Squirrel Census Methods

This page will be updated in summer 2015

This is a general introduction to squirrel census methods. For more detailed information, see the following publications:

  • Gurnell, J., Lurz, P.W.W., McDonald, R. & Pepper, H. (2009) Practical techniques for surveying and monitoring squirrels. Forestry Commission Practice Note FCPN011. Forestry Commission, Edinburgh. 12pp. (PDF)
  • Gurnell, J., Lurz, P. W. W., McDonald, R., Cartmel, S., Rushton, S. P., Tosh, D., Sweeney, O. & Shirley, M. D. F. (2007) Developing a monitoring strategy for red squirrels across the UK. pp. 1-40. London: Queen Mary, University of London. (PDF)
  • Gurnell, J., Lurz, P.W.W., Shirley, M.D.F., Magris, L. & Steele, J. (2004) A critical look at methods for monitoring red and grey squirrels. Mammal Review 34: 51-74.
  • Gurnell, J., Lurz, P.P.W. & Pepper, H. (2001) Practical techniques for surveying and monitoring squirrels. Forestry Commission Practice Note 11. Forestry Commission, Edinburgh. 12pp.
  • Gurnell, J., Lurz, P.W.W., Shirley, M.D.F., Magris, L. & Steele, J. (2004) A critical look at methods for monitoring red and grey squirrels. Mammal Review 34: 51-74.
  • Gurnell, J., McDonald, R., & Lurz, P. W. (2011). Making red squirrels more visible: the use of baited visual counts to monitor populations. Mammal Review, 41, 244-250.

Why census squirrels?

These notes refer to carrying out field surveys of tree squirrels in Britain, but they will be of use to those people who wish to study squirrels in any country. There are three main reasons for censusing squirrels:

  • To find out whether red and/or grey squirrels are present in a woodland site at one particular time (i.e. to record presence or the lack of confirmed presence).
  • To estimate the relative or absolute densities of squirrels in different woodland sites at one particular time.
  • To monitor changes in absolute or relative numbers of red and/or grey squirrels through time.

Red squirrel/grey squirrel identification

Red Squirrel

It is not always easy to distinguish red and grey squirrels by visual appearance alone, even though adult grey squirrels are about a third larger than red squirrels. In particular, general fur colour is not a reliable guide.

The dorsal coat of red squirrels varies considerably in colour from grey to brown or dark brown, or from sandy to bright red; frequently the tail may be darker than the rest of the coat. The underside is whitish. During the autumn the ears grow long tufts which are at their finest in mid-winter; during the spring they thin and by the summer the tufts have largely disappeared.

Picture here
Grey squirrel

The underside of grey squirrels is white or pale grey and the sides, limbs and paws are often reddish-brown. The underfur is grey but the longer guard hairs have characteristic salt-and-pepper appearance from their grey bases, black and brown shafts and white tips. Grey squirrels never have prominent ear tufts.

Methods

There are four indirect (i.e. they do not involve trapping or handling squirrels) methods of studying squirrels in the field, but only the first two of them can distinguish red from grey squirrels, and feeding transects can only be carried out in conifer woodland:

These notes offer practical advice on carrying out these methods in the field. It is important to standardise methods so that different surveys in space or time can be compared. However, they can be modified to some extent to take into account local circumstances, such as the area of woodland to be surveyed and the number of people who can help with the work. Once a method has been adopted, however, it should not be changed. It should be remembered that the red squirrel is a fully protected species under the Wildlife and Countryside Act 1981, and a licence is required from the appropriate country agency (English Nature, Countryside Council for Wales, Scottish Natural Heritage) if any planned study is likely to interfere in any way with red squirrels or their nests.

Method Can it distingush red and grey squirrels? Type of woodland
VISUAL COUNTS YES ALL
HAIR TUBE SURVEYS YES ALL
DREY COUNTS NO ALL
SQUIRREL FEEDING TRANSECTS NO CONIFER ONLY

Visual counts

These standardised time-area counts are carried out in woodland by one or more observers. Surveys carried out each season for a number of years can monitor changes over time.

  1. Basic MethodThis involves recording all the animals seen.
    • Mark out several (ideally between 6 and 12) survey lines at a density of about one line per 100 ha.
    • Each line should be between 500 m and 1000 m long, and situated along rides or inspection racks, or between rows of trees within suitable squirrel habitat.
    • A single observer should walk a line on specified days of the year, starting as soon after first light as possible (this is the time when squirrels are most likely to be active).
    • Make 2 to 5 minute stops at 100 m intervals along each line and take about 5 minutes to walk each 100 m.
    • Make a record of all squirrels seen including time, place and behaviour.
    • Postpone the survey is the weather is unsuitable, e.g. squirrels are unlikely to be very active in heavy rain, strong winds or when it is very cold.
    • Try and repeat the surveys three times within a two week period to take into account variations in the activity of squirrels.
  2. More complex method.This involves estimating the perpendicular distance of each squirrel seen to the survey line. BY knowing these distances for all the animals seen the width of the sampling ‘belt’ (termed p) can be estimated. Then, from p x the length of the line (L), the area from which the squirrels were sampled. Knowing the sampling area and the number of squirrels seen within that area, then squirrel density can be estimated. More on this below.
    Estimate and note down the perpendicular distance (p) of each squirrel seen to the survey line or the distance from the squirrel to the observer (the sighting distance, h) and the angle between the direction of the squirrel and the survey line (the sighting angle, alpha); the perpendicular distance can then be worked out from p = h sin(alpha)(see Figure). Note that it is difficult to subjectively measure distances or angles in woodland, and it is a good idea to measure out some distances accurately with a tapemeasure until the observer gets his or her ‘eye in’. The problem with this is that it could distrub other squirrels in the area, a point we shall return to later.
Red circles are squirrels in a wood (irregular outline) and a survey line is walked in the direction of the arrow. The observer, O (blue square) sees a squirrel (S) and notes down either the distance p, or the angle alpha and the distance h.

