2021

Tianshu Li (tl6gk@virginia.edu), University of Virginia
Mohamad Alipour, University of California, Los Angeles
Bridget Donaldson, Virginia Transportation Research Council
Devin Harris, University of Virginia

Bat inventory surveys on bridges, structures, and dwellings are an important step in protecting threatened and endangered bat species that use the infrastructure as roosting locations. Observing guano droppings and staining is a common indicator of bat presence, but it can be difficult to verify whether certain stains originated from bats or other sources such as water seeps, rust staining, asphalt leaching, or other structural deterioration mechanisms. While bat indicators are hard to distinguish by humans without training, from a computer vision perspective they show different features that, coupled with expert opinion, can be used for automated detection of bat presence. To facilitate bat presence detection and streamline bat surveys, this paper leverages recent advances in visual recognition using deep learning to develop an image classification system that identifies bat indicators. An array of state-of-the-art convolutional neural networks were investigated. To overcome the shortage of data, parameters previously trained on large-scale datasets were used to transfer the learned feature representations. Using a pool of digital photographs collected through the Virginia Department of Transportation (1), a visual recognition model was developed and achieved 92.0% accuracy during testing. To facilitate the application of the developed model, a prototype web application was created to allow users to interactively upload images of stains on structures and receive classification results from the model. The web application is being deployed by VDOT in a pilot study and the success of the proposed approach is expected to help facilitate bat inventory surveys and the resulting conservation efforts.

Basak Bektas, Minnesota State University, Mankato
Julie Blanchong, Iowa State University
Katelyn Freeseman, Iowa State University
Ahmed AlBughdadi, Iowa State University

White-Nose Syndrome and increased disturbance of habitats used by bats for roosting and foraging has led to a growing concern about the bat population in the US in the last decade. Bridge repair and replacement projects are required to follow additional regulatory requirements to avoid and minimize impacts to bats when protected bat species are present on bridges. Some of these requirements (e.g., timing restrictions) are challenging to implement, given Minnesota’s short construction season. The objective of this project was to evaluate the feasibility and efficacy of deploying non-lethal ultrasonic acoustic devices in the field to temporarily deter bats from roosting on bridges ahead of construction or maintenance activities, while minimizing harm to bats and non-target species. The technology was evaluated at two test sites located in Minnesota, one short term and one long term, during the summer of 2019. Considering the findings from both acoustic monitoring data and the field inspections, acoustic deterrents appear to effectively work to temporarily deter bats from select abutments and present an option to temporarily deter bats at construction sites. The implementation of this technology by any transportation agency will require the development of an agency protocol, through collaboration of relevant offices.

Damon Fick, Montana State University
Robert Ament, Western Transportation Institute (WTI)
Matthew Bell, Western Transportation Institute (WTI)
Nina Marie Lister, Montana State University, Bozeman

There are decades of studies demonstrating the effectiveness of wildlife overpasses and underpasses in conjunction with fencing to reduce wildlife-vehicle collisions and provide ecological connectivity. Current methods to build these structures with concrete and steel result in high costs due to their need to support large loads (e.g., soils, vegetation), lengthy construction times, and maintenance costs during their service life. Recently, engineers and architects are finding new applications of fiber-reinforced polymer (FRP) composites due to their high strength-to-weight ratio, long service life, and low life-cycle costs that might provide an alternative material. FRP is better suited to withstand environmental stresses compared to concrete and metal, are more corrosion resistant and are more durable over longer periods of time, and therefore, can reduce life-time maintenance costs. The carbon footprint of the FRP materials, combined with transportation and construction efficiencies can be lower compared to concrete and metal structures. Improvements continue to be developed for FRP materials to lower the carbon footprint of both fibers and resins by increasing the use of bio-based and recycled materials.

Currently FRP bridges can be designed to support the static and dynamic loads of vehicle traffic; making it more than capable to support the loads required for wildlife crossing infrastructure. Adapting FRP bridges and other FRP structural components for wildlife crossings can contribute to the multiple goals of improving motorist safety, conserving wildlife, lowering infrastructure costs, reducing greenhouse gas emissions, and reducing plastics in the environment.

