Analysis of the Accuracy of Radiotelemetry Station Locations
John A. VanLaningham
Introduction
My study involves analysis of radiotelemetry data from a highly studied and documented group of white-tailed deer on Huntington Wildlife Forest (HWF) of the State University of New York College of Environmental Science and Forestry (SUNY-ESF). Radiotelemetry is a procedure that is utilized by wildlife biologists as a way to track and monitor the animals that they are studying. The technique behind radiotelemetry is locating a beacon on a known radio frequency and determining the direction from which the strongest signal is coming. 
This is repeated at several known locations. The compass bearings, or azimuths, are then plotted on a map or in a computer program to show where the azimuths cross. That point is generally considered the point at which the animal was located at the time the azimuths were taken. The average error in the telemetry locations is determined by using a beacon that is placed in the study area in a location unknown to the data collectors but known to the principle researcher. Bearings to the beacon are collected and plotted in the same manner as with the deer. The researcher then calculates the actual distance from the plotted location of the beacon to its actual location and uses this to calculate the average error of the telemetry readings.
There are studies (Sanderson 1966, Masters 1978, White and Garrott 1990) that compare different telemetry techniques and there are others (Schmutz and White 1990, Aebischer et al. 1993) that discuss the sources and amount of error involved in this process. One source of error that is generally overlooked is determining the location of the telemetry station, that is, the location from which the azimuth was taken. Most station locations are determined using a compass, protractor and a U.S. Geological Survey (USGS) topographical map. This is generally due to the availability and low cost of this method. Another way to determine the location of the stations is to use a Global Positioning System (GPS) unit to determine the coordinates, either in latitude and longitude, Universe Transverse Mercator (UTM), or one of several other coordinate systems in use today. Another possible way to determine the location of a station is using a computer Geographical Information System (GIS) program and satellite imagery.
The white-tailed deer at HWF have been studied since the 1960's and several discoveries have been made concerning their movements and social behavior using radiotelemetry data. It was found that female white-tailed deer establish home ranges that overlap those of their mother's and their female siblings. This philopatric nature of deer was studied further and it was found that the deer form social groups consisting of female relatives (ie. Mothers, sisters, daughters and aunts). The overlap in the home ranges of these social groups has been termed the Rose Petal Metaphor by Matthews (1989). To test the rate of dispersion of female deer into 
surrounding areas, a "void" of low deer density was created on HWF by Hill (1995). One social group of deer was captured and moved to another location in the Adirondack Mountains. More recent graduate students have monitored this void to determine the rate of dispersal of deer into the area.
Throughout all of this research, radiotelemetry has been necessary to determine the home ranges of the deer in and around the study area. Telemetry was and still is conducted at chainage markers (henceforth called 
telemetry stations) along the road system on HWF. The coordinates of the chainage markers were determined from USGS maps at some unknown point in time and have been used for several years. A comparison of the accuracy of the map derived coordinates with coordinates obtained with a GPS unit showed that there was a fairly large difference. This raised a question as to how much the variation would affect the plotting of the locations of the deer and the calculation of their home ranges.
Methods
The original coordinates were entered into the GIS program ArcView, using the NAD83 projection, and plotted over satellite imagery of the HWF in the autumn to determine how closely they followed the roads. Autumn imagery was used because the lack of leaves on the trees made it easier to see the roads. It was apparent from looking at the points that there was error because a good deal of the points deviated from the roads. A Trimbleâ GeoExplorer III GPS was used to collect the new coordinates of the telemetry stations. This unit uses an average of several X and Y coordinates (points) to determine location and calculates a standard deviation to determine the error of the coordinates. For consistency we used a standard of 20 points per location with the unit set to take one point every three seconds. The data collected was downloaded from the unit using the Trimble GPS Pathfinder Office software and converted to shape and dbf files for use in ArcView and LOCATEII.
For this initial analysis, only the home ranges of the eight deer that lie on the perimeter of the social group were used. Due to the differences in computer technology and digital information storage, using the exact same information that Hill (1995) did was not possible. The original MS-DOS input files for LOCATEII were converted from the NAD27 GIS projection to the NAD83 projection. The stations file that was originally used has since been revised and the coordinates of the stations had to be converted to the NAD83 projection. The bearings were then analyzed in LOCATEII and the locations of the deer were analyzed with Home Range to determine the 50% core area home range of each of the eight deer.
