Conservation concern for the northern goshawk (Accipiter gentilis) reflects evidence that goshawks may abandon nest sites or suffer from reduced nesting success in response to some forms of timber harvest. However, this evidence is mixed and has yet to be reviewed systemically and quantitatively. Therefore, we conducted a meta-analysis to assess the extent to which timber harvest and tree size explain variation in goshawk productivity and site occupancy. Goshawk productivity was not significantly explained by the presence of nearby timber harvest nor by the average size of nearby trees either in North America or in Eurasia or when averaged across all studies. Effect sizes differed dramatically among studies and the average effect size was close to zero (Zr = 0.04). However, timber harvest and tree size together more strongly explained goshawk occupancy of nest sites or territories. Within studies, goshawk nest sites or territories with less timber harvest nearby or relatively larger trees were, in most cases, more likely to be occupied. When we estimated average effect sizes separately for the two continents, the averages were moderate, consistent (Zr = 0.23–0.27), and significantly > 0. When we combined studies from North America and Eurasia, average effect sizes for timber harvest (Zr = 0.24) and tree size (Zr = 0.25) were similar in strength and both significantly > 0. Thus taken together, our results suggest that although both timber harvest and a lack of large trees are associated with lower occupancy by nesting goshawks, pairs that nest near timber harvest or in small trees have indistinguishable nesting success from pairs nesting in large trees or farther from timber harvest. We found substantial heterogeneity in results among studies, especially within North America, which is not surprising given that studies differed greatly in research methods, forest type, and forest management. In conclusion, our results suggest goshawk nest sites in populations of conservation concern, such as A. g. laingi, may need more protection from timber harvest than they are currently receiving. Equally important, to better understand effects of forest management on goshawks, we recommend additional studies designed to: (1) better identify the spatial and temporal extent of the effect of timber harvest on goshawk site occupancy; and (2) determine what goshawks do and where they go after a timber harvest.
Data for meta-analysis derived from studies comparing timber harvest or tree size to nest site occupancy or productivity in northern goshawks (Accipiter gentilis)
These are data derived from published literature for a meta-analysis assessing the degree to which timber harvest and tree size explain productivity and site occupancy in northern goshawks (Accipiter gentilis)
We present these data in two alternative formats.
(A) An Excel spreadsheet with multiple tabs. This spread sheet includes many comments linked to individual cells explaining the derivation of individual values.
(B) A series of .csv files, each corresponding to a different tab in the Excel file. These are the same data as in the Excel sheet, but without the comments linked to individual cells.
The Excel tabs / CSV files are as follows:
(1) repro_raw
These are the studies that compared productivity (mostly number of fledged young per pair or per nest) to either timber harvest or tree size. We report the individual effects reported in each study and our derivation of the standard effect size ‘r’.
(2) repro_per-study_average
For each study listed in (1), we have calculated an average effect size (both r and Zr), corresponding average sample size (n) and average variance corresponding with Zr (vz).
(3) occupancy_raw
These are the studies that compared the occupancy of nest sites or clusters of nest sites to either timber harvest or tree size. We report the individual effects reported in each study and our derivation of the standard effect size ‘r’.
(4) occupancy_continent_study_av
For each study listed in (3), we have calculated an average effect size (both r and Zr), corresponding average sample size (n) and average variance corresponding with Zr (vz).
(5) occupancy_variable_study_av
For each study listed in (3), we have calculated an average effect size (both r and Zr), corresponding average sample size (n) and average variance corresponding with Zr (vz). There is only one small difference with (4) above. One of the studies examined the correlations between occupancy and both timber harvest history and average tree size. Thus for this study, we calculated two average effects on this sheet, one corresponding to timber harvest and the other corresponding to average tree size. This facilitates comparing effects of tree size to effects of timber harvest.
Goshawk_Dryad_data_20160322.xlsx
repro_raw
These are the studies that compared productivity (mostly number of fledged young per pair or per nest) to either timber harvest or tree size. We report the individual effects reported in each study and our derivation of the standard effect size ‘r’.
repro_per-study_average
For each study listed in repro_raw, we have calculated an average effect size (both r and Zr), corresponding average sample size (n) and average variance corresponding with Zr (vz).
Occupancy_raw
These are the studies that compared the occupancy of nest sites or clusters of nest sites to either timber harvest or tree size. We report the individual effects reported in each study and our derivation of the standard effect size ‘r’.
occupancy_continent_study_av
For each study listed in occupancy_raw, we have calculated an average effect size (both r and Zr), corresponding average sample size (n) and average variance corresponding with Zr (vz).
occupancy_variable_study_av
For each study listed in occupancy_raw, we have calculated an average effect size (both r and Zr), corresponding average sample size (n) and average variance corresponding with Zr (vz). There is only one small difference with occupancy_continent_study_av above. One of the studies examined the correlations between occupancy and both timber harvest history and average tree size. Thus for this study, we calculated two average effects on this sheet, one corresponding to timber harvest and the other corresponding to average tree size. This facilitates comparing effects of tree size to effects of timber harvest.