Allington, George Washington University; Ryan Engstrom, George Washington University; Mia Murray, George Washington University
Description: High-resolution CORONA imagery acquired by the United States through spy missions in the 1960’s may provide critical quantitative insight into historic land cover change in prior decades, particularly in regions such as northern China, which have experienced significant grassland degradation and desertification. Since the 1960’s, the region has undergone many policy changes that have led to periods of intensified cultivation and, in recent years, restoration. However, growing demand for resources from China’s grasslands and the impacts of climate change continue to threaten the long-term health of these vital ecosystems. Most large scale studies of land cover change only extend back to the 1980s, due to the availability of remotely sensed data, but this may be after significant changes had already occurred. CORONA imagery presents an opportunity to expand the timeline of available data on land use and land change in northern China. However, CORONA imagery is single-band, so we cannot derive the same information on surface conditions that is available from multi-spectral sources. Here, we evaluate the feasibility and accuracy of using derived texture and spatial contextual features from single band images for classification. Contextual features generate bands from statistical quantification of features such as edge patterns, pixel groups, gaps, and texture. This approach has the potential to increase the amount of quantitative information derived from CORONA imagery, furthering our understanding of historic land cover types.
In this study, four CORONA images were georeferenced using ground control points and Landsat TM imagery. Contextual features were derived at several scales from a mosaic of the georeferenced images and combined with a Landsat MSS NDVI to create a multi-band image. We used an unsupervised clustering algorithm, paired with cluster statistics to derive a final set of classes. Preliminary results suggest that there are several distinct classes which were accurately classified by the contextual features, particularly urban and agricultural areas. These classes, along with visual interpretation of the mosaic and modern imagery, reveal that urban areas, agricultural areas and exposed dunes have increased significantly since the 1960’s, while bodies of water and vegetated dunes have decreased. The georeferenced mosaic and extracted land cover classes from this study can also inform future change detection analyses of Naiman Banner.
Mapping Vegetation Types at Lake Nakuru National Park in Kenya's Rift Valley and Quantifying Landcover Change
Authors: John Maingi, Miami University; McNichol Kaloki, Kenya Wildlife Service
Description: Lake Nakuru National Park (LNNP) located in Kenya's Rift Valley is known for its high biodiversity and proximity to major urban areas. Biodiversity has, however, been in decline from urban expansion, habitat loss, invasive species, climate change, and pollution from agriculture and industry. A 1986 floristic survey coupled with a visual interpretation of aerial photographs and Landsat MSS prints delineated 24 vegetation types in LNNP. The Park has since become more insularized through expanding urbanization and agricultural settlements in the park surrounds, and completion of a high voltage fence around the park. Populations of ungulates in the park have expanded, increasing risk of overgrazing and land degradation. There is a need to update the 1986 vegetation map of the park and document how habitats within the park have changed. In the current study, we evaluate the utility of multitemporal Landsat OLI, ASTER, and Sentinel-2 imagery to delineate ten broad vegetation types in LNNP. We then use a combination of Landsat MSS, TM, ETM+, and OLI images to document land use and land cover change in the LNNP area and surrounds for the 1978-2021 period.
Authors: John Maingi, Miami University; McNichol Kaloki, Kenya Wildlife Service
Description: Lake Nakuru National Park (LNNP) located in Kenya's Rift Valley is known for its high biodiversity and proximity to major urban areas. Biodiversity has, however, been in decline from urban expansion, habitat loss, invasive species, climate change, and pollution from agriculture and industry. A 1986 floristic survey coupled with a visual interpretation of aerial photographs and Landsat MSS prints delineated 24 vegetation types in LNNP. The Park has since become more insularized through expanding urbanization and agricultural settlements in the park surrounds, and completion of a high voltage fence around the park. Populations of ungulates in the park have expanded, increasing risk of overgrazing and land degradation. There is a need to update the 1986 vegetation map of the park and document how habitats within the park have changed. In the current study, we evaluate the utility of multitemporal Landsat OLI, ASTER, and Sentinel-2 imagery to delineate ten broad vegetation types in LNNP. We then use a combination of Landsat MSS, TM, ETM+, and OLI images to document land use and land cover change in the LNNP area and surrounds for the 1978-2021 period.
Drivers of multi-trophic diversity in forested ecosystems across CONUS
Student-Graduate
Authors: Ayanna St. Rose; Kusum J. Naithani
Description: Understanding the effects of structural complexity of forest canopy on multi-trophic biodiversity is critical for biodiversity conservation and sustainable land management. With an overarching goal of understanding the relationship between structural complexity of forest canopy and multi-trophic biodiversity at continental scale, we asked (1) does forest canopy structural complexity show trends at the multi-trophic levels of primary producers, herbivores, and carnivores, and (2) to what extent does forest canopy structural complexity affect these layers? We used beetle pitfall trap, plant presence, bird count, and high density LiDAR data from the National Ecological Observatory Network (NEON) to investigate the relationship between forest structural complexity and bird, plant and beetle shannon diversity in CONUS. We hypothesized that higher forest canopy structural complexity will result in higher multitrophic biodiversity. Our study showed that multi-trophic Shannon diversity and species richness generally increased with rugosity, canopy height, rumple, vertical complexity index, vegetative area index, cover fraction, but it decreased with deep-gap fraction. These results show that, not only is it possible to employ high density LiDAR data to study the link between multitrophic diversity in forested ecosystems, but also that forests must sustain a level of complexity to sustain multi-trophic diversity. It is imperative that we study forested ecosystems as a whole with many interacting parts so we can develop better forest management and sustainable forest restoration practices that promote diversity at all levels of the food chain.
Student-Graduate
Authors: Ayanna St. Rose; Kusum J. Naithani
Description: Understanding the effects of structural complexity of forest canopy on multi-trophic biodiversity is critical for biodiversity conservation and sustainable land management. With an overarching goal of understanding the relationship between structural complexity of forest canopy and multi-trophic biodiversity at continental scale, we asked (1) does forest canopy structural complexity show trends at the multi-trophic levels of primary producers, herbivores, and carnivores, and (2) to what extent does forest canopy structural complexity affect these layers? We used beetle pitfall trap, plant presence, bird count, and high density LiDAR data from the National Ecological Observatory Network (NEON) to investigate the relationship between forest structural complexity and bird, plant and beetle shannon diversity in CONUS. We hypothesized that higher forest canopy structural complexity will result in higher multitrophic biodiversity. Our study showed that multi-trophic Shannon diversity and species richness generally increased with rugosity, canopy height, rumple, vertical complexity index, vegetative area index, cover fraction, but it decreased with deep-gap fraction. These results show that, not only is it possible to employ high density LiDAR data to study the link between multitrophic diversity in forested ecosystems, but also that forests must sustain a level of complexity to sustain multi-trophic diversity. It is imperative that we study forested ecosystems as a whole with many interacting parts so we can develop better forest management and sustainable forest restoration practices that promote diversity at all levels of the food chain.
Development of a satellite-based water resource monitoring system specific to semi-arid systems
Student-Graduate
Authors: Nicholas Kolarik, Boise State University; Jodi Brandt, Boise State University; Amy Pickens, University of Maryland; Anand Roopsind, Conservation International
Description: Semi-arid systems are becoming increasingly drier but current maps of water resource dynamics do not capture all of the types of resources that sustain these systems. Available maps focus on surface water, wetlands, or mesic vegetation, but none represent all water resources in aggregate. Further, these products rely on the Landsat time series or single date aerial photographs that are too coarse either spatially or temporally to effectively monitor dwindling water resources, and do not consider other data sources known to be helpful for mapping them. In this study, we produced monthly water resources maps at 10-m spatial resolution using freely available data on an open access platform. We analyzed data from the Sentinel-2 and Sentinel-1 time series in tandem with topographic variables. We trained random forest classifiers to capture extents of surface water and mesic areas at the monthly scale. We compared the maps of our Sentinel Fusion (SF) product with a leading publicly available Landsat surface water product at three archetype sites, and revealed improvements in all water class metrics (93.52% producer’s accuracy (PA), 98.62% user’s accuracy (UA) vs. 28.32% PA, 70.78% UA). During the driest months of the year, our SF product of mesic resources had higher PAs compared to National Wetlands Inventory estimates (19.54% - 50.47%) and increased UA over the Sage Grouse Initiative mesic resources maps (51.05% - 67.34%). We also found that inclusion of SAR data from Sentinel-1 did not meaningfully improve our estimates when compared to Sentinel-2 data fused with only topographic variables. Overall, SF maps better capture and describe intra-annual dynamics of small water resource areas integral to ecosystem functioning in semi-arid systems than the leading products, and are suitable for monitoring water resources in the context of conservation efforts, land use, and disturbances.
Student-Graduate
Authors: Nicholas Kolarik, Boise State University; Jodi Brandt, Boise State University; Amy Pickens, University of Maryland; Anand Roopsind, Conservation International
Description: Semi-arid systems are becoming increasingly drier but current maps of water resource dynamics do not capture all of the types of resources that sustain these systems. Available maps focus on surface water, wetlands, or mesic vegetation, but none represent all water resources in aggregate. Further, these products rely on the Landsat time series or single date aerial photographs that are too coarse either spatially or temporally to effectively monitor dwindling water resources, and do not consider other data sources known to be helpful for mapping them. In this study, we produced monthly water resources maps at 10-m spatial resolution using freely available data on an open access platform. We analyzed data from the Sentinel-2 and Sentinel-1 time series in tandem with topographic variables. We trained random forest classifiers to capture extents of surface water and mesic areas at the monthly scale. We compared the maps of our Sentinel Fusion (SF) product with a leading publicly available Landsat surface water product at three archetype sites, and revealed improvements in all water class metrics (93.52% producer’s accuracy (PA), 98.62% user’s accuracy (UA) vs. 28.32% PA, 70.78% UA). During the driest months of the year, our SF product of mesic resources had higher PAs compared to National Wetlands Inventory estimates (19.54% - 50.47%) and increased UA over the Sage Grouse Initiative mesic resources maps (51.05% - 67.34%). We also found that inclusion of SAR data from Sentinel-1 did not meaningfully improve our estimates when compared to Sentinel-2 data fused with only topographic variables. Overall, SF maps better capture and describe intra-annual dynamics of small water resource areas integral to ecosystem functioning in semi-arid systems than the leading products, and are suitable for monitoring water resources in the context of conservation efforts, land use, and disturbances.
Fine resolution remote sensing improves estimates of gross primary production in agricultural and grassland landscapes
Student-Graduate
Authors: Gabriela Shirkey, Department of Geography, Environment & Spatial Sciences, Michigan State University; Ranjeet John, Department of Biology, University of South Dakota; Kyla Dahlin, Department of Geography, Environment & Spatial Sciences, Michigan State University; Michael Abraha, Great Lakes Bioenergy Research Center, Michigan State University; Pietro Sciusco, Department of Geography, Environment & Spatial Sciences, Michigan State University; Cheyenne Lei, Department of Geography, Environment & Spatial Sciences, Michigan State University; David Reed, Department of Environmental Science, University of Science and Arts of Oklahoma; Jiquan Chen, Center for Global Change and Earth Observations, Department of Geography, Environment & Spatial Sciences, Michigan State University
Description: Gross primary production (GPP) is a fundamental measure of the terrestrial carbon cycle that informs land management decisions at global, regional and landscape processes. As many actively managed landscapes are composed of areas less than 500m and 250m, MODIS GPP estimates may be composed of a mix of land covers and therefore be incorrect. This study demonstrates the capability of high-resolution imagery to capture GPP across managed landscapes, including seven sites located at the Great Lakes Biological Research Station. Four sites exist on previous USDA conservation reserve program (CRP) converted in 2009; and the other three on land conventionally farmed with corn-soybean-wheat rotation until 2009. Sites converted include (1) no-till corn; (2) native prairie; and (3) switchgrass. Sites planted on traditionally farmed land include (4) no-till corn; (5) switchgrass; and (6) native prairie; and (7) historically preserved CRP land. We compare eddy-covariance GPP estimates with five satellite models: three vegetation photosynthesis models (VPMs) – built with Sentinel-2 (S2; 5m), Landsat 8 (LS8; 30m), and MODIS (500m) – as well as conventional products Landsat CONUS (30m) and MOD17A (500m). High-resolution (30m and 20m) imagery integrated within VPM generally agrees with tower-based GPP in heterogeneous landscapes in daily estimates and study period sums more so than MODIS VPM (500m) or conventional GPP products from MOD17A or Landsat 8 CONUS. Notably, VPM with Landsat 8 best agreed with tower estimates in corn systems, whereas VPM with Sentinel-2 agreed best with prairie and switchgrass sites. While existing methods using MODIS-derived GPP models serve as an important baseline for large spatial extent studies, future endeavors to estimate GPP in managed landscapes with greater frequency and improved accuracy are accessible and affordable at 30m and 20m resolutions.
Student-Graduate
Authors: Gabriela Shirkey, Department of Geography, Environment & Spatial Sciences, Michigan State University; Ranjeet John, Department of Biology, University of South Dakota; Kyla Dahlin, Department of Geography, Environment & Spatial Sciences, Michigan State University; Michael Abraha, Great Lakes Bioenergy Research Center, Michigan State University; Pietro Sciusco, Department of Geography, Environment & Spatial Sciences, Michigan State University; Cheyenne Lei, Department of Geography, Environment & Spatial Sciences, Michigan State University; David Reed, Department of Environmental Science, University of Science and Arts of Oklahoma; Jiquan Chen, Center for Global Change and Earth Observations, Department of Geography, Environment & Spatial Sciences, Michigan State University
Description: Gross primary production (GPP) is a fundamental measure of the terrestrial carbon cycle that informs land management decisions at global, regional and landscape processes. As many actively managed landscapes are composed of areas less than 500m and 250m, MODIS GPP estimates may be composed of a mix of land covers and therefore be incorrect. This study demonstrates the capability of high-resolution imagery to capture GPP across managed landscapes, including seven sites located at the Great Lakes Biological Research Station. Four sites exist on previous USDA conservation reserve program (CRP) converted in 2009; and the other three on land conventionally farmed with corn-soybean-wheat rotation until 2009. Sites converted include (1) no-till corn; (2) native prairie; and (3) switchgrass. Sites planted on traditionally farmed land include (4) no-till corn; (5) switchgrass; and (6) native prairie; and (7) historically preserved CRP land. We compare eddy-covariance GPP estimates with five satellite models: three vegetation photosynthesis models (VPMs) – built with Sentinel-2 (S2; 5m), Landsat 8 (LS8; 30m), and MODIS (500m) – as well as conventional products Landsat CONUS (30m) and MOD17A (500m). High-resolution (30m and 20m) imagery integrated within VPM generally agrees with tower-based GPP in heterogeneous landscapes in daily estimates and study period sums more so than MODIS VPM (500m) or conventional GPP products from MOD17A or Landsat 8 CONUS. Notably, VPM with Landsat 8 best agreed with tower estimates in corn systems, whereas VPM with Sentinel-2 agreed best with prairie and switchgrass sites. While existing methods using MODIS-derived GPP models serve as an important baseline for large spatial extent studies, future endeavors to estimate GPP in managed landscapes with greater frequency and improved accuracy are accessible and affordable at 30m and 20m resolutions.
Application of ECOSTRESS evapotranspiration in a paired catchment analysis following the 2018 Holy Fire in California
Student-Graduate
Authors: Brenton A. Wilder, Boise State University; Alicia M. Kinoshita, San Diego State University
Description: Ecohydrological processes such as evapotranspiration (ET) and streamflow are highly variable after fire in Mediterranean systems and require accurate assessments to improve long-term risk mitigation and revegetation strategies, especially at the small catchment scale. Using the case of the 2018 Holy Fire in southern California, we characterized 1) pre-fire rainfall and evapotranspiration conditions and 2) recovery of ecohydrological processes using a paired analysis between an unburned (Santiago) and burned (Coldwater) catchment. ECOsystem Spaceborne Thermal Radiometer Experiment on Space Station (ECOSTRESS), Operational Simplified Surface Energy Balance Model (SSEBop), vegetation indices, and local rainfall-runoff data were used to characterize the sites and investigate spatial and temporal patterns of post-fire ET. Consistent with the drought conditions in California, we observed low precipitation and ET prior to the fire. Additionally, compared to other vegetation types, montane hardwood species were more likely to be classified as high soil burn severity. We also found that the high spatial and temporal resolution of ECOSTRESS provided more information about the general ET patterns. After the fire, ECOSTRESS ET was sensitive to parameters such as slope aspect, soil burn severity, and vegetation species, which has implications for post-fire vegetation recovery and water storage. This work demonstrates opportunities to apply ECOSTRESS ET across globally diverse ecoregions and small catchment scales to identify potentially high-risk areas and improve fire risk and vegetation recovery assessments.
