Gelatinous filter feeders increase ecosystem efficiency

Abstract

Gelatinous filter feeders (e.g., salps, doliolids, and pyrosomes) have high filtration rates and can feed at predator:prey size ratios exceeding 10,000:1, yet are seldom included in ecosystem or climate models. We investigated foodweb and trophic dynamics in the presence and absence of salp blooms using traditional productivity and grazing measurements combined with compound-specific isotopic analysis of amino acids estimation of trophic position during Lagrangian framework experiments in the Southern Ocean. Trophic positions of salps ranging 10-132 mm in size were 2.2 ± 0.3 (mean ± std) compared to 2.6 ± 0.4 for smaller (mostly crustacean) mesozooplankton. The mostly herbivorous salp trophic position was maintained despite biomass dominance of ~10-µm-sized primary producers. We show that potential energy flux to >10-cm organisms increases by approximately an order of magnitude when salps are abundant, even without substantial alteration to primary production. Comparison to a wider dataset from other marine regions shows that alterations to herbivore communities are a better predictor of ecosystem transfer efficiency than primary-producer dynamics. These results suggest that diverse consumer communities and intraguild predation complicate climate change predictions (e.g., trophic amplification) based on linear food chains. These compensatory foodweb dynamics should be included in models that forecast marine ecosystem responses to warming and reduced nutrient supply.

Fig. 1. Conceptual food web diagrams for a size-structured ecosystem or an ecosystem with diverse omnivores and variable predator:prey size ratios.

Conceptual food web diagrams for a size-structured ecosystem with fixed predator:prey size ratios (a, b) or diverse omnivores with highly variable predator:prey size ratios (c, d) in a large-phytoplankton-dominated system (a, c) or a small-phytoplankton-dominated system (b, d). The color of circles is proportional to the production (primary or secondary) of a functional group. The trophic amplification hypothesis is based on conventional size-structured ecosystem models (a, b). Thus, shifts toward small phytoplankton in a future climate (represented by moving from a to b) lead to food chain elongation through the insertion of additional protistan trophic levels. In contrast, the compensatory foodweb dynamic suggests that bottom–up processes driving a shift from large (c) to small (d) phytoplankton would be accompanied by a shift of metazoan communities toward filter feeders with large predator:prey ratios (e.g., salps). This conceptualization of the foodweb involves high functional diversity amongst consumer trophic levels and substantial intraguild predation. These processes could stabilize ecosystem functions in response to climate change disruptions of nutrient supply.

Fig. 2. Ecosystem bulk stable isotopes.

Ecosystem bulk stable isotopes for size-fractionated zooplankton (a) and salps (b). Color represents the Lagrangian experiment number (C1–C5). Polygons (a, b) represent the isotopic signatures of suspended particulate organic matter (POM) for the corresponding experiment. Depth profiles are POM δ¹³C (c) and δ¹⁵N (d). Error bars are ±1 standard error of measurements made on different days of the Lagrangian experiment (n = 3–6). SA Subantarctic, SA-Sc Subantarctic-Southland-Current-influenced, ST Subtropical.

Fig. 3. Zooplankton and salp trophic positions.

Trophic positions of size-fractionated zooplankton samples (a) and salps (b). In boxplot (a), central red line indicates median, box indicates one quartile above and below median and whiskers extend to most extreme non-outlier samples. Outliers (1.5 times the interquartile range above or below the 25th or 75th percentile) are plotted as “+” symbols. In (b), colors represent the Lagrangian experiment and shapes represent salp species (Salpa thompsoni, Thetys vagina, Pegea confoedereata, and Soestia zonaria).

Fig. 4. Ecosystem production as a function of organism size with and without salp blooms.

Production as a function of size for different Lagrangian experiments. a Total biomass production (primary production + secondary production). b Secondary production of a size bin divided by total net primary production. To account for slightly different widths of size bins, the production of each size bin was normalized by dividing by the number of octaves (factors of 2) covered by the size bin. For phytoplankton and heterotrophic protists, we assumed that production in a size bin was proportional to biomass in that size bin (note that a gap in the 0.1–0.2 mm size bin exists for cycles in which no organisms in this size class were enumerated by FlowCam). For higher trophic level (predators of zooplankton) calculations, see the “Methods” section.

Fig. 5. Ecosystem transfer efficiency.

Ecosystem transfer efficiency as a function of surface chlorophyll a (x-axis) and the proportion of metazoan herbivory conducted by gelatinous filter feeders (thaliaceans = salps, doliolids, and pyrosomes), which is depicted on the color axis. Results are from this study (STF) and the Costa Rica Dome (CRD), Equatorial Pacific (EqP), North Pacific Subtropical Gyre (NPSG), California Current Ecosystem (CCE), and Gulf of Mexico (GoM). All studies had similar approaches to estimating the NPP and grazing rates of all herbivores. Trophic position was assessed via either CSIA-AA or food web models.