Shifting patterns of nest density and predation risk: A spatial analysis of 15 years of duck nesting data from California

Background/Question/Methods

One of the best known adaptive habitat selection rules is the win-stay, lose-shift rule of thumb that many bird species appear to use to select nest sites in consecutive years.  While this is an individual behavior, it may be possible to detect shifting patterns of nest density and predation risk at a population level.  Here, we conducted a spatial analysis using 15 years of waterfowl nesting data (~11,000 nests) to assess whether there were areas of consistently high or low predation risk (temporal correlation in habitat quality) at our site, and whether ducks used previous risk landscapes to make current habitat selection decisions.  To diagnose spatio-temporal correlation in predation risk and win-stay, lose-shift dynamics, we compared site-level kernel density maps of successful and unsuccessful nests between years.  We also examined between-year correlations in percent nest success and nest density at three smaller spatial scales: individual nesting fields, 400 m² grid cells, and 200 m² grid cells.  Finally, we used banding data from individual birds to determine if there were win-stay, lose-shift dynamics occurring at the individual level that were undetectable at a population scale.

Results/Conclusions

We found no evidence for temporal correlation in predation risk at any spatial scale, and no evidence for population-level win-stay, lose-shift at any scale.  At the individual level, we found only a weak relationship between nest success and dispersal distance.  Taken together, these results were surprising because many studies have documented win-stay, lose-shift dynamics in a variety of avian systems, including waterfowl.  We conclude that in our system, because there is little year-to-year spatial correlation in predation risk, there is no advantage to win-stay, lose-shift.  Interestingly, at all spatial scales, we found spatial correlations between nest density in consecutive years, but not at longer time scales.  This may suggest that birds use short-lived habitat cues and/or conspecific attraction to select nest sites.  Given the importance of adaptive nest clustering in our particular system, we suggest that winners (especially older, early-nesting birds) stay, and later-nesting losers cluster around them, adaptively “shifting” to new social environments, but not new geographical locations per se.  In unpredictable environments, current social cues, such as the presence of active conspecific nests, may be especially valuable in selecting nest sites.