Student-Graduate
Authors: Brenton A. Wilder, Boise State University; Alicia M. Kinoshita, San Diego State University
Description: Ecohydrological processes such as evapotranspiration (ET) and streamflow are highly variable after fire in Mediterranean systems and require accurate assessments to improve long-term risk mitigation and revegetation strategies, especially at the small catchment scale. Using the case of the 2018 Holy Fire in southern California, we characterized 1) pre-fire rainfall and evapotranspiration conditions and 2) recovery of ecohydrological processes using a paired analysis between an unburned (Santiago) and burned (Coldwater) catchment. ECOsystem Spaceborne Thermal Radiometer Experiment on Space Station (ECOSTRESS), Operational Simplified Surface Energy Balance Model (SSEBop), vegetation indices, and local rainfall-runoff data were used to characterize the sites and investigate spatial and temporal patterns of post-fire ET. Consistent with the drought conditions in California, we observed low precipitation and ET prior to the fire. Additionally, compared to other vegetation types, montane hardwood species were more likely to be classified as high soil burn severity. We also found that the high spatial and temporal resolution of ECOSTRESS provided more information about the general ET patterns. After the fire, ECOSTRESS ET was sensitive to parameters such as slope aspect, soil burn severity, and vegetation species, which has implications for post-fire vegetation recovery and water storage. This work demonstrates opportunities to apply ECOSTRESS ET across globally diverse ecoregions and small catchment scales to identify potentially high-risk areas and improve fire risk and vegetation recovery assessments.
Finding Remote Sensing Solutions for Mapping Fires in Kenya: Active or Passive?
Authors: Mary C. Henry, John Maingi - Department of Geography, Miami University
Description: In our previous work to study wildfire and agricultural burning in Kenya, we have found that Landsat data (TM, ETM+, OLI) and active fire data (MODIS, VIIRS) do not capture all fires and can also have date mismatches. Smaller, short lived fires might be missed by active fire detection methods- even with daily coverage. Landsat repeat coverage of 16 days with one satellite and eight with two can also miss burn scars given frequent cloud cover in the region. In this study, we test the use of Sentinel-2A/2B MultiSpectral Instrument (MSI) to capture burn scars given the finer spatial resolution (10m and 20m) and more frequent repeat coverage (5 days with two satellites). Our goal is to see which fires are missed by VIIRS and also determine if Sentinel is a better option for monitoring fire activity in Kenya. We obtained Sentinel data from late in the dry season preceding the Short Rains in 2021 for a portion of Machakos County, Kenya (east of Nairobi) which contains a mixture of landcover including agriculture and savanna bushland. Our assumption is that any dry season fires will continue to be detected by Sentinel until the end of the dry season since only rain will cause vegetation recovery and green-up. We also have some ground observation data between August and November 2021 to further verify fire locations.
Authors: Mary C. Henry, John Maingi - Department of Geography, Miami University
Description: In our previous work to study wildfire and agricultural burning in Kenya, we have found that Landsat data (TM, ETM+, OLI) and active fire data (MODIS, VIIRS) do not capture all fires and can also have date mismatches. Smaller, short lived fires might be missed by active fire detection methods- even with daily coverage. Landsat repeat coverage of 16 days with one satellite and eight with two can also miss burn scars given frequent cloud cover in the region. In this study, we test the use of Sentinel-2A/2B MultiSpectral Instrument (MSI) to capture burn scars given the finer spatial resolution (10m and 20m) and more frequent repeat coverage (5 days with two satellites). Our goal is to see which fires are missed by VIIRS and also determine if Sentinel is a better option for monitoring fire activity in Kenya. We obtained Sentinel data from late in the dry season preceding the Short Rains in 2021 for a portion of Machakos County, Kenya (east of Nairobi) which contains a mixture of landcover including agriculture and savanna bushland. Our assumption is that any dry season fires will continue to be detected by Sentinel until the end of the dry season since only rain will cause vegetation recovery and green-up. We also have some ground observation data between August and November 2021 to further verify fire locations.
Spatial patterns of conifer encroachment into meadows of Lassen Volcanic National Park since 1941
Authors: Olivia Duane, U.S. Geological Survey; Miguel L. Villarreal, U.S. Geological Survey; Jerry Davis, San Francisco State University; Leonhard Blesius San Francisco State University; Steven Buckley, National Park Service
Description: Within the last century, there has been widespread establishment of conifers in mountain meadows throughout mountain ranges in western North America. Understanding the factors influencing conifer invasion and how they vary in importance across gradients in climate, vegetation, and disturbance regime is crucial for managing and predicting future changes in dynamic meadow ecosystems. We examined the rate and pattern of conifer encroachment in Lassen Volcanic National Park (LVNP) using ~1,250 historical aerial photos that include near complete coverage of the park for the years 1941, 1952-56, 1966, 1973, 1978, 1981, 1988 and 2004. We implemented an automated image mosaic and orthorectification process using Structure-from-Motion (SfM) photogrammetry algorithms. Since most of the camera calibration parameters are unknown for these photo sets, SfM provided a more efficient workflow compared to traditional photogrammetric techniques. High-resolution orthomosaics were used to map individual trees at the meadow ecotone and measure changes in tree density and spatial patterns over time. To assess biotic interactions in different environments, we used marked point pattern analysis in a subset of forest-meadow ecotones that vary in elevation, vegetation, and landscape context. Preliminary results show that the timing and magnitude of encroachment into meadows varied with landscape context. Recent wildfires appear to have reduced tree density at the meadow-forest ecotone, but not within the meadow interior. Modeling the spatial pattern and timing of conifer establishment in different meadow environments over an 80-year period will improve our ability to predict which conditions (biotic and abiotic) might enhance recruitment in the future.
Authors: Olivia Duane, U.S. Geological Survey; Miguel L. Villarreal, U.S. Geological Survey; Jerry Davis, San Francisco State University; Leonhard Blesius San Francisco State University; Steven Buckley, National Park Service
Description: Within the last century, there has been widespread establishment of conifers in mountain meadows throughout mountain ranges in western North America. Understanding the factors influencing conifer invasion and how they vary in importance across gradients in climate, vegetation, and disturbance regime is crucial for managing and predicting future changes in dynamic meadow ecosystems. We examined the rate and pattern of conifer encroachment in Lassen Volcanic National Park (LVNP) using ~1,250 historical aerial photos that include near complete coverage of the park for the years 1941, 1952-56, 1966, 1973, 1978, 1981, 1988 and 2004. We implemented an automated image mosaic and orthorectification process using Structure-from-Motion (SfM) photogrammetry algorithms. Since most of the camera calibration parameters are unknown for these photo sets, SfM provided a more efficient workflow compared to traditional photogrammetric techniques. High-resolution orthomosaics were used to map individual trees at the meadow ecotone and measure changes in tree density and spatial patterns over time. To assess biotic interactions in different environments, we used marked point pattern analysis in a subset of forest-meadow ecotones that vary in elevation, vegetation, and landscape context. Preliminary results show that the timing and magnitude of encroachment into meadows varied with landscape context. Recent wildfires appear to have reduced tree density at the meadow-forest ecotone, but not within the meadow interior. Modeling the spatial pattern and timing of conifer establishment in different meadow environments over an 80-year period will improve our ability to predict which conditions (biotic and abiotic) might enhance recruitment in the future.
Developing Spectral Signatures for South Texas Native Grasses Using Multispectral Sensors
Student-Graduate
Authors: Annalysa M. Camacho, Caesar Kleberg Wildlife Research Institute, Texas A&M University-Kingsville; Amanda L. Montemayor, Caesar Kleberg Wildlife Research Institute, Texas A&M University-Kingsville; Walter E. Gless, Caesar Kleberg Wildlife Research Institute, Texas A&M University-Kingsville; Michael T. Page, Caesar Kleberg Wildlife Research Institute, Texas A&M University-Kingsville; Melanie A. Ramirez, Caesar Kleberg Wildlife Research Institute, Texas A&M University-Kingsville; Humberto L. Perotto, Caesar Kleberg Wildlife Research Institute, Texas A&M University-Kingsville; J. Alfonso Ortega-S., Caesar Kleberg Wildlife Research Institute, Texas A&M University-Kingsville; Evan P. Tanner, Caesar Kleberg Wildlife Research Institute, Texas A&M University-Kingsville; Anthony D. Falk, Caesar Kleberg Wildlife Research Institute, Texas A&M University-Kingsville; Dwain Daniels, USDA-NRCS Central National Technology Support Center, Fort Worth; Tony Kimmet, USDA-NRCS National Geospatial Center of Excellence; Shad D. Nelson, Mary Lewis Kleberg College of Agriculture & Natural Resources, Texas A&M University–Kingsville
Description: Multispectral sensors are becoming part of the drone toolbox for vegetation monitoring in rangelands. Multispectral sensors can capture spectral bands in the visible and non-visible regions of the light spectrum which can be analyzed to identify plant species. Using drone mounted multispectral sensors allows for less than 1-inch pixel size image acquisition in 5 bands: blue, green, red, rededge, and near infrared. This could potentially help the classification and quantification of grass cover over larger areas in rangelands. Our goal is to develop methods to develop spectral signatures for native grasses in South Texas, USA. Our objectives are to (1) develop spectral signatures of 15 bunch grass species and (2) determine the best time of the year to separate grass species using spectral signatures. I have acquired monthly data from monoculture plots at the USDA NRCS E. “Kika” de la Garza Plant Materials Center (one species) and the South Texas Natives Project Farm (15 species) in Kingsville, Texas. I am comparing the spectral signature between months and between species. Our initial results show that the blue and red bands tend to be similar across months while the green, rededge and near infrared bands differ between months. The highest and most variable reflectance values occur in the near infrared band with Cenchrus myosuroides Kunth having the highest values in August (August 2021; x = 0537; SE = 0.0053;) and the lowest values in December (December 2020; x = 0.1676; SE=0.00027). Our results are the first step to identify, monitor and quantify the presence of native grass species on South Texas rangelands.
Student-Graduate
Authors: Annalysa M. Camacho, Caesar Kleberg Wildlife Research Institute, Texas A&M University-Kingsville; Amanda L. Montemayor, Caesar Kleberg Wildlife Research Institute, Texas A&M University-Kingsville; Walter E. Gless, Caesar Kleberg Wildlife Research Institute, Texas A&M University-Kingsville; Michael T. Page, Caesar Kleberg Wildlife Research Institute, Texas A&M University-Kingsville; Melanie A. Ramirez, Caesar Kleberg Wildlife Research Institute, Texas A&M University-Kingsville; Humberto L. Perotto, Caesar Kleberg Wildlife Research Institute, Texas A&M University-Kingsville; J. Alfonso Ortega-S., Caesar Kleberg Wildlife Research Institute, Texas A&M University-Kingsville; Evan P. Tanner, Caesar Kleberg Wildlife Research Institute, Texas A&M University-Kingsville; Anthony D. Falk, Caesar Kleberg Wildlife Research Institute, Texas A&M University-Kingsville; Dwain Daniels, USDA-NRCS Central National Technology Support Center, Fort Worth; Tony Kimmet, USDA-NRCS National Geospatial Center of Excellence; Shad D. Nelson, Mary Lewis Kleberg College of Agriculture & Natural Resources, Texas A&M University–Kingsville
Description: Multispectral sensors are becoming part of the drone toolbox for vegetation monitoring in rangelands. Multispectral sensors can capture spectral bands in the visible and non-visible regions of the light spectrum which can be analyzed to identify plant species. Using drone mounted multispectral sensors allows for less than 1-inch pixel size image acquisition in 5 bands: blue, green, red, rededge, and near infrared. This could potentially help the classification and quantification of grass cover over larger areas in rangelands. Our goal is to develop methods to develop spectral signatures for native grasses in South Texas, USA. Our objectives are to (1) develop spectral signatures of 15 bunch grass species and (2) determine the best time of the year to separate grass species using spectral signatures. I have acquired monthly data from monoculture plots at the USDA NRCS E. “Kika” de la Garza Plant Materials Center (one species) and the South Texas Natives Project Farm (15 species) in Kingsville, Texas. I am comparing the spectral signature between months and between species. Our initial results show that the blue and red bands tend to be similar across months while the green, rededge and near infrared bands differ between months. The highest and most variable reflectance values occur in the near infrared band with Cenchrus myosuroides Kunth having the highest values in August (August 2021; x = 0537; SE = 0.0053;) and the lowest values in December (December 2020; x = 0.1676; SE=0.00027). Our results are the first step to identify, monitor and quantify the presence of native grass species on South Texas rangelands.
Integrating remote sensing and ecological forecasting into decision support for beaver rewilding
Authors: Jodi Brandt, Boise State University; Nick Kolarik, Boise State University; Nancy Glenn, Boise State University; Joe Wheaton, Utah State University; Phillip Bailey, North Arrow Research; Wally Macfarlane, Utah State University
Description: Throughout history, humans have engineered riverscapes (rivers, streams and their floodplains) to control the amount and timing of water availability with a relentless focus on efficient conveyance of water. Paradoxically, hydrologic inefficiency is a hallmark of a healthy river, and actually provides more reliable water supply, cleaner water, and better resiliency of natural and human infrastructure from floods, fires and droughts. Human interventions have degraded the health of riverscapes and led to the deterioration of mesic resources (riparian zones, and wet meadows) and the ecosystem services they provide. Furthermore, with climate change, the frequency of extreme events like fires, droughts and floods has exposed the vulnerability of water resources, leading to a new legislative focus on “natural infrastructure” through efforts like beaver rewilding. In this poster, we present our newly-funded NASA Applied Sciences project, which focuses on the impacts of beaver rewilding for ecosystem restoration. In particular, we ask: Are beaver rewilding projects having the desired effect and how can we maximize effectiveness? To answer this question, we will create MRRMaid “Mesic Resource Restoration Monitoring aid”, an Earth Observation based adaptive management decision support system that uses NASA’s Landsat archive and the upcoming NISAR mission; the European Space Agency’s Sentinel 1 and 2 missions; and the Harmonized Landsat-Sentinel (HLS) dataset. MRRMaid development will focus on the western U.S., and enable the monitoring of beaver rewilding and other mesic restoration projects at broad spatial and temporal scales using freely-available satellite imagery and cloud-based analysis platforms.
Authors: Jodi Brandt, Boise State University; Nick Kolarik, Boise State University; Nancy Glenn, Boise State University; Joe Wheaton, Utah State University; Phillip Bailey, North Arrow Research; Wally Macfarlane, Utah State University
Description: Throughout history, humans have engineered riverscapes (rivers, streams and their floodplains) to control the amount and timing of water availability with a relentless focus on efficient conveyance of water. Paradoxically, hydrologic inefficiency is a hallmark of a healthy river, and actually provides more reliable water supply, cleaner water, and better resiliency of natural and human infrastructure from floods, fires and droughts. Human interventions have degraded the health of riverscapes and led to the deterioration of mesic resources (riparian zones, and wet meadows) and the ecosystem services they provide. Furthermore, with climate change, the frequency of extreme events like fires, droughts and floods has exposed the vulnerability of water resources, leading to a new legislative focus on “natural infrastructure” through efforts like beaver rewilding. In this poster, we present our newly-funded NASA Applied Sciences project, which focuses on the impacts of beaver rewilding for ecosystem restoration. In particular, we ask: Are beaver rewilding projects having the desired effect and how can we maximize effectiveness? To answer this question, we will create MRRMaid “Mesic Resource Restoration Monitoring aid”, an Earth Observation based adaptive management decision support system that uses NASA’s Landsat archive and the upcoming NISAR mission; the European Space Agency’s Sentinel 1 and 2 missions; and the Harmonized Landsat-Sentinel (HLS) dataset. MRRMaid development will focus on the western U.S., and enable the monitoring of beaver rewilding and other mesic restoration projects at broad spatial and temporal scales using freely-available satellite imagery and cloud-based analysis platforms.
