Daily survival rate.
All years. The first round of model building included temporal variables with nests from all years (2016, 2018, 2019, 2021; NOCC=63; Table 2). Daily survival probability for Swainson’s thrush across the four-year sampling period was 23%. We did not find support that elevation affected daily survival probability (b =0.0005, 95% CI [-0.0008, 0.0019], 90% CI [0.0017, -0.0005]). The model with the most support suggested that nests have a higher probability of failing as the season progresses (Table 2; b =-0.02, 95% CI [-0.05, 0.0092]). When we examined linear time within 90% confidence intervals, the model suggests some support (b =-0.02, 90% CI [-0.003, -0.051]). We examined all models using AIC and found two models that were within 2 ΔAICc values of one another and demonstrated the strongest support above the null model, which were the quadratic time trend and linear time trend (Table 3). Finally, we ran an interaction effect between time and elevation (b = -0.00004, 95% CI [-0.0001, 0.00006]) but this model did not perform well and was removed from our model comparison output.
2019 and 2021. In the second stage of our analysis, we built models for nests from 2019 and 2021 (Table 2) and included temperature and precipitation covariates, as well as effects with initiation date against daily survival rate. There was no evidence that nest survival was affected by either initiation date (b = -0.007, 95% CI [-0.047, 0.032]) or elevation (b = 0.0003, 95% CI [-0.0009, 0.0016]). Nest survival was negatively related to linear time (b =-0.03, 90% CI [-0.002, -0.057]), indicating some support that nest survival declined over the 2019 and 2021 nesting season.
We ran an interaction effect between elevation and several other meaningful parameters, including initiation date (b = -0.00003, 95% CI [-0.0001, 0.00008]), linear time (b = -0.00003, 95% CI [-0.0003, 0.00006]) and nest age (b = 0.00004, 95% CI [-0.00008, 0.00016]) but we did not find support for these relationships. We also included an interaction between linear time and year (Table 1, model 15) because precipitation varied markedly among years (Fig. 1A&B), and we were interested in evaluating year for other interannual variation, but this relationship did not demonstrate an effect (Table 2; b = -0.002, 95% CI [-0.07, 0.06]).
Environmental covariates (2019 and 2021 only). Our models did not show support that average daily temperature (b = -0.02, 95% CI [-0.15, 0.11]), minimum daily temperature (b = -0.02, 95% CI [-0.13, 0.08]) or daily temperature range (b =0.06, 95% CI [-0.04, 0.16]) affected daily nest survival in Swainson’s thrush (Table 2). We found strong support that an interaction between elevation and minimum daily temperature effected daily survival rate (b = -0.007, 95% CI [-0.001, -0.0002]). We also found strong support for an interaction between elevation and average daily temperature (b = -0.0006, 95% CI [-0.001, -0.000009]). Because our other temperature variable (i.e., daily temperature range) demonstrated collinearity with elevation, we did not test interaction effects of these relationships.
We examined “light” and “heavy” rain events and found support that light rain (i.e., 0-6.9mm) negatively affected the daily survival rate of Swainson’s thrush nests (b = -1.39, 95% CI [-2.72, -0.05]) but we did not find an effect of “heavy” rain events (>7mm; b = 3.51, 95% CI [-144.70, 151.73]) (Table 2). We further investigated our rain metrics by quantifying rain intensity as millimeters per hour and found a strong negative relationship on daily survival rate (b =-0.09, 95% CI [-0.18, -0.01]; Figure 3). We did not find an effect of cumulative daily precipitation on daily nest survival probability (b = 0.01, 95% CI [-0.09, 0.11]). We did not find any support for an interaction effect between rain intensity and cumulative daily precipitation (b = 0.004, 95% CI [-0.02, 0.03]), rain intensity and nest age (b = 0.003, 95% CI [-0.007, 0.015]), and elevation and rain intensity (b = -0.00005, 95% CI [-0.0007, 0.0006]).
In the final stage of our analysis for nests from 2019 and 2021, we compared models of interest using AIC. This revealed two competing models against DSR, the most parsimonious model to be the interaction between elevation and minimum daily temperature against daily survival rate (Table 3). We found some support for the second most competitive model, suggesting that as rain intensity increases, the daily survival rate of Swainson’s thrush decreases (Δ AICc <4; Table 3), though it should be noted that this model was within 2Δ AICc values of the null model.