What to do with the data

The number of animals seen can be used as a simple relative index of population size. More precisely, densities can be estimated by working out the area of the visual belt within which the animals have been seen. The width of the belt to either side of the transect line will depend on habitat, and vary along the survey line according to habitat structure and ‘visibility’.

Thera are two methods for estimating the density of squirrels:

  1. Plot the cumulative number of sightings (from most distant to least distant) against distance, and calculate the fall-off distance (the distance beyond which squirrels cannot reliably be seen). This can arbitrarily be decided by drawing two trend lines through the data points by eye, the first through the series of steeply rising points on the curve and the second through the points on the curve where it starts to level off as the detection rate falls away. The intersection of the two lines is taken as the width of the fall-off distance and hence the belt width (w).More objectively, this can be taken as the distance when the number of detections drops to approximately half the previous interval (Hoodless & Morris 1988). Squirrel density (D) is then estimated by:D = n/(2Lw)where n is the number of animals seen, L is the length of the transect line and w the belt width to either side of the line.
  2. Method 2 An assumption with Method 1 is that all animals will be seen within the belt width to either side of the line. This is unlikely to be the case (see below) and more sophisticated methods can be used to estimate the detection curve (i.e. the probability of detecting an animal that is a certain distance from the transect line) mathematically.

These methods have the following assumptions, in order from most to least critical:

  1. Squirrels directly on the line will never be missed (i.e., they are seen with a probability of 1.0). If a squirrel is disturbed, it may move away, or move out of sight behind a branch or the trunk of a tree, or , just as likely, is may freeze and remain motionless for some time. Thus, it is quite easy to miss a squirrel, even if it is sitting on a branch over the observer’s head, thus violating this assumption.
  2. Squirrels do not move in reaction to the observer before being detected and none are counted twice.This assumption is violated if a squirrel moves to avoid the observer and is not seen, or will be seen at a larger perpendicular distance than it should have been. (It is unlikely that squirrels will be attracted to the observer, except maybe where they take food from the hand, such as at the National Trust Reserve in Formby, Merseyside).
  3. Distances and angles are measured without bias. It is difficult to subjectively measure distances or angles in woodland, hence assumption this assumption may be violated.
  4. Sightings are independent events.

In a more complex manner, density can be estimated by first estimating the detection function g(x)which is obtained by fitting a curve to the frequency distribution of detection distances. The definite integral of this function is referred to as the effective strip width and defines the width at which the number of animals seen outside the strip equals the number missed inside it. Fitting this function and calculating densities and confidence intervals is most easily done using the computer program DISTANCE, obtainable from www.ruwpa.st-and.ac.uk/distance/.

Hair tube surveys

The collection of squirrel hairs in tubes is a simple and inexpensive way of identifying the presence of red or grey squirrels in an area. The tubes are 300 mm long and made out of 65 mm diameter round, or 65 mm x 65 mm square ended, PVC drain pipe. They should be strapped to branches at a convenient height and baited by placing sunflower seed and maize inside them. Two wooden or opaque plastic blocks (2.5 x 2.5 x 0.5 cm) are covered by double sided sticky tape (e.g. Scotch ‘pressure sensitive’ tape; Stock reference no: 465; North British Tapes Ltd; Killingworth, Tyne & Wear) and placed on the inside roof at either end of each tube, approximately 3 cm in from the entrance. Sticky blocks are retrieved and numbered after 14 days. As the animals enter the tubes to get the food, they leave some of their hairs on the tapes which are removed for later examination under a microscope. The sticky blocks should be protected at collection to prevent the hairs being damaged.

Tube variations include blocking the tubes at one end, or using food hoppers with a tunnel entrance.

The hairs contained on the blocks can be identified using a reference collection of red and grey squirrel hair and by staining a sample of hairs with ink if this is required (Gurnell & Pepper 1994).

It is not possible to separate red and grey squirrel hairs on the basis of colour, and the hairs have similar cuticle scale patterns and medullas. However, the cross-section is different in red and grey squirrels; red squirrel hairs have a concave or a dumb-bell shaped cross section whereas grey squirrels have a round one.

