A rare oasis effect for forage fauna in oceanic eddies at the global scale

Oceanic mesoscale eddies are coherent and transient swirling structures, ubiquitous in the world's oceans, and often considered oceanic oases that aggregate marine life in pelagic deserts. Evidence from fishing catch and satellite-tracking data reveals a preference for Anticyclonic Eddies (AE) by a variety of marine predators. Recent studies demonstrated higher abundance of predatory fish and fishing activity in AE across regional and global scales. The prevailing explanation is a bottom-up structuring effect, where increased food resources within eddies attract predators feeding on forage fauna (small fishes, crustaceans, molluscs). However, the response of forage fauna to eddies appears nuanced: some studies report aggregation in AE, others in Cyclonic Eddies (CE), no discernible effect, or variability depending on eddy characteristics such as age, size, amplitude, or lifespan. Observational constraints hinder comprehensive assessment: while satellites sense surface temperature (SST) and phytoplankton, direct observations of forage fauna often require net tows or ship-borne echosounders. Acoustic echosounders offer high-resolution vertical insights but limited taxonomic resolution. Two seminal publications with small eddy samples inferred an oasis effect in AE, but limited sample sizes preclude generalization. Therefore, this study systematically evaluates how forage fauna respond to eddies by merging a global set of acoustic vertical profiles (upper 750 m) with a global eddy database (2001–2020), examining 999 eddies. By extracting key eddy characteristics (size, amplitude, lifespan, trapping ability, temperature and chlorophyll signatures), the study tests the eddy-enhanced forage biomass hypothesis across diverse regions.

Prior work has suggested that mesoscale eddies act as biological oases for higher trophic levels, with multiple studies reporting predator use of AE and increased fishing activity within them. For forage fauna, findings have been mixed: increases in AE, increases in CE, no effect, and dependence on eddy properties (age, size, amplitude, lifespan) have all been reported. Hypotheses include bottom-up structuring (enhanced surface chlorophyll in CE attracting zooplankton and cascading up to forage fauna), physical trapping of micronekton within eddies leading to distinct communities from surrounding waters, and physical convergence in AE aggregating micronekton. Earlier acoustic studies underpinning the oasis hypothesis analyzed very few eddies (e.g., 13 and 4), limiting inference. The literature also documents well-known SST and chlorophyll signatures of eddies (cold, chlorophyll-enriched CE; warm AE), with regional variability including cold-core AE and warm-core CE. Many positive forage-fauna effects were reported in regions with strong eddy amplitudes and energetic boundary currents, raising the possibility of spatial sampling bias in prior conclusions.

Datasets: The study collated georeferenced single-beam shipborne acoustic data at 38 kHz from public databases spanning 2001–2020 (Pacific, Atlantic, Indian Oceans), plus three Mozambique Channel surveys. Backscatter was integrated to compute Nautical Area Scattering Coefficient (NASC; m^2 nmi^−2), a proxy for forage fauna density, primarily detecting gas-bearing organisms (e.g., swimbladdered fishes, siphonophores). Acoustic profiles were constrained to 20–750 m and interpolated to 10 m vertical resolution. Profiles during astronomical twilight (sun 0° to 18° below horizon) were removed to avoid diel vertical migration transitions; shallow shelf regions (seabed <1000 m) were excluded. In total, 118,813 vertical profiles were available. Eddy characterization used the altimetric Mesoscale Eddy Trajectories Atlas (META3.x DT, AVISO/DUACS), with daily eddy detection/tracking on absolute dynamic topography (ADT). Provided attributes include eddy type (AE/CE), amplitude (ADT difference between edge and center), effective area, effective radius, and exact effective and speed-edge contours. Eddies with lifespan ≥14 days (1993–2021) were considered. A trapping (non-linearity) metric was computed as the ratio of eddy translational distance to rotational speed over a 5-day window. Collocation and anomaly computation: Each acoustic observation was associated to the nearest eddy center, and its position classified as inside the exact effective eddy contour or outside. The outside control region was defined as the ribbon from the effective border to twice the mean eddy radius, sampled in the same day/night period as inside profiles. For inside observations, subregions (core, intern, speed border, effective border) were also assigned for specific analyses. For each eddy, mean inside and mean outside NASC vertical profiles were computed separately for coherent day and night samples. Anomalies were defined as (inside − outside)/outside and evaluated for two depth layers: epipelagic (0–200 m) and mesopelagic (200–750 m), and across depth. Satellite SST (ESA SST CCI/C3S L4, ~5 km, daily) and surface chlorophyll-a (Copernicus GlobColour L4, ~4 km, daily) were extracted along acoustic tracks and similarly used to compute inside vs outside anomalies per eddy. Statistical analysis: For each eddy, Wilcoxon rank-sum tests (two-sided, 95% confidence) compared the distributions of inside vs outside mean values for NASC (for each layer), SST, and chlorophyll. Eddies were classified per variable as Increase (inside > outside), Decrease (inside < outside), or Null (no significant difference). Aggregated proportions were calculated by eddy type (AE, CE) and depth layer. Mean anomalies and 95% confidence intervals (1.96 SD/√N) were reported. Dependence of NASC response on six eddy characteristics (amplitude, trapping metric, effective area, lifespan, SST anomaly, chlorophyll anomaly) was assessed using Wilcoxon tests between Increase/Decrease vs Null groups. Regional analyses summarized responses within Longhurst biogeochemical provinces and mapped the fraction of affecting eddies on a 3° grid, comparing with global maps of mean eddy amplitude. All analyses were performed in R (v4.2.1) with custom code; coastlines from rnaturalearth.

