Nest site selection across environmental gradients is a major driver of early life demographic rates and can regulate populations when habitat is limiting. Several modes of nest site selection have been proposed in studies of salmonid spawning behavior including sequential nesting, size-based hierarchies, superimposition, bet hedging, and random, non-adaptive behavior. Each mode likely has distinct outcomes for average larval survival; however, the individual-level data required to evaluate the evidence for each mode isn’t always available. Here we report the results of a study on a spawning aggregation of wild brook trout (Salvelinus fontinalis) in the Dead Diamond River, New Hampshire, USA. Pre-spawn environmental data— including hyporheic temperature and dissolved oxygen, sediment size, and depth — were collected in a 50m X 20m reach at a 2m X 2m resolution to define available spawning habitat. Snorkel surveys were performed three times each day during the spawning season (9/30-11/03) to identify recently constructed nests, and standardized video recordings captured information on female identity, size, and behavior associated with each nest. We used these data to test several hypothesized relationships between environmental gradients and the distribution, timing, and size-dependency of nest site selection.
Results/Conclusions
We found evidence for adaptive, non-random nest site selection, with the probability of nest occurrence increasing with temperature and dissolved oxygen (log odds O2 = 0.9 ± 0.5 SE, p = 0.05; log odds temperature = 1.7 ± 0.6, p = 0.006). Nest sites filled sequentially as a function of temperature but not dissolved oxygen (linear regression: F2,28 = 3.1, p = 0.06, R2 = 0.18); yet both nest timing and environment weren’t related to female body size, perhaps because we observed several larger females digging multiple nests across temperature and oxygen gradients. Qualitatively, superimposition appeared to occur more frequently than expected given the available habitat, though the frequency of superimposition was not related to environmental gradients. A thermal model of embryonic development indicated that spring hatching phenology spanned 81 days relative to the 15-day fall nesting window in this spawning reach. Our preliminary results demonstrate that this brook trout population exhibits multiple nest site selection behaviors, and this behavioral diversity across habitat quality gradients may affect our estimates of young-of-year production and eventual recruitment in unanticipated ways.