Mapping riparian corridors with aerial lidar and high-resolution remotely sensed data: a variable width approach for sagebrush steppe ecosystems
Student-Undergraduate
Authors: Claire Vaage, Boise State University; Trevor Caughlin, Boise State University; Anna Bergstrom, Boise State University
Description: Riparian vegetation is critically important for dryland ecosystem functions including maintaining water temperatures for resident fish populations, enhancing carbon sequestration, stabilizing stream banks and flow, and supplying and retaining nutrients within water systems. Development of management and conservation strategies for these vital areas is dependent on mapping the extent of the characteristic riparian vegetation. The primary step in modeling hydrological ecosystem dynamics includes defining the riparian buffer through fixed-width or variable-width approaches. Fixed-width buffers do not accurately capture smaller, unique riparian areas because they only account for the watercourse, ignoring inputs such as stream order or geomorphology. A variable-width buffer reflects spatial variability in riparian vegetation by accounting for landscape complexities, such as fine-scale variation in topography that complicate proper estimations of riparian zones. Using a 1-m spatial resolution digital elevation model (DEM) derived from aerial lidar, we mapped the stream network of the southwestern section and lower elevations of the Dry Creek Experimental Watershed in Idaho, USA. This watershed encompasses a topographically complex/diverse ecosystem gradient, from sagebrush steppe to evergreen forest. Then, we generated random points along the predicted stream network to digitize (n = 1,500) and ground truth (n = 150) riparian/streamside vegetation. We developed a regression model to predict riparian vegetation width using the collected digitized and ground truth measurements, and additional covariates (i.e., slope, aspect, elevation, vegetation). Our results provide a way to accurately map riparian buffers in sagebrush steppe ecosystems through a novel variable-width approach and allow for additional research on the geomorphological, hydrological, and ecological characteristics of riparian areas within drylands.
Student-Undergraduate
Authors: Claire Vaage, Boise State University; Trevor Caughlin, Boise State University; Anna Bergstrom, Boise State University
Description: Riparian vegetation is critically important for dryland ecosystem functions including maintaining water temperatures for resident fish populations, enhancing carbon sequestration, stabilizing stream banks and flow, and supplying and retaining nutrients within water systems. Development of management and conservation strategies for these vital areas is dependent on mapping the extent of the characteristic riparian vegetation. The primary step in modeling hydrological ecosystem dynamics includes defining the riparian buffer through fixed-width or variable-width approaches. Fixed-width buffers do not accurately capture smaller, unique riparian areas because they only account for the watercourse, ignoring inputs such as stream order or geomorphology. A variable-width buffer reflects spatial variability in riparian vegetation by accounting for landscape complexities, such as fine-scale variation in topography that complicate proper estimations of riparian zones. Using a 1-m spatial resolution digital elevation model (DEM) derived from aerial lidar, we mapped the stream network of the southwestern section and lower elevations of the Dry Creek Experimental Watershed in Idaho, USA. This watershed encompasses a topographically complex/diverse ecosystem gradient, from sagebrush steppe to evergreen forest. Then, we generated random points along the predicted stream network to digitize (n = 1,500) and ground truth (n = 150) riparian/streamside vegetation. We developed a regression model to predict riparian vegetation width using the collected digitized and ground truth measurements, and additional covariates (i.e., slope, aspect, elevation, vegetation). Our results provide a way to accurately map riparian buffers in sagebrush steppe ecosystems through a novel variable-width approach and allow for additional research on the geomorphological, hydrological, and ecological characteristics of riparian areas within drylands.
Using remote sensing data to characterize bird habitat in an urban ecosystem
Student-Graduate
Authors: Christian Benitez, California State University Los Angeles; Dr. Eric M. Wood, California State University Los Angeles; Dr. Michael C. Beland, California State University Los Angeles
Description: For decades, urban ecosystems have been viewed as biological deserts. However, recently, biodiversity research within urban ecosystems has surged, highlighting the potential of cities to harbor a high diversity of flora and fauna. While there has been a plethora of work detailing the ecology of biodiversity within cities, most land within urban areas is private, presenting an obstacle for a robust biological survey. Therefore, there is a need to use methodologies that allow for examining biodiversity on lands where a ground-based biological survey is problematic. One such approach that holds promise is remote sensing. Remote sensing, the acquisition of information via remote sources (e.g., aerial photographs), has been extensively used to characterize wildlife habitat within natural settings. However, the use of remote sensing to characterize habitat for wildlife in urban areas has mostly been unexplored, which presents a gap in our knowledge of biodiversity monitoring within cities. In this study, we tested the utility of using remote sensing methods coupled with ground-based measures to characterize habitat and the wintering avifauna across L.A. Avifaunal communities and data from remote sensing co-varied across the socioeconomic gradient, with birds affiliated with trees being more common in high-income parts of L.A. Similarly, remote sensing data designed to capture vegetation structure and greenness were also more pronounced in high-income areas of the city. We found that data from light detection and ranging (LiDAR), which is a fine-resolution remote sensing technique useful in characterizing vertical vegetation structure was a strong predictor of avian communities throughout L.A. being positively related with forest-affiliated species and negatively with urban-affiliated species. Given our results, we recommend that urban ecologists prioritize the use of fine-resolution remote sensing data along with field surveys of a study wildlife taxa to improve our understanding of the distribution of urban biodiversity within other cities.
Student-Graduate
Authors: Christian Benitez, California State University Los Angeles; Dr. Eric M. Wood, California State University Los Angeles; Dr. Michael C. Beland, California State University Los Angeles
Description: For decades, urban ecosystems have been viewed as biological deserts. However, recently, biodiversity research within urban ecosystems has surged, highlighting the potential of cities to harbor a high diversity of flora and fauna. While there has been a plethora of work detailing the ecology of biodiversity within cities, most land within urban areas is private, presenting an obstacle for a robust biological survey. Therefore, there is a need to use methodologies that allow for examining biodiversity on lands where a ground-based biological survey is problematic. One such approach that holds promise is remote sensing. Remote sensing, the acquisition of information via remote sources (e.g., aerial photographs), has been extensively used to characterize wildlife habitat within natural settings. However, the use of remote sensing to characterize habitat for wildlife in urban areas has mostly been unexplored, which presents a gap in our knowledge of biodiversity monitoring within cities. In this study, we tested the utility of using remote sensing methods coupled with ground-based measures to characterize habitat and the wintering avifauna across L.A. Avifaunal communities and data from remote sensing co-varied across the socioeconomic gradient, with birds affiliated with trees being more common in high-income parts of L.A. Similarly, remote sensing data designed to capture vegetation structure and greenness were also more pronounced in high-income areas of the city. We found that data from light detection and ranging (LiDAR), which is a fine-resolution remote sensing technique useful in characterizing vertical vegetation structure was a strong predictor of avian communities throughout L.A. being positively related with forest-affiliated species and negatively with urban-affiliated species. Given our results, we recommend that urban ecologists prioritize the use of fine-resolution remote sensing data along with field surveys of a study wildlife taxa to improve our understanding of the distribution of urban biodiversity within other cities.
UNMANNED AERIAL VEHICLES TO EVALUATE GRAZING INTENSITY AND DISTRIBUTION FOR WILDLIFE HABITAT MANAGEMENT
Student-Graduate
Authors: J. Silverio Avila-Sanchez, Caesar Kleberg Wildlife Research Institute; Bradley K. Johnston, Caesar Kleberg Wildlife Research Institute; Humberto L. Perotto-Baldivieso, Caesar Kleberg Wildlife Research Institute; J. Alfonso Ortega-S. , Caesar Kleberg Wildlife Research Institute; Leonard A. Brennan, Caesar Kleberg Wildlife Research Institute; Fidel Hernandez, Caesar Kleberg Wildlife Research Institute; Jason W. Karl, Department of Forestry, Rangeland, and Fire Sciences, University of Idaho
Description: Northern bobwhite (Colinus virginianus) populations have been affected by dense stands of undisturbed vegetation that result in low plant diversity and limited bare ground. Bobwhites require a diverse plant composition and arrangement of woody, herbaceous cover and bare ground for nesting, brooding, feeding, resting, and roosting. Cattle reduces vegetations density and height through grazing and trampling, and in return create a patch-mosaic vegetation structure of different plants. The objective of this research is to compare vegetation structure (height and density) in pastures that have been grazed versus control pastures with no grazing. We will use unmanned aerial vehicles and collect multispectral imagery from three 100-ha plots in each pasture at an altitude of 50 m above ground level in grazed and control pastures. We will compare density, height, and grass species spatial heterogeneity between pastures. We will evaluate vegetation density and height, by creating a normalized digital surface model. We will also collect field data to evaluate plant species richness and conduct an accuracy assessment of the models developed with the drone. We will focus on the volume and spatial arrangement of the vegetation at the pasture scale level, evaluating the utilization of the pasture by cattle. This research will improve our understanding of how spatial heterogeneity in vegetation structure can be improved with grazing to generate bobwhite habitat while maintaining a ranch operation.
Student-Graduate
Authors: J. Silverio Avila-Sanchez, Caesar Kleberg Wildlife Research Institute; Bradley K. Johnston, Caesar Kleberg Wildlife Research Institute; Humberto L. Perotto-Baldivieso, Caesar Kleberg Wildlife Research Institute; J. Alfonso Ortega-S. , Caesar Kleberg Wildlife Research Institute; Leonard A. Brennan, Caesar Kleberg Wildlife Research Institute; Fidel Hernandez, Caesar Kleberg Wildlife Research Institute; Jason W. Karl, Department of Forestry, Rangeland, and Fire Sciences, University of Idaho
Description: Northern bobwhite (Colinus virginianus) populations have been affected by dense stands of undisturbed vegetation that result in low plant diversity and limited bare ground. Bobwhites require a diverse plant composition and arrangement of woody, herbaceous cover and bare ground for nesting, brooding, feeding, resting, and roosting. Cattle reduces vegetations density and height through grazing and trampling, and in return create a patch-mosaic vegetation structure of different plants. The objective of this research is to compare vegetation structure (height and density) in pastures that have been grazed versus control pastures with no grazing. We will use unmanned aerial vehicles and collect multispectral imagery from three 100-ha plots in each pasture at an altitude of 50 m above ground level in grazed and control pastures. We will compare density, height, and grass species spatial heterogeneity between pastures. We will evaluate vegetation density and height, by creating a normalized digital surface model. We will also collect field data to evaluate plant species richness and conduct an accuracy assessment of the models developed with the drone. We will focus on the volume and spatial arrangement of the vegetation at the pasture scale level, evaluating the utilization of the pasture by cattle. This research will improve our understanding of how spatial heterogeneity in vegetation structure can be improved with grazing to generate bobwhite habitat while maintaining a ranch operation.
Quantifying Erosion Susceptibility as a Function of Geomorphic Variables, Trail Type, and Use with Implications for Trail Planning in the City of Boulder Open Space and Mountain Parks
Student-Graduate
Authors: Ara Metz, University of Northern Colorado; Dr. Sharon Bywater-Reyes, University of Northern Colorado; Dr. Chelsie Romulo, University of Northern Colorado
Description: The City of Boulder Open Space and Mountain Parks (OSMP) in Boulder, Colorado maintains trail systems underlain by sedimentary rocks, some of which are prone to erosion. Our research objectives were to 1) assess trail conditions with drone imagery; 2) produce a process domain map that identifies dominant erosional processes in different lithologic units; and 3) quantify erosion rates as a function of process domains and geologic units with the goal to understand how erosional processes influence the trail system. For (1) we used 1-m LiDAR and produced physiographic variables (e.g., slope, TPI, curvature) as a function of lithology and trail type (designated and undesignated). For (2) and (3) we produced high-resolution imagery and three-dimensional models using Agisoft Metashape for zoomed-in areas of interest (AOIs), highlighting areas within our AOIs where active erosion is occurring on designated and undesignated trails.
Data obtained by these methods show that the Boulder hillslopes are transport limited landscapes with diffusive and advective processes occurring. Designated trails occurred more frequently in diffusive processes with a mean curvature of -0.015349 m-1 compared to 0.000491 m-1 for undesignated trails showing advective processes. The data also showed that the average slope of designated trails was 28.49◦ compared to 24.30◦ for undesignated trails suggesting the movement of material is higher on designated trails. Field observations indicate that mass movements are likely correlated to slope and localized process domains.
This research suggests that trail design is the largest contributor to poor trail conditions in the front range when compared to other geomorphic processes. To reduce the impact of erosion on the OSMP trail system, management practices such as switchbacks, reducing the grade of the trail, and following natural contours can reduce erosion. Future work includes comparing the data to background erosion rates obtained from cosmogenic nuclide dating.
Student-Graduate
Authors: Ara Metz, University of Northern Colorado; Dr. Sharon Bywater-Reyes, University of Northern Colorado; Dr. Chelsie Romulo, University of Northern Colorado
Description: The City of Boulder Open Space and Mountain Parks (OSMP) in Boulder, Colorado maintains trail systems underlain by sedimentary rocks, some of which are prone to erosion. Our research objectives were to 1) assess trail conditions with drone imagery; 2) produce a process domain map that identifies dominant erosional processes in different lithologic units; and 3) quantify erosion rates as a function of process domains and geologic units with the goal to understand how erosional processes influence the trail system. For (1) we used 1-m LiDAR and produced physiographic variables (e.g., slope, TPI, curvature) as a function of lithology and trail type (designated and undesignated). For (2) and (3) we produced high-resolution imagery and three-dimensional models using Agisoft Metashape for zoomed-in areas of interest (AOIs), highlighting areas within our AOIs where active erosion is occurring on designated and undesignated trails.
Data obtained by these methods show that the Boulder hillslopes are transport limited landscapes with diffusive and advective processes occurring. Designated trails occurred more frequently in diffusive processes with a mean curvature of -0.015349 m-1 compared to 0.000491 m-1 for undesignated trails showing advective processes. The data also showed that the average slope of designated trails was 28.49◦ compared to 24.30◦ for undesignated trails suggesting the movement of material is higher on designated trails. Field observations indicate that mass movements are likely correlated to slope and localized process domains.
This research suggests that trail design is the largest contributor to poor trail conditions in the front range when compared to other geomorphic processes. To reduce the impact of erosion on the OSMP trail system, management practices such as switchbacks, reducing the grade of the trail, and following natural contours can reduce erosion. Future work includes comparing the data to background erosion rates obtained from cosmogenic nuclide dating.
DAY TWO - Characterizing and Interpreting Landscapes
Tuesday, April 12, 2022
Diverse effects of filters on soundscape indicators in different landscapes and seasons
Authors: Emilia Hyland, Furman University; Hannah Lee, Furman University; Annie Schulz, Furman University; Calla Pederson, Furman University; John Quinn, Furman University
Description: Automated recording units (ARUs) are a time and cost effective method to measure and monitor biodiversity in varied ecosystem types. It is also a way to collect a large amount of data for extensive and reliable analysis, including with soundscape indices. However, with the breadth of data, there may be sounds that skew soundscape indicators and their relationship to more traditional measures of biodiversity including traffic noise and static associated with microphones. However, there is limited research and precedent on appropriate filters to apply to acoustic data to minimize this potential bias. The purpose of this study is to fill the literature gap and highlight the effect of 80, 500, 1k, 2k Hz filters on various acoustic indices (e.g., ADI, AEI, NDSI). Acoustic data used in this study was collected in different temperate and tropical regions and in different ecosystem types (grassland, forest, tropical, agricultural) and during different times of the year (summer and winter). We found the effects on indices varied by filter, region, and season. For example, in the winter of Nebraska there was no difference in ACI while the response of ADI flipped between treatments with a 1k filter. In the summer in both the tropics and temperate AEI and ADI indices changed by filter. For AEI in the tropics, there was no difference between a 500 and 1k filter, suggesting that there is no need to apply a 500 filter in this context. These results suggest careful consideration of the spatial and seasonal context when applying a filter in acoustic data analysis.