This active research project will report on the progress it has made to adapt FRP materials into North America’s first wildlife overpass structure and related crossing design elements, such as sound barriers and fencing. A site for the crossing has been selected in northern California in partnership with the California Department of Transportation (Caltrans). To date, a literature review and an evaluation of commercial FRP manufacturer products available in the North American market have been completed. Exploratory structural designs are underway and a team of experts has been empaneled to provide additional design recommendations.  In addition, future plans and challenges associated with implementing the FRP wildlife overpass will be described.

Fraser Shilling, University of California, Davis
Tricia Nguyen, University of California, Davis
Malek Saleh, University of California, Davis
Min Khant Kyaw, University of California, Davis
Karla Tapia, University of California, Davis
Gabrielle Trujillo, University of California, Davis
Mireya Bejarano, University of California, Davis
David Waetjen, University of California, Davis
Jon Peterson, Washington State Department of Transportation
Glen Kalisz, Washington State Department of Transportation
Raquel Sejour, Washington State Department of Transportation
Sarah Croston, Washington State Department of Transportation
Eric Ham, Maine Department of Transportation

High traffic volume is a primary contributor to wildlife-vehicle conflict (WVC) and wildlife mortality on roads. Government shelter-in-place (SIP) orders have been used to help mitigate the spread of COVID-19, resulting in unprecedented reductions in global traffic volumes. Our objective was to investigate how these orders may have affected traffic and wildlife. Using traffic and collision data from four US states (California, Idaho, Maine, and Washington), we investigated changes in total WVC, following the state and local SIP orders. From early March to mid-April, 2020, these orders have resulted in up to 71%, 63%, 73%, and 72% reduction in driving, as measured by vehicle miles traveled (VMT), in CA, ID, ME, and WA respectively. The daily WVC rates declined 34% between the 4 weeks prior to SIP orders going into effect, to the 4 weeks after, with 21, 36, 44, and 33% declines for CA, ID, ME, and WA, respectively. The reductions in WVC from 1 month pre-SIP orders to 1 month post-order only occurred in 2020 and not 2015, 2016, 2017, 2018, or 2019, suggesting that the reductions were associated with the reductions in traffic. Mountain lions (Puma concolor) are proposed for listing as endangered in CA and for them there was a 58% decline in mortality during the traffic reduction. The measured declines in WVC reversed in ME and WA during May, June and July, 2020, paralleling reversals in traffic volumes. A 34% reduction in WVC would potentially equate to 10s of millions fewer vertebrates killed on US roadways during one month of traffic reduction, representing an unintentional conservation action unprecedented in modern times.

Thomas Yamashita, Texas A&M University, Kingsville
Zachary Wardle, Texas A&M University, Kingsville
Jason Lombardi, Texas A&M University, Kingsville
Michael Tewes, Texas A&M University, Kingsville
John Young, Texas Department of Transportation

Roads can have major impacts on wildlife, causing habitat fragmentation and direct mortality from wildlife-vehicle collisions. Therefore, wildlife crossings are often constructed on roads to help promote connectivity and reduce wildlife road mortality. In order to assess the effectiveness of these structures over long time periods, it is important to establish baseline rates of use along the highway before the mitigation structures are constructed. In South Texas, wildlife crossings are being built to reduce threats of roads to the endangered ocelot (Leopardus pardalis). However, these crossings benefit a wide variety of species in addition to ocelots. In this study, we assessed baseline usage of future wildlife mitigation structures on Farm-to-Market (FM) 1847 in eastern Cameron County, Texas. The pre-construction preparation for the construction of wildlife mitigation structures on FM 1847 began in February 2020. We used camera traps set at the locations where the mitigation structures will be built to monitor wildlife use of those sites from July 2019 through February 2020. We determined the number of individual photographic events, camera trap success rates, and diel activity for all large and medium-sized mammal species. We identified between six and 12 unique species at each of the five crossing locations. Bobcats were detected at four of the crossings and were most prominent at an underpass along a resaca (crossing 2). Most wild species showed nocturnal activity, while domestic cats and domestic dogs showed diurnal and nocturnal activity. These results will help establish baseline usage of future wildlife crossing locations on FM 1847 and allow us to assess their effectiveness at reducing road mortality and increasing landscape connectivity.

