Density dependence can obscure nonlethal effects of disturbance on life history of medium-sized cetaceans

Nonlethal disturbance of animals can cause behavioral and physiological changes that affect individual health status and vital rates, with potential consequences at the population level. Predicting these population effects remains a major challenge in ecology and conservation. Monitoring fitness-related traits may improve detection of upcoming population changes, but the extent to which individual traits are reliable indicators of disturbance exposure is not well understood, especially for populations regulated by density dependence. Here we study how density dependence affects a population’s response to disturbance and modifies the disturbance effects on individual health and vital rates. We extend an energy budget model for a medium-sized cetacean (the long-finned pilot whale Globicephala melas) to an individual-based population model in which whales feed on a self-replenishing prey base and disturbance leads to cessation of feeding. In this coupled predator-prey system, the whale population is regulated through prey depletion and the onset of yearly repeating disturbances on the whale population at carrying capacity decreased population density and increased prey availability due to reduced top-down control. In populations faced with multiple days of continuous disturbance each year, female whales that were lactating their first calf experienced increased mortality due to depletion of energy stores. However, increased prey availability led to compensatory effects and resulted in a subsequent improvement of mean female body condition, mean age at first reproduction and higher age-specific reproductive output. These results indicate that prey-mediated density dependence can mask negative effects of disturbance on fitness-related traits and vital rates, a result with implications for the monitoring and management of marine mammal populations.

Compared to a constant environment, seasonality in productivity led to a strong densitydependent decrease in survival during the first years of life with no and ten days of disturbance per year (Fig. 7). With 30 days of disturbance this density-dependent effect of seasonality was relaxed, 24 which led to increased survival during the first six to eight years of life, depending on whether disturbance happened in summer or winter. Young mature females experienced starvation-induced mortality during lactating of their first calf with 30 days of disturbance in seasonal environments 27 and the resulting drop of survival nullified the survival advantage that was present at younger ages.
Ten days of disturbance per year did not change the age-specific survival pattern compared to no S2 Supplemental results Hin et al.
disturbance. The pattern of age-specific expected reproductive output was not affected by seasonal 30 variation in prey productivity. With seasonal variation in prey productivity, there was a small effect of disturbance on mean life expectancy and mean reproductive output of females older than age 10 yrs (Fig. 8). With summer disturbance, mean life expectancy increased from 10.6 to 11.5 33 yrs between no and 40 days of disturbance and mean reproductive output of females beyond age 10 yrs slightly decreased (from 3.14 to 3.07). When disturbance happened in winter there was no noticeable effect of increased disturbance duration on these life history statistics. 36 Without disturbance, females living in seasonal environments had a lower AfR and a higher AfW than females living in non-seasonal environments (Fig. S8). This response to seasonality can be explained by the lower calf survival associated with seasonality. As for constant prey productivity, 39 there was an overall downward trend in length at first reproduction, AfR and AfW with increasing disturbance duration in seasonal environments (Fig. 8  Figure 3: Functions related to prey feeding, lactation and starvation mortality and the effect of changes in parameters on the shape of these functions. In a), the feeding level 1 1 + e −η( ρW /F −1) determines both prey and milk assimilation rate and is plotted as a function of body condition ( F /W ). In b) and c), the feeding efficiency a γ T γ R + a γ affects resource assimilation rate. In d), milk provisioning represents the component Note that the female entered the waiting period during the end of the lactation period of calf #5.
Prey productivity is seasonal (A = 0.25) and other parameters at default values (Table 1 of S1: Model parameterization).   Figure 9: Sensitivity of whale and prey density to parameter T R , which modifies the increase in resource feeding efficiency with age and sets the age at which resource feeding efficiency is 50%.
Default value is T R = 500 (triangles), for which feeding efficiency at weaning (a = 1223) is 86%.
Population response to disturbance is altered at T R = 1000, for which feeding efficiency at weaning is 60%. This led to starvation mortality among young males and changed the sex ratio towards more female-dominated.  Figure 10: Sensitivity of whale and prey density to parameter γ, which modified the non-linearity in the increase in resource feeding efficiency with age. Default value is γ = 2 (triangles), for which feeding efficiency at weaning (a = 1223) is 86%. Population response to disturbance is altered at γ = 1, for which feeding efficiency at weaning is 70%.  Figure 13: Sensitivity of whale and prey density to parameter ξ c , which modified the shape of the decline in milk feeding efficiency by the calf with calf age. Default value is ξ c = 0.9 (triangles). For ξ c = 0.0, milk feeding declined linearly with calf age, and for ξ c = 0.999 milk feeding was unaffected by calf age up to the age at weaning when it rapidly declined.  Figure 14: Sensitivity of whale and prey density to parameter T N , which sets the age beyond which the milk feeding efficiency starts to decline with calf age. Default value is T N = 365 (triangles).
For T N = 1, milk feeding starts to decline with calf age when the calf is one day old. For T N = 730 this only happens when the calf is two years old.  Figure 19: The effect of changes in η, γ, T R , ξ m , ξ c , T N and µ s on the response of the mean lifetime reproductive output of females that die beyond 10 yrs of age to increasing disturbance duration.
Points and lines as in Fig. 16.