Authors: Emilia Hyland, Furman University; Hannah Lee, Furman University; Annie Schulz, Furman University; Calla Pederson, Furman University; John Quinn, Furman University
Description: Automated recording units (ARUs) are a time and cost effective method to measure and monitor biodiversity in varied ecosystem types. It is also a way to collect a large amount of data for extensive and reliable analysis, including with soundscape indices. However, with the breadth of data, there may be sounds that skew soundscape indicators and their relationship to more traditional measures of biodiversity including traffic noise and static associated with microphones. However, there is limited research and precedent on appropriate filters to apply to acoustic data to minimize this potential bias. The purpose of this study is to fill the literature gap and highlight the effect of 80, 500, 1k, 2k Hz filters on various acoustic indices (e.g., ADI, AEI, NDSI). Acoustic data used in this study was collected in different temperate and tropical regions and in different ecosystem types (grassland, forest, tropical, agricultural) and during different times of the year (summer and winter). We found the effects on indices varied by filter, region, and season. For example, in the winter of Nebraska there was no difference in ACI while the response of ADI flipped between treatments with a 1k filter. In the summer in both the tropics and temperate AEI and ADI indices changed by filter. For AEI in the tropics, there was no difference between a 500 and 1k filter, suggesting that there is no need to apply a 500 filter in this context. These results suggest careful consideration of the spatial and seasonal context when applying a filter in acoustic data analysis.
Using Freely Available LANDFIRE Data to Explore Past & Present Vegetation
Student-Undergraduate
Authors: Erin Matula, Northern Michigan University, The Nature Conservancy, Conservation Data Lab; Randy Swaty, The Nature Conservancy, Conservation Data Lab
Description: Ecologists and natural resource managers face a challenge of finding contiguous datasets for broad-scale planning and assessing of land management. LANDFIRE is a synergistic U.S. federal program comprised of a partnership between the U.S. Department of Interior, the U.S. Department of Agriculture Forest Service, and partners such as The Nature Conservancy. To demonstrate its potential on a statewide scale, I conducted an ecosystem analysis of Vermont through maps and graphs. As with any dataset there were challenges and opportunities. Here I discuss strategies to overcome the sheer size and depth of the data, and explore ways to present “combined” data to assess, for example, the canopy cover characteristics of the Laurentian Acadian Northern Hardwoods Forest. The summarized vegetative analysis prompted productive conversations with the Vermont chapter of The Nature Conservancy. This vegetative analysis of Vermont is a blueprint for vegetative analysis anywhere across the U.S. and insular islands. This usage of LANDFIRE is just the starting point to the vast potential of spatial data analysis.
Student-Undergraduate
Authors: Erin Matula, Northern Michigan University, The Nature Conservancy, Conservation Data Lab; Randy Swaty, The Nature Conservancy, Conservation Data Lab
Description: Ecologists and natural resource managers face a challenge of finding contiguous datasets for broad-scale planning and assessing of land management. LANDFIRE is a synergistic U.S. federal program comprised of a partnership between the U.S. Department of Interior, the U.S. Department of Agriculture Forest Service, and partners such as The Nature Conservancy. To demonstrate its potential on a statewide scale, I conducted an ecosystem analysis of Vermont through maps and graphs. As with any dataset there were challenges and opportunities. Here I discuss strategies to overcome the sheer size and depth of the data, and explore ways to present “combined” data to assess, for example, the canopy cover characteristics of the Laurentian Acadian Northern Hardwoods Forest. The summarized vegetative analysis prompted productive conversations with the Vermont chapter of The Nature Conservancy. This vegetative analysis of Vermont is a blueprint for vegetative analysis anywhere across the U.S. and insular islands. This usage of LANDFIRE is just the starting point to the vast potential of spatial data analysis.
Identifying Land Cover Change Trajectories in Xilinhot, Inner Mongolia using the LandTrendr Segmentation Algorithm
Student-Graduate
Authors: Mia Murray, George Washington University; Ginger Allington, The George Washington University; Jeremiah Sjoberg, University Corporation for Atmospheric Research
Description: Current modeling of land use and land cover of rangelands is limited by shortcomings in the ontological and methodological approaches to classification. For example, current classification schemes often limit rangeland cover types into one of a small number of categories, typically grassland, shrubland, or barren, and change is only measured when a pixel shifts to a new class. These coarse categories can mask the complex dynamics happening within a given land cover type, such as greening or browning that might occur prior to a state shift. A more nuanced classification of grasslands that represents the dynamics occurring within a class would contribute meaningfully to policy and management practices throughout the region. In this study, we assess a novel method for classifying land cover change, focussed on Xilinhot, a provincial city in central Inner Mongolia, located in one of the largest remaining arid grassland regions of the world. The region has faced rapid development and land use conversion in recent decades and grassland condition is declining in many areas, but the mechanisms of these changes are difficult to assess with traditional change detection approaches. In this study we apply the LandTrendr (LT) segmentation algorithm to a time series of Landsat-derived imagery of vegetation condition to capture changes in 30+ years of cover in Xilinhot. After accounting for effects of climate, we then classify the resulting segmented data within an unsupervised k-means clustering scheme. We identified ten unique classes in the data; these classes represent unique pixel trajectories resulting from past changes in management and land use. These data on the influence of past intervention on pixel-level responses can be used to help build predictive models to better understand the potential impact of future policy and management changes.
Student-Graduate
Authors: Mia Murray, George Washington University; Ginger Allington, The George Washington University; Jeremiah Sjoberg, University Corporation for Atmospheric Research
Description: Current modeling of land use and land cover of rangelands is limited by shortcomings in the ontological and methodological approaches to classification. For example, current classification schemes often limit rangeland cover types into one of a small number of categories, typically grassland, shrubland, or barren, and change is only measured when a pixel shifts to a new class. These coarse categories can mask the complex dynamics happening within a given land cover type, such as greening or browning that might occur prior to a state shift. A more nuanced classification of grasslands that represents the dynamics occurring within a class would contribute meaningfully to policy and management practices throughout the region. In this study, we assess a novel method for classifying land cover change, focussed on Xilinhot, a provincial city in central Inner Mongolia, located in one of the largest remaining arid grassland regions of the world. The region has faced rapid development and land use conversion in recent decades and grassland condition is declining in many areas, but the mechanisms of these changes are difficult to assess with traditional change detection approaches. In this study we apply the LandTrendr (LT) segmentation algorithm to a time series of Landsat-derived imagery of vegetation condition to capture changes in 30+ years of cover in Xilinhot. After accounting for effects of climate, we then classify the resulting segmented data within an unsupervised k-means clustering scheme. We identified ten unique classes in the data; these classes represent unique pixel trajectories resulting from past changes in management and land use. These data on the influence of past intervention on pixel-level responses can be used to help build predictive models to better understand the potential impact of future policy and management changes.
Measuring Ecosystem Services Supply and Demand Trade-Offs Associated with Projected Urbanization in the Western United States
Student-Graduate
Authors: Sarah Halperin, Boise State University; Dr. Jodi Brandt, Boise State University
Description: Globally, as the human population increases, so does the demand for developed land to meet housing needs. Frequently, agricultural land is converted to meet this demand. Therefore, balancing housing needs with food and fiber production is an important societal issue. In this study, we measured how the supply and demand of Ecosystem Services (ES) changes with development of agricultural land in the Boise Metropolitan Area (BMA). We use the BMA as a region representative of agricultural landscapes of the western United States that have limited arable land and that face severe development threats. We quantified the supply of six ES (carbon sequestration, nutrient retention, habitat quality, food production, pasture production, and recreation) under historical (2001), current (2016), and projected land uses (2030, 2050) and we included a metric of population density to represent ES demand. We used Spearman rank correlation method, Principal Component Analysis, and K-means cluster analysis to analyze spatial trade-offs and synergies. We found that the projected landscape in 2050 will have a higher demand for ES, but a lower supply of four of the six ES when compared to land uses in 2001. We identified a tendency for trade-offs between provisioning and regulating services. Lastly, we identified three bundle types which showed areas with high levels in a) carbon storage and population b) recreation, carbon storage, and habitat quality c) nutrient retention and pasture production. In this study, we present a spatially explicit approach for analyzing interactions among multiple ES that take into account social and ecological dynamics and investigate temporal variability. Overall, our results suggest the importance of managing to sustain multiple ES and can provide practical information for land use managers to balance the protection of our natural resources and future development needs.
Student-Graduate
Authors: Sarah Halperin, Boise State University; Dr. Jodi Brandt, Boise State University
Description: Globally, as the human population increases, so does the demand for developed land to meet housing needs. Frequently, agricultural land is converted to meet this demand. Therefore, balancing housing needs with food and fiber production is an important societal issue. In this study, we measured how the supply and demand of Ecosystem Services (ES) changes with development of agricultural land in the Boise Metropolitan Area (BMA). We use the BMA as a region representative of agricultural landscapes of the western United States that have limited arable land and that face severe development threats. We quantified the supply of six ES (carbon sequestration, nutrient retention, habitat quality, food production, pasture production, and recreation) under historical (2001), current (2016), and projected land uses (2030, 2050) and we included a metric of population density to represent ES demand. We used Spearman rank correlation method, Principal Component Analysis, and K-means cluster analysis to analyze spatial trade-offs and synergies. We found that the projected landscape in 2050 will have a higher demand for ES, but a lower supply of four of the six ES when compared to land uses in 2001. We identified a tendency for trade-offs between provisioning and regulating services. Lastly, we identified three bundle types which showed areas with high levels in a) carbon storage and population b) recreation, carbon storage, and habitat quality c) nutrient retention and pasture production. In this study, we present a spatially explicit approach for analyzing interactions among multiple ES that take into account social and ecological dynamics and investigate temporal variability. Overall, our results suggest the importance of managing to sustain multiple ES and can provide practical information for land use managers to balance the protection of our natural resources and future development needs.
Toward Understanding Degradation in Mountain Pastures: Current Perspectives and New Insights from the Highlands of Kyrgyzstan
Student-Graduate
Authors: Munavar Zhumanova, Center for Global Change and Earth Observation, Michigan State University; Monika A. Tomaszewska, Center for Global Change and Earth Observation, Michigan State University; Aliaskar Mambetov, Sagynbek Orunbaev, Zheenbek Kulenbekov, Department of Applied Geology Environmental Sustainability, and Climate Science, Environment, and American University of Central Asia, Bishkek, Kyrgyz Republic; Geoffrey M. Henebry, Center for Global Change and Earth Observation, Department of Geography, Environment, and Spatial Sciences, Michigan State University, East Lansing, MI, USA
Description: Mountain pastures of Central Asia are productive arid ecosystems experiencing climate change and biodiversity hotspots threatened by human activity. Disturbances in montane landscapes rarely occur in isolation, making it significant to recognize interactions between them. Besides direct grazing impacts, vegetation belts along elevations were formed by interactions among highly heterogeneous microclimatic, topographic, and edaphic conditions. Disturbances overlap in space and time, creating complex mosaicked landscapes and challenging to classify pasture degradation. Moreover, despite ongoing climate change, there has been a tendency to evaluate pasture conditions by referring to their historic condition and prescribing management to restore pastures to resemble that reference state. It is not easy to find undisturbed pastures. Moreover, projected future climates—warmer temperatures, high precipitation variability, and altered snow seasonality— will affect the functioning of pasture ecosystems even though mountain vegetation is dominated by types formed under arid conditions. A further complication is shifting seasonality in pastures. Herders already alter seasonal use practices to compensate for shifts in phenology. Pasture location and access are prioritized over pasture condition or optimal seasonality because livestock products remain the foundational rural economy. Here we present a framework for pasture degradation classification. Our dataset documents pasture conditions across seasonal pastures representing key vegetation types of the Central Tien-Shan mountains. During July 2021, we sampled 48 pastures in Naryn and At-Bashy rayons. We took 102 near-nadir digital photographs at each site along a pair of 100 m orthogonal transects. Grazing impacts were characterized by four variables: (i) seasonality of grazing; (ii) total fractional cover; (iii) composition by palatability and life form; and (iv) apparent recovery from earlier grazing involved weather conditions. We identified six types of pasture degradation and their inter-linkages to illustrate how semiarid landscapes undergoing vegetation transitions may function in future climatic conditions under unchanged grazing practices.
Student-Graduate
Authors: Munavar Zhumanova, Center for Global Change and Earth Observation, Michigan State University; Monika A. Tomaszewska, Center for Global Change and Earth Observation, Michigan State University; Aliaskar Mambetov, Sagynbek Orunbaev, Zheenbek Kulenbekov, Department of Applied Geology Environmental Sustainability, and Climate Science, Environment, and American University of Central Asia, Bishkek, Kyrgyz Republic; Geoffrey M. Henebry, Center for Global Change and Earth Observation, Department of Geography, Environment, and Spatial Sciences, Michigan State University, East Lansing, MI, USA
Description: Mountain pastures of Central Asia are productive arid ecosystems experiencing climate change and biodiversity hotspots threatened by human activity. Disturbances in montane landscapes rarely occur in isolation, making it significant to recognize interactions between them. Besides direct grazing impacts, vegetation belts along elevations were formed by interactions among highly heterogeneous microclimatic, topographic, and edaphic conditions. Disturbances overlap in space and time, creating complex mosaicked landscapes and challenging to classify pasture degradation. Moreover, despite ongoing climate change, there has been a tendency to evaluate pasture conditions by referring to their historic condition and prescribing management to restore pastures to resemble that reference state. It is not easy to find undisturbed pastures. Moreover, projected future climates—warmer temperatures, high precipitation variability, and altered snow seasonality— will affect the functioning of pasture ecosystems even though mountain vegetation is dominated by types formed under arid conditions. A further complication is shifting seasonality in pastures. Herders already alter seasonal use practices to compensate for shifts in phenology. Pasture location and access are prioritized over pasture condition or optimal seasonality because livestock products remain the foundational rural economy. Here we present a framework for pasture degradation classification. Our dataset documents pasture conditions across seasonal pastures representing key vegetation types of the Central Tien-Shan mountains. During July 2021, we sampled 48 pastures in Naryn and At-Bashy rayons. We took 102 near-nadir digital photographs at each site along a pair of 100 m orthogonal transects. Grazing impacts were characterized by four variables: (i) seasonality of grazing; (ii) total fractional cover; (iii) composition by palatability and life form; and (iv) apparent recovery from earlier grazing involved weather conditions. We identified six types of pasture degradation and their inter-linkages to illustrate how semiarid landscapes undergoing vegetation transitions may function in future climatic conditions under unchanged grazing practices.
Changes in Ecosystem Services in the Catawba River Watershed From 2001 to 2016
Student-Graduate
Authors: Barbara Teague, Appalachian State University; Megan Holland, University of Toronto; Steve Seagle, Appalachian State University
Description: Watersheds are effective analysis units for the evaluation of ecosystem services, especially when water-related services are being considered or when multiple interacting ecosystem services are being analyzed. The Catawba River Watershed (CRW) headwaters are in the Blue Ridge physiographic province of western North Carolina. The CRW then extends through the Western Piedmont, leaving NC in the Charlotte metropolitan area. Land and water quality of the CRW are impacted by many anthropogenic issues, including deforestation for development and agriculture, nutrient and sediment runoff from agriculture, fragmentation of wildlife habitat, and introduction of invasive species. Historical and ongoing forest loss is one of the most significant impacts on production of CRW ecosystem services, though its effects may vary widely across the watershed. For purposes of land use planning and conservation, this project (1) quantifies changes in land use and patterns of land use within the CRW from 2001 to 2016, (2) examines how land use change has altered the production of several ecosystem services within the CRW, and (3) determines locations within the CRW where production of multiple ecosystem services is concentrated. Spatial optimization analyses was employed to map and quantify the extent and location of “hotspots” of ecosystem services for 2001 and 2016. Because of the concentration of protected forestland in the Blue Ridge, the loss of ecosystem services in the CRW uplands were localized from 2001 to 2016, but still of significant impact. Loss of ecosystem services was greatest in the Piedmont portion of the CRW. Nonetheless, hotspots of services remain in the Piedmont and should be considered as hotpots for sustainable production of ecosystem services.
