One of the most basic questions to be asked by an ecologist is how many animals there are, of a species in any one area. For some mammals or birds this may not be too much of a problem, but otters present difficulties. In Shetland, where Eurasian otters are diurnal, it is possible to recognize individuals over stretches of coast of a few kilometres, after intensive observation over a few weeks there. The throat patches are variable and often easy to see (see Fig. 3.2). In South America giant otters are similarly distinctive (see Fig. 5.7) and active in daylight, as are spotted-necked otters (see Fig. 2.27) in many areas of Africa. These species, therefore, can be counted somewhat more easily. For animals such as the Cape clawless otter this is much more difficult, and often the only evidence of their presence is tracks in the sand (Fig. 11.1), or scats.
Sea otters cannot hide themselves as well as many other species, and populations have been counted routinely in surveys from small boats, from aeroplanes and from the shore. Obviously often some animals will be missed, and it is therefore necessary to calibrate census methods, but such techniques are now well established (Estes and Jameson 1988; Udevitz et al. 1995). However, these kinds of field condition are exceptional amongst otter species and populations. For almost all others, including the inland populations of river otters and Eurasian otters, it is impossible to get any idea of numbers without using some indirect way of census, an index.
The most common census method, on which most of our knowledge of Eurasian otter numbers in
inland waters is based, and from which we deduce most of the pattern of decline and fall in various countries, is the survey of otter scats, the 'spraints'
(Mason and Macdonald 1986). Spraints are often the only evidence of the animals' presence; they can be quite conspicuous (Fig. 11.2) and are frequently found in convenient spots, such as under bridges. In one of the standard methods, a carefully chosen 600-metre stretch of bank is surveyed (so not randomly selected). If no spraints are found, it scores a negative, etc. At the end of a survey the percentage of 'positive' sites is taken as a measure of the strength of the otter population. In earlier days such surveys were used to estimate actual numbers of otters; for instance, it was concluded from spraint distribution that in 1975 there were 36 otters in Suffolk (Mason and Macdonald 1986).
However, there are serious problems with this. Clearly, when spraints are present, there are otters, and in areas with many spraints there are likely to be more otters than in regions where spraints are few. But Eurasian otters at least (and probably also the others) also defaecate when swimming, as one can see in both wild and captive animals. The spraints that are deposited on the bank have a specific biological function, apart from elimination (see Chapter 6). Absence of spraints, therefore, does not necessarily mean absence of otters. Along inland waters, where we followed otters with radio-tracking, we found areas where the animals spent a great deal of time (in reed marshes, for instance, far away from open water, and along very small streams), but where despite intensive effort we could not find any spraints. My students and I tried hard, because we wanted to study the otters' food in those sites. Even more worrying was that some of the otters hardly ever sprainted on land at all (which we knew because their spraints were labelled with an isotope).
The implication is that, not only does absence of spraints tell us little about otter activity, but also, when spraints are present, numbers do not necessarily correlate with otters (Kruuk et al. 1986). In Shetland we surveyed 21 350-metre stretches of coast for spraints, while also counting the number of directly observed otter visits. There was no significant relationship. In an apparent contradiction later, when studying actual sprainting behaviour, I saw otters deposit more spraints in sections where they also spent more time. The most likely explanation for the discrepancy was that many spraints were deposited on seaweed below the high-tide mark, and these were soon washed off again. Such problems would not arise in freshwater areas, along rivers.
There is also the problem of seasonality. In Britain, otters spraint on land in winter many times more often than in summer (Conroy and French 1987; Kruuk 1992; Mason and Macdonald 1986), and this annual fluctuation is by no means regular, with peaks, troughs and main increases or decreases occurring during different months (Conroy and French 1987). Other regions and other species are likely to show a different seasonality. This makes spraint density, even if controlled for time of year, a tenuous measure for utilization of an area by otters.
