Testing the surveys and visual and track censuses of Eurasian otters (Lutra lutra)

1. What they did

Since the late 1970s, otters across Europe have been surveyed by checking short stretches of riverbank (usually 600 m) for their tracks and spraints (otter droppings), and recording each site as positive or negative. Counting otters is harder, and researchers have argued for years about how much these surveys and censuses really tell us. The authors set out to test them against otters whose whereabouts were known from radio-tracking.

• Radio-tracking. As part of a reintroduction of Eurasian otters to the Muga and Fluvià river basins in north-east Spain, otters caught in Extremadura were fitted with radio transmitters and released from November 1995. Wild otters on the River Bergantes were also tagged. In all, 22 otters were radio-tracked, located daily from the ground, and by small plane when the signal was lost. Some were followed in 24-hour sessions, with a location at least every 30 minutes.
• Otter surveys. A network of 83 sites, each 600 m long, was surveyed five times between spring 1996 and summer 1997 in the reintroduction area. The results were compared with the otters’ home ranges, and with three earlier surveys from 1984, 1989 and 1994, before the otters were brought back.
• Visual censuses. Teams of 18 to 24 observers, spaced about 500 m apart along 9 to 12 km of river, watched for otters for about two hours at dawn and two hours at dusk, between mid-May and mid-July.
• Track censuses. On the same days, otter footprints along the same stretches were measured (forefoot length with and without the nail, and heel pad length) to tell individuals apart. Fresh prints (under one day old, sometimes two) were separated from old ones.

In total, 24 visual and track censuses were carried out between 1995 and 1998: eight where radio-tracking showed at least one otter had been present the day before, and 16 on Pyrenean rivers with no tagged otters.

2. What they found

Surveys and the reintroduced otters

• Before the reintroduction, none of the surveyed sites were positive. With only four otters, 23.2% of sites were positive in early 1996, spread over about 80 km of river and 940 ha of marsh. By winter 1996 to 1997 this had risen to 39%, over 135 km.
• Releasing a further 14 otters did not raise the share of positive sites further (39.2% in spring 1997). A few otters spread their signs over large areas, and new arrivals tended to use the same places.
• The otters always left spraints and prints over large distances (dozens of kilometres). One female used 35.5 km of river in 233 days, and one male 90 km in 588 days.

Surveys compared with home ranges

• Survey results and radio-tracked home ranges matched closely, but not exactly. 37.5% of sites outside the estimated home ranges were positive, and 24.5% of negative sites lay inside home ranges.
• Where only one otter lived, 70.9% of site surveys were positive. Where two or more lived, 96.9% were.
• Detection was lowest on medium rivers (15 to 40 m wide), at 55.5%. On small rivers it was 93.3%, and in lakes and marshes 90.5%. Young and older otters did not differ.
• Surveying two or three consecutive sites was enough to find otters every time, except on medium rivers, which needed seven (three sites found them 92% of the time, four sites 97%).
• The average number of spraints per latrine (1.25 to 2.5) did not change across the five surveys.

Visual censuses

• Every radio-tagged otter present in a census stretch (7 otters) was seen during the standard visual censuses.
• Tagged otters were out of their rest sites during 76.2% of dawn watches and 95.5% of dusk watches, so a census with both a dawn and a dusk watch would have seen every otter in the observers’ field of view.

Visual and track censuses compared

• The number of otters seen and the number told apart by footprint size were closely related.
• Using only fresh prints (from the night before) gave almost the same density (otters per km) as the visual census. Using old and new prints together overestimated the number of otters by about two times.
• Some otters were found only by their prints (they were inactive that day, moved outside the watch times, or were missed by an observer), and some were seen but not found by their prints.

3. What they concluded

• In small and medium waterways, a single otter marks so many kilometres that they can be found at most survey sites. Several negative sites in a row can be read as the otters being absent from that area.
• Otters passing through, or using a stretch only to move between core areas, are not picked up well by surveys. A single spraint in an unusual place should be treated with caution until it is repeated.
• The number of positive sites and the number of spraints do not rise in step with the number of otters; spraints quickly reach saturation. They give only a rough idea of abundance.
• Fresh-footprint and visual censuses give numbers close to the real number of otters. Visual censuses need experienced census takers working at set times with a standard method.
• At high densities (over 0.6 otters per km), footprint censuses may undercount, because two otters of the same sex and similar size can have prints that are confused. At low densities (under 0.1 per km), visual censuses may miss otters that move out of the census stretch.
• Ten-kilometre stretches seem adequate as census units when only one-day-old tracks or one activity period are used, though 20 to 30 km may be more accurate. Counting every stretch with tracks a few days old could suggest up to 10 otters where only one lives.

4. How tracking was or was not used

Tracking was central to the study. The otter surveys relied on finding and identifying otter tracks and spraints at each site. In the track censuses, several prints in each otter’s trail were measured (forefoot length with and without the nail, and heel pad length) to tell individuals apart, following Sidorovich (1992). Fresh prints were told from old ones by (a) sometimes doing a footprint census before the visual census, (b) noting whether prints were in new mud after rain or on a wet morning, and (c) checking whether prints were dry, since in late spring and summer prints there dry within one or two days. Prints were also used to follow a radio-tagged mother and her cub after her transmitter failed: spraints and adult and cub footprints showed them living apart from the other otters. The authors note that at low density telling individuals apart by prints “caused no difficulty”, but at high density prints of two otters of the same sex and similar size may be confused.

5. Tracker or observer certification

Not reported. The paper does not say who did the sign surveys or measured the footprints, or how experienced they were. The authors stress that visual censuses need experienced census takers and cite earlier work in which an experienced observer almost doubled the chance of seeing otters, but they give no test of the observers’ or trackers’ skill in this study.

6. Other technologies, skills or innovations

• Radio transmitters (Telonics IMP/300/L, ATS 203-12 and Wagener) weighing 36 to 40 g, located from four-wheel-drive vehicles and, when the signal was lost, from small planes.
• Using a reintroduced population of known size as a test of survey and census methods.
• Conditional probability to estimate how many consecutive sites need surveying to find otters.

7. Implications for conservation (from the paper)

• On the edges of the species’ range, a single otter can be easily detected in small and medium waterways, so repeated failure to find otters in an area over time suggests they are absent.
• When otters recolonize a river, naturally or after reintroduction, they reach most waterbodies in the basin quickly, often within one to four years, because of their large home ranges.
• The traditional image of a single otter, or a pair, living in a short and hard-to-find stretch of mountain stream is not realistic.