David Waetjen, University of California, Davis
Fraser Shilling, University of California, Davis

Animal observation data on and near roads are critical to understanding wildlife vehicle conflict (WVC). The California Highway Patrol (CHP), which helps to monitor and maintain safety on California roads, also report incidents of “animal hazards”, “live or dead animal” and crashes involving animals, providing a continuous source of WVC records. These incidents often include collisions with mule deer (Odocoileus hemionus), which pose serious risks during collisions due to their large size. CHP report their observations in real-time through their own public-facing web-system. We developed a web-system to harvest and automatically ingest these real-time CHP reports called the California Highway Incident Processing System (CHIPS). We used a custom data-scraping protocol to collect the data every 15 minutes from the CHP site and update the CHIPS database. The data were maintained in an SQL database and is queried periodically providing data to automated “pipelines” for use in various tools and services. One online tool is the Real-time Deer Incidents & Wildlife-Vehicle Conflict (WVC) Hotspots map (https://roadecology.ucdavis.edu/hotspots/map), which shows real-time and recent animal incidents on top of a hotspots analysis map (generated from the Road Ecology Center’s hotspots tool). The location data is accurate to approximately 10 to 20 meters, which enables the individual records to be associated with a county and highway. The CHP officers often include detailed descriptions about the animals’ fates, including whether they were injured, or had to be dispatched. Using these data, we have met a variety of needs in California: 1) approximately once per month we receive and meet ad hoc requests for data from agency, consultant and academic scientists; 2) we produce an annual WVC hotspots report; and 3) we have developed the only continuously-maintained repository for WVC in the state. Most recently, we used these data to demonstrate that the rate of WVC declined during the COVID-19 shelter at home orders from the state government. These tools and data are critical aids to decision making; for example, clustered hotspots of incidents can indicate where wildlife mitigation measures can be taken, such as building a wildlife overpass.

David Waetjen, University of California, Davis
Christian John, University of California, Davis
Fraser Shilling, University of California, Davis

State departments of transportation are increasing the rate at which they collect data and improve interactions between wildlife and highways. Our objective was to build upon previous research and relationships with states DOTs and their partners to develop a data management website (http://wildlifeobserver.net) that also provides analytical services. Organizations operating camera trap arrays to monitor wildlife have stressed that several aspects of data collection and analysis are particularly challenging: 1) managing many thousands of images during collection, processing, and analysis; 2) false triggers of camera traps creating many hours of unneeded work for analysts; 3) being able to identify species automatically, especially common ones; 4) having the ability to annotate video records for wildlife behavior; and 5) determining if wildlife transit crossing structures, the measure of ecological effectiveness. We developed two primary data collection and management services to: a) ingest and automatically process images from cell and wifi communicating cameras and b) allow bulk uploading of all images from a camera’s memory card. To identify false positive images, we also developed an image processing service (“ImageID”) based on artificial intelligence to identify images that contain an animal (or human, or vehicle). We estimate that the 2 instances of the tool available, running at full capacity, can save >$800,000 in analyst costs per year. Each additional instance would add another ~$400,000/year in potential cost-savings. We built services for annotating videos of animal behavior (“BehaviorID”), increasingly seen as important for understanding animal interaction with roads. Finally, we created a tool (“EventID”) that analyzes groups of images to determine if they are related (i.e., of the same animal) and if an animal traversed a wildlife crossing, the most critical ecological outcome of these structures. These combined services provide a user the potential to carry out end-to-end automation of most steps from field data collection to species and individual-level analysis. They also could provide significant cost-savings to managers and funders of camera trap systems.

