Author: Melanie Steinkamp, Wetlands International, 4401 North Fairfax Drive, Room 730, Arlington, Virginia 22203, melanie_steinkamp@fws.gov, 703-358-1953

Species list
Description of technique
       Survey design
       Data collection
       Data analysis
       Parameter
       Sample size
Things that could bias your counts
       Time of year/time of day
Advantages and disadvantages
Approaches to analyzing your data
Send a comment on this technique (this takes you to another page)
See existing comments (this takes you to another page)

Bermuda Petrel, Manx Shearwater, Black-vented Shearwater, Fork-tailed Storm-Petrel, Leach's Storm-Petrel, Ashy Storm-Petrel, Black Storm-Petrel, Ancient Murrelet, Cassin's Auklet, Rhinoceros Auklet, Atlantic Puffin, Horned Puffin, Tufted Puffin

Definition of burrow - Any hole in the ground.

There are a few factors which must be considered before beginning to design your monitoring program for burrow nesters.

Species Identification - Many burrow-nester colonies contain more than one species; the most diverse containing 5-6 species. Where index plots are subjectively placed, an effort should be made to minimize diversity, especially of species similar in size. When transects are used, multiple species often cannot be avoided. In such cases, it will be necessary to assign burrows to species groups based on entrance sizes. The following divisions are suggested:

Occupancy - Many burrows have tunnels so long that observers cannot reach the nest chamber with their arm, or they are curved which also may restrict access. Furthermore, it is frequently impossible to see the nest chamber from a burrow entrance with a bright light. Typically, excavating entry ports to tunnels will not be used for routine monitoring. Remote camera set ups are probably the best tools, particularly for puffin and Rhinoceros Auklet burrows, to determine the contents of burrows quickly. Nevertheless, for storm-petrels, Ancient Murrelets, and Cassin's Auklets reaching into each burrow is the recommended method. Indirect evidence of occupancy should be used for puffins and Rhinoceros Auklets when cameras are not available, and Cassin's auklets when burrows are too deep to reach the chambers. Indirect evidence of an occupied burrow would be the presence of droppings, feathers, or egg fragments near the burrow entrance.

Placement of transects - We need to consider the influence annual monitoring has on vegetation succession at colony sites. Choosing random plots along transect lines instead of going back to the same burrows each year will have less influence on the vegetation and should be considered.

It is important to note that accuracy and precision are both problems for methods developed to count burrow nesters and methods will be improved as studies are completed to determine the best methods for counting burrow nesters. We need to better understand the relationship between the number of burrows and the number of birds, and how much error is associated with these numbers. In the meantime, the following protocols are recommended for estimating population trends using population indices.

Burrow-nesting seabirds are hidden from view at their nest sites and colony attendance is either highly variable among days and hours (Tufted Puffin) or it occurs only at night (other species). As a result, it is not feasible to monitor population trends by counting birds. Instead, the target population is the number of burrow entrances either in an entire colony or in a series of index plots.

It is seldom possible for observers to get to all burrows on an island without technical climbing gear, because burrow-nesters often use steep slopes or bluff edges. Therefore, the sampling "universe" would usually be the portion of the colony that is accessible to observers.

Survey Unit: Strip Transects - If the objective is to estimate entire colony populations (at least the accessible portion), strip transects are the preferred sample plot. The width of the transects would depend upon the species involved. For example storm-petrels may occur at such high densities that 2 m wide transects would be ideal, whereas wider transects may be appropriate for tufted puffins where burrow densities are lower. Transects would vary in length depending upon the "depth" of the colony in a particular spot.

Whether the objective calls for estimating the total population in the colony, the approach is to record the number of burrow entrances in each transect or plot. For the purpose of this procedure, a burrow entrance is defined as a hole at least 10 cm long apparently dug by a bird (no burrowing mammals are present at most locations where burrow-nesting birds would be monitored). For transects, a systematic sampling design should be employed. As indicated above, 5%-10% of the colony should be sampled. The steps are as follows:

For estimating the entire colony population, systematic sampling estimators would be used to calculate the mean density of burrows per transect. The mean and variance would be used to estimate the total number of burrows in the colony. In a similar way, the mean occupancy rate would be calculated. It is important to keep track of counts for each transect, so that investigators will be able to determine whether future changes are restricted to certain portions of the colony. Inter-year comparisons would be made by analysis of variance techniques and long-term trends would be characterized using regression.

Two parameters are of interest: 1) The number of burrow entrances in the colony, and 2) the proportion of burrows that are occupied. A burrow is considered occupied if an egg or a chick is present. This statistic will vary among plots. It is essential to record the units being used.

To estimate overall population size, in relatively small colonies, select enough transects to sample approximately 10% of the area occupied by burrow nesters. The proportion could be as low as 5% in larger colonies, but in any case at least 20 transects should be sampled.

Things that could bias your counts

Observing burrows is much easier before vegetation becomes rank, but evidence of occupancy (see below) is not as prevalent as later checks. Most species of burrow-nesting seabirds are prone to abandon nests if disturbed early in incubation, but then tolerate disturbance much better after mid-incubation. Ideally, burrows should be counted in plots as early as possible, but not until incubation is well underway (for those species where observers will reach into nest chambers, and then rechecked after chicks hatch).

Advantages: The advantage is that such data are useful for assessing absolute impacts from local perturbations.
Disadvantages: A disadvantage of estimating overall populations is that it is labor intensive and therefore costly. Furthermore, it is probably not necessary for monitoring trends.

Once you've used this technique to gather data, decide which approach to use to analyze your data.