3.2 Chick growth
We modeled the growth of 166 brant goslings (99 females, 67 males) from
105 nests, 390 snow goose goslings (195 females, 195 males) from 215
nests, 250 observations of 188 semipalmated sandpiper chicks from 94
nests, and 118 observations of longspur broods from 56 nests. Excluding
the intercept-only null models, conditional R 2values ranged from ≥0.73 (snow goose) to ≥0.97 (longspur) and marginalR 2 ranged from ≥0.31 (snow goose) to ≥0.92
(longspur) across all model sets (Tables S1–S4), indicating that the
combinations of fixed and random variables in our model sets
satisfactorily accounted for variation in avian growth. For all species,
null models received no support (wi = 0) in
multimodel comparisons, and models with only age and sex (herbivores) or
age alone (insectivores) as covariates were likewise poorly supported
(wi = 0; Tables S1–S4) except for longspurs,
where a model fitting only age had the second-highest model weight
(wi = 0.14; Table S4) in the model set. For all
species, the model-averaged parameter estimates for juvenile age were
positive (Table 4), unsurprisingly indicating the strong influence of
age on body mass. Additionally, the model-averaged parameter estimates
of sex for both goose species (Table 4) indicated that male goslings
weighed more than females of the same age, as expected (Ankney, 1980;
Hupp et al., 2017).
TABLE 4 Model-averaged predictor variables (±95% confidence
interval) from analysis of factors affecting the mass of chicks of black
brant (BLBR), snow geese (LSGO), semipalmated sandpipers (SESA), and
Lapland longspurs (LALO) at the Colville River, Alaska, 2012–2017.
Values in bold highlight variables with confidence intervals that do not
overlap 0. All predictor variables except age and sex were standardized
prior to analysis; see Table 2 for species-specific variables