The type of cross-section can be most easily seen using the technique of negative staining (Dagnall et. al.,1995).

  • Place the tapes in warm water containing a strong detergent and leave to soak overnight.
  • With forceps, remove at least 10 representative hairs from each identifiable cluster of hairs, avoiding very fine, small underfur hairs.
  • Only use complete hairs and measure the length from bulb to tip. Make a note of the colour bands along the hair. A binocular microscope at x80 is useful here. Hairs <1 5 mm long should be discarded; these will be from mice or voles.
  • Make up a 5:1 solution of Indian ink:water. Place two or three hairs on a slide together with a few drops of ink solution. Cover with a coverslip and examine using a light microscope (x400 is best). Look at the hairs at their widest part (i.e. the shield region). Ninety-five percent of mounts which show a continuous dark band along the hair are red squirrel hairs. Hairs which show no banding could be grey squirrel, polecat, pine marten, stoat or weasel. Note that cracked or damaged hairs may give misleading results, so take care when removing the hairs from the tapes.

Drey counts

Drey in an oak tree

The presence of active dreys can be used reliably to indicate the presence of squirrels, but there is no obvious difference between a drey built by a red squirrel and a drey built by a grey squirrel. If dreys look unkempt as if they are not in use, they should not be counted. Dreys are >50 cm diameter and >30 cm deep and are usually built close to the main stem of a tree with support from one or more side branches, and at heights of anything from 3 m upwards.

Dreys tend to be semi-permanent, and thus the number of dreys tends to reflect squirrel numbers over a season, year or even longer. However, if dreys are not found, then it does not mean that squirrels are absent, because dreys in the canopy may be difficult to see in some woodland habitats. This is particularly the case when they are built among thick clumps of ivy or honeysuckle covering the trees, and sometimes squirrels use holes in trees to nest in, called dens.

Cone line transect in young Corsican pine

Squirrels use several dreys at one time, and it is necessary to use established relationships between the density of dreys and the density of squirrels from the literature:

  • Red squirrels: Number of red squirrels per hectare = number of dreys per hectare x 0.26 (From data from a variety of conifer and broadleaf woodlands in Belgium. Counts were mostly made in November.) This relationship has also be used for grey squirrels in conifer forest (Gurnell et al., in press).
  • Grey squirrels: Number of grey squirrels per hectare = number of dreys per hectare x 0.74 (From a variety of broadleaf woodlands in southern England. Counts were made between January and May.)

Squirrel feeding transects

In conifer forests, a systematic assessment of the remains of complete cones and cone cores fed on by squirrels can provide good information on the timing and spatial distribution of conifer seed availability, and on habitat-use by squirrels. There is little difference between feeding signs made by red and grey squirrels.

To sample cone cores, quadrats or line transects are marked out on the forest floor. Transects require less effort to manage and are therefore more frequently used. Each transect consists of a measured line, 50 m x 1 m marked with small sticks. The lines do not necessarily have to be straight, and should not avoid natural objects such as logs or stumps. The transects are raked to bare earth to remove all complete cones and cone cores. Each transect line is visited at regular intervals to remove and count all fresh complete cones and cone cores. The number of cores taken at each visit is recorded and this provides a measure of the intensity of feeding between successive inspection dates. After each visit the transect line is raked clean.

An estimate of the quantity of seed eaten can be obtained by measuring the length of the cone cores, as there is a relationship between the number of seeds in a cone and the cone length for each tree species in a locality. It will be necessary, initially, to establish this relationship by collecting a sample of up to 100 ripe cones, measuring their length and extracting and counting the number of seeds contained in each.

The number of line transects required for an assessment will depend on the area of forest, the mix of tree species and the resources available. One line every four hectares is a reasonable guide. Cone producing trees can be patchily distributed throughout a plantation, and edge trees generally produce more cones than centre trees. Where possible, lines should cross the most important cone producing area. Squirrels will use different feeding areas at different times of the year, but this will become clear as the studies progress.

Note that the presence of squirrels can be determined from the remains of yellow whole maize put out as feed or bait for traps or hair tubes. Squirrels, both red and grey, are the only species to remove the germ from the maize grain and discard the rest intact.

Further information

Dagnall, J.L., Duckett, J.G. and Gurnell, J. (1995). A simple negative staining technique for the identification of mammal hairs. Journal of Zoology 237:670-675.

Don, B.A.C. (1985). The use of drey counts to estimate grey squirrel populations. Journal of Zoology, .206,282286.

Gurnell, J. and Pepper, J. (1994). Red squirrel conservation: field study methods. Research Information Note 191. Forestry Commission, Edinburgh.

Pepper, H.W. (1990). Grey squirrels and the law. Research Information Note 191. Forestry Commission, Edinburgh.

Terrink, BA. (1991). Hair of west European mammals. Cambridge University Press, Cambridge.

Wauters, L. and Dhondt, A.A. (1988). The use of red squirrel (Sciurus vulgaris) dreys to estimate population density. Journal of Zoology, 214, 179-187.