Audubon Christmas Bird Counts

This page is available for historic purposes but is not maintained, and many of the links no longer work. For up to date information on the Christmas Bird Count, please visit Audubon's web site at:


The first Christmas Bird Counts (CBCs) were conducted in 1900 in response to a suggestion made in the National Audubon Society's magazine Bird-Lore. Twenty-six individuals responded by spending several hours on Christmas Day in their neighborhood, counting all of the birds they observed or heard. This survey has been coordinated by the National Audubon Society ever since, and has become the oldest continuous wildlife survey in North America.

CBCs quickly expanded into entire day censuses, when single observers covered as much territory as possible counting all of the individuals that could be positively identified. While the number of hours in the field and miles covered varied from individual to individual, the intention was that repetition of these surveys would allow for the monitoring of trends in bird populations of each area.

These single observer counts quickly evolved into a more organized effort, where groups of people would count birds in a larger area during a single day. By the 1920s, most of the CBCs consisted of group counts rather than censuses by individuals. Unfortunately, as CBCs evolved into group efforts, there was little consistency in some of the aspects of the methodologies followed by these counts. The area covered varied from CBC to CBC, as did the count date and level of coverage. While these differences prevented meaningful comparisons of the results between CBCs, the data from single CBCs could still provide useful information on the status and trends of populations within each area.

In order to provide more useful results from CBCs, the National Audubon Society implemented uniform standards beginning in the late 1950s. This methodology is still in use today.

All CBCs occur within a 15-mile diameter circle. These circles are not randomly located, but are chosen by the compiler. The only restriction is that they should not overlap any other CBC circle in the area. Each count is conducted during a single calendar day within two weeks of Christmas; the census periods are currently determined by the National Audubon Society but the actual census day is chosen by the count coordinator. There are no limits to the number of people who can participate in a CBC, or qualifications for their abilities to identify birds. The proportion of the circle that is actually covered varies from CBC to CBC, but none of the counts has even covered an entire 15-mile diameter circle.

Any variety of methods may be used to collect data. Most data are collected on foot, in cars, or by people watching birds coming to feeders. However, some counts have employed bicycles, boats, canoes, snowmobiles, and even airplanes to count birds. Effort used to census nocturnal birds is kept separate from diurnal effort. Hopefully, the same methods will be utilized each year in order to provide comparable results. Such consistency is seldom possible, and yearly differences in methodology can significantly influence the results of the counts.

The results from the CBCs are submitted to the National Audubon Society, which has compiled the data since the inception of the survey. The results are currently published in an issue of the National Audubon Society Field Notes (formerly American Birds).

The growth in CBC effort was summarized by Butcher (1990). While the number of CBCs increased between 1900 and 1910, this number then remained fairly constant through the mid-1940s. Between 150-200 counts were conducted during most of these years, although the numbers increased to 200-250 counts during the 1940s. The increase in the number of CBCs started during the 1950s, and continued into the 1990s. The 1000-count plateau was surpassed during the early 1970s, and nearly 1500 CBCs were conducted annually by the mid-1980s. The rate of increase has slowed during the past 10 years, with a total of 1692 CBCs conducted during the 1995-1996 season.

Prior to 1970, all of the CBCs were conducted in the United States and Canada. In the 1970s, this approach was extended into Mexico, Central America, and Caribbean islands. The number of CBCs outside of the United States and Canada has slowly increased, but still represents a very small proportion of the total number conducted each year. During 1995-1996 for example, only 47 of the 1692 CBCs were conducted outside of the United States and Canada.

Increases in the number of participants on CBCs reflects the growth in the total number of counts. Fewer than 500 people participated during most years prior to 1940, and the exponential growth in the number of participants did not begin until the 1950s. The 10,000-participant plateau was surpassed by 1965, while more than 40,000 individuals were participating by 1985. The rate of increased has also slowed after 1985, with slightly more than 45,000 participants recorded during the 1995-1996 counts.

Please see the Audubon Christmas Bird Count site.

CBC trend data from (1959-1988) are available on John Sauer's web site.

CBC data has received considerable use during recent decades, primarily directed towards two topics: winter distribution patterns and population trends. Many other potential uses exist for these data, such as showing patterns of winter species diversity over large geographic scales, but these studies would require a better understanding of the biases associated with the data before they could be conducted.

These data provide a wealth of information on the early winter distributions of birds in North America, and geographic patterns in their relative abundances. Root (1988) used CBC data in preparation of an atlas of winter bird distribution. Other distribution maps have been developed by John Sauer at Patuxent Wildlife Research Center.