Student-Graduate
Authors: Barbara Teague, Appalachian State University; Megan Holland, University of Toronto; Steve Seagle, Appalachian State University
Description: Watersheds are effective analysis units for the evaluation of ecosystem services, especially when water-related services are being considered or when multiple interacting ecosystem services are being analyzed. The Catawba River Watershed (CRW) headwaters are in the Blue Ridge physiographic province of western North Carolina. The CRW then extends through the Western Piedmont, leaving NC in the Charlotte metropolitan area. Land and water quality of the CRW are impacted by many anthropogenic issues, including deforestation for development and agriculture, nutrient and sediment runoff from agriculture, fragmentation of wildlife habitat, and introduction of invasive species. Historical and ongoing forest loss is one of the most significant impacts on production of CRW ecosystem services, though its effects may vary widely across the watershed. For purposes of land use planning and conservation, this project (1) quantifies changes in land use and patterns of land use within the CRW from 2001 to 2016, (2) examines how land use change has altered the production of several ecosystem services within the CRW, and (3) determines locations within the CRW where production of multiple ecosystem services is concentrated. Spatial optimization analyses was employed to map and quantify the extent and location of “hotspots” of ecosystem services for 2001 and 2016. Because of the concentration of protected forestland in the Blue Ridge, the loss of ecosystem services in the CRW uplands were localized from 2001 to 2016, but still of significant impact. Loss of ecosystem services was greatest in the Piedmont portion of the CRW. Nonetheless, hotspots of services remain in the Piedmont and should be considered as hotpots for sustainable production of ecosystem services.
Stepping or Sinking Stones: The role of small habitat patches in fragmented landscapes
Student-Undergraduate
Authors: Eli Shaw, University of Florida School of Natural Resources and the Environment; Robert J. Fletcher Jr., University of Florida Department of Wildlife Ecology and Conservation; Thomas A. H. Smith, University of Florida Department of Wildlife Ecology and Conservation
Description: Habitat fragmentation is a major driver of species declines worldwide and poses a threat to the continued stability and security of global biodiversity. Increasing landscape connectivity is proposed as a way to mitigate the effects of fragmentation typically by two major conservation strategies: habitat corridors or stepping stones. While corridors have been extensively tested, stepping stones, small patches of habitat connecting larger patches, lack direct empirical evidence. However, small habitat patches may act as attractive sinks, absorbing dispersing individuals and creating a “shadow” effect on larger patches, decreasing disperser’s ability to reach them. This shadow effect presents the possibility that small patches actually do not function as stepping stones but instead decrease landscape connectivity. We conducted a field experiment as one of the first empirical tests of the function of small habitat patches in fragmented landscapes utilizing the cactus bug Chelinidea vittiger - prickly pear Opuntia mesacantha spp. lata study system. We examined the rate of dispersal to large habitat patches across in situ experimental arenas containing one of three configurations of intervening small patches or a control that lacked small patches. We found that both likelihood of reaching the target habitat and overall time taken, including time spent at the release site, was significantly higher in the treatment with multiple small patches. In addition, time spent actively dispersing through the matrix was significantly greater with the presence of any small habitat patch in a landscape. Connectivity models showed a similar jump in average commute distance with the introduction of any small patch to a landscape. However, an observed increase in survival for treatments with any small patch poses the question of whether small patches may still benefit dispersing species in fragmented landscapes even if these patches delay dispersal.
Student-Undergraduate
Authors: Eli Shaw, University of Florida School of Natural Resources and the Environment; Robert J. Fletcher Jr., University of Florida Department of Wildlife Ecology and Conservation; Thomas A. H. Smith, University of Florida Department of Wildlife Ecology and Conservation
Description: Habitat fragmentation is a major driver of species declines worldwide and poses a threat to the continued stability and security of global biodiversity. Increasing landscape connectivity is proposed as a way to mitigate the effects of fragmentation typically by two major conservation strategies: habitat corridors or stepping stones. While corridors have been extensively tested, stepping stones, small patches of habitat connecting larger patches, lack direct empirical evidence. However, small habitat patches may act as attractive sinks, absorbing dispersing individuals and creating a “shadow” effect on larger patches, decreasing disperser’s ability to reach them. This shadow effect presents the possibility that small patches actually do not function as stepping stones but instead decrease landscape connectivity. We conducted a field experiment as one of the first empirical tests of the function of small habitat patches in fragmented landscapes utilizing the cactus bug Chelinidea vittiger - prickly pear Opuntia mesacantha spp. lata study system. We examined the rate of dispersal to large habitat patches across in situ experimental arenas containing one of three configurations of intervening small patches or a control that lacked small patches. We found that both likelihood of reaching the target habitat and overall time taken, including time spent at the release site, was significantly higher in the treatment with multiple small patches. In addition, time spent actively dispersing through the matrix was significantly greater with the presence of any small habitat patch in a landscape. Connectivity models showed a similar jump in average commute distance with the introduction of any small patch to a landscape. However, an observed increase in survival for treatments with any small patch poses the question of whether small patches may still benefit dispersing species in fragmented landscapes even if these patches delay dispersal.
Examining the effect of novel background point selection methods on species distribution model accuracy
Student-Undergraduate
Authors: Anna Whitford, Benjamin Shipley, Jenny McGuire
Description: Understanding how species’ ranges shift under changing climates is key to making informed conservation and policy decisions. Presence-only species distribution models (SDM) are a common tool; however, they are highly influenced by modeling inputs, including the number and distribution of background points. Many strategies for background point sampling exist, although more research describing their effects on modelled species ranges is necessary. One recent approach to background sampling includes the use of a buffered polygon around occurrence points that spatially weighs background values adjacent to species ranges. Here, we systematically test background sampling strategy. We modelled the distributions of 21 North American mammal, reptile, and amphibian species with small and large range sizes, using a variety of SDM inputs. Those inputs included number of background points used, proportion of background points restricted to the buffer (spatial weighting), size of the background buffer, and shape of the background buffer. Our goal is to determine optimal modeling inputs for species of varying range sizes and taxa and to evaluate the influence of these inputs on the modelled species ranges.
Our preliminary results showed higher AUC values with decreased spatial weighting and low AUC values when using IUCN ranges as a buffer opposed to other, larger buffers. In addition, AUC values were lowest when the number of background points equaled the number of occurrence points. Finally, AUC values were consistently higher for small-ranged species than large-ranged species.
These conclusions suggest that models preform best when points are not weighted, many background points are used, and background buffers are large. However, AUC values are susceptible to overfitting and may not effectively describe the accuracy of the model to the true species range. Therefore, we will expand this analysis to include the Boyce Index and the concordance between predicted and expert-generated ranges to assess model accuracy.
Student-Undergraduate
Authors: Anna Whitford, Benjamin Shipley, Jenny McGuire
Description: Understanding how species’ ranges shift under changing climates is key to making informed conservation and policy decisions. Presence-only species distribution models (SDM) are a common tool; however, they are highly influenced by modeling inputs, including the number and distribution of background points. Many strategies for background point sampling exist, although more research describing their effects on modelled species ranges is necessary. One recent approach to background sampling includes the use of a buffered polygon around occurrence points that spatially weighs background values adjacent to species ranges. Here, we systematically test background sampling strategy. We modelled the distributions of 21 North American mammal, reptile, and amphibian species with small and large range sizes, using a variety of SDM inputs. Those inputs included number of background points used, proportion of background points restricted to the buffer (spatial weighting), size of the background buffer, and shape of the background buffer. Our goal is to determine optimal modeling inputs for species of varying range sizes and taxa and to evaluate the influence of these inputs on the modelled species ranges.
Our preliminary results showed higher AUC values with decreased spatial weighting and low AUC values when using IUCN ranges as a buffer opposed to other, larger buffers. In addition, AUC values were lowest when the number of background points equaled the number of occurrence points. Finally, AUC values were consistently higher for small-ranged species than large-ranged species.
These conclusions suggest that models preform best when points are not weighted, many background points are used, and background buffers are large. However, AUC values are susceptible to overfitting and may not effectively describe the accuracy of the model to the true species range. Therefore, we will expand this analysis to include the Boyce Index and the concordance between predicted and expert-generated ranges to assess model accuracy.
A Reproducible Workflow for Computing Connectivity of Protected Areas
Student-Graduate
Authors: Hejun Quan, Arizona State University; Wenxin Yang, Arizona State University; Peter Kedron, Arizona State University; Amy Frazier, Arizona State University
Description: Achieving the historic and ambitious mission of conserving at least 30% of U.S. lands and waters by 2030 is a serious challenge given the current levels of biodiversity loss, habitat destruction, and human overexploitation. Based on the research from Protected Planet, only 13.02% of terrestrial areas and 19.05% of marine areas are presently protected in the U.S. Therefore, a greater national effort is needed to increase and maintain protected areas (PAs). Rather than indiscriminately expanding the coverage of PAs, the structural connectivity among PAs is a key factor to increase the practicability and sustainability of PA estates, which affect regional biodiversity and further climate change. In this paper, we present a reproducible workflow in Python and R to compute the protected connected indicator (ProtConn; Saura et al. 2017) and partitions to measure how well the PAs in the U.S. are connected. We developed a database of terrestrial protected areas by updating the World Database of Protected Areas, with information from the current USGS Protected Areas Database. We selected areas managed for biodiversity and merged overlapping regions to prevent double counting. To minimize distance-based distortions, we simplified polygons and reprojected PAs at each analysis scale. We then computed the ProtConn metric and all partitions. The reproducible workflow utilizes Python and R.
Saura, S., Bastin, L., Battistella, L., Mandrici, A., & Dubois, G. (2017). Protected areas in the world’s ecoregions: How well connected are they?. Ecological indicators, 76, 144-158.
Student-Graduate
Authors: Hejun Quan, Arizona State University; Wenxin Yang, Arizona State University; Peter Kedron, Arizona State University; Amy Frazier, Arizona State University
Description: Achieving the historic and ambitious mission of conserving at least 30% of U.S. lands and waters by 2030 is a serious challenge given the current levels of biodiversity loss, habitat destruction, and human overexploitation. Based on the research from Protected Planet, only 13.02% of terrestrial areas and 19.05% of marine areas are presently protected in the U.S. Therefore, a greater national effort is needed to increase and maintain protected areas (PAs). Rather than indiscriminately expanding the coverage of PAs, the structural connectivity among PAs is a key factor to increase the practicability and sustainability of PA estates, which affect regional biodiversity and further climate change. In this paper, we present a reproducible workflow in Python and R to compute the protected connected indicator (ProtConn; Saura et al. 2017) and partitions to measure how well the PAs in the U.S. are connected. We developed a database of terrestrial protected areas by updating the World Database of Protected Areas, with information from the current USGS Protected Areas Database. We selected areas managed for biodiversity and merged overlapping regions to prevent double counting. To minimize distance-based distortions, we simplified polygons and reprojected PAs at each analysis scale. We then computed the ProtConn metric and all partitions. The reproducible workflow utilizes Python and R.
Saura, S., Bastin, L., Battistella, L., Mandrici, A., & Dubois, G. (2017). Protected areas in the world’s ecoregions: How well connected are they?. Ecological indicators, 76, 144-158.
Modeling prescribed fire in the Siskiyou Mountains, USA
Student-Graduate
Authors: Alison Deak, University of Oregon; Lucas C.R. Silva, University of Oregon; Shelby Weiss, University of Oregon; Michael Coughlan, University of Oregon; Melissa Lucash, University of Oregon
Description: Climate change, fire exclusion and suppression, and forest management practices have resulted in larger and more severe wildland fires throughout the western United States. Prior to Euro-American settlement and the subsequent forced displacement of Indigenous peoples, the Siskiyou Mountains were actively managed through the practice of Indigenous fire stewardship. Using fire, Indigenous peoples lowered forest stand densities, reduced fire hazard through the removal of hazardous fuels, and maintained a diversity of old growth habitats. Today, land managers and scientists increasingly stress the need to increase the pace and scale of restoration using prescribed fire to restore ecosystems and reduce wildfire risk to communities at the same time that Indigenous communities are seeking to assert their sovereignty and traditions using cultural burning practices.
Using the dynamic forest succession model LANDIS-II, natural, human-accidental, and prescribed fire ignitions were simulated across historical, low-emission (RCP 4.5), and high-emission (RCP 8.5) climate warming scenarios. Our results indicate that by reintroducing small and low-severity fires inspired by Indigenous fire stewardship practices onto the landscape at frequent intervals, the reoccurrence of large-scale, high-severity fires and the area burned by high-severity wildfire is decreased under all climate scenarios. Increased aboveground biomass also results from these lowered wildfire severities under all scenarios due to decreased mortality from repetitive high severity fires. Together these processes preclude the transition from forested to shrub ecosystems that is predicted by other ecosystem models. These results illustrate how indigenous communities and land managers can work together to add additional prescribed fire to landscapes to moderate the effects of climate change in western forests.
Student-Graduate
Authors: Alison Deak, University of Oregon; Lucas C.R. Silva, University of Oregon; Shelby Weiss, University of Oregon; Michael Coughlan, University of Oregon; Melissa Lucash, University of Oregon
Description: Climate change, fire exclusion and suppression, and forest management practices have resulted in larger and more severe wildland fires throughout the western United States. Prior to Euro-American settlement and the subsequent forced displacement of Indigenous peoples, the Siskiyou Mountains were actively managed through the practice of Indigenous fire stewardship. Using fire, Indigenous peoples lowered forest stand densities, reduced fire hazard through the removal of hazardous fuels, and maintained a diversity of old growth habitats. Today, land managers and scientists increasingly stress the need to increase the pace and scale of restoration using prescribed fire to restore ecosystems and reduce wildfire risk to communities at the same time that Indigenous communities are seeking to assert their sovereignty and traditions using cultural burning practices.
Using the dynamic forest succession model LANDIS-II, natural, human-accidental, and prescribed fire ignitions were simulated across historical, low-emission (RCP 4.5), and high-emission (RCP 8.5) climate warming scenarios. Our results indicate that by reintroducing small and low-severity fires inspired by Indigenous fire stewardship practices onto the landscape at frequent intervals, the reoccurrence of large-scale, high-severity fires and the area burned by high-severity wildfire is decreased under all climate scenarios. Increased aboveground biomass also results from these lowered wildfire severities under all scenarios due to decreased mortality from repetitive high severity fires. Together these processes preclude the transition from forested to shrub ecosystems that is predicted by other ecosystem models. These results illustrate how indigenous communities and land managers can work together to add additional prescribed fire to landscapes to moderate the effects of climate change in western forests.
Ecosystem Service Production: Relationships between Forest Age, Fragmentation, and Food Security
Student-Graduate
Authors: Aeryn Ng, The University of British Columbia; Sarah Gergel, The University of British Columbia
Description: Numerous populations across the globe are forest-dependent, whereby the livelihoods of these communities rely on forests or forest products. Through the provision of ecosystem services (ES), forests can directly and indirectly support the food security and nutrition (FSN) of local communities. In this study, we aim to explore the impact of forest age and fragmentation on the FSN of communities reliant on the Mau forest in Narok, Kenya. These communities face FSN challenges – specifically vitamin A, iron, and zinc deficiencies. The Mau forest acts as a source of these crucial nutrients; however, significant exploitation has led to forest fragmentation and loss. Currently, reforestation plans are underway. Information on the impact of both the historical forest loss and future regeneration is required to better support the long-term FSN of the local forest-dependent communities. We investigate a two-part research question with regard to the direct provision of fruits and vegetables: Does ES production differ amongst forest patches of varying (1)age?; (2)measures of edge-to-interior ratio?
