Fig 1.
Proportion (Mean ± SE) of samples that contained at least one raptor call. Sample represents a 30 minute acoustic recording. Data for individual owl species are least means squares (see text); means/standard error otherwise derived from raw data.
Undisturbed site data is represented by the left bar (dark grey), mid-disturbed site data is represented by the middle bar (light grey), and most-disturbed site data are represented by the right bar (second lightest shade of grey).
Table 1.
Results for Type III Tests of fixed effects for raptor call rate models.
Significant P values are bolded. ‘NS’: non-significant interaction that was excluded from the fixed effect structure (see Methods).
Fig 2.
Mean (± SE) home range size per site.
Home range size was significantly smaller at the undisturbed site compared to the most-disturbed site during the nonbreeding season.
Fig 3.
Probability distribution of fat scores for (a) parid species (CACH and TUTI) and for (b) satellite species (WBNU and DOWO).
Means + SD are given. Fat score probabilities are averaged across season and species. Both groups have a higher probability of a fat score of ‘0’ (no visible fat) and a lower probability of fat accumulation (scores 1–4) in the most-disturbed site (closed circles) compared to the mid-disturbed site (closed triangles) or the undisturbed site (X’s). The values within each fat score are offset to facilitate comparisons between sites.
Fig 4.
Least mean squares (LMS±SE) fledgling count estimates.
For each site, TUTI data are represented by the left bar (black), while CACH data are represented by the right bar (grey).
Fig 5.
Site-specific apparent survival estimates (± unconditional SE).
(left) CACH survival estimates, and (right) TUTI survival estimates. The darker shaded bar represents non-breeding season and the lighter color bar represents breeding season in each graph. Note: no survival estimates are available for CACH at the undisturbed site.
Table 2.
Sample summary and Poisson mark-resight model-averaged derived estimates for CACH and TUTI abundance.
The total number of marked individuals resighted at least once and known to be in the population (n*j), the total number of unmarked individual sightings (Tuj), and the total number of times an individual was sighted and identified as marked, but not identified to individual identity (∊j). Estimate of ‘super population’ size (), overall mean resighting rate estimate (Lambda-hat), overall probability of being captured 1 or more times (
).
Fig 6.
Mean ± SE derived from a Poisson regression of number of satellite members per flock at each site during breeding and nonbreeding seasons.
Average number for the most-disturbed site is represented by the circle, while the triangle and ‘x’ represents the average number for the mid-disturbed and undisturbed sites, respectively. Note: numbers less than 1 suggest that not all flocks had satellites species.
Table 3.
Results for Type III Tests of fixed effects for flock composition/association models.
Significant P values are bolded. ‘NS’: non-significant interaction that was excluded from the fixed effect structure (see Methods). ‘Time interval’: breeding vs. non-breeding time intervals (see Methods).
Fig 7.
Flow diagram depicting a mechanistic chain leading from degraded forest structure to altered flock structure in the most-disturbed site.
Core species respond to poor-quality foraging habitat by increasing home range and foraging time budget, which are energetically costly behaviors that can reduce the value of caching. The resultant energetic deficit in turn can increase the cost of reproduction and risk of winter starvation, potentially influencing abundance and satellite species flocking associations. Indeed, the number of satellite members in mixed-species flocks was unexpectedly high during the breeding season. Photo credit: Katherine E. Gentry.