These data have also been widely used to establish the population trends of birds. Given the lack of consistency of coverage, temporal changes in effort, and other problems, the estimation of trends using CBC data has remained controversial (see Analytical Considerations section below). Some preliminary analyses and graphs have been conducted and created at Patuxent, but additional studies are needed in order to better understand the potential biases associated with CBC data and how to estimate population trends with these data.

With nearly 100 years of data, the CBC is a valuable source of information on historic and recent changes in the status and distribution of birds during the early winter period in the United States and Canada. Unfortunately, the absence of standardization of many aspects of the CBC complicates any analyses of these data. A number of factors need to be considered whenever CBC data are used. Some of these factors are briefly discussed below, and are described in greater detail in some of the cited references.

Representativeness of CBC Locations
While CBCs represent approximately 4.3% of the land mass north of the Mexican border, they are hardly representative of the landscape as a whole. Two factors influence the location of CBC circles: the distribution of participants and the concentrations of birds. Hence, CBCs tend to be concentrated near cities. The circles tend to avoid urban areas, since these areas are frequently devoid of birds, but suburban areas tend to be over-represented on CBCs. The count circles also reflect the distribution of the human population, with a preponderance of counts in heavily populated areas such as the Atlantic coast of New England and the Mid-Atlantic states, but relatively few counts where the human population is sparse. For example, most of the Canadian counts are in located in the southern quarter of the country, while very few counts are made in central and northern Canada.

Additionally, CBCs tend to be located near areas where large numbers of birds are concentrated during that season of the year. These locations tend to be coastal, along the Great Lakes or other large bodies of water, or at wildlife refuges and other similar areas that attract birds.

These biases do not reflect on the counts of individuals reported on CBCs. However, drawing inferences on changes in status and distribution from CBC data to overall populations may be risky, since the habitats surveyed on CBCs may not be representative of a region or continent as a whole.

Even within CBCs, coverage of residential areas or other developed sites is generally poor or non-existent, so most of the data tend to come from the "natural" habitats. As count circles become more developed over time, coverage tends to become more concentrated in the less- developed areas. These subtle changes in coverage could influence the apparent population trends of some species within a count, but could never be detected by the effort data collected for each CBC.

Changes in Coverage
While the locations of the count circles has remained fairly constant during recent years, some of the circles have been shifted somewhat yet retained the same name. Some of these shifts are relatively minor changes in the boundaries to avoid overlap with neighboring counts, while other changes are more significant. Depending upon the extent of these changes, comparison of the results before and after the modifications may not be valid. Unfortunately, these changes are not necessarily well documented in the count results, and recent compilers may not be aware of changes that occurred decades ago.

Documenting observer effort undertaken each year also poses some problems. Each count provides a summary of the effort, but these summaries are frequently incorrect. Estimates of total party-hours or total party-miles on foot may be very unlikely given the number of observers participating on a CBC, and these estimates should be carefully examined prior to conducting any analyses of the data. Additionally, the summaries of effort have changed over time. For example, the amount of time spent counting birds at feeders was only recently separated from the other effort categories of hours spent in the field. There are other potential problems as well, such as people participating in the count but not paying the count fee and appearing in the published list of participants. Hence, the number of "official" participants and their effort may under-represent the true effort expended on the CBC.

No qualification criteria exist for CBC participants, and it is up to the individual count compilers to decide which people can participate in a CBC. For most CBCs, anybody who wants to participate will be allowed to do so, whether they are expert birders or beginners. Hence, a whole range of bird identification expertise is usually represented in the participants of each count.

The results from each CBC will reflect both the identification skills of the participants and the knowledge of the compiler. Counts from adjacent areas may produce very different results, reflecting the expertise of the participants rather than actual changes in the distributions of birds. In some cases, the addition of a few expert birders can significantly improve the results from a CBC. All of these factors are very difficult to assess without considerable familiarity with the count participants, but can influence the conclusions drawn from the results.

As better field guides and optical equipment have been developed, bird identification skills have generally improved in the United States and Canada. This improvement in bird identification skills has been most apparent within the past 20 years, and poses some significant challenges in the analyses of CBC data. As these skills improve, birders generally count greater numbers of individuals and species within the area they cover each year. Even if overall effort within a count remains constant, this improvement in skill levels will probably result in slightly higher totals over time, and analyses based on counts adjusted by party-hour may be positively biased as a consequence. These factors need to be considered when analyzing CBC data, but trying to quantify the extent of these potential biases may prove to be impossible.

