Technical Comments

Response to Comment on “Demographic dynamics of the smallest marine vertebrates fuel coral reef ecosystem functioning”

See allHide authors and affiliations

Science  20 Dec 2019:
Vol. 366, Issue 6472, eaaz1301
DOI: 10.1126/science.aaz1301


Allgeier and Cline suggest that our model overestimates the contributions of cryptobenthic fishes to coral reef functioning. However, their 20-year model ignores the basic biological limits of population growth. If incorporated, cryptobenthic contributions to consumed fish biomass remain high (20 to 70%). Disturbance cycles and uncertainties surrounding the fate of large fishes on decadal scales further demonstrate the important role of cryptobenthic fishes.

In their comment on our study, which highlighted the role of cryptobenthic fishes in coral reef trophic dynamics (1), Allgeier and Cline (2) question the value of consumed biomass and consumption turnover as components of reef fish productivity and posit that the contributions of cryptobenthic fishes to consumed biomass decrease greatly when our theoretical model is extended to ~20 years, rather than the 365 days used originally. Their comment provides a welcome opportunity to elaborate on the dynamics that underpin the role of cryptobenthic fishes on coral reefs.

We fully agree that consumed biomass is not the same as traditional productivity; rather, it represents the component of production that is transferred to the next trophic level. In fact, we noted that “cryptobenthics appear to make negligible contributions to net productivity and standing fish biomass” when highlighting their role as rapidly replenishing vectors of energy to larger-bodied fishes, a role we quantify as consumed biomass (1). Similarly, consumption turnover (our metric) and net turnover [traditionally defined as production/biomass, the P/B ratio (3)] quantify distinct aspects of how quickly particles move through the food web. Net turnover describes how quickly newly produced particles are stored as standing biomass, whereas consumption turnover quantifies how rapidly they are passed to larger consumers.

Allgeier and Cline present a series of models to show that (i) at a hypothetical endpoint, when all individuals have died, consumed and produced biomass are equal (this is axiomatic); (ii) without replenishment, cryptobenthic populations rapidly decrease in net produced and consumed biomass [this is also axiomatic, given that small sizes and short lives are the hallmarks of cryptobenthic life history (4)]; and (iii) when our model is extended over ~20 years, the contribution of cryptobenthic fishes decreases even when populations are replenished.

Extending short-term simulations of individual growth and mortality to decadal scales introduces a range of unaccounted meta-community, community, and population dynamics and, as such, should be interpreted with caution. Most important, however, Allgeier and Cline’s simulations disregard basic size-dependent limits on population growth. After 20 years, the abundance of large reef fishes in Allgeier and Cline’s model (~70% of individuals) greatly surpasses that of cryptobenthics (~30%) (Fig. 1A), which sharply contrasts with empirical data. Empirical densities from the three distinct locations in our original paper (1) show that cryptobenthic fishes account for ~66% of individual fishes on reefs [but can, in some cases, account for up to 95% (5)], which reflects the positive relationship between body size and self-thinning rates in populations that constrain larger-bodied individuals to lower densities (6). This suggests that Allgeier and Cline’s simple temporal extension of our model greatly exceeds the ecologically plausible population growth of large reef fish species. The well-documented temporal clustering of large reef fish recruitment (and resulting density-dependent mortality) compared to the continuous settlement, growth, and death of cryptobenthics (7) further aggravates this issue. Our paper highlighted the omissions of density dependence and recruitment pulses in our original, short-term model as simplifications that likely reduce the contribution of cryptobenthics to consumed biomass [supplementary materials of (1)].

Fig. 1 Contribution of cryptobenthic and large reef fishes to energy and nutrient fluxes over time.