Sean McDowell, University of California, Davis
Adetayo Oke, University of California, Davis
Aaron Bass, University of California, Davis
David Waetjen, University of California, Davis
Travis Longcore, University of California, Los Angeles
Winston Vickers, University of California, Davis
Fraser Shilling, University of California, Davis

Transportation and other agencies and organizations are increasingly planning and building crossing structures for wildlife to traverse busy highways. Several criteria are key for the success of these structures – sufficient safety and/or conservation need, cost, location, and anticipated use by wildlife. Although wildlife biologists are aware of the potential for traffic noise and light to impede structure use by wildlife, existing guidance is inadequate to inform their advice to designers, engineers, architects and habitat designers of the structural and vegetation elements that could reduce disturbance. Each structure represents a significant investment in time, money, and effort. Therefore, if wildlife are hesitant or refuse to approach structures due to noise, light and other factors, then the structures’ benefit-cost ratios may be diminished. In previous research we have highlighted the importance that traffic noise and light has in reducing species presence and impacting individual animal behavior when approaching crossings. We describe a solution to this problem, where field measurements and modeling of traffic-sourced light and noise were used to improve and test wildlife crossing design. We developed wildlife-responsive designs for two crossings being considered or planned by the California Department of Transportation: 1) the proposed Liberty Canyon wildlife over-crossing (LCOC) across US 101, and 2) an over-crossing at a very early stage of consideration across Interstate 15, south of Temecula (TOC). For LCOC, we developed and tested 3 different designs (sound wall, sound wall + berm, and sound wall + multiple berms) for modeled propagation of traffic noise and light. For TOC, we developed and tested one design. We identified key limitations and opportunities for each design approach. We concluded that a combination of berms and walls could be used to decrease disturbance in the crossing structure approach zones, increasing the wildlife-responsiveness of the designs.

Fraser Shilling, University of California, Davis

Globally, nationally and locally, wildlife-vehicle conflict (WVC) fragments wildlife populations (due to road/traffic-aversion), kills and injures individual animals, can cause wildlife population declines, may eventually contribute to local or total extinction of certain species, and can harm vehicles and drivers. WVC is defined here as any negative interaction between vehicles and wildlife. Preventing WVC begins with recording locations of conflict, such as vehicle crashes, animal carcasses (roadkill), or animal behavior around roads, such as avoidance of roads or crossing-behavior. These data are ideally used to inform transportation policy and planning and to retrofit roadways and their structures to reduce WVC. The goal of this study was to develop standards for WVC data collection and management with partners at multiple scales. A companion paper is In Press in Biological Conservation. My objective is to use these decision-support roles to describe ways that WVC data, metadata, data collection, and reporting could be standardized across all states. The standards are based on review of systems for recording WVC by volunteers and agency staff at different geographical scales, based on existing literature, experience with several state-scale systems, and partnerships with other countries engaged in a similar process. There is currently a wide range of WVC data collection methods, relying primarily on web-systems and apps. Although the “technology stacks” employed by these systems vary somewhat in the specific software, there is a common range of variables collected among all of them. These common themes and elements could be used to update existing systems, or replicated to create new systems. There are several types of groups, volunteers, and agencies involved in the different systems, each of which could require different data input mechanisms. I will describe methodological specifications based on current best-practice for collecting and using WVC data to inform transportation and conservation decisions. I will conclude with a vision of next steps toward a national network of independent WVC reporting systems that have clear and practical applications for improved conservation research and management of road networks.

Autumn Iverson, University of California, Davis
David Waetjen, University of California, Davis
Fraser Shilling, University of California, Davis

Habitat linkages are hypothetical objects developed in geographic information systems (GIS) proposed to connect areas of habitats in fragmented landscapes. Assuming they are used by organisms in nature, linkages could be an important tool for the maintenance of viable wildlife populations and biodiversity conservation.  The Desert Renewable Energy Conservation Plan (DRECP) is a landscape-level planning effort in the California desert, and includes the long-term conservation of species and natural resources. A connectivity model was developed for the area, identifying protected areas and proposed linkages. However, these linkages have not been evaluated using wildlife detection data.

We used twenty years of opportunistic wildlife detection data collected from six biodiversity databases: Biodiversity in Service of Our Nation (BISON, US Geological Survey), California Natural Diversity Database (CNDDB, California Department of Fish and Wildlife), HerpMapper (private database maintained by professional biologists), DRECP assessments by Dudek, as well as wildlife carcass and crash data from the California Highway Incident Processing System (CHIPS, Road Ecology Center) and the California Roadkill Observation System (CROS, Road Ecology Center). Using these detections and a 1-km squared grid over the California desert region, we employed occupancy modeling in R to assess the effect of variables such as major roads and proposed linkages on the occupancy probability of 20 native reptiles, amphibians, and mammals species/species groups. Occupancy modeling estimates “true” occurrence probability and the factors affecting it, while correcting for imperfect detection.    