Despite such reservations the method of spraint survey to assess otter populations remains popular, because the fieldwork is easy. However, it is controversial, and the merits and objections against it have been argued at length (e.g. Mason and Macdonald 1987 versus Kruuk and Conroy 1987). In the end, one should accept statements on otter density with caution if they are based on spraint surveys. When applied on a large scale, however, such as in the repeated national spraint surveys in Britain, they must be useful, and are likely to indicate trends. Experienced field-workers, using a somewhat vague correction for seasonality, found repeatedly that the region in Britain where spraints occur in an area has gradually enlarged, first in Wales and from Wales further into England (Andrews and Crawford 1986; Crawford 2003; Crawford et al. 1979; Mason and Macdonald 2004; Strachan etal. 1990). There is no doubt about the conclusion from the spraint surveys that otters have returned to the country, after being absent for decades. How strong the come-back is, is still open to surmise.
For our own studies we decided that spraint abundance would not be sufficiently reliable as an index for Eurasian otter numbers and distribution, both in Shetland and elsewhere. We resorted to two alternative procedures to arrive at a figure for otter density, one to be used in Shetland and one in freshwater areas. In Shetland the otters use holts on a daily basis, and holts were easy to find and could be used for population assessment. Elsewhere in freshwater areas this method was unsuitable, and we developed a method based on radio-isotope labelling in spraints from otters with radio-transmitters, by calculating the proportion of spraints that was deposited in an area by the known number of radio-otters.
I will describe these methods in some detail, but, more recently, genetic methods have made a new breakthrough. Developments in the extraction of DNA from otter spraints have made it feasible to identify the sprainter from as many as 65% of spraints found along streams on Kinmen island, off the coast of China (Hung etal. 2004). This enables an estimate of the number of individuals in any one area (with some reservations), and it showed one otter per 0.5-0.7 km of stream—a high density, possibly caused by the fact that these animals were feeding in the nearby sea, and only resting along the streams.
In Shetland we estimated the number of used otter holts and the number of adult otters associated with them, in a 100-metre strip along the coast (Kruuk etal. 1989; Moorhouse 1988). We used the number of individually known resident adult females in each female group range (see Chapter 6), and related that to numbers of used holts. Separately, and over larger areas, we estimated how many males and how many vagrants (non-residents) there were for every resident female.
We used a different method in mainland Scotland to assess otter numbers in inland streams and rivers (Kruuk et al. 1993). Otters were fitted out with radio-transmitters and injected with a harmless radio-nuclide (65Zn), which enabled us to recognize their spraints with a scintillation counter. In principle, we established the home range of 'focal animals', the ones with transmitters, and collected all spraints in a given part of that area, for example along a single stream. We assumed that other otters in that stream would be equally likely to spraint along any one bank at any one time. This assumption was based on the observation that there were no significant differences in sprainting rates and seasonality between otters of different sex or social status (Kruuk 1992). Together, the proportion of spraints with the 65Zn, and the proportion of time spent in that stream by the focal otter, allowed us to make an estimate of time spent in that area by all otters present over the study period.
As an example, one male otter (8.0 kg) was a focal animal from November 1989 until June 1990. He spent most time along the River Dee (69 of 125 nights of radio-tracking, or 55%), and foraged along the Sheeoch Burn, a small tributary of the Dee, during 47 nights (38%). He used 11.6 km of the Sheeoch (or 5.1 ha of water), and spent the equivalent of 47/125 X 365 = 137.2 nights per year there. Of 700 spraints collected along the Sheeoch during this period, 78% contained zinc-65 (77.2 ± 5.4% over five collecting periods). This suggests that the Sheeoch Burn was used by otters for a total number of 100/78 X 137.2 = 175.9 'otter nights' per year, with a mean nightly otter biomass of 4.6 kg, or 0.09 g/m2 water.
The overall, median density of otters in the Rivers Dee and Don and their tributaries was one otter per 15.1 km of stream, but it varied between one per 3 km and one per 80 km of stream, or one otter (8.5 kg) per 2-50 ha water (or 0.02-0.4 g/m2 water). As shown in Chapter 4, most of this variation in density could be related to the width of streams (Kruuk etal. 1993a).
Comparison of the freshwater data with those on otters along sea coasts in Shetland, at first glance, suggests that Shetland otters live at much higher densities. However, if otter numbers are expressed per area of water rather than length of bank, the discrepancy largely disappears. If we assume the strip of water used by coastal otters to be 80 m wide (enclosing 98% of all otter dives; Kruuk and Moorhouse 1991), the estimate for good otter habitat in Shetland is one animal (6 kg) per 10 ha water (mean 0.06 g/m2 water), of the same order as densities in streams and lakes.
