CHANGES IN THE WINTER DISTRIBUTION OF THE ROUGH-LEGGED HAWK 
CBC count year to identify time. For example, count year 79 = winter of 
1978-1979, count year 80 = winter of 1979-1980, etc. We chose a 28- 
year period from 1978 to 2006 because we found that the number of CBC 
circles (CBCs) within the range of the Rough-legged Hawk decreased rapidly 
prior to count year 79, resulting in large geographical gaps in the data set. 
We used data from CBC circles (CBCs) that met the following two criteria: 
• Conducted during at least 8 of the 14 years between count year 79 
and count year 92 and at least 8 of the 14 years between count year 
93 and count year 106. 
• Averaged more than three Rough-legged Hawks per year for at least 
one of the two 14-year periods so defined. 
We chose the first criterion to ensure that the CBCs used were conducted 
on a fairly regular basis over the entire 28-year period used for this analysis. 
A more stringent criterion (e.g., >8 of 14 years), would reduce the number 
and the geographic coverage of the CBCs substantially. We used the average 
of at least three Rough-legged Hawks per year to ensure that each CBC was 
within the normal winter range of the Rough-legged Hawk and to exclude 
CBCs where this species occurs rarely. A higher threshold for inclusion 
would have required dropping many CBCs that record the Rough-legged 
Hawk nearly every year. Of approximately 2000 CBCs in North America, 
293 met these criteria. We normalized CBC data for the Rough-legged and 
Red-tailed Hawks by party hour. 
Regional Analyses 
We divided the Rough-legged Hawk’s winter range into ten geographic 
regions (Figure 1). The line separating the two Pacific regions from the two 
Intermountain West regions is the axis of the Cascade Range-Sierra Ne- 
vada, the line separating the two Intermountain West regions from the two 
Plains regions is the axis of the Rocky Mountains, and the line separating 
the Atlantic Coast region is the axis of the Appalachian Mountains. For the 
line dividing these regions into northern and southern segments, we used 
latitude 41° N from the Pacific Ocean to Illinois in the west and the border 
between Pennsylvania and New York in the east. For the area between Il- 
linois and Pennsylvania we defined a border placing CBCs within 50 km 
of one of the Great Lakes into one of the Great Lakes regions. The border 
between the North Plains and West Great Lakes regions is the western bor- 
der of Wisconsin, deviating from that border in the north to place any CBC 
within 50 km of a Great Lake into the West Great Lakes region. The border 
between the two Great Lakes regions was chosen to divide the number of 
CBCs between those two regions roughly in half. 
We analyzed data by region by comparing results for two consecutive 
periods of 14 years each (count years 79-92 vs. count years 93-106). We 
also performed linear regression analyses to compare the Rough-legged 
Hawk’s trends in each region versus various predictor variables. In addition, 
we analyzed data by comparing four consecutive intervals of 7 years each 
in order to display trends for each region graphically. 
211 