The distance to a major road was an important detection predictor for four species, with greater distance from a road resulting in higher detection probability for one species (bighorn sheep) and lower detection for three (deer, gopher snake and western banded gecko). Proposed linkages were not important predictors for the occupancy or detection of most species, with three showing positive (glossy snake, desert tortoise, deer) and two showing negative (zebratail lizard, bighorn sheep) relationships to the linkages. These results show that hypothetical habitat linkages are likely not a one-size fits all strategy, and caution should be taken when planning them based only on landscape-level habitat variables. Next steps will include connectivity modeling based on the opportunistic wildlife detection data.

Wijesooriya Abeyrathna, Clarkson University
Tom Langen, Clarkson University

About 40% of the countries in the world practice Daylight Saving Time (DST), which reduces the quality of sleep, causing a higher number of motor vehicle accidents. The White-tailed Deer (Odocoileus virginianus) -Vehicle Collision (DVC) rates in the New York state is about 60,000-70,000 annually, which causes fatal accidents and costs over $300 million. Because of DST, the driving time of people is shifted towards the dawn in the morning in Spring and towards dusk in the evening in Fall, which will cause low visibility and driver fatigue. We hypothesized that due to the DST change, driver fatigue, and traveling shifts towards the dawn and dusk would result in a higher number of DVCs. To find out the effect of DST on DVC rates, we analyzed 35,167 accident reports of DVCs from 2005 to 2007 in New York State, USA, collated by the New York State Department of Transportation. We predicted that there would be an increase of DVC after the DST, which will later decline over time as drivers get used to the time change. We compared the data before DST, and after DST, we found a significantly higher rate of  DVCs after the DST in Fall, but not in the Spring. The DVC rates were also much higher in the Fall than Spring. Nevertheless, we also analyzed the temporal patterns of the DVCs, which showed an increase in the Fall starting before the DST and continuing after the DST. This pattern indicates that the higher rates in the Fall are due to the change of deer activity patterns associated with seasonal changes in deer breeding phenology and not because of the time change itself. We recommend that drivers should be cautious at all times, but especially when driving in Fall near Deer-road crossing hotspots.

Renee Callahan, ARC Solutions

This poster will summarize opportunities to advance wildlife crossings within the federal legislative process for reauthorizing the nation’s current surface transportation law, the Fixing America’s Surface Transportation (FAST) Act. Among other things, the poster will highlight wildlife-related provisions in the 2020 House bill H.R. 2, The Moving Forward Act, a $1.5 trillion infrastructure package that passed the full House in July 2020, and the 2019 Senate bill S. 2302, America’s Transportation Infrastructure Act, which unanimously passed the Senate Committee on Environment and Public Works in July 2019. These provisions are expected to inform the new Congress when it begins to craft its own legislation in 2021.

Although their approaches differ, the most recent House and Senate reauthorization bills included – for the first time ever – dedicated funding for wildlife crossings. The House bill would allocate $300 million in National Highway Performance Program funds over 4 years to reduce wildlife-vehicle collisions (WVCs) or restore and maintain connectivity within terrestrial or aquatic habitats affected by roads. Taking a different tactic, the Senate bill would establish a 5-year, $250 million competitive Wildlife Crossing Pilot grant program that would similarly have funded infrastructure projects to reduce WVCs and maintain or improve connectivity, with 60% of funding going to projects in rural areas.

The poster will also showcase a variety of other provisions included in both bills. These shared provisions include:

 Expand the list of projects eligible for funding under the existing Surface Transportation Block Grant program to include wildlife infrastructure projects;

 Update the WVC Reduction Study: 2008 Report to Congress;

 Develop guidance for voluntary statewide transportation and wildlife action plans;

 Create a series of workforce development and technical training courses;

 Develop a standardized methodology for WVC and carcass data collection and voluntary standardized data template; and

 Establish a voluntary threshold for when to evaluate highways for mitigation.

