Institute of Agriculture, Natural Resources and Extension
Non Technical Summary
Alaska has the range and forage resources to produce substantial quantities of red meat through the production of reindeer, a livestock commodity in great demand and commanding a high price. Many reindeer producers lost 75-100 percent of their herds through commingling and out-migration with wild caribou in the late 1990's-early 2000's. Producers on the Seward Peninsula now wish to reintroduce reindeer herds to many of the vacant grazing allotments. Reindeer exhibit strong site fidelity to home ranges and if relocated will exhibit a strong behavior to return to their home range. Travel on these ranges to monitor reintroduced animals is extremely difficult as there are no roads or trails and the landscape is mountainous with numerous water and geographical obstacles to overcome for overland travel. So there is a real danger reintroduced reindeer will be lost because of the difficulty in constant on the ground monitoring. In this project we will develop a system using GPS satellite telemetry to establish a virtual home range and an alert system if animals are moving off range. This will help reindeer producers to monitor the locations and movement of their animals around the clock and intervene when immediate action is needed.
Animal Health Component
Research Effort Categories
Goals / Objectives
The goals of this project are to test the capabilities of the two commercially available satellite GPS radio-collar systems for use as an intensive animal management tool on remote Alaskan reindeer ranges. We also plan to develop the algorithms and software on a server to generate a GPS boundary and alert system. We will also develop an estimate of the startup up and maintenance costs of this system to provide a producer the information on both the capabilities and cost of the proposed system.
Initial development and testing of the virtual boundary warning system will be conducted at the University of Alaska Fairbanks campus.Battery and Signal Quality Testing (Phase 1)One of each of the 4 models of GPS collars will be selected for battery and signal quality testing.The data from each of the four collar types will be analyzed to look for gaps in the data, as well as poor quality of locations.Virtual Fence Warning System Development and Testing (Phase1)Two of each model of collars will be deployed on two reindeer at the Large Animal Research Station, University of Alaska Fairbanks for a period of 30 days for each model type.Virtual Fence Warning System Development and Testing (Phase 2)The reindeer herds at Stebbins/St Michael and Savoonga, St Lawrence Island were selected to participate in the free range testing of the system.The managers of these herds have expressed a great need for satellite telemetry-managed reindeer systems. There are over 2,000 animals in each herd and the animals are allowed to free-range over large remote areas. Each group has offered logistical assistance with the project and housing and snowmachines are available for project personnel.We will work with the management team of each herd to develop one or more virtual fence boundaries and to configure the virtual fence warning system settings for their herd. In June of 2021 during annual summer reindeer handlings, two collars of each type will be deployed on reindeer of each of the participating herds. The collars will remain deployed on the reindeer until their removal during the 2022 summer reindeer handlings. All location data generated by the collar systems will be sent to herd managers in a format that can be integrated with a Google Maps TM application on their Smart phones. In this way the herd managers can continuously monitor locations of the collared animals. During the year the active virtual fence will be arbitrarily changed to locations likely to be approached by moving reindeer, but any generated warnings will only be sent to project personnel. These test cases will allow more rigorous testing of the system than reindeer managers may implement.Pre-Determined Software Parameters for calculating distance, speed, and bearing, are used in the data collection methods.Closest Point on Virtual Fence (Polygon) to current GPS location.Each virtual fence polygons will have corresponding ASCII multi-segment line.xy file to be used with Generic Mapping Tools (GMT) map project command determine how close a reindeer's location is to the virtual boundary.Point in Polygon DeterminationGPS locations will need to be tested to verify whether they are inside the virtual fence or outside. Point in polygon code (developed and used previously by the RRP on a former project) will be modified to work with the virtual fence warning system application and used to make this determination.Communicating Boundary Proximity of Animal to Producer. The server essentially defines boundary zones as points on the surface of the earth. Thus, boundaries are infinite lines with a one-half plane defined as being desirable for animals. Virtual boundaries can be added or removed at any time, and several of them can be created at once from definitions stored in the database. Three fences can be combined to create convex polygonal shapes. When the GPS readings indicate that an animal has crossed the boundary, an alert is triggered.Minimizing the cost of the system is important for the producer as they will have to weigh the benefits against initial startup and annual maintenance costs. By obtaining at least one location per day we can use cumulative days locations to determine average speed and average direction and distance from the boundary at the time that location was taken. On the vast Seward Peninsula ranges, this is the minimal data needed to implement the virtual fence warning system for most producers. We are testing the reliability of the system vs. cost as it is critical to obtain at least one location per day. The collar model that transmits two locations a day is the most economical collar for the producer if it is able to reliably transmit at least one of those locations every day. The next level up model collar transmits 4 locations daily so there is increased probability to obtain at least one accurate location per day. The other two models of collar allow a user-programmable number of locations per day, but the more locations, the higher the data costs and the faster the battery drains making a big impact on the overall cost of the system. More locations would allow a tighter virtual fence application with closer to real time warnings, but producers of Seward Peninsula free range reindeer herds on the large ranges may better served with large virtual grazing areas, and collars that have long battery lives and reliably report at least one location per day.The study findings will be presented through various venues.Workshops will be conducted in Savoonga, St Paul, and Stebbins in association with the IRA Council to include community members, to demonstrate the application of the developed range management tool.Presentation of study finding at the Reindeer Herders Association Annual Meeting.Presentation of results at the Alaska Sustainable Agriculture, Research and Education Annual Meeting.Integration of project results into the HLRM curriculumDevelopment of a webpage that will highlight project products that will be hosted on the Reindeer Research Program, UAF website.Publication of study findings in an academic journal and Cooperative Extension Bulletins.Presentation to forums that focus on Alaska meat producers and retailers.