Abundance and distribution of birds from comprehensive surveys of the Canadian Arctic, 1994-2018
Data files
Sep 09, 2025 version files 1.04 MB
-
Mean_counts_per_plot.csv
1.03 MB
-
README.md
4.01 KB
-
Strata.csv
1.92 KB
Abstract
Abundance and distribution are critical metrics of population status and foundational information for conservation. However, even these basic metrics have been difficult to obtain for birds in the Canadian Arctic, where costs are high and access is difficult. Here, we present estimates of population size and distribution for 24 species of shorebirds and 37 other species of birds across the Canadian Arctic, using ground-based surveys of 2,528 plots surveyed between 1994 and 2018. Plot locations were selected by stratified random sampling, within wet, moist, or dry habitat categories defined based on satellite remote sensing. Double-sampling was used to determine detection rates and correct the estimates to reflect birds breeding within the plots. Because this was the first large-scale survey effort across Arctic Canada, many of the results provide new insights with conservation implications. Nearly all estimated population sizes for shorebird species exceed prior estimates, many by several-fold. The six most abundant shorebird species were Calidris pusilla (Semipalmated Sandpiper), Phalaropus fulicarius (Red Phalarope), Calidris fuscicollis (White-rumped Sandpiper), Calidris melanotos (Pectoral Sandpiper), Pluvialis dominica (American Golden-Plover), and Phalaropus lobatus (Red-necked Phalarope), all with estimated population sizes exceeding 3 million (CVs for these species ranged from 0.12 to 0.24). Although shorebird densities were much higher in wet and moist habitats, the large areal extent of dry habitats meant that 42% of all shorebirds were estimated to occupy these dry areas, previously considered to be of marginal importance. The estimates of abundance and distribution for the 37 non-shorebird species are, in many cases, the first ever estimates of population size from the Arctic breeding grounds. This new understanding of population sizes will have important implications for several current conservation programs, such as those where areas are identified as key habitat on the basis of the percentage of populations using them.
Dataset DOI: 10.5061/dryad.08kprr5fz
Description of the data and file structure
See Smith et al. (2025) for a complete description of methods. In brief, the study area is all of Arctic Canada (2,629,390 km²) as defined in the Northern Landcover of Canada (NLCC; Olthof 2009, NLCC 2011). The study area is divided into 12 regions, defined by physiographic boundaries, each of which is further subdivided into subregions (32 in total) of a size that can be feasibly surveyed by one helicopter-supported crew in a single year. Each subregion was then split into zones, contiguous areas that included ~12 plots of possibly different habitats. Plots were typically 12 ha (300m x 400m) but were 16 ha (400 m x 400 m) in the early years of the survey. Sampling effort was allocated unevenly among strata (defined as areas of like habitat within subregions), with more plots surveyed from within strata believed a priori to support the most shorebirds based on habitat information. The NLCC has 15 landcover categories, which were condensed into 3 habitat categories (wet, moist, and dry) and 2 non-land categories (water and snow/ice). Habitat categories of wet, moist, and dry (believed to support high, medium, and low densities of shorebirds, respectively) were assigned at the scale of plots, based on interpretation of remotely sensed landcover data from the NLCC, which uses 30m pixels derived from Landsat data. Plots of the same habitat type in a given zone were referred to as a cluster. Thus, a given zone might have clusters of plots in different strata (e.g., Wet and Dry habitat plots within the same zone). All selected plots were covered by “rapid surveys”, in which observers move quickly on foot, counting all birds encountered. Rapid surveys were conducted during the birds’ late courtship or early incubation period, which varies regionally between late May and early July. If a plot was surveyed more than once, a mean count was used in analyses.
In addition to counts of birds, the analyses rely on information related to plots and strata, including cluster and plot areas, numbers of clusters, and area survey per subregion and stratum. These data are best accessed directly through a GIS, but are provided here for reference. Similarly, the PRISM dataset is continually growing and is accessible in full by contacting the authors.
Northern Land Cover of Canada, NLCC (2011). Northern Land Cover of Canada - circa 2000 description. Canada Centre for Remote Sensing, Natural Resources Canada, Ottawa, ON.
Olthof, I., R. Latifovic, and D. Pouliot (2009). Development of a circa 2000 land cover map of northern Canada at 30 m resolution from Landsat. Canadian Journal of Remote Sensing 827 35:152–165.
Files and variables
Files: Mean_counts_per_plot.csv, Strata.csv
Description: The files contain rapid survey counts of birds as well as plot, cluster, and stratum information used in analyses.
Variables
Mean Counts per Plot
Stratum: The stratum name, composed of subregion and habitat numbers. For example, 1.2_1 is region 1, subregion 2, habitat 1
Habitat: The habitat category assigned at the scale of the plot, based on remotely sensed landcover. 1 = wet, 2 = moist and 3 = dry habitat
Cluster ID: The unique identifier for the cluster, within the stratum
Clus area: The land area of the cluster in km²
Plot area: The land area of the plot in km²
Plot: The unique name of the plot
AMGP, BASA, BBPL, etc., the mean count per survey of the plot for birds of various species following AOU 4-letter codes
Strata
Stratum ID: The stratum name, composed of subregion and habitat numbers. For example, 1.2_1 is region 1, subregion 2, habitat 1
Area Surveyed: The land area (in km²) within all potential plots in surveyed clusters
Total Number of Clusters: Total number of clusters within the stratum