Using historical aerial photographs and high-resolution satellite imagery, we age forest patches and track forest fragmentation over five decades, beginning in 1967. Fruit and vegetable production is then measured in forests of varying ages along a gradient of fragmentation. Data is collected via field work and the analysis of remotely-sensed Planetscope imagery (3m resolution). Finally, we conduct statistical regression analyses, whereby ES production is examined as a function of forest age and edge-to-interior ratio.
We expect ES production to increase with age but decrease with fragmentation. These findings will inform the value of newly-established, small, or isolated forest patches in supporting the FSN of forest-dependent communities. Overall, this research will aid in the development of forest conservation and restoration strategies that not only provide ecological benefits, but also support local FSN and ensure sustainable ES use.
Student-Graduate
Authors: Aeryn Ng, The University of British Columbia; Sarah Gergel, The University of British Columbia
Description: Numerous populations across the globe are forest-dependent, whereby the livelihoods of these communities rely on forests or forest products. Through the provision of ecosystem services (ES), forests can directly and indirectly support the food security and nutrition (FSN) of local communities. In this study, we aim to explore the impact of forest age and fragmentation on the FSN of communities reliant on the Mau forest in Narok, Kenya. These communities face FSN challenges – specifically vitamin A, iron, and zinc deficiencies. The Mau forest acts as a source of these crucial nutrients; however, significant exploitation has led to forest fragmentation and loss. Currently, reforestation plans are underway. Information on the impact of both the historical forest loss and future regeneration is required to better support the long-term FSN of the local forest-dependent communities. We investigate a two-part research question with regard to the direct provision of fruits and vegetables: Does ES production differ amongst forest patches of varying (1)age?; (2)measures of edge-to-interior ratio?
Using historical aerial photographs and high-resolution satellite imagery, we age forest patches and track forest fragmentation over five decades, beginning in 1967. Fruit and vegetable production is then measured in forests of varying ages along a gradient of fragmentation. Data is collected via field work and the analysis of remotely-sensed Planetscope imagery (3m resolution). Finally, we conduct statistical regression analyses, whereby ES production is examined as a function of forest age and edge-to-interior ratio.
We expect ES production to increase with age but decrease with fragmentation. These findings will inform the value of newly-established, small, or isolated forest patches in supporting the FSN of forest-dependent communities. Overall, this research will aid in the development of forest conservation and restoration strategies that not only provide ecological benefits, but also support local FSN and ensure sustainable ES use.
Mechanisms of Knowledge Transfer in Forest Landscape Management
Student-Graduate
Authors: Emlyn Crocker, Department of Natural Resources and the Environment, University of Connecticut; Anita T. Morzillo, Department of Natural Resources and the Environment, University of Connecticut
Description: Diverse land uses and ownerships, expanding exurban development, multiple forest stressors (e.g., severe weather, pathogens), and combinations thereof, present unique challenges for forest landscape management. Forest managers navigate these challenges by sharing knowledge and collaborating with each other throughout the forest management process. However, constraints on the flow of knowledge among stakeholders may augment such challenges and, therefore, undermine efforts to implement forest management decision-making at the landscape level. Little is known about the mechanisms of knowledge transfer among the forest management community, and how shared information facilitates decision-making processes pursued to address forest management concerns. The northeastern state of Connecticut (US) contains a combination of densely forested roadsides, extensive wildland-urban interface, and a multitude of land uses and ownerships. Our objectives were: 1) to evaluate the mechanisms of knowledge transfer used in decision-making for forest landscape management, and 2) to examine the role of forest stressors, roadside forests, management objectives, and management challenges in constraining or catalyzing knowledge transfer. Semi-structured interviews were completed with 39 stakeholders in the forest management community within the state . Interview topics included stakeholder characteristics (e.g., professional tenure ), management objectives, forest stressors, collaboration, and knowledge sought. Results of qualitative analysis suggested that stressors of greatest concern among respondents were forest pests, invasive plants, and severe weather. Knowledge sought by respondents varied but included occurrence and impacts of individual forest stressors and corresponding successful management techniques. Information needs varied across respondents based on existing knowledge, as well as individual-based management objectives and challenges. Our results suggested that management information needs are not universal and vary among managers. Therefore, decision-making processes and management outcomes may be enhanced by focusing specific knowledge to individual managers.
Student-Graduate
Authors: Emlyn Crocker, Department of Natural Resources and the Environment, University of Connecticut; Anita T. Morzillo, Department of Natural Resources and the Environment, University of Connecticut
Description: Diverse land uses and ownerships, expanding exurban development, multiple forest stressors (e.g., severe weather, pathogens), and combinations thereof, present unique challenges for forest landscape management. Forest managers navigate these challenges by sharing knowledge and collaborating with each other throughout the forest management process. However, constraints on the flow of knowledge among stakeholders may augment such challenges and, therefore, undermine efforts to implement forest management decision-making at the landscape level. Little is known about the mechanisms of knowledge transfer among the forest management community, and how shared information facilitates decision-making processes pursued to address forest management concerns. The northeastern state of Connecticut (US) contains a combination of densely forested roadsides, extensive wildland-urban interface, and a multitude of land uses and ownerships. Our objectives were: 1) to evaluate the mechanisms of knowledge transfer used in decision-making for forest landscape management, and 2) to examine the role of forest stressors, roadside forests, management objectives, and management challenges in constraining or catalyzing knowledge transfer. Semi-structured interviews were completed with 39 stakeholders in the forest management community within the state . Interview topics included stakeholder characteristics (e.g., professional tenure ), management objectives, forest stressors, collaboration, and knowledge sought. Results of qualitative analysis suggested that stressors of greatest concern among respondents were forest pests, invasive plants, and severe weather. Knowledge sought by respondents varied but included occurrence and impacts of individual forest stressors and corresponding successful management techniques. Information needs varied across respondents based on existing knowledge, as well as individual-based management objectives and challenges. Our results suggested that management information needs are not universal and vary among managers. Therefore, decision-making processes and management outcomes may be enhanced by focusing specific knowledge to individual managers.
DAY THREE - Landscape Processes and Biodiveristy
Wednesday, April 13, 2022
Small Plants vs. Big Energy: Plant Conservation and Energy Development in the Colorado Plateau
Student-Graduate
Authors: Josh Carrell, Utah State University; Thomas C. Edwards, Utah State University; Edd Hammill, Utah State University
Description: The Colorado Plateau has abundant oil, gas, and alternative energy potential. This energy potential is scattered among a patchwork of land ownership, with private, tribal, and public lands being actively developed for energy extraction. Elements of biodiversity (e.g., listed and sensitive plant and animal species) are distributed among all land tenures, yet the laws protecting them can vary as a function of land tenure. Therefore, it is imperative to understand the spatial distributions of threatened, endangered, and sensitive species in relation to land tenure to preserve habitat and conserve species populations in areas undergoing energy development. I seek to explore the interactions and relationships among land ownerships, existing and potential energy extraction, and sensitive, threatened, and endangered plant species in the Colorado Plateau region of Western North America. I will accomplish this by modelling the spatial distribution and habitat of selected threatened, endangered, and sensitive plant species using species distribution models (SDMs). SDMs portray likely spatial locations of modelled species. I will next link the SDMs with land tenure in the Colorado Plateau in a geographic information system (GIS). Next, potential energy extraction locations will be overlaid with the SDMs for the plants and land tenure. Last, spatially explicit optimization models will be developed using MARXAN that will depict various land management strategies designed to simultaneously minimize impacts on private lands and plants while maximizing energy extraction. Each scenario represents a different attitude towards the value of rare plants and the risk of energy development. Comparing these results will give insight into the financial consequences of plant species protection and quantify biodiversity costs of energy development across landscapes.
Student-Graduate
Authors: Josh Carrell, Utah State University; Thomas C. Edwards, Utah State University; Edd Hammill, Utah State University
Description: The Colorado Plateau has abundant oil, gas, and alternative energy potential. This energy potential is scattered among a patchwork of land ownership, with private, tribal, and public lands being actively developed for energy extraction. Elements of biodiversity (e.g., listed and sensitive plant and animal species) are distributed among all land tenures, yet the laws protecting them can vary as a function of land tenure. Therefore, it is imperative to understand the spatial distributions of threatened, endangered, and sensitive species in relation to land tenure to preserve habitat and conserve species populations in areas undergoing energy development. I seek to explore the interactions and relationships among land ownerships, existing and potential energy extraction, and sensitive, threatened, and endangered plant species in the Colorado Plateau region of Western North America. I will accomplish this by modelling the spatial distribution and habitat of selected threatened, endangered, and sensitive plant species using species distribution models (SDMs). SDMs portray likely spatial locations of modelled species. I will next link the SDMs with land tenure in the Colorado Plateau in a geographic information system (GIS). Next, potential energy extraction locations will be overlaid with the SDMs for the plants and land tenure. Last, spatially explicit optimization models will be developed using MARXAN that will depict various land management strategies designed to simultaneously minimize impacts on private lands and plants while maximizing energy extraction. Each scenario represents a different attitude towards the value of rare plants and the risk of energy development. Comparing these results will give insight into the financial consequences of plant species protection and quantify biodiversity costs of energy development across landscapes.
The Role of Wildfire, Restoration, and Climate in the Distributions of Northern Great Basin Plant Species
Student-Graduate
Authors: Fiona C. Noonan, Boise State University, Human-Environment Systems & Department of Geosciences; T. Trevor Caughlin, Boise State University, Department of Biological Sciences; Jodi S. Brandt, Boise State University, Human-Environment Systems; Megan E. Cattau, Boise State University, Human-Environment Systems
Description: Distributions of important rangeland plant species in the Northern Great Basin—including sagebrush (Artemisia tridentata sspp.), conifers (Juniperus spp., Pinus spp.), and invasive annual grasses (e.g. Bromus tectorum)—are shifting due to changes in fire regimes, invasive species dynamics, human land use, and the climate. Characterizing how these overlapping disturbances influence species abundance and distribution is critical for both advancing ecological theory and informing land management decision-making, but the specific interactions between these disturbances and sagebrush systems remain poorly understood. To address this gap, my research employs a joint species distribution model that incorporates anthropogenic factors to both improve predictions and better represent the social-ecological interactions at play in the Northern Great Basin’s sagebrush shrublands and sagebrush-steppe. I use a mixed-effects Bayesian conditional autoregressive modeling approach to predict distributions for multiple species with known interactions in sagebrush ecosystems, such as cheatgrass invasion and juniper expansion. The model also incorporates a suite of fire variables and restoration treatments, and uses vapor pressure deficit as a fire-relevant proxy for droughtiness. Joint predictions of Northern Great Basin plant species capture the consequences of competition and post-disturbance successional dynamics under changing fire regimes, highlighting plausible vegetation transitions. This work also points to opportunities for supporting ecological resilience in the face of ongoing disturbance interactions, particularly on Bureau of Land Management lands.
Student-Graduate
Authors: Fiona C. Noonan, Boise State University, Human-Environment Systems & Department of Geosciences; T. Trevor Caughlin, Boise State University, Department of Biological Sciences; Jodi S. Brandt, Boise State University, Human-Environment Systems; Megan E. Cattau, Boise State University, Human-Environment Systems
Description: Distributions of important rangeland plant species in the Northern Great Basin—including sagebrush (Artemisia tridentata sspp.), conifers (Juniperus spp., Pinus spp.), and invasive annual grasses (e.g. Bromus tectorum)—are shifting due to changes in fire regimes, invasive species dynamics, human land use, and the climate. Characterizing how these overlapping disturbances influence species abundance and distribution is critical for both advancing ecological theory and informing land management decision-making, but the specific interactions between these disturbances and sagebrush systems remain poorly understood. To address this gap, my research employs a joint species distribution model that incorporates anthropogenic factors to both improve predictions and better represent the social-ecological interactions at play in the Northern Great Basin’s sagebrush shrublands and sagebrush-steppe. I use a mixed-effects Bayesian conditional autoregressive modeling approach to predict distributions for multiple species with known interactions in sagebrush ecosystems, such as cheatgrass invasion and juniper expansion. The model also incorporates a suite of fire variables and restoration treatments, and uses vapor pressure deficit as a fire-relevant proxy for droughtiness. Joint predictions of Northern Great Basin plant species capture the consequences of competition and post-disturbance successional dynamics under changing fire regimes, highlighting plausible vegetation transitions. This work also points to opportunities for supporting ecological resilience in the face of ongoing disturbance interactions, particularly on Bureau of Land Management lands.
Evaluation of the key factors maintaining alternative mating strategies using a spatially explicit individual-based model
Authors: Thais Bernos, University of Toronto; Sarah Chang, University of British Columbia; Kaeli Davenport, University of Montana; Rachael Giglio, USDA-APHIS; Jeff Fisher, Seattle City Light; Erin Lowery, Seattle City Light; Marie-Josee Fortin, University of Toronto; Casey Day, University of Montana (presenting); Erin Landguth, University of Montana
Description: Alternative mating strategies (AMS), defined as discontinuous variations in behaviours, morphology, and life histories in the context of intrasexual competition, are widespread in the animal kingdom. The literature suggests that the extent of frequency-dependent and condition-dependent selection will influence the coexistence of AMS in wild populations. We examined these relationships using simulations of an empirical system produced with CDMetaPop, a spatially explicit individual-based simulation model. The simulated system resembled a metapopulation of Bull Trout (Salvelinus confluentus), a native fish species of conservation concern, in the Skagit River. We then varied the extent of density-dependent selection, AMS-specific attributes (i.e. maturation time, mortality), female mating preferences, and the initial proportion of each AMS. Finally, we used a global sensitivity analysis to test how variation in these parameters influenced the proportion of sneaker male phenotypes and sneaker allele frequency at the population and the metapopulation-level. Our preliminary results suggest that, while all the factors influenced the proportion of sneaker male phenotypes, the strength of female mating preference had the strongest influence on the number of sneaker males and sneaker allele frequencies. They also indicate that the proportion of sneaker males initially present in the system generated variability in the outcome of density-dependent selection, AMS-specific attributes, and female mating preferences, on the maintenance of AMS in the metapopulation. As AMS directly influence effective population size and population dynamics, understanding factors influencing their coexistence in wild populations has important ecological, evolutionary, and conservation ramifications.
Authors: Thais Bernos, University of Toronto; Sarah Chang, University of British Columbia; Kaeli Davenport, University of Montana; Rachael Giglio, USDA-APHIS; Jeff Fisher, Seattle City Light; Erin Lowery, Seattle City Light; Marie-Josee Fortin, University of Toronto; Casey Day, University of Montana (presenting); Erin Landguth, University of Montana
Description: Alternative mating strategies (AMS), defined as discontinuous variations in behaviours, morphology, and life histories in the context of intrasexual competition, are widespread in the animal kingdom. The literature suggests that the extent of frequency-dependent and condition-dependent selection will influence the coexistence of AMS in wild populations. We examined these relationships using simulations of an empirical system produced with CDMetaPop, a spatially explicit individual-based simulation model. The simulated system resembled a metapopulation of Bull Trout (Salvelinus confluentus), a native fish species of conservation concern, in the Skagit River. We then varied the extent of density-dependent selection, AMS-specific attributes (i.e. maturation time, mortality), female mating preferences, and the initial proportion of each AMS. Finally, we used a global sensitivity analysis to test how variation in these parameters influenced the proportion of sneaker male phenotypes and sneaker allele frequency at the population and the metapopulation-level. Our preliminary results suggest that, while all the factors influenced the proportion of sneaker male phenotypes, the strength of female mating preference had the strongest influence on the number of sneaker males and sneaker allele frequencies. They also indicate that the proportion of sneaker males initially present in the system generated variability in the outcome of density-dependent selection, AMS-specific attributes, and female mating preferences, on the maintenance of AMS in the metapopulation. As AMS directly influence effective population size and population dynamics, understanding factors influencing their coexistence in wild populations has important ecological, evolutionary, and conservation ramifications.
Do landscape structure and composition affect deer herbivory in suburban forests?