Elsewhere in Europe, Ruiz-Olmo (1998), using radio-tracking combined with direct observations in northern Spain, estimated Eurasian otter densities in the lower altitudes of the region (300 m above sea level) of one per 2.2 km of stream, declining to almost zero at altitudes over 700 m. Per area of water, these otters occurred at about one per 10 ha at lower altitudes, which translates into about 0.08 g/m2 water. Eurasian otters are reported in 'densities' of one per 2 km of lake shore and one per 4-5 km of stream in Sweden (Erlinge 1968), and one per 1 km of bank in eastern Germany (Hauer et al. 2002a). However, these last figures were based largely on snow tracking, and it is possible that they did not consider the many very small tributaries to main rivers, which are frequently used by otters in summer but not when covered in snow.
In general, such estimates have to be very approximate, but the results from the various studies are somewhat similar. We do have to keep in mind, however, that animals tend to be studied most intensively in the most profitable places, that is, high-density areas.
Densities of the river otter in the mountains of Idaho were estimated by a combination of radio-tracking, direct observation and snow tracking, along streams (Melquist and Hornocker 1983). The researchers found a mean of one otter per 3.7 km of stream, over an area of 158 km, with more than twice as many females as males. Similarly, by radio-tracking in Alberta, Canada, Reid et al. (1994b) found one river otter per 5.7 km of lake shore-line.
Using a different method along the coasts of Kelp Bay in Alaska, Woolington (1984) estimated numbers of river otters from the proportion of individually known animals that he encountered in the field. He found one river otter per 1.2 km of coast. In Prince William Sound, Larsen (1983) radio-tracked river otters and estimated one per 2 km; in that same area Testa et al. (1994) concluded, from a study of radio-tracking and radio-istopes from the radio-otters in spraints, that there was one otter per 1.2-3.6 km. The numbers of river otters per length of coast are comparable to those of Eurasian otters.
The sea otter, along those same Pacific shores, occurs (or occurred) in sometimes very large numbers, and there have been reports of aggregations ('rafts') of up to 2000 (Estes 1980). The mean number of sea otters along the shores, from visual counts of 50-340 km in California between 1938 and 1984, was one sea otter per 7.0 ± 1.1 km of coast (Riedman and Estes 1990). Along Amchitka coasts, numbers were as high as about one sea otter per 0.2 km during 1940-1970, but density then declined steadily to one-tenth of that in 2000. Similar densities and declines were observed for the other Aleutian islands, where at present there is about one sea otter per 2 km of coast (Doroff etal. 2003). However, these animals use a wide area of water, so the figures should not be compared with those for other otter species.
There are few data on numbers of any of the other otter species, mostly of the giant otter, which is easily observed in daytime. In Nicole Duplaix's (1980) main study river in Surinam there was approximately one otter per 0.5 km of river, or one per hectare of water, at least in the dry season. However, during the rains the forests flooded, and the giant otters moved all over the place. Laidler
(1984) estimated one giant otter per 5.6 km of river in Guyana, and Schenck (1997) in the Rio Manu, in the Peruvian Amazon, one per 5.7 km. Schenck found that in the large oxbow lakes near the river there was about one giant otter for every 14 ha (27 kg, or 0.2 g/m2 water), a biomass considerably higher than that of otters in most temperate regions.
Along oceanic coasts of Chile, several researchers also based estimates of marine otter numbers on direct observations in daytime. Densities were often very high, varying from one per 25 km in the south (the Beagle Channel) to one per 0.15 km and one per 0.23 km in the more central areas of Chile (see references in Ebensperger and Castilla 1992; Medina 1995). In Africa, another diurnal species, the spotted-necked otter, was estimated independently by Procter (1963) and Kruuk and Goudswaard (1990) at about one per 1 km along the shores of Lake Victoria, but only where their numbers had not been affected by people.
In conclusion, numbers of otter of all species studied along banks and shores in 'optimal' areas vary between one animal per 0.2-15 km. The variation within species is at least as large as it is between species. Where estimates exist, otter biomass varies between 0.02 and 0.2 g/m2 water, but not all parts of such areas may have been suitable and used by the animals, for instance where it was too deep.
Was this article helpful?