In March 1998, we began locating and marking the intersection points. A map provided by base headquarters provided a scale with which we could measure our grid intervals. A starting point was chosen from this map, and grid lines drawn. In the field, a two-person team used a measuring tape and a compass to walk a magnetic course for a distance of 250 m from the chosen starting point. At the end point, the map was checked to verify that the location was correct by referencing recognizable landmarks such as fence lines, road intersections, and prominent tree stands. (The topography of the area was relatively flat with minor elevation changes arranged in ridges oriented north and south. Using a two dimensional approach resulted in some degree of distortion of actual ground distance between points, but we always maintained a minimum ground distance of 250 m.) A wood stake with a plastic identification tag was placed at each intersection point. The tag was labeled with the point’s (x, y) coordinates, with the (1, 1) point occurring in the southwest corner of the grid. A non-differentially corrected GPS fix was obtained for each point and written on the identification tag. Subsequent checks with an independent differentially corrected receiver showed that the accuracy of the original unit was acceptable with agreement to 0.05 minutes latitude/longitude.
Small Mammal Mark-Recapture Live-trap plots were established in all of the coarse scale variations of the beachgrass and conifer woodland plant communities. A plot consisted of 25 Tomahawk® small mammal traps placed in a 5 x 5 grid at 10 m spacing. Two to three plots were located within each plant community, and placed to avoid habitat interfaces by at least 50 m. The location of each plot was random within the habitat type, but was representative of the overall plant community and did not include anomalous or unique attributes. Certain areas of special interest and areas previously unavailable due to flooding, were sampled with line transect placement of small mammal traps. These were small areas with unique plant communities, or sites that due to their physical shape or limited access, did not lend themselves to plot sampling. Examples of transects include those placed along wetland edges, or in willow wetlands. In addition, in September we placed a line transect in each of the plant communities already being sampled by plots to examine the relative effectiveness of each trapping method. Line transects consisted of 10-14 traps positioned at 10 m intervals. Capture methods and data recorded were the same for plot and transect trapping.
Sampling was conducted in late March 1998 and repeated in late September 1998. Training activities limited access to certain areas in the fall, thus only one plot from each plant community was re-sampled in September. All traps were positioned near the corresponding wire flag denoting the 10m spacing, and were placed facing runways, burrows, or adjacent to structures such as logs. Traps were baited with a mix of crimped oats, rolled oats, rolled wheat flakes and peanut butter, and the doors locked in the open position. In order to document species which may have been wary of the traps, an acclimation period spanning a minimum of six days preceded trapping, after which time the traps were re-baited, set, and placed in exactly the same position (as during the acclimation period) for four trap nights. Traps were checked each day at least two hours after daylight, ensuring that diurnal and crepuscular species had capture opportunities. Each captured animal was identified to species, sex, and age class, and a numbered #1 monel ear tag was attached (National Band and Tag Co, Newport KY). All information was recorded on field data forms. Shrews (Sorex sp.) were not aged, sexed, or tagged. The survival rate for shrews was very low, and the additional stress (on the live shrews) of handling, combined with the difficulty in marking and sex identification, did not warrant these procedures. Live captured shrews were only recorded to species, and released. Processing time for newly captured animals averaged three minutes from trap check until release. Subsequently, all recaptured animals were released after recording of tag number and capture coordinates. All mortalities were retained for species verification, either in the field or later in the laboratory.
Precautions were taken to reduce the possibility of pathogen transmission to field personnel. These included using gloves when appropriate, minimizing the direct handling of animals by using capture bags, and applying anti-microbial agents periodically during and after working with animals and traps.
Infrared Monitoring We placed active infrared monitors (Trailmaster® TM1500) and 35 mm camera units (fully automatic 35mm Olympus® DLX loaded with 24 exposure, ASA 400 slide film) at nine locations on Camp Rilea. Monitoring sites included woodland, grassland, and wetland plant communities. Specific monitoring sites were chosen after initial reconnaissance, and consisted of game trails, vehicle trails, gravel roads, and water sources. Monitoring began in August, and continued intermittently through February. A monitoring period predicted to be free from extreme weather conditions was selected. Although the trail monitor units were reported to be weatherproof, many false event recordings were experienced during periods of rain while initially testing the units. The transmitter and the receiver with camera unit were attached to trees, existing posts, or stakes on opposite sides of the target area, and partially concealed with natural vegetation. The infrared beam was adjusted to pass approximately seven inches above the ground. The receiver was programmed to record an event when the pulsed infrared beam was interrupted for 0.20 seconds. In theory, this timing was to allow the legs of deer and elk to usually pass through the beam without recording an event or activating the camera, yet still record the passing of a body approximately 10 to 20 inches tall. The receiver was also programmed for a delay of two minutes between transmitted camera signals to eliminate multiple photographs of one slow moving target. The monitors recorded events 24 hours/day, but cameras were programmed to only accept input signals from the receivers from late afternoon until mid-morning. We allowed certain cameras to be active 24 hours/day, if human traffic was not an issue. Monitors were activated for three to five nights per session. No bait or scent was used for the first one or two nights; later, scent lure and/or bait was placed in the path of the infrared beam. The scent was a commercially prepared compound (M&M Fur Co), placed on low vegetation. The bait was a fish or chicken carcass wired to the ground. The use of scent and/or bait was not intended to attract animals, but rather to cause an animal to hesitate in passing through the infrared beam and facilitate a clear photograph. Each unit was checked daily to check alignment and to record event data from the receiver onto data sheets. The receivers recorded date and time of each event, and the camera recorded this information on the film, as well. By synchronizing clocks on the receiver and the camera, the two systems could be integrated to match a recorded event with the corresponding photograph. This provided a back-up method for identifying the location, date, and time of each photograph.