Student-Graduate
Authors: Philip P. Johnson, University of Illinois at Chicago; Wendy Leonard, Forest Preserve District of DuPage County; Scott Kobal, Forest Preserve District of DuPage County; Emily S. Minor, University of Illinois at Chicago
Description: White-tailed deer (Odocoileus virginianus) are common in suburban forest preserves of the American Midwest. Deer can have a considerable impact on plant community composition and the overall structure of forests by both browsing woody species and grazing forest herbs. The effect of deer on forest herbs is of management concern because herbs comprise the majority of plants in forests, and increased levels of herbivory can alter the viability of their populations, impeding conservation efforts. Landscape factors, such as the structure and composition of deer habitat and forage, are known to impact deer herbivory in rural forested and agricultural landscapes, but we know little about how landscape factors affect localized herbivory in suburban landscapes. It is likely that the abundant alternate food sources available to deer in suburban landscapes alter their home range size, foraging behavior, and herbivory on forest herbs. Here we present preliminary analyses of deer herbivory damage on 15 common understory herb species along 104 transects in 22 suburban forest preserves across DuPage County, Illinois, USA collected over multiple years. Along each transect, the number of grazed stems for each species was recorded, and surrounding land use and land cover data were measured to evaluate the effect of landscape composition and configuration on deer herbivory. Specifically, we use regression analyses to evaluate how preserve size and shape, along with composition of potential forage within and surrounding the forest preserves, relate to the amount of herbivory damage. This research provides insight into the role of landscape factors on deer herbivory pressure in suburban forests and can help inform future conservation and management actions.
Student-Graduate
Authors: Philip P. Johnson, University of Illinois at Chicago; Wendy Leonard, Forest Preserve District of DuPage County; Scott Kobal, Forest Preserve District of DuPage County; Emily S. Minor, University of Illinois at Chicago
Description: White-tailed deer (Odocoileus virginianus) are common in suburban forest preserves of the American Midwest. Deer can have a considerable impact on plant community composition and the overall structure of forests by both browsing woody species and grazing forest herbs. The effect of deer on forest herbs is of management concern because herbs comprise the majority of plants in forests, and increased levels of herbivory can alter the viability of their populations, impeding conservation efforts. Landscape factors, such as the structure and composition of deer habitat and forage, are known to impact deer herbivory in rural forested and agricultural landscapes, but we know little about how landscape factors affect localized herbivory in suburban landscapes. It is likely that the abundant alternate food sources available to deer in suburban landscapes alter their home range size, foraging behavior, and herbivory on forest herbs. Here we present preliminary analyses of deer herbivory damage on 15 common understory herb species along 104 transects in 22 suburban forest preserves across DuPage County, Illinois, USA collected over multiple years. Along each transect, the number of grazed stems for each species was recorded, and surrounding land use and land cover data were measured to evaluate the effect of landscape composition and configuration on deer herbivory. Specifically, we use regression analyses to evaluate how preserve size and shape, along with composition of potential forage within and surrounding the forest preserves, relate to the amount of herbivory damage. This research provides insight into the role of landscape factors on deer herbivory pressure in suburban forests and can help inform future conservation and management actions.
Impacts of Mowing and Two Causes of Isolation on Urban Ground Beetles
Student-Graduate
Authors: Michael B Roberts, University of Illinois at Chicago, Department of Biological Sciences; Dr. Crystal A. Maier, Museum of Comparative Zoology, Harvard University; Dr. Emily Minor, University of Illinois at Chicago, Department of Biological Sciences
Description: Studies on urbanization’s effect on ground beetles have increased over the last twenty years, becoming a common model for the effects of urbanization on insects. Disturbance and isolation are important variables within the urban environment that can filter for ground beetle morphology. However, most previous research uses a single qualitative category for sites that are “urban” or “not urban” rather than examining disturbance and isolation as a continuum. This makes it difficult to explain inconsistent results and move the field forward.
Here we compare the effects of two causes of isolation, distance to long term habitat and impermeable surface in the matrix, along with one ubiquitous cause of urban disturbance, mowing. We found that different causes of isolation have different effects on biodiversity: impervious surfaces in the matrix appear to filter for flight capable individuals, while distance from a long-term refuge is a better predictor of species composition at a site. Disturbance via mowing is associated with smaller average species size and larger legs relative to body size. Our study adds to a body of research quantitatively describing the habitats that ground beetles are responding to in urbanization. It also highlights that different causes of ground beetle isolation are not interchangeable and should be studied comparatively.
Student-Graduate
Authors: Michael B Roberts, University of Illinois at Chicago, Department of Biological Sciences; Dr. Crystal A. Maier, Museum of Comparative Zoology, Harvard University; Dr. Emily Minor, University of Illinois at Chicago, Department of Biological Sciences
Description: Studies on urbanization’s effect on ground beetles have increased over the last twenty years, becoming a common model for the effects of urbanization on insects. Disturbance and isolation are important variables within the urban environment that can filter for ground beetle morphology. However, most previous research uses a single qualitative category for sites that are “urban” or “not urban” rather than examining disturbance and isolation as a continuum. This makes it difficult to explain inconsistent results and move the field forward.
Here we compare the effects of two causes of isolation, distance to long term habitat and impermeable surface in the matrix, along with one ubiquitous cause of urban disturbance, mowing. We found that different causes of isolation have different effects on biodiversity: impervious surfaces in the matrix appear to filter for flight capable individuals, while distance from a long-term refuge is a better predictor of species composition at a site. Disturbance via mowing is associated with smaller average species size and larger legs relative to body size. Our study adds to a body of research quantitatively describing the habitats that ground beetles are responding to in urbanization. It also highlights that different causes of ground beetle isolation are not interchangeable and should be studied comparatively.
Integrating Human Dimensions, Encounter, and Landscape Data to Evaluate Human-Timber Rattlesnake Conflict
Student-Graduate
Authors: Abigail R. Dunn, University of Connecticut; Anita T. Morzillo, University of Connecticut; Lindsay S. (Keener-Eck) Larson, Housatonic Valley Association; Rebecca A. Christoffel, Snake Conservation Society
Description: There are increased occurrences of human-wildlife encounters among the expanding exurban landscape as human development contributes to fragmentation of wildlife habitat. To date, most human-wildlife conflict research has focused on larger wildlife species, with less attention to smaller, less-charismatic species. Here we focus on the timber rattlesnake (Crotalus horridus), a state-endangered species that has two populations in the state of Connecticut (US). Our objective was to integrate human dimensions, human-timber rattlesnake encounters, and land cover and landscape characteristics data to evaluate spatial distribution of human-timber rattlesnake conflicts. Data were collected in 2016 using a mail survey (n = 583) and reports of human-timber rattlesnake encounters. Two human attitudes variables were derived: coexistence with timber rattlesnakes and concern about presence of timber rattlesnakes. Landscape variables were associated with land cover, road density, distance to forest, and parcel characteristics. Encounter locations and attitudes variables exhibited spatial clustering, as well as relationships with several landscape variables. Individual human-timber rattlesnake encounters were more likely to be associated with parcels with greater proportions of grassland, less forest and shrub cover, and closer to forested areas. Clusters of human-timber rattlesnake encounters were more likely to be associated with greater areas of impervious surface, less forest and developed open space, and in closer proximity to forested areas. Variation existed in the relationship between encounters and road density at fine versus coarse scales. Respondents who were less favorable to coexistence with timber rattlesnakes were more likely to be located near areas containing more reported human-timber rattlesnake encounters. Results illustrated that different ecological management and public outreach techniques may be warranted in different locations across the landscape to mitigate human-timber rattlesnake conflicts.
Student-Graduate
Authors: Abigail R. Dunn, University of Connecticut; Anita T. Morzillo, University of Connecticut; Lindsay S. (Keener-Eck) Larson, Housatonic Valley Association; Rebecca A. Christoffel, Snake Conservation Society
Description: There are increased occurrences of human-wildlife encounters among the expanding exurban landscape as human development contributes to fragmentation of wildlife habitat. To date, most human-wildlife conflict research has focused on larger wildlife species, with less attention to smaller, less-charismatic species. Here we focus on the timber rattlesnake (Crotalus horridus), a state-endangered species that has two populations in the state of Connecticut (US). Our objective was to integrate human dimensions, human-timber rattlesnake encounters, and land cover and landscape characteristics data to evaluate spatial distribution of human-timber rattlesnake conflicts. Data were collected in 2016 using a mail survey (n = 583) and reports of human-timber rattlesnake encounters. Two human attitudes variables were derived: coexistence with timber rattlesnakes and concern about presence of timber rattlesnakes. Landscape variables were associated with land cover, road density, distance to forest, and parcel characteristics. Encounter locations and attitudes variables exhibited spatial clustering, as well as relationships with several landscape variables. Individual human-timber rattlesnake encounters were more likely to be associated with parcels with greater proportions of grassland, less forest and shrub cover, and closer to forested areas. Clusters of human-timber rattlesnake encounters were more likely to be associated with greater areas of impervious surface, less forest and developed open space, and in closer proximity to forested areas. Variation existed in the relationship between encounters and road density at fine versus coarse scales. Respondents who were less favorable to coexistence with timber rattlesnakes were more likely to be located near areas containing more reported human-timber rattlesnake encounters. Results illustrated that different ecological management and public outreach techniques may be warranted in different locations across the landscape to mitigate human-timber rattlesnake conflicts.
California Rangelands: Impacts of Drought on Net Primary Productivity (NPP)
Student-Undergraduate
Authors: Jeremy James, Cal Poly SLO Natural Resources and Environmental Sciences Department; Jack Alexander, Cal Poly SLO Biology Department; Mary McCafferty, Cal Poly SLO Mathematics Department; Andrew Fricker, Cal Poly SLO Social Sciences Department
Description: California rangelands, including grasslands, shrublands, and oak woodlands, comprise 57% of California’s land area and play a key role in the state’s overall carbon budget. A drying climate and increased precipitation variability pose a serious threat to the ability of these ecosystems to store carbon. We quantified spatiotemporal variation in net primary production (NPP) and asymmetry of NPP response to periods of anomalously low and high water availability in California rangelands. Using Generalized Boosted Models (GBMs), we compared NPP with several biotic and abiotic predictors at three spatial scales (4 km, 270 m, and 30 m) to identify variables strongly associated with NPP and NPP asymmetry. We incorporated these variables into a linear mixed-effects model to model NPP dynamics across seven key rangeland cover types in California. Our preliminary results indicate that variation in cumulative NPP was most strongly associated with mean early spring precipitation (R2 = 0.25, P < 0.001), mean mid-spring maximum temperature (R2 = 0.18, P < 0.001), and mean annual maximum temperature (R2 = 0.15, P < 0.001), whereas variation in asymmetry of NPP was most strongly associated with standard deviation of functional group evenness (R2 < 0.01, P < 0.001), standard deviation of annual maximum temperature (R2 < 0.01, P < 0.001), and mean functional group evenness (R2 = 0.02, P < 0.001). Coastal oak woodlands yielded the highest mean cumulative NPP, while annual grasslands yielded the lowest mean cumulative NPP. Coastal scrub, coastal oak woodland, and montane hardwood cover types demonstrated NPP gains in high water availability years that balanced NPP losses in low water availability years, while all other cover types had greater gains in NPP during high water availability years than losses during low water availability years. Our results are consistent with previous findings that precipitation and maximum temperature are the primary drivers of NPP dynamics, but we highlight important differences at fine spatial scales in NPP response to environmental conditions.
Student-Undergraduate
Authors: Jeremy James, Cal Poly SLO Natural Resources and Environmental Sciences Department; Jack Alexander, Cal Poly SLO Biology Department; Mary McCafferty, Cal Poly SLO Mathematics Department; Andrew Fricker, Cal Poly SLO Social Sciences Department
Description: California rangelands, including grasslands, shrublands, and oak woodlands, comprise 57% of California’s land area and play a key role in the state’s overall carbon budget. A drying climate and increased precipitation variability pose a serious threat to the ability of these ecosystems to store carbon. We quantified spatiotemporal variation in net primary production (NPP) and asymmetry of NPP response to periods of anomalously low and high water availability in California rangelands. Using Generalized Boosted Models (GBMs), we compared NPP with several biotic and abiotic predictors at three spatial scales (4 km, 270 m, and 30 m) to identify variables strongly associated with NPP and NPP asymmetry. We incorporated these variables into a linear mixed-effects model to model NPP dynamics across seven key rangeland cover types in California. Our preliminary results indicate that variation in cumulative NPP was most strongly associated with mean early spring precipitation (R2 = 0.25, P < 0.001), mean mid-spring maximum temperature (R2 = 0.18, P < 0.001), and mean annual maximum temperature (R2 = 0.15, P < 0.001), whereas variation in asymmetry of NPP was most strongly associated with standard deviation of functional group evenness (R2 < 0.01, P < 0.001), standard deviation of annual maximum temperature (R2 < 0.01, P < 0.001), and mean functional group evenness (R2 = 0.02, P < 0.001). Coastal oak woodlands yielded the highest mean cumulative NPP, while annual grasslands yielded the lowest mean cumulative NPP. Coastal scrub, coastal oak woodland, and montane hardwood cover types demonstrated NPP gains in high water availability years that balanced NPP losses in low water availability years, while all other cover types had greater gains in NPP during high water availability years than losses during low water availability years. Our results are consistent with previous findings that precipitation and maximum temperature are the primary drivers of NPP dynamics, but we highlight important differences at fine spatial scales in NPP response to environmental conditions.
Identifying Optimal Railway Crossing Locations for the One-horned Rhinoceros in Nepal
Student-Graduate
Authors: Nathaniel Arringdale, University of Michigan School for Environment and Sustainability; Neil Carter, University of Michigan School for Environment and Sustainability
Description: Rapidly expanding transport infrastructure, especially in biodiverse areas of the world, increases the extinction risk for endangered species by increasing mortality via animal-vehicle collisions, fragmenting their habitats, and isolating their populations. In Nepal, there are plans for a new, nation-wide railway that will likely pass through critical habitats of endangered species, including the One-horned Rhinoceros (Rhinoceros unicornis). A better understanding of how the proposed railway might impact endangered species is vitally important to mitigate the impacts before it is too late—i.e., after construction has begun. We used GPS collar data on 4 One-horned Rhinoceroses—a critically endangered species—to assess how the proposed railway would affect their habitats and movements. We used an integrated step-selection function to model fine-scale habitat selection and movement patterns, which in turn were used to estimate least-cost paths rhinos would use to cross the planned railway. We identified areas best suited for wildlife crossing based on paths rhinos are currently taking. These areas were in forested areas right along the main river in the area. Protecting these areas along the river will be critical along the planned highway. The mean step length for the rhinos was 181.02 meters. The habitat characteristics of where rhinos moved were areas of low slope (coefficient: -0.025, p-value: 8.7e-5), dense vegetation (coefficient: 5.29, p-value: <2e-16), and near water (coefficient: 5.20, p-value: <2e-16). While the land cover types positively associated with rhino movements were closed forests, both deciduous broadleaf (coefficient: 0.25, p-value:2.56e-12) and evergreen broadleaf (coefficient: 0.47, p-value: <2e-16). It is critical to use field data to identify how these rhinos choose habitat to move through. Conducting research before construction begins can provide opportunities to properly plan areas to mitigate habitat alterations and implement crossing structures.
Student-Graduate
Authors: Nathaniel Arringdale, University of Michigan School for Environment and Sustainability; Neil Carter, University of Michigan School for Environment and Sustainability
Description: Rapidly expanding transport infrastructure, especially in biodiverse areas of the world, increases the extinction risk for endangered species by increasing mortality via animal-vehicle collisions, fragmenting their habitats, and isolating their populations. In Nepal, there are plans for a new, nation-wide railway that will likely pass through critical habitats of endangered species, including the One-horned Rhinoceros (Rhinoceros unicornis). A better understanding of how the proposed railway might impact endangered species is vitally important to mitigate the impacts before it is too late—i.e., after construction has begun. We used GPS collar data on 4 One-horned Rhinoceroses—a critically endangered species—to assess how the proposed railway would affect their habitats and movements. We used an integrated step-selection function to model fine-scale habitat selection and movement patterns, which in turn were used to estimate least-cost paths rhinos would use to cross the planned railway. We identified areas best suited for wildlife crossing based on paths rhinos are currently taking. These areas were in forested areas right along the main river in the area. Protecting these areas along the river will be critical along the planned highway. The mean step length for the rhinos was 181.02 meters. The habitat characteristics of where rhinos moved were areas of low slope (coefficient: -0.025, p-value: 8.7e-5), dense vegetation (coefficient: 5.29, p-value: <2e-16), and near water (coefficient: 5.20, p-value: <2e-16). While the land cover types positively associated with rhino movements were closed forests, both deciduous broadleaf (coefficient: 0.25, p-value:2.56e-12) and evergreen broadleaf (coefficient: 0.47, p-value: <2e-16). It is critical to use field data to identify how these rhinos choose habitat to move through. Conducting research before construction begins can provide opportunities to properly plan areas to mitigate habitat alterations and implement crossing structures.
