A Comparison of Population Survey Techniques for Swift Foxes (Vulpes velox) in New Mexico

1. What they did

Swift foxes are small foxes (under 3 kg) of the shortgrass prairie, from eastern New Mexico and northwestern Texas north to southern Alberta and Saskatchewan. They declined sharply after settlement and were once a candidate for endangered species listing, so monitoring them is a key part of their conservation strategy. The authors set out to find which survey techniques work best in New Mexico.

Working in the Kiowa National Grasslands, they compared:

• Relative abundance techniques (which show where foxes are and whether numbers are rising or falling): live trapping, scent stations, collecting scats, searching for tracks, spotlighting, and calling foxes with recorded sounds.
• Absolute abundance techniques (which count the actual number of foxes): trapping and marking foxes, then “resighting” them with automatic cameras at bait stations; and counting individuals from the DNA in collected scats.

They trapped and radio-collared foxes, and marked each one with a unique patch of dyed fur so it could be recognised in photos. Scent stations (cleared patches of sand and oil with bait and lure, watched by infrared-triggered cameras) were set inside the known home ranges of collared foxes for six nights, to find how many stations and nights are needed. Scats were collected at conspicuous places along roads, such as road, trail and fence intersections, and only scats up to 20 mm across were taken, to avoid coyote scats. The species was confirmed with mitochondrial DNA, and individual foxes were identified with nuclear (microsatellite) DNA. They also tested whether putting lure at a site makes foxes leave more scats there. Track searches, spotlighting and calling were only tried in preliminary studies.

2. What they found

Trapping

• They caught 33 swift foxes plus 20 recaptures in 793 trap-nights, and 3 more at dens. In all, 36 foxes (18 males, 18 females) were radio-collared.
• Capture success was highest in autumn and early winter.

Scat collection

• In the 2000 survey, scats were found at 36 of 48 sites; 141 of 194 scats (72.7%) were swift fox.
• In the 2001 survey, scats were found at 25 of 39 sites; 89 of 137 scats (65.0%) were swift fox and 4 were coyote. The rest could not be identified.
• Swift fox scats were found in every known home range and in all gaps between home ranges where no foxes had been trapped.
• Scats identified as swift fox and unidentified scats looked the same and had the same average diameter (about 14 mm).
• Lure put out 7 months earlier meant scats were found at more sites (63.6% with lure, 35.5% without), but not more scats per site.

Comparing techniques at the same sites

• Swift foxes were detected at 61.9% (2000) and 66.7% (2001) of locations by scats, 31.4% and 47.1% by bait/camera stations, and 11.5% and 8.4% by trapping (per trap-night). In both years, scats detected significantly more than bait/camera stations, and bait/camera stations significantly more than trapping.

Scent stations

• Collared and uncollared foxes made 112 visits over 420 station-nights; 75% of stations were visited within four nights.
• Detection nearly reached its highest after three nights with three stations per home range (all foxes), or four nights and four stations (collared foxes only).

Tracks, spotlighting and calling

• Only one clear swift fox track was seen on unprepared ground in the whole study. None were found at 64 lure-test sites, along 31 km of road searched by vehicle, or 12.8 km searched on foot.
• Spotlighting along 187 km of road through the home ranges of about 15 foxes found none.
• In 11 calling attempts, only one fox came to the vehicle and called back.

Counting foxes

• Trapping and camera resighting estimated 23.9 foxes (November 1999 to January 2000), 18.6 (February 2000) and 16.2 (January to March 2001) in the study area, or about 0.07 to 0.1 foxes per km².
• Scat DNA gave 63 and 27 different genotypes, far more than these estimates. Survey scats had fewer heterozygotes than blood, tissue and control scats, a sign of allelic drop-out (genes failing to show up in poor-quality DNA).

Costs

• A statewide scat survey for relative abundance was estimated at $20,000 to $30,000 and at most two months of fieldwork; a scent-station or trapping survey at $15,000 but 6 to 10 months of fieldwork.
• For absolute abundance, camera units cost about $675 per station; a statewide trapping and resighting survey was estimated at about $90,000, and a scat DNA survey at $30,000 to $50,000.

3. What they concluded

• In New Mexico, collecting scats and confirming the species with DNA is the most efficient way to measure relative abundance, followed by scent stations, then trapping. Collecting at least 10 scats per site would have confirmed swift foxes at 98% of sites.
• Scat species must be confirmed, because scat sizes of coyotes, red foxes and swift foxes overlap.
• If scent stations are used, stations 1.6 km (1 mile) apart operated for three nights are the most practical for statewide surveys. Scent stations remain an important option because they cost less.
• About 65 to 90 independent transects would be needed to detect a 20% decline between years.
• Trapping is the least efficient technique and risks injuring the animals, so the authors rule it out as a survey technique.
• Track, spotlight and calling surveys are not efficient in New Mexico. The soils are mostly too hard and dry, and rain too irregular, to take and hold clear tracks.
• In wetter parts of the swift fox range, the value of scat and track searches may be reversed: mould breaks down scat DNA quickly, while wetter soils show tracks more easily.
• For counting foxes, trapping with camera resighting worked well but is expensive, so it suits small areas. Scat DNA overestimated numbers because of poor-quality DNA and because old scats include foxes just passing through. Collecting only fresh scats and repeating the DNA tests could help but would raise costs. Scat DNA was useful for studying the population’s genetic structure.

4. How tracking was or was not used

• Scat searches: the main tracking-based technique. Scats were searched for by eye at conspicuous places along roads, picked by size (20 mm across or less) to leave out most coyote scats, and then identified by DNA. No effort was made to collect only fresh scats.
• Track searches: a preliminary test along unpaved roads (on foot and by slow driving), around wet areas at livestock water tanks, and during scat collection. Only one clear swift fox track was seen, so the authors judged track surveys unsuitable for New Mexico’s hard, dry soils.
• Scent stations had a prepared sand surface, but visiting foxes were identified with cameras.
• Radio telemetry was used to locate foxes and map home ranges.

5. Tracker or observer certification

Not reported. The paper does not say who searched for scats and tracks or what their experience was. Scat species identification was checked by DNA, so the scat results do not depend on the observers’ identification skills; the track search results may.

6. Other technologies, skills or innovations

• Infrared-triggered cameras (Trailmaster 1500) in boat dry boxes at scent and bait stations, combined with dyed fur patches for recognising individual foxes in photos.
• Mitochondrial DNA (cytochrome b) to identify the species from scats, and seven microsatellite markers to identify individual foxes, checked against matching blood and scat from captive swift foxes.
• Radio collars tracked from a vehicle with a roof-mounted antenna pair.
• Recorded rabbit distress calls and swift fox vocalisations for calling.

7. Implications for conservation (from the paper)

• Monitoring swift fox populations is an essential part of the multi-state conservation strategy, and the choice of survey technique should suit local soils and climate.
• In conservation, failing to detect a real decline matters more than a false alarm, so surveys should be designed to keep that error low.
• Population counts from scat DNA can be much higher than counts from conventional methods, so they need care.