- Dataset and coverage: 999 eddies sampled globally with collocated acoustics (473 AE, 526 CE). Acoustic data spanned >350,000 km^2, depths 20–750 m (2001–2020). - SST and chlorophyll signals: Clear eddy signatures were recovered. Mean SST anomaly inside CE was −0.1 °C relative to outside (p < 2×10^−16); AE showed +0.05 °C (p ≈ 4×10^−? as reported). Across eddies, 87% showed significant SST responses. Chlorophyll exhibited modest signals: CE had a mean +3.2% increase (p = 2×10^−3); AE anomalies were marginal (p = 0.051). Overall, 82% of eddies showed significant chlorophyll responses. - Forage fauna (NASC) response: In stark contrast, most eddies showed no significant change in NASC. In the epipelagic layer (0–200 m), 86% of CE and 90% of AE were Null; only 5% (AE) and 7% (CE) showed significant increases (oasis effect). In the mesopelagic (200–750 m), 84% (CE) and 88% (AE) were Null; 9% of AE showed increases, while CE showed a slightly larger proportion (11%) of decreases. No robust mean NASC anomaly was detected overall; CE exhibited a significant ~10% mean NASC decrease between 450–600 m. - Overall rarity of oasis effect: Globally, only a minority (~13%) of eddies exhibited any significant forage-fauna effect, with about 6% showing an oasis (increase) effect. Oasis effects occurred in both AE and CE. - Drivers of NASC response: Among six tested eddy characteristics, only amplitude and trapping capacity (non-linearity) differed significantly between affecting (increase/decrease) and Null eddies. Stronger, more trapping eddies were more likely to show either increases or decreases, consistent with a physical barrier/trapping mechanism rather than simple vertical isopycnal displacement or bottom-up effects. SST and chlorophyll anomalies did not distinguish affecting from Null eddies. - Even among strong eddies, effects remained uncommon: In the top 20% by amplitude (>0.1 m), 82% showed no detectable NASC effect; in the top 5% (>0.22 m), 70% showed no effect, with only 35% of AE displaying significant mesopelagic increases. - Regional patterns: Eddies with NASC effects clustered in subtropical and southern latitudes with strong mean eddy amplitudes and energetic currents (e.g., South Subtropical Convergence). In regions with high mean amplitude, the fraction of affecting eddies rose to ~35% on average, with mixed positive and negative responses. The publicly available acoustic dataset undersampled some high-energy regions (e.g., northwest Pacific, northwest Atlantic, northern Indian Ocean), potentially biasing global proportions toward Null effects.

The findings challenge the prevalent view that mesoscale eddies generally act as biological oases for forage fauna. While SST and surface chlorophyll display expected and frequent eddy signatures, forage-fauna density as detected acoustically typically does not differ inside versus outside eddies. The rare cases of significant changes are associated with high-amplitude eddies and strong water-mass trapping in both AE and CE, supporting a trapping/barrier mechanism whereby eddies isolate and advect distinct micronekton communities, sometimes increasing or decreasing biomass relative to surroundings. The lack of differences in SST and chlorophyll between affecting and Null eddies argues against a generalized bottom-up pathway from surface phytoplankton to forage fauna, at least as captured by surface chlorophyll. The study’s results align with prior localized reports of strong eddy effects in high-energy current systems, but suggest earlier global extrapolations may have been biased by regional focus. The rarity of forage-fauna aggregation implies that predator use of eddies may also be influenced by physical habitat modifications (e.g., deeper thermoclines in warm AE expanding thermal niches and enabling prolonged deep foraging) rather than prey biomass increases alone. Regional analyses and the persistence of Null effects even among the strongest eddies underscore the nuanced, context-dependent nature of eddy–fauna interactions.

By integrating a global eddy trajectory atlas with an extensive archive of shipborne acoustic profiles, this study provides a systematic global assessment of eddy impacts on forage fauna. It shows that significant forage-fauna responses are uncommon globally and that the canonical oasis effect is rare, occurring in both AE and CE and largely confined to high-amplitude, strongly trapping eddies. Amplitude and trapping capacity emerge as the principal predictors of any effect, whereas SST and surface chlorophyll signals are not predictive of forage-fauna changes. These results imply that predator aggregation in eddies may often arise from physical habitat alterations rather than widespread prey aggregation. Future research should: (i) broaden geographic sampling to underrepresented high-energy regions; (ii) employ multi-frequency or broadband acoustics and integrate midwater trawls to resolve species composition and scattering biases; (iii) resolve vertical phytoplankton structure beyond surface chlorophyll; and (iv) leverage autonomous platforms (e.g., sailing drones, instrumented floats) for sustained, depth-resolved biological and physical observations across eddy life cycles.

- Acoustic constraints: Single-frequency (38 kHz) data primarily detect gas-bearing organisms, limiting taxonomic resolution and potentially biasing biomass inference; surface layer (0–20 m) is unsampled due to bubble/surface interference. - Depth range: Acoustic sampling limited to 750 m prevents assessment of deeper diel migrant distributions and responses below this depth. - Use of surface chlorophyll only: The study could not evaluate eddy-induced changes in the vertical chlorophyll structure (e.g., deep chlorophyll maxima), which may be relevant for forage fauna. - Spatial sampling bias: Public acoustic data under-sampled several high-energy regions (e.g., northwest Pacific, northwest Atlantic, northern Indian Ocean), potentially inflating the proportion of Null effects in the global analysis. - Statistical design: Day/night segregation mitigates, but does not eliminate, potential confounding from diel vertical migration; eddies without sufficient inside/outside regional sampling were excluded, possibly affecting representation. - Eddy strength focus: Even among strong eddies, effects were heterogeneous, indicating unmeasured factors (e.g., local environmental context, species composition) may modulate responses.