Statistical Analyses:
Statistical analyses were performed using R, version 3.6.3 (Team, 2013). The 3D energy metrics were treated as continuous variables and were log-transformed so that residuals fitted a normal distribution. The energy of the nest and null sites were compared at each spatial scale with the lme4 package using linear mixed models derived from the following formula:
lmer (log (Energy) ~ poly(Scale,2)*ROI+(1|Site)+(1|ScanID) +(1|NestID) ) , where ROI specifies nest vs null measurement, ScanID specifies the label for the individuals can and NestID is the shared ID for the nest and its corresponding null.
The influence of 3D energy and habitat type on visibility was also assessed using linear mixed models, while visibility’s effect on predation was analysed using generalised mixed models with predation as a binary (yes/no) variable. We also compared the effect of broad scale (agri-environment and arable)) and fine-scale (wet grassland, gravel, crop, fallow, cattle-grazed and sheep-grazed) on 3D energy using Tukey’s post-hoc tests.