The nesting period is critical to avian population dynamics and often assessed using daily survival rate (DSR) of nests. DSR is commonly estimated using ecological and behavioral cues from field surveys. Storms increase the uncertainty of DSR estimates in beach-nesting birds because evidence of true nest fate may be washed away and failure erroneously attributed to the storm. I used simulated data to investigate the following questions:
1. How does changing storm duration and frequency affect the uncertainty of DSR estimates?
2. What is the relative performance of different models to recover DSR in the presence of storms?
I simulated nest histories using a Python script with n=50, 100, and 200 nests, with nest initiation dates randomly assigned. The model had a true DSR and a conditional daily failure probability from multiple causes (predation, abandonment, and flooding), plus an observation process that included a discovery probability and a varying observation interval (# days between observations). I simulated storms at frequencies of 1x/season, 1x/month, and 2x/month and durations of 1, 2, and 4 days. I compared the bias and variance of DSR estimates from three different models under each scenario.
I found that storm frequency has a larger effect on variance than storm duration, although increased sample size does mitigate bias in DSR estimates. Bias was also significantly lowered by reducing the observation interval, and by using a Bayesian framework to analyze data. As climate change continues to increase the frequency and duration of coastal storm events, and coastal nesting areas become more vulnerable to storms, it is important to think about the potential bias to nesting models. 24/7 monitoring via trail cameras can reduce bias, but for studies that lack the budget or appropriate location for camera use (e.g. public beaches) these models may prove useful. For studies where sample size is necessarily small, the effects of storm events on bias must be considered.