Once the original home range of each deer was recreated as closely as possible with the original station locations, the exact same points were used to recalculate the home ranges with the new station location coordinates. The perimeter points of the two different home ranges for each deer were compared with the Multiresponse Permutation Procedure (MRPP) in the Blossom Statistical Package from the U.S. Geological Survey in Ft. Collins, Colorado. The void was then recreated by linking the outer points of the eight deer with both sets of data in ArcView and analyzed with MRPP.
Results
Although the initial view of the difference in the locations of the stations seems significant, MRPP results show that there is no significant difference in the locations of the home ranges plotted with either the new or old coordinates of the telemetry stations.
Discussion
The initial purpose of this study was to determine if the void needed to be re-plotted for accurate analysis of the repopulation of the area. There was also a possibility that all of the deer telemetry data from previous studies would need to be re-analyzed and re-plotted to ensure that the findings of these studies are still valid. Any significant change in the location of the void may have caused there to be a shift in the actual locations of the home ranges of the deer from the surrounding social groups. This could have altered the actual make up of the social groups if a neighboring deer was now plotted closer to the void. This could also have affected the results of other studies are taking place that are 
calculating the deer densities in and out of the void. Two of the ways that this is accomplished is through roadside observation (S. A. Hauver, Roosevelt Wild Life Station, unpublished data) and trap capture data (J. L. Isabelle, Roosevelt Wild Life Station, unpublished data). Any findings that would have been interpreted as significant due to the location of an observation or a trap would now be invalid.
The initial thought that the differences in the coordinates would cause a significant shift in the plotted location of the home ranges of the deer, and therefore the void, was not the case. To illustrate this I have plotted the different home range boundary points for deer #0159. Even though the location and orientation of the perimeter of her home range changed somewhat, there is a large amount of overlap in the areas of both plotted home ranges. Even with the shift in the location of her home range it is not significant enough to say that using the GPS coordinates of the telemetry stations is more accurate.
Deer maintain summer home ranges that average roughly 500 acres and tend to return to the same home range every summer. With an animal that maintains a home range of that size, the differences that we are seeing in these locations isn't significant. For studies of smaller organisms or organisms that don't tend to move a great deal, increasing the accuracy of individual plotted telemetry locations may be more important. Studies that involve analysis of habitat use may find that any shift in the actual plotted area of use may be extremely significant as the individual locations of the subject may change from one habitat type to another within a matter of meters. Even though the findings of this study were not significant, the telemetry station coordinates determined with the Trimble GPS unit are going to be used in the current deer study (A. M. Oyer, personal communication). The ability of the unit to calculate the standard deviation in the coordinates of the stations will increase the accuracy in calculating the average error of the telemetry locations.
About The Author: John A. VanLaningham
I am currently an undergraduate at SUNY-ESF pursuing a B.S. in Wildlife Biology. Research interests include the population dynamics, behavior and habitat use of the mountain lion in the Rocky Mountains. I am also a certified MP in the New York Army National Guard and interested in conservation law enforcement.
Literature Cited
Aebischer, N. J., P. A. Robertson, and R. E. Kenward. 1993. Compositional analysis of habitat use from animal radio-tracking data. Ecology 74(5):1313-1325.
Hill, J. A. 1995. Creation of a geographic void is a white-tailed deer population in northern New York: implications for management. Thesis, State University of New York College of Environmental Science and Forestry, Syracuse, New York, USA.
Masters, R. D. 1978. Deer trapping, marking and telemetry techniques. State University of New York College of Environmental Science and Forestry, Syracuse, New York, USA.
Matthews, N. E. 1989. Social structure, genetic structure and anti-predator behavior of white-tailed deer in the central Adirondacks. Dissertation, State University of New York College of Environmental Science and Forestry, Syracuse, New York, USA.
Sanderson, G. C. 1966. The study of animal movements - a review. Journal of Wildlife Management 30(1):215-235.
Schmutz, J. A., G. A. White. 1990. Error in telemetry studies: effects of animal movement on triangulation. Journal of Wildlife Management 54(3):506-510.
White, G. C., R. A. Garrott. 1990. Analysis of wildlife radio-tracking data. Academic Press, San Diego, California, USA.