Northern Prairie Wildlife Research Center
|Table 1. Number, type and size (km²) of activity ranges of Sandhill Cranes during staging at the Platte River, Nebraska|
|Year||Bird||Number of ranges||Activity range size (days observed)|
|3||2||30.3 (8), 23.0 (12)|
|4||2||47.1 (18)||8.9 (8)|
|3||2||70.5 (26), 5.9 (9)|
|11||2||2.5 (2), 6.0 (6)|
|12||2||44.6 (21), 34.5 (8)|
|13||3||25.3 (3)||10.2 (8), 12.6 (21)|
|14||2||32.1 (13)||23.5 (2)|
|15||2||49.6 (9)||17.5 (6)|
|19||2||21.2 (22), 4.9 (6)|
| a Only one activity range
had sufficient observations to estimate size and determine type.
b Insufficient observations to estimate size or type.
|Fig. 1. Examples of activity ranges and habitat locations for staging Sandhill Cranes. A. Bird 79-2 with a centrally located roost, B. Bird 79-13 with three discrete activity ranges, two of which have peripheral roosts, C. Bird 79-18 with a peripheral roost. Habitat types are riparian closed circles; native grassland squares; planted hayland diamonds; cornfields triangles; other open circles. Coordinates are based on a common meridian in the universal transverse mercator system. Approximately 33% of all observations were randomly deleted to improve clarity of figures.|
We tested whether activity range size was dependent on type of roost, year, or interaction between type and year. Because accuracy of estimating activity range size is related to the number of observations for an activity range, we weighted the analysis by multiplying the squared residual for each observation by the square root of the number of observations within each range (SAS 1987). Size of activity range differed between roost types (P = 0.049). The weighted mean activity range area for individuals with central roosts (39.6 ± 64.3 km²) exceeded that of peripheral roosts (17.4 ± 29.0 km²). The unweighted mean for central roosts was 38.7 ± 17.1 km² and for peripheral roosts was 17.4 ± 8.3 km². There were no differences between years or in interaction between roost type and year (P > 0.11).
To test whether the type of roost affected travel distances to foraging locations, we compared the means of distances for first flights in the morning from roosts to feeding locations between years and type of roost. If a crane had two activity ranges of the same type, a combined mean was obtained by weighting each separate mean by its sample size. Means of actual distances from roosts to initial feeding areas did not differ between activity ranges with peripheral roosts (1285 ± 388 m) and those with central roosts (1408 ± 407 m) nor between years or interaction between years and roost type (P > 0.40).
We also examined changes in activity range size and in daily travel distance through the staging period. Activity range types were pooled to assure adequate sample size. No differences were noted between years, among weeks, or in the interaction of the two terms (P > 0.26).
Minimum daily flight distance varied among weeks (P = 0.019). In 1978, flight distances increased from the first to the third week and then declined. In 1979, the increase lasted through the fourth week (Fig. 2). Minimum daily distance was also greater in 1979 (7858 ± 3950 m) than in 1978 (5705 ± 3577 m) (P = 0.023). None of the interactions between weeks, years, or roost type was significant (P > 0.14).
|Fig. 2. Minimum daily flight distance through the Sandhill Crane staging periods of 1978 (solid line) and 1979 (dashed line). Birds with peripheral and central roosts are combined within each year. Bars represent ± 2 SE.|
Information sharing and associations among roosting cranes. We looked at two major sources of data to determine if cranes followed each other to specific foraging sites. First, we examined synchrony of departure from nocturnal roosts to feeding sites. Synchrony could mean that birds follow each other or that they use a common extrinsic stimulus such as sunrise to time departures. We used a χ² goodness-of-fit test to compare data presented by Lewis (1974) on crane departures to a Poisson distribution (Fig. 3). Lewis' data indicated that crane departures were more clustered than would be expected by chance (P < 0.0001).
|Fig. 3. Observed (filled) frequencies for departure intervals of Sandhill Cranes leaving roost sites compared to expected (open) based on a Poisson distribution of observed values. Data are from Lewis (1974), N = 5850 observations.|
Second, we noted that most telemetry locations occurred south of the roosts, suggesting a tendency for roost mates to flock together. For example, five of seven cranes in 1978 with sufficient observations and 12 of 13 in 1979 left roosts in a southern direction. Seventy-five percent of 3102 telemetry locations were south of the communal roosts. The frequency of telemetry locations to the south was higher for peripheral roosts (90.3%) than for central roosts (66.2%) (P = 0.0035), for 1979 (78.1%) than for 1978 (70.4%) (P < 0.0001), and for interaction between year and type of roost (P < 0.0001) (PROC CATMOD, SAS 1987). In 1979, peripheral roosts had a higher percentage of locations south of the communal roosts relative to central roosts than in 1978.
