Demographic Theory of Coral Reef Fish Populations with Stochastic Recruitment: Comparing Sources of Population Regulation

Abstract

The effects of three forms of density-dependent regulation were explored in model coral reef fish populations: top-down (predation), bottom-up (competition for food), and pelagic (non-reef-based mechanisms) control. We describe the demographic responses of both biomass and numbers of adult fish, predicting the mean and the variance of temporal fluctuations resulting from stochastic recruitment of juveniles. We find that top-down control acts by suppressing variability of numbers of fish, which in turn suppresses the variability of biomass. Bottom-up control has no effect on fluctuations of numbers of fish, though it strongly reduces fluctuations of biomass. Because fecundity of fish is directly linked to body mass, the regulation of biomass tightly regulates reproductive output independently of the number of individuals in the population. Finally, populations under pelagic control experience bounded fluctuations of biomass and numbers directly proportional to the bounded fluctuations of recruitment. The demographic signatures predicted from both bottom-up and pelagic control are consistent with current evidence supporting the recruitment limitation hypothesis in reef fish ecology. We propose tests to discriminate the dominant mode of density-dependent regulation using qualitative trends in time series demographic data across environmental clines.

Effects of monsoon-driven wave action on coral reefs of Guam and implications for coral recruitment

“…Likewise, larvae from corals that reproduce multiple times throughout the year, as many brooding corals do, may also have a better chance of finding suitable substrate to settle on. Recruitment shows high levels of spatiotemporal variation in many species including corals (Caley et al 1996;Glassom et al 2004;Sale et al 2005), and stochastic processes may be important for coral larval recruitment and coral reef community development (Palmer et al 1996;Sale 2004;Sandin and Pacala 2005). For example, recruitment patterns might differ depending on the actual conditions at the time of recruitment, which might in turn be affected by El Nino/Southern Oscillation (ENSO) events or variations in the monsoon system.…”

Predictability of Reef Fish Recruitment in a Highly Variable Nursery Habitat

“…However, the relative importance of pre-and postrecruitment processes (such as predation, competition, and habitat selection) has been subject to a lengthy debate regarding the unpredictability of recruitment (Hixon andJones 2005, Sandin andPacala 2005).…”

Oceanographic coupling across three trophic levels shapes source–sink dynamics in marine metacommunities

“…The goal of the latter effort was to determine whether the analytical model predictions were robust to the more complex dispersal patterns and nonlinear interactions present in real coral reef metacommunities. In general, the analytical and numerical models make similar predictions about the effects of tri‐trophic coupling on planktivore density, though the numerical model predictions are more nuanced because the perceived outcome depends on the demographic currency (numbers or biomass) under consideration (also see Sandin and Pacala 2005). The specific predictions of this analysis depend on the parameter values and dispersal scenarios chosen, and may not readily extrapolate to very different systems.…”

“…in terms of patch value). Theoretical explorations such as this one provide one avenue for investigating how processes that are empirically tractable on small spatial scales exhibit effects on larger spatial scales (also see Sandin and Pacala 2005, White 2008).…”

“…where δ N is a background mortality rate and ξ P represents a linear predator attack rate, a reasonable assumption for a generalist fish predator over typical ranges of prey densities (Sandin and Pacala 2005). We assume that large planktivores have a size refuge from gape‐limited predators (Wainwright and Bellwood 2002), so for age classes with L a > L gape (where L a is length at age a), ξ P = 0.…”