Comparing the efficacy and cost-effectiveness of sampling methods for estimating population abundance and density of a recovering carnivore: the European pine marten (Martes martes)

1. What they did

The European pine marten is recovering in Britain after a severe historical decline. Interest in their numbers is growing, partly because they may help control invasive American grey squirrels and partly because some people worry about their effect on vulnerable prey such as capercaillie. Many methods exist for counting wildlife, but few have been compared with each other.

In 2017 the researchers surveyed pine martens in the Fleet Basin of Galloway Forest, Scotland, with three methods at the same time, and compared how well each one worked and what each one cost:

• Live trapping. Forty baited cage traps, about 2 km apart, were pre-baited for 2.5 weeks and then set in two phases of 20 traps, each for four nights and days, and checked twice a day. Each captured marten gave a small hair sample for DNA and was marked with coloured stock-marker spray, so that later recaptures could be recognised by eye.
• Hair tubes. Ninety-nine plastic tubes, one per square kilometre, were baited with chicken wings and fitted with a sticky patch that caught hair as a marten climbed in. They were tied to trees at least 1.5 m up (or to fence posts in felled areas), within 30 m of forest tracks, at the same places as in a 2014 survey. They were checked every 7 days over five sessions (36 days in all).
• Scat collection. One or two experienced surveyors searched both sides of forest tracks (75.8 km in total) for fresh marten scats. Each track was searched once, and scats were also picked up while checking traps and tubes.

DNA from the hair and scats identified each marten’s sex and individual identity. The researchers then estimated the number of martens with the capture-recapture program Capwire, and their density per square kilometre, for each method alone and for the methods combined.

2. What they found

Samples and individuals

• Live trapping gave 48 captures in 160 trap days and identified 24 individuals (17 males and 7 females). Every hair sample from a captured marten could be genotyped.
• The hair tubes collected 157 hair samples. Only 44% could be genotyped, which identified 15 individuals (10 males and 5 females). Flies stuck to the glue patches sometimes stopped them catching enough hair. No new individuals turned up in the fifth and last session.
• The surveyors collected 103 scats. Only 32% could be genotyped, which identified 16 individuals (6 males and 10 females).
• Altogether at least 31 different martens were found (19 males and 12 females).

How the methods compared

• Live trapping found the most individuals of any single method (77% of known martens). Hair tubes found 48% and scats 52%.
• Hair tubes and scats together found 81% of known martens.
• Only five martens (16%) were found by all three methods; 14 (45%) by two methods and 12 (39%) by one method. Six martens were only ever caught in traps, three only in hair tubes and three only from scats.
• Population estimates: 30 for live trapping, 15 for hair tubes, 26 for scats, 27 for hair tubes and scats together, and 32 (95% confidence interval 31 to 35) for all three methods combined. This gives a density of 0.27 martens per km².

Sex bias

• The number of detections by each method depended strongly on sex: males were over-represented in traps and hair tubes, and females in scats. The number of individuals per method did not differ significantly by sex.

Costs

• Hair tubes were the most expensive method (£22,992, including 456 personnel hours), because they ran for five weeks and produced the most samples to genotype.
• Live trapping cost £11,574.50 and took 8 days of trapping.
• Scat collection was by far the cheapest (£3940, with only 20 personnel hours).

Population change

• The population appears to have almost doubled since 2014 (32 martens compared with 18), and density rose from 0.15 to 0.27 martens per km², even though about 7% of the forest was felled in between. Counting only hair tubes and scats, to match the 2014 effort, the number of individuals found still rose from 15 to 25. Only three martens (two males and one female) were found in both surveys.

3. What they concluded

• Of the three methods, combining hair tubes and scat collection works best for estimating pine marten numbers. It finds the most individuals, balances the sex bias of each method, avoids trapping and does not need a protected species licence.
• Hair tubes alone under-sample the population and should not be used on their own. Their cost could be cut by running four sessions instead of five, since the fifth found no new martens.
• Live trapping is better justified when samples such as blood are needed, or when animals must be handled for health checks, tagging or radio-collaring.
• The low genotyping success of hair and scat samples means more martens were probably sampled than could be identified. Better laboratory methods would make non-invasive surveys more accurate.
• Bold, exploratory martens may be more likely to enter traps and hair tubes, so shy individuals may be found mainly through their scats.
• The authors suggest the rise in martens may be linked to 50 artificial den boxes put up in autumn 2014, in a plantation where natural tree cavities are scarce. Vole numbers seemed high in both years, so are unlikely to explain the change, and some martens may have been missed in 2014.
• Camera traps should be compared with these methods in future.

4. How tracking was or was not used

Tracks and signs identification was used in the scat survey: one or two experienced surveyors searched both sides of forest tracks for fresh pine marten scats, and also collected scats they came across while checking traps and hair tubes. All scats seen were recorded for another study, but only fresh ones were collected for DNA. No trailing was reported. How scats were identified as pine marten in the field is not described; the DNA results confirmed the individuals, but the paper does not report whether any collected scats turned out to be from other species. The introduction mentions snow tracking as one of the methods others have used to estimate marten numbers.

5. Tracker or observer certification

Reported but not validated. The scat collectors are described as “one or two experienced surveyors”, with no certification or test of their skill reported. Volunteers who helped with fieldwork are thanked by name in the acknowledgements.

6. Other technologies, skills or innovations

• A home-made hair tube design: plastic cable trunking with holes so martens can grip the tree, a hinged front, a wire hook to hold the bait, and a patch of mouse glue on Velcro to catch hair.
• Coloured stock-marker spray applied with a paintbrush in a distinctive pattern, so recaptured martens could be recognised without taking more samples.
• Individual identification from DNA with up to 10 microsatellite markers.
• Capwire, a capture-recapture program suited to small populations where some individuals are caught more often than others.
• A detailed cost breakdown of personnel hours, vehicle mileage, genotyping and equipment for each method.

7. Implications for conservation (from the paper)

• As pine martens continue to recover and spread in Britain, combined non-invasive sampling (hair tubes and scats) should be used to monitor them and to guide conservation and management.
• Non-invasive methods avoid the possible discomfort, distress, injury or death that trapping can cause.
• Artificial den boxes may help martens in commercial plantations where natural den sites are scarce.