To address further whether cranes from the same roost flew to identical feeding areas, we correlated the angles of morning departures from roosts among all birds. In 1978, only three of 16 possible correlations proved significant (P < 0.05). In 1979, relationships were even weaker with three of 69 possible correlations significant (P < 0.05). The number of significant correlations in 1979 was not greater than would be expected by chance based on our accepted value for significance.
Habitat preferences. Usage relative to availability gives some indication of preferred habitats. We characterized activity ranges used by six cranes in 1978 and 14 in 1979. Average habitat composition of these ranges was 44.3% cornfields (range 35.7-51.0%), 19.7% native grassland (12.6-25.8%), 9.6% planted haylands (7.1-14.2%), 17.9% riverine (7.6-29.2%), and 8.5% other (3.4-20.3%). We used the method of Byers et al. (1984) to compare number of locations within each habitat with availability based on overall habitat composition in the study area (Table 2). There was a highly significant difference (P < 0.0001) between use and availability. Sandhill Cranes used riverine habitats, native grasslands, and planted haylands more often than expected and cornfields and other habitats less than expected.
|Table 2. Proportion of Sandhill Crane visits and mean flight distances (m) to different habitats during staging in Nebraskaa|
|Habitat||Percentage of locations||Bonferonni
|Flight distance (SD)|
|Riverine||26.4||11.6||23.4-29.9||1839 (1137)||2053 (1467)|
|Native grassland||24.3||17.3||21.0-27.6||1162 (901)||1645 (1321)|
|Planted hayland||14.8||10.8||12.1-17.5||976 (893)||1440 (1268)|
|Cornfield||31.7||49.8||30.5-32.9||916 (800)||1589 (1251)|
|Other||2.8||10.5||2.4-3.2||712 (620)||1286 (755)|
| a A total of 1506 observations
were used in the analysis; χ²
for overall analysis = 531, df = 4, P< 0.0001.
b If expected values fall outside of the Bonferonni interval, the difference between observed and expected is significant at P≤0.05 (Byers et al. 1984).
Habitat usage varied throughout the day. Before 08:00 h and after 20:00 h 53.1% of the telemetry locations occurred in riverine habitats. Between 08:00-18:00 h cranes primarily used cornfields (51.6% of all locations during this period), native grasslands (28.0%) and planted haylands (16.3%). Use of native grasslands (35.6%) and planted haylands (17.5%) increased from 10:00-16:00 h as cranes gathered in traditional resting areas; cornfield use (43.7%) declined during this period. Use of cornfields increased late in the afternoon (56.2%) before cranes returned to nocturnal roosts.
We tested whether average flight distance differed among habitats, years, or interaction between year and habitat. To simplify interpretation, we included only cranes that had flown to all five habitats at least once (N = 19) and ignored any effect due to type of activity range. Flight distance depended upon destination (P < 0.0001) (Table 2) and year (P = 0.003) but not on the interaction between year and habitat (P > 0.38). Flight distances between habitats were greater in 1979 (1741 ± 1354 m) than in 1978 (1156 ± 974 m). In both years, flight distances to riparian roost sites were the longest, followed by those to native grassland. Distances to planted hayland and cornfield were intermediate, and distances to other habitats were shortest.
During the day, cranes fed primarily in native grassland (pasture), hayland, and cornfield (Table 3). Locomotion, which may reflect food searching behavior, was greatest in native grasslands and haylands. Resting occurred primarily in native grasslands and cornfields. Cranes were least often alert in plowed fields (Table 3). Courtship, although rare, was most common in native grasslands and haylands.
|Table 3. Diurnal activity budgets (% of time) of staging Sandhill Cranes by type of habitat along the Platte River, Nebraska|
| a NG = native grassland;
PHAY = planted hayland; CULT = cultivated other than corn; PLOW = plowed
b GHeterogeneity = 1580, df = 32, P < 0.001.
Group size and activities. Frequencies of behaviors during the day varied with flock size (Table 4). Groups of 1-5 and 100-199 individuals fed most, rested least, and were alert most often. Cranes in flocks > 200 tended to rest and engage in comfort behaviors such as preening more frequently than those in flocks of other sizes.
Table 4. Diurnal activity
budgets (% of time) for Sandhill Cranes staging along the Platte River,
Nebraska, by flock size
1762, df = 35, P < 0.001; GHeterogeneity
= 1760, df = 31, P < 0.001.
Previous Section -- Study Area and Methods
Return to Contents
Next Section -- Discussion