(A) Abundance of cryptobenthic reef fishes (blue line) and large reef fishes (light gray line) over 8000 days (~22 years). In Allgeier and Cline’s simulation, large reef fishes greatly outnumber cryptobenthic individuals (~70% versus 30%), which is at odds with empirical data and ecological theory. The dark gray trajectory shows constrained large reef fish densities based on empirical data from the supplementary materials of (1). (B) Daily contributions to consumed biomass over ~22 years under the density-constrained simulations. Cryptobenthic reef fishes account for a minimum of ~20% consumed biomass, which is more than three times the estimate of Allgeier and Cline. (C) Net turnover of cryptobenthic and large reef fishes, averaged for each year (starting at the end of year 1). Contrary to Allgeier and Cline’s interpretation, cumulative cryptobenthic turnover far exceeds that of large reef fishes after 1 year and increases through time.

Nonetheless, Allgeier and Cline’s comments provide a valuable opportunity to devise a more nuanced model that integrates additional real-world processes. We reran their 20-year simulations with a simple, empirically supported density constraint on large reef fish densities relative to cryptobenthics (based on the proportions detailed above) where recruits experience instantaneous mortality if the population threshold is reached. The extension of the model time frame results in contributions of cryptobenthics to consumed biomass after 2, 3, 10, and 20 years of 43%, 32%, 20%, and 19%, respectively; these proportions are invariably more than three times the estimates of Allgeier and Cline (2) (Fig. 1B). Furthermore, Allgeier and Cline misinterpret figure S6E of (1) when claiming that large reef fishes have higher net turnover than cryptobenthics (2). The figure tracks single cohorts over 365 days, rather than summarizing cumulative contributions that incorporate replenishment. In the extended model, the P/B ratio (i.e., net turnover) of cryptobenthics increases from ~300% to ~390% year–1 after 22 years, whereas that of large reef fish declines from 130% to 50% year–1 (Fig. 1C).

Although the temporal extension of our model indeed reduces the contribution of cryptobenthics to consumed biomass, even when simple group-specific population thresholds are imposed, ecological dynamics operating on reef fish communities over larger temporal scales cannot be disregarded (Fig. 2). Given sufficient time without disturbance, large-bodied, mature individuals perish and supply a higher contribution to consumed biomass than do cryptobenthics. However, this contribution hinges on spatial and temporal uncertainties. Spatially, in the absence of fishing, most large, mature fishes (e.g., large groupers) will eventually fall prey to large, transient predators [e.g., (8)], which transfer energy and nutrients across seascapes (9) and represent a net export for reefs. Where fishing occurs, the largest fishes are preferentially caught and consumed (10), which again removes energy and nutrients from the reef and diminishes the contribution of large reef fishes to consumed biomass. Temporally, cyclical disturbances (e.g., tropical storms) are likely to alter population dynamics by causing mortality for large individuals [e.g., (11)]. Such disturbances occur at much higher frequencies than the 20-year time frame of our model (12), thus bringing into question the steady equilibrium of reef fish assemblages in extended simulations.

Fig. 2 Conceptual diagram highlighting spatial and temporal factors that determine the trophic contribution of cryptobenthic fishes.

Over short time frames (as created by frequent pulse disturbances or fishing pressure) and within a spatially constrained reefscape, cryptobenthics dominate the production of consumed fish biomass. Over longer time frames that result in mature, steady-state fish assemblages and across entire seascapes, the role of larger species may increase, but little is known about the ultimate fate of large-bodied species. Cryptobenthic reef fishes, in contrast, do not die of old age.

Overall, through their demonstrated population dynamics, cryptobenthic reef fishes provide a continuously available and predictable daily source of energy and nutrients for predators at local scales despite having a relatively small direct contribution to the traditional metric of secondary production. Over longer time frames and entire seascapes, the contribution of larger, longer-lived species can increase. However, even in the most extreme (and unlikely) scenarios, the daily contributions of cryptobenthic fishes still amount to 20% of consumed fish biomass, and our revised model yields a maximum contribution of 68% after 70 to 80 days. Our original estimate lies firmly within this range. Thus, regardless of the temporal and spatial scales considered, cryptobenthic reef fishes provide a sustained and sustaining resource on coral reefs.


View Abstract

Stay Connected to Science


Navigate This Article