Distribution of the Eastern Grey Squirrel Along an Urban-Rural Gradient in Southeastern Massachusetts
Student-Undergraduate
Authors: Bobbi Scully, Department of Biological Sciences1, Bridgewater State University; Thilina D. Surasinghe, Department of Biological Sciences1, Bridgewater State University; Jahaziel Garrido-Daly, Department of Natural Science and Mathematics, Lesley University; Amy Mertl2, 2Department of Natural Science and Mathematics, Lesley University; Christopher Richardson, Department of Natural Science and Mathematics, Lesley University; Brendan P. Keegan, Arnold Arboretum of Harvard University; Jeffrey Taylor, Zoo New England
Description: Documenting wildlife distribution along urban to rural gradients (URGs) creates opportunities to comparatively analyze wildlife habitat associations. In urban environments, the wildlife diversity is expected to be impoverished and limited to invasive and generalist species while a much diverse wildlife assemblages are expected in rural environments. Using motion-triggered game cameras, we documented presence of the Eastern Grey Squirrel (EGS) along an URG in Southeastern Massachusetts, to determine the effects of urbanization on the EGS. We hypothesized EGS abundance to increase within increasing built-up land-cover where competition and predation can plummet in species-depauperated urban landscapes. Their generalist life-history strategies help them capitalize human-subsidized resources plentiful in urban settings. Our study covered 27 sites, that radiate southward from Greater Boston area with decreasing urban land cover. Our survey period spanned across a year (January 2021 to January 2022) where a single sampling month represented a different season: January months for Winter, April for Spring, July for Summer, and October for Fall. Each photograph was annotated with species identification. The EGS aside, we identified 10 mammalian species collectively from all sites. Our preliminary analyses (data from spring and summer) indicated significant variations in the EGS sightings along an URG. A negative-binomial generalized linear model suggested that both predominant land-cover type at the survey site (χ2 = 7.45, p = 0.02) and the survey month (χ2 = 29.84, p = 3.31×10-7) are significant predictors of the EGS sightings, while the influence of the percent impervious surfaces per survey site was insignificant. In the future, we will continue our analyses with the inclusion of occupancy modeling to test hypotheses on ESG distribution as well as distribution of other predators (i.e., coyote) and prey (i.e., eastern cottontail) to understand how predator-prey dynamics changes along the URG.
Student-Undergraduate
Authors: Bobbi Scully, Department of Biological Sciences1, Bridgewater State University; Thilina D. Surasinghe, Department of Biological Sciences1, Bridgewater State University; Jahaziel Garrido-Daly, Department of Natural Science and Mathematics, Lesley University; Amy Mertl2, 2Department of Natural Science and Mathematics, Lesley University; Christopher Richardson, Department of Natural Science and Mathematics, Lesley University; Brendan P. Keegan, Arnold Arboretum of Harvard University; Jeffrey Taylor, Zoo New England
Description: Documenting wildlife distribution along urban to rural gradients (URGs) creates opportunities to comparatively analyze wildlife habitat associations. In urban environments, the wildlife diversity is expected to be impoverished and limited to invasive and generalist species while a much diverse wildlife assemblages are expected in rural environments. Using motion-triggered game cameras, we documented presence of the Eastern Grey Squirrel (EGS) along an URG in Southeastern Massachusetts, to determine the effects of urbanization on the EGS. We hypothesized EGS abundance to increase within increasing built-up land-cover where competition and predation can plummet in species-depauperated urban landscapes. Their generalist life-history strategies help them capitalize human-subsidized resources plentiful in urban settings. Our study covered 27 sites, that radiate southward from Greater Boston area with decreasing urban land cover. Our survey period spanned across a year (January 2021 to January 2022) where a single sampling month represented a different season: January months for Winter, April for Spring, July for Summer, and October for Fall. Each photograph was annotated with species identification. The EGS aside, we identified 10 mammalian species collectively from all sites. Our preliminary analyses (data from spring and summer) indicated significant variations in the EGS sightings along an URG. A negative-binomial generalized linear model suggested that both predominant land-cover type at the survey site (χ2 = 7.45, p = 0.02) and the survey month (χ2 = 29.84, p = 3.31×10-7) are significant predictors of the EGS sightings, while the influence of the percent impervious surfaces per survey site was insignificant. In the future, we will continue our analyses with the inclusion of occupancy modeling to test hypotheses on ESG distribution as well as distribution of other predators (i.e., coyote) and prey (i.e., eastern cottontail) to understand how predator-prey dynamics changes along the URG.
Broad-scale Population Trend for the California Quail in California: Inferences from a Cycling Population
Student-Graduate
Authors: Sarah K. Jacobson, Caesar Kleberg Wildlife Research Institute, Texas A&M University Kingsville; Leonard A. Brennan, Caesar Kleberg Wildlife Research Institute, Texas A&M University Kingsville; Humberto L. Perotto-Baldivieso, Caesar Kleberg Wildlife Research Institute, Texas A&M University Kingsville; Evan P. Tanner, Caesar Kleberg Wildlife Research Institute, Texas A&M University Kingsville; Katherine S. Miller, California Department of Fish and Wildlife
Description: The California quail (Callipepla californica) is an important upland gamebird for California and the Pacific region of the United States. During the past half century, large regions in California have become increasingly fragmented and degraded as a result of urban development, altered forest and rangeland management, and large-scale agriculture. These and other factors could threaten the persistence of California quail in some areas. Our goal was to determine the long-term population trend of the California quail throughout its geographic range in California. We used data from the North American Breeding Bird Survey (1968-2019) to create relative abundance maps and developed 5-year-averages to account for annual variability in abundance. We then established 50 spatially balanced random points to calculate mean birds/route from 1970-2017. The California quail population displayed a clear cycle-driven pattern over time, with a small peak in the early 1980s and larger peaks in the mid-1990s and late 2000s. Fitting a time series autocorrelation function indicated that the California quail population was stationary with an endogenously generated periodicity. Possible explanations include broad-scale responses to climatic patterns, especially widespread drought, and a combination of increasing and decreasing populations at local scales. Currently we are identifying areas with declining and increasing populations and will compare road density, human population density, and land use between them. The information gathered from this study will provide a quantitative basis for wildlife biologists and stakeholders to prioritize areas for quail population and habitat conservation on a state-wide basis for California.
Student-Graduate
Authors: Sarah K. Jacobson, Caesar Kleberg Wildlife Research Institute, Texas A&M University Kingsville; Leonard A. Brennan, Caesar Kleberg Wildlife Research Institute, Texas A&M University Kingsville; Humberto L. Perotto-Baldivieso, Caesar Kleberg Wildlife Research Institute, Texas A&M University Kingsville; Evan P. Tanner, Caesar Kleberg Wildlife Research Institute, Texas A&M University Kingsville; Katherine S. Miller, California Department of Fish and Wildlife
Description: The California quail (Callipepla californica) is an important upland gamebird for California and the Pacific region of the United States. During the past half century, large regions in California have become increasingly fragmented and degraded as a result of urban development, altered forest and rangeland management, and large-scale agriculture. These and other factors could threaten the persistence of California quail in some areas. Our goal was to determine the long-term population trend of the California quail throughout its geographic range in California. We used data from the North American Breeding Bird Survey (1968-2019) to create relative abundance maps and developed 5-year-averages to account for annual variability in abundance. We then established 50 spatially balanced random points to calculate mean birds/route from 1970-2017. The California quail population displayed a clear cycle-driven pattern over time, with a small peak in the early 1980s and larger peaks in the mid-1990s and late 2000s. Fitting a time series autocorrelation function indicated that the California quail population was stationary with an endogenously generated periodicity. Possible explanations include broad-scale responses to climatic patterns, especially widespread drought, and a combination of increasing and decreasing populations at local scales. Currently we are identifying areas with declining and increasing populations and will compare road density, human population density, and land use between them. The information gathered from this study will provide a quantitative basis for wildlife biologists and stakeholders to prioritize areas for quail population and habitat conservation on a state-wide basis for California.
Investigating Prescribed Fire Effects on Mid-Atlantic Tick Populations Through a Landscape Lens
Student-Graduate
Authors: Olivia Spencer, Department of Geography, The Pennsylvania State University; Erica A.H. Smithwick, Department of Geography, The Pennsylvania State University
Description: To address fire risk and promote critical wildlife habitat in Pennsylvania, prescribed fire burning has increased five-fold since 2009, but knowledge of forest response from this practice has been limited. Recent research at Penn State found that a top perceived benefit of prescribed burning among local forest users in Pennsylvania is the potential for controlling tick populations, though this relationship is not yet established in the area. Landscape ecology approaches are useful for understanding the complexities associated with tick phenology through time and space. To date, landscape ecology studies of ticks primarily focus on the northeast and do not directly investigate firescape mosaics. Studies in the southern and mid-western U.S. have found varied effects of prescribed burning on tick abundance, noting low abundance immediately following burn and varying effects with long-term burning, but have not considered landscape-scale covariates. This work aims to characterize black-legged tick (Ixodes scapularis) abundance and distribution in Pennsylvania in response to prescribed fire impacts on ecosystems through a landscape ecology lens. We account for landscape metrics including habitat fragmentation, patch density, and distance to roads alongside ecologic and geologic gradients associated with tick habitat suitability including elevation, slope, aspect, soil type, land cover, and hydrology. This landscape approach allows for the isolation of time-since-burn as a potential explanatory variable of tick abundance, a relationship that is not yet understood despite important implications for community concerns of tick control. Following tick collection in the spring and summer of 2022, abundance patterns will be modeled across varying burn regimes. We anticipate black-legged tick abundance will be low shortly following fire and will gradually increase with time as viable animal hosts seek out regenerating vegetation in post-burn forest patches and reintroduce ticks to those areas. These results will ultimately provide valuable insight into tick distributions across firescapes.
Student-Graduate
Authors: Olivia Spencer, Department of Geography, The Pennsylvania State University; Erica A.H. Smithwick, Department of Geography, The Pennsylvania State University
Description: To address fire risk and promote critical wildlife habitat in Pennsylvania, prescribed fire burning has increased five-fold since 2009, but knowledge of forest response from this practice has been limited. Recent research at Penn State found that a top perceived benefit of prescribed burning among local forest users in Pennsylvania is the potential for controlling tick populations, though this relationship is not yet established in the area. Landscape ecology approaches are useful for understanding the complexities associated with tick phenology through time and space. To date, landscape ecology studies of ticks primarily focus on the northeast and do not directly investigate firescape mosaics. Studies in the southern and mid-western U.S. have found varied effects of prescribed burning on tick abundance, noting low abundance immediately following burn and varying effects with long-term burning, but have not considered landscape-scale covariates. This work aims to characterize black-legged tick (Ixodes scapularis) abundance and distribution in Pennsylvania in response to prescribed fire impacts on ecosystems through a landscape ecology lens. We account for landscape metrics including habitat fragmentation, patch density, and distance to roads alongside ecologic and geologic gradients associated with tick habitat suitability including elevation, slope, aspect, soil type, land cover, and hydrology. This landscape approach allows for the isolation of time-since-burn as a potential explanatory variable of tick abundance, a relationship that is not yet understood despite important implications for community concerns of tick control. Following tick collection in the spring and summer of 2022, abundance patterns will be modeled across varying burn regimes. We anticipate black-legged tick abundance will be low shortly following fire and will gradually increase with time as viable animal hosts seek out regenerating vegetation in post-burn forest patches and reintroduce ticks to those areas. These results will ultimately provide valuable insight into tick distributions across firescapes.
Using spatially explicit models to assess the effects of socio-ecological covariates on the occurrence of an invasive aquatic plant in Great Lakes coastal wetlands
Student-Graduate
Authors: Louis Jochems, Jodi Brandt, Andrew Monks, Shane Lishawa, Trevor Caughlin, Juan Requena-Mullor, Don Uzarski
Description: Invasive plant species pose a major threat to wetland ecosystems. One effective way to control the spread of invasive plants is to catch them early in the invasion process, and species distribution models (SDMs) provide a means of predicting where new invasions are likely to occur. Since the species has not yet filled its niche during the initial stages of invasion, proximity to known presences and/or vectors (spatial autocorrelation) may control its spread as much as habitat suitability. Yet, many SDMs assume that the species has filled its niche, incorporate only biophysical predictors, and do not consider spatial autocorrelation. For this study, we fit Integrated Nested Laplace Approximation (INLA) SDMs of the occurrence (presence and absence) and abundance (% cover) of the invasive aquatic plant, Hydrocharis morsus-ranae (European frogbit; EFB), as a response to socio-ecological covariates and spatial autocorrelation in Great Lakes coastal wetlands. We modeled a time-series of 1,853 field observations recorded across Michigan from 2011-2018 and produced marginal posterior distributions of model parameters, including spatial autocorrelation. The most parsimonious model included the following important predictors on occurrence: distance to boat launches (km), emergent plant cover, fetch (wave energy), and water depth (cm), and altogether they explained 29 % variance in EFB occurrence, with r2 increasing to 40 % with spatial autocorrelation. The most parsimonious abundance model included only water depth, which explained 50 % of variance in EFB abundance, but spatial autocorrelation decreased r2 to 43 %. Moreover, including spatial autocorrelation improved information criteria for both models. Overall, our results indicate that incorporating spatial autocorrelation is necessary for obtaining reliable estimates of the socio-ecological drivers of invasive species. Our INLA approach also enables prediction of uninvaded plots to become invaded in future years. We conclude that INLA SDMs can inform early detection and management of expanding invasive plant species.
Student-Graduate
Authors: Louis Jochems, Jodi Brandt, Andrew Monks, Shane Lishawa, Trevor Caughlin, Juan Requena-Mullor, Don Uzarski
Description: Invasive plant species pose a major threat to wetland ecosystems. One effective way to control the spread of invasive plants is to catch them early in the invasion process, and species distribution models (SDMs) provide a means of predicting where new invasions are likely to occur. Since the species has not yet filled its niche during the initial stages of invasion, proximity to known presences and/or vectors (spatial autocorrelation) may control its spread as much as habitat suitability. Yet, many SDMs assume that the species has filled its niche, incorporate only biophysical predictors, and do not consider spatial autocorrelation. For this study, we fit Integrated Nested Laplace Approximation (INLA) SDMs of the occurrence (presence and absence) and abundance (% cover) of the invasive aquatic plant, Hydrocharis morsus-ranae (European frogbit; EFB), as a response to socio-ecological covariates and spatial autocorrelation in Great Lakes coastal wetlands. We modeled a time-series of 1,853 field observations recorded across Michigan from 2011-2018 and produced marginal posterior distributions of model parameters, including spatial autocorrelation. The most parsimonious model included the following important predictors on occurrence: distance to boat launches (km), emergent plant cover, fetch (wave energy), and water depth (cm), and altogether they explained 29 % variance in EFB occurrence, with r2 increasing to 40 % with spatial autocorrelation. The most parsimonious abundance model included only water depth, which explained 50 % of variance in EFB abundance, but spatial autocorrelation decreased r2 to 43 %. Moreover, including spatial autocorrelation improved information criteria for both models. Overall, our results indicate that incorporating spatial autocorrelation is necessary for obtaining reliable estimates of the socio-ecological drivers of invasive species. Our INLA approach also enables prediction of uninvaded plots to become invaded in future years. We conclude that INLA SDMs can inform early detection and management of expanding invasive plant species.