Managers' Monitoring Manual | Colonial Waterbirds


Patuxent Wildlife Research Center	Managers' Monitoring Manual

Home > Taxonomic Groups > Colonial Waterbirds > Cliff (Ledge) Nesting Species > Land Counts

Land Counts for Cliff (Ledge) Nesting Colonial Waterbirds

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
       Parameters
       Sample size
Things that could bias your counts
       Time of year/time of day
Advantages and disadvantages
Measuring detectability
Approaches to analyzing your data
Literature cited
Send a comment on this technique (this takes you to another page)
See existing comments (this takes you to another page)

Species list

Northern Fulmar, Northern Gannet, Brandt's Cormorant, Double-crested Cormorant, Great Cormorant, Red-faced Cormorant, Pelagic Cormorant, Herring Gull, Thayer's Gull, Iceland Gull, Glaucous-winged Gull, Black-legged Kittiwake, Red-legged Kittiwake, Common Murre, Thick-billed Murre, Razorbill

Note: Species in bold text are those for which the recommended protocol were developed. However, these survey methods may be applied to all of the species within the respective group.

Description of technique

This technique is from Byrd et al.

The following protocol was developed for Common Murres and Kittiwakes in Alaska, but can be applied to other cliff-nesting species. The nest is defined as any structure to which vegetation has been added that year.

Survey design

For many seabird colonies, it is too expensive to conduct replicate counts of large colonies. Therefore, index plots consistently measured over time may provide a reasonable basis for assessing trends. A sample plot is defined as a segment of cliff-nesting habitat which; (1) may be viewed from the same location repeatedly, (2) has readily identifiable boundaries by any person conducting the survey, and (3) contains fewer than 300 birds. It is not necessary to attempt to randomly select plots from a colony, but, where the viewable portion of a colony allows the option, plots should be selected systematically for thorough geographic coverage. On small colonies, it may be desirable to count the entire viewable population, but at large colonies, as little as 10% of the viewable population might be included in the survey. Cliff sections, viewable from above or below on a beach are good candidates for plots. Continuous coverage is fine, but the cliffs should be subdivided to create segments supporting fewer than 300 birds. It is ideal to have 20-30 plots per monitoring site; if you have a small number of plots, the loss of any one plot over the course of your monitoring reeks havoc on the statistical effectiveness of a long-term monitoring scheme. Observation points must be carefully marked to allow year-to-year repeatability of counts from the same location. The most reliable method is to physically mark the location and record the GPS coordinates. For boat counts, GPS coordinates should be recorded for observation points.

Data collection

To minimize variability and standardize counts the following protocols should be followed:

	1.	Make counts during the mid-incubation to early chick-rearing stage of the reproductive cycle.
	2.	Make all counts between 1100 - 1800 hours ( if conditions necessitate counting outside these periods it may be worth doing, but variability likely will increase. Note: This may change regionally!)
	3.	Complete 5 - 10 separate counts (replicate samples) of birds on all plots during the periods listed above. Ideally, complete counts of all plots on a single day, however, if that is not possible, complete one replicate within as short a period as possible, and finish it before starting a second. An exception may be where most (e.g. 90%) of the plots or birds have been counted, but fog persistantly precludes finishing the replicate. In such cases missing values may be estimated, and it would be wise to start a second replicate.
	4.	At each plot, record the number of birds and the number of nests for cormorants and kittiwakes (a nest is defined as a structure to which vegetation has been added in the current year). Count birds and nests at least twice at each visit to insure that counts are within 5% of each other (generally the only reason for more divergent counts is observer error). Note that this standard is more precise than that suggested for the Relative Abundance Procedure above. Record the average of clustered counts for all plots on each count date or dates on appropriate forms. Gaston et al. (1983) determined that there was little improvement to increased observation hours above two hours per day at a site.
	5.	Don't conduct counts when winds are severe, during fog, or during heavy rain.

Data analysis

A completed data set for each year would be composed of daily counts of each plot, and daily totals for the system of plots. It is important to keep track of counts for each plot in case one or more plots or observation points are lost in the future, making it necessary to use a smaller subset for multi-year comparisons.

Parameters

Due to variability in attendance of ledge-nesting seabirds at cliffs, the average number of birds present during the "count" period on the index plots is the parameter of interest for population trend monitoring. Day to day changes in counts at plots is one of the most important sources of variation in counts. This is influenced by daily changes in attendance at plots. Components of variation include hourly differences and daily differences within a given year, but the daily differences are much greater than hourly differences as long as counts are conducted during the middle portion of the day.

Sample size

A complete count of birds on all plots is one sample. Multiple counts on the same day likely are not independent and thus should not be considered as additional samples. It is better to count again on a different day. To attain the objective of detecting between-year differences as small as 20%, 5-10 replicate samples (complete counts of all plots) are needed. The exact number of replicates needed depends on the variability among count days which may vary among sites and years.

Things that could bias your counts:

Time of year/time of day

Surveys should coincide with peak nest numbers, which for most cliff nesters, occurs during the mid-incubation to early nest-rearing stage of the breeding cycle. The timing of counts is best determined by the bird's biology at each individual colony (Rothery et al. 1988). For example, in California, one survey per colony per year in late May or early June has been sufficient to obtain a peak or near peak count of nests of Common Murres, Brandt's Cormorants, and Double-crested Cormorants. Keep in mind that peak or near peak counts actually under-represent the total number of nests over the entire breeding season as they fail to include nest failures prior to the survey and/or egg laying that occurs after counts are conducted, and that species will peak at different times.

Advantages and disadvantages of this technique

Advantages: The advantage is that such data as these are useful in ecosystem monitoring. In fact, monitoring for population trends in seabirds is becoming an objective of conservation biologists concerned with marine ecosystems worldwide. Seabirds are seen as relatively inexpensive indicators of change in this complex ecosystem. In Alaska, ledge-nesting seabirds have been identified as important indicators. As such, trends will be used to track the response of target populations to natural and man-caused events. These data will be used with information gathered on other components of the ecosystem (e.g., forage fish, marine mammals, oceanography) to try to understand processes, an integral part of ecosystem management.
Disadvantages: A disadvantage of this type of monitoring is that it is labor intensive and therefore costly.

Measuring detectability

A double-observer approach can be used to develop detection probabilities for this method. During the first count at each plot, one observer is designated as the "primary" observer, who verbally describes the specific portion of the plot where they are counting and the number of individuals of each species that are present. At the same time, the second observer counts individuals at the same portion of the plot, and independently records individuals that are observed and missed by the "primary" observer. This process is followed throughout the entire count. The observers switch roles during the second count at each plot. By recording both the individuals that are detected by both observers and those missed by each observer, detection probabilities can be calculated for both observers.

Approaches to analyzing your data

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

Literature cited

Byrd, G.V., R.H. Day, and E.P. Knudtson. 1983. Condor: 85:274-280. Patterns of colony attendence and censusing of auklets at Buldir Island, Alaska.

Rothery, P., S. Wanless, and M.P. Hrris. 1988. Journal of Animal Ecology: 57:1-19. Analysis of counts from monitoring guillemots in Britain and Ireland.