Recommended methods: Absolute abundance of woodpeckers can be determined through traditional territory mapping, or through potentially less labour intensive encounter transect distance methods. Spot mapping is similar to a total count of breeding individuals in a population. This is not a total count of all individuals since non-breeders or birds without territories can rarely be counted. Therefore, the results of spot mapping should be given as absolute abundance of breeding or birds with territories as opposed to the total bird population. In addition, spot mapping counts will be an unbiased and accurate estimate of absolute abundance of breeding birds only if all birds are sighted during a survey effort. This assumption might be violated if birds are sitting tight in the nest, or if heavy foliage influences sightability at some sites. The conditions also apply to applications of distance sampling and inherent problems of different probabilities of sighting different ages and sexes of birds. Detection of secretive woodpeckers can be improved through use of tape call playbacks (described in section 3.3.1). However, the observer must avoid attracting birds in towards the transect line, as may happen with frequent and prolonged use of call playback. If the tape call playback creates excessive disturbance, primary assumptions of distance sampling will be violated (see below).
Sampling design and effort: Proper stratification by habitat types is needed if the results are to be extrapolated to larger areas. For spot mapping, the optimal sampling effort would depend on how many birds change territories and would have to be determined adaptively. Also for spot mapping, sites must be visited six to ten times during the breeding season, with one week to 10 days between visits.
Appropriate sample sizes can be estimated using Power Analysis techniques for improving the sampling design of both relative and absolute abundance methods for woodpeckers. This kind of evaluation that makes use of previously collected data is an important first step in specifying the details of how to inventory woodpeckers in particular habitats.
The spot-mapping method, first proposed by Williams (1936), is widely accepted as the preferred and most accurate method for measuring absolute abundance of territorial species such as woodpeckers. It generates detailed knowledge of the absolute abundance, distribution, and potentially of the territory size of individuals or breeding pairs within a given area.
Spot mapping requires repeated surveys of measured plots, where each woodpecker detected within the plot is registered on a base map. The data recorded include exact location within the plot, species, sex, behaviour, and number. After each survey, registrations are transferred to a separate summary map for each species, and clusters of registrations are identified that are assumed to represent one individual territory. Each territory is assumed to contain one breeding pair. For a more detailed description of spot mapping refer to Franzreb (1974), Manuwal and Carey (1991) and Bibby et al. (1992).
The assumptions of the spot-mapping method are as follows:
Figure 4. Sample spot-map and associated symbols for delineating bird territories during the breeding season (from Bibby et al. 1992; Arbib 1970*).
Determination of locations and number of territories
This is intended only as a brief description of distance methods. Due to the potential complexity of these surveys it is recommended that a biostatistician be consulted prior to conducting this (or any other) survey. An experienced biostatistician will provide assistance with estimating sample size (e.g., transect line length or number of points) based on the desired level of precision of a project. They may also be able to suggest specific field protocol that will be necessary for appropriate data collection and analysis to meet the objectives of a project.
Encounter transect distance methods may be a viable alternative to labour intensive spot territory-mapping in some situations. As discussed in Section 5.4.1 of the RIC Species Inventory Fundamentals manual, distance methods allow for estimates of absolute density from transects. Encounter transect distance methods requires that the distance from the observer on the transect be recorded, which might be less time consuming then recording coordinates for spot mapping. The advantages of distance methods are:
The disadvantage is the complexity of the analysis, which can be overcome if a statistician is consulted in the design of studies, and for the analysis of data. Another difficulty is the large sample size required (e.g., number of individual woodpecker observations) and difficulties related to estimating distances in dense forests.
The primary assumptions of distance sampling methods, in order of greater to lesser importance are: all woodpeckers present on a transect line or point are detected; woodpeckers do not move prior to being detected; and linear distances form a transect line or point to detected woodpeckers are accurately measured (Buckland et al. 1993). Further to the assumptions, further restrictions to using distance sampling include the number of observations necessary for appropriate analysis. Within a given strata, the number of individual woodpecker observations may have to be at least 60-80 (Buckland et al., pp. 296). However, recent studies have shown that a minimum of 30 observations may be sufficient for abundance analysis (J. Boulanger, Integrated Ecological Research, pers. comm.).
The following is a synopsis of suggested methods for line transects and point counts using distance sampling methods (Buckland et al. 1993, pp. 313-316):
Line transects - Well marked and straight lines are required for line transect surveys. Upon detection of a woodpecker, the observer must know the exact position of the line in order to determine the distance from the line or point to the observation. Lines should be walked slowly, emphasizing search efforts on and near the line, and search effort or efficiency should decrease smoothly with increasing distance from the line. The observer is allowed to walk slowly in dense cover and faster in open or less suitable cover, and can stop frequently to observe. Surveyors can leave the centerline, provided they record the distances to observations from the centerline of the transect. Observers should occasionally look behind in case a hidden object was not observed. Even though it will be difficult in dense forest, observers can ensure that they are collecting reliable data by moving as quietly as possible to avoid creating disturbance. If this is not possible, consider using point surveys (described below). Methods for calculating sample sizes are provided in Buckland et al. (1993, pp. 301-308).
Point counts - Many of the considerations discussed above for line transects apply to point counts. To avoid disturbance to animals prior to conducting a point count, in some instances it may be useful to observe a point from some distance and record observations. Another option, as described for call playback technique, is to wait at the point for a few minutes to allow any disturbed woodpeckers to settle prior to conducting the count. The longer that observers remain at a point, the greater the probability of detections. However, for standardization, point duration should be limited to 10 minutes unless another duration is agreed upon among a study group.
All detections of woodpeckers should be recorded with estimated distance to the observation, and the sex, age and behaviour of the bird. If the distance can not be reliably estimated during the point count, the distance can be measured more accurately at the end of the count session (not during as this may create further unnecessary disturbance). Birds first detected flying over a point should either be ignored, or counted only if they land within detection distance (Buckland et al. 1993, pp. 342).
General Considerations - Distance estimation will be difficult for multiple-species surveys in dense forest habitat. As an alternative to estimating the distance to each observation, the use of 5-7 distance categories is recommended (Buckland et al. 1993, pp. 328), but distances assigned to just one of two distance intervals may also be possible (Buckland et al. 1993, pp. 316). However, an experienced biostatistician should be consulted prior to developing field protocol of grouping distance data. Generally, for flexibility in analysis procedures, it is best to estimate distances as accurately as possible in the field (Buckland et al. 1993), and to let the data analyst group and truncate observations as required.