Butcher (1990) listed the following recommendations for the use of CBC data:

Hence, knowledge of CBC methodologies, the potential for identification problems, and the biology of the species involved is necessary for the appropriate analysis of CBC data.

Most analyses of population trends from CBC data have involved adjusting the total counts by observer effort, usually party-hours but occasionally party-miles for some species. These adjusted counts are then used in the estimation of trends. The potential biases associated with these effort adjustments, or with other aspects of the analyses of CBC data, have never been the subject of thorough statistical review. Hence, considerable skepticism remains concerning the utility of CBC data for the purposes of trend estimation. Such statistical studies may be required before the use of CBC data reaches its full potential.

The following references comprise a partial list of publications related to the CBC, the analyses of CBC data, and using CBC data in selected studies :

Arbib, R.S. Considering the Christmas census. Aud. Field Notes 21:39-42.

Arbib, R.S. 1981. The Christmas Bird Count: constructing an "ideal model". Pages 30-33 in C.J. Ralph and J.M. Scott, eds. Estimating numbers of terrestrial birds. Stud. Avian Biol. No. 6.

Bock, C.E., and L.W. Lepthien. 1974. Winter patterns of birds species diversity and abundance in the United States and southern Canada. Am. Birds 28:556-562.

Bock, C.E., and L.W. Lepthien. 1975. A Christmas Count analysis of woodpecker abundance in the United States. Wilson Bull. 87:355-366.

Bock, C.E., and L.W. Lepthien. 1975. Patterns of bird species diversity revealed by Christmas Counts versus Breeding Bird Surveys. West. Birds 6:95-100.

Bock, C.E., and L.W. Lepthien. 1976. A Christmas Count analysis of the Fringillidae. Bird- Banding 47:264-272.

Bock, C.E., and T.L. Root. 1981. The Christmas Bird Count and avian ecology. Pages 17-23 in C.J. Ralph and J.M. Scott, eds. Estimating numbers of terrestrial birds. Stud. Avian Biol. No. 6.

Butcher, G.S. 1990. Audubon Christmas Bird Counts. Pages 5-13 in J.R. Sauer and S. Droege, eds. Survey designs and statistical methods for the estimation of avian population trends. U.S. Fish Wildl. Serv. Biol. Rept. No. 90(1).

Butcher, G.S., and C.E. McCulloch. 1990. Influence of observer effort on the number of individual birds recorded on Christmas Bird Counts. Pages 120-129 in J.R. Sauer and S. Droege, eds. Survey designs and statistical methods for the estimation of avian population trends. U.S. Fish Wildl. Serv. Biol. Rept. No. 90(1).

Butcher, G.S., M.R. Fuller, C.S. McAllister, and P.H. Geissler. 1990. An evaluation of the Christmas Bird Count for monitoring population trends of selected species. Wildl. Soc. Bull. 18:129-134.

Daniels, G.G. 1975. An inquiry into Christmas Bird Count Accipiter reports. Am. Birds 29:634- 637.

Drennan, S.R. 1981. The Christmas Bird Count: an overlooked and underused sample. Pages 24-29 in C.J. Ralph and J.M. Scott, eds. Estimating numbers of terrestrial birds. Stud. Avian Biol. No. 6.

Dunn, E.H. 1995. Bias in Christmas Bird Counts for species that visit feeders. Wilson Bull. 107:122-130.

Falk, L.L. 1979. An examination of observers' weather sensitivity in Christmas Bird Count data. Am. Birds 33:688-697.

Haney, J.C. 1983. The use of reference speciers as a technique in evaluating Christmas Bird Count data. Am. Birds 37:363-364.

Mark, D.M. 1981. Thayer Gulls from western Christmas Bird Counts: a cautionary note. Am. Birds 35:898-900.

Peterson, A.P. 1994. Erroneous party-hour data and a proposed method of correcting observer effort in Christmas Bird Counts. Jour. Field Ornith. 66:385-390.

Plaza, P.D. 1978. Distribution of selected North American picids determined by computer mapping. Am. Birds 32:912-922.

Raynor, G.S. 1975. Techniques for evaluating and analyzing Christmas Bird Count data. Am. Birds 35:898-900.

Root, T. 1988. Atlas of wintering North American birds: an analysis of Christmas Bird Count data. Univ. Chicago Press, Chicago, IL. 312 pp.

Smith, K.G. 1979. The effect of weather, people, and time on 12 Christmas Bird Counts, 1954- 1973. Am. Birds 33:698-702.

Stewart, P.A. 1954. The value of the Christmas Bird Counts. Wilson Bull. 66:184-195.