Ultimately, this poster aims to raise awareness among practitioners and other interested stakeholders about potential opportunities to include wildlife infrastructure within the next generation of federal legislative solutions.

Shannon Crossen, ICF
Jon Walker, ICF International Inc
Matthew Townley, ICF International Inc
Martin Fisher, ICF International Inc

Habitat connectivity is essential to maintaining and conserving fish and wildlife species populations and important ecological processes. For many species, connectivity is constrained by infrastructure which can result in population declines. Methods to remediate fish and wildlife movement barriers along roadway infrastructure have been implemented globally for decades. Wildlife barrier remediation typically involves constructing a combination of wildlife crossings, combined with wildlife fencing. Fish passage barrier remediation involves engineering culverts or bridges with appropriate size, slope, velocity, and substrate to accommodate fish passage. These measures ensure fish and wildlife can access historical habitats and allow safe movement across roadways. For both fish and wildlife this remediation results in improved access to feeding, breeding, and shelter and constitute important conservation actions. Despite our understanding of the importance of habitat connectivity and our global progress in incorporating fish and wildlife connectivity considerations into standard practices, there is room for improvement to ensure that fish and wildlife connectivity are adequately considered and integrated into fish passage barrier remediation projects. Interestingly, fish and wildlife barrier remediation efforts have largely been conducted within separate subdisciplines and largely uncoordinated with one another. For instance, many fish passage barriers also experience adjacent wildlife-vehicle collisions (WVCs) and are in locations which warrant improved wildlife crossing design considerations. Additionally, infrastructure development projects may also miss opportunities to remediate WVCs and fragmentation if proper assessments have not been conducted. These opportunities are demonstrated with our Ecosystem Connectivity Planning Tool which identifies locations that warrant such considerations. The web-based GIS tool identifies locations where fish and wildlife movement are impeded and where wildlife-vehicle collisions (WVCs) and public safety are a concern. The tool allows for deep exploration and visualization of connectivity data including multi-scale hot spot and cluster analytics, ecological data, infrastructure planning data, and more. The tool is highly customizable, scalable, and transferrable to capture the objectives and priorities of each user. Such tools provide simple to use and cost-effective way to identify and prioritize connectivity opportunities early in the planning process to facilitate connectivity for more species, improve road safety, and prevent lost opportunities in enhancing and restoring connectivity.

Bridget Donaldson, Virginia Transportation Research Council
Kaitlyn Elliott, Virginia Transportation Research Council

Given the established success of wildlife crossings with fencing in reducing wildlife crashes and connecting habitat, information is needed on cost-effective means of implementation for departments of transportation.  Given limited transportation budgets, more research is needed on the effects of adding fencing to existing underpasses, particularly those that are too far from one another to be connected with contiguous fencing.  In this study we evaluated the effectiveness of this measure when applied to isolated underpasses. One mile of 2.4-m high wildlife fencing was added to a large bridge underpass and a large box culvert approximately 8 km apart from one another on Virginia’s Interstate 64. We conducted a 2-year post-fencing camera monitoring study and compared the findings from a 2-year pre-fencing study with regard to collision frequencies with white-tailed deer (Odocoileus virginianus) and black bear (Ursus americanus); the use of the underpasses by wildlife; and roadside deer activity. We also evaluated deer activity data to compare different fence end designs applied at the study sites. Our study found that the addition of wildlife fencing to certain existing isolated underpasses can be a highly cost effective means of increasing diver safety and enhancing habitat connectivity for wildlife. After fencing installation, deer crashes were reduced by 92% on average and there were no increases in DVCs within 1 mile of the fence ends. Deer crossings increased 410% at the box culvert and 71% at the bridge underpass. Use of the culvert and bridge underpasses by other mammals increased 81% and 165%, respectively. Although deer use of the underpasses was much greater than their activity at any of the fence ends, there was relatively high deer activity at the fence ends that did not tie into a feature such as right-of-way fencing. The benefits from crash reduction exceeded the fencing costs in 1.8 years, and fencing resulted in an average saving of more than $2.3 million per site. Our results add to the growing body of knowledge about the most effective ways we can use existing infrastructure to broaden the scale of wildlife crossing implementation.