Millions of seafloor pits, not pockmarks, induced by vertebrates in the North Sea

The study addresses the origin of abundant seafloor depressions in the German Bight (North Sea) that were previously interpreted as methane-related pockmarks. Pockmarks are globally widespread depressions typically attributed to fluid (often methane) venting, yet similar features can arise from other processes including current scouring and megafaunal foraging. The authors hypothesize that the North Sea features are not classic pockmarks but biologically induced pits created by harbor porpoises during benthic foraging, which subsequently act as nuclei for current-driven scouring. The work aims to evaluate this alternative, non-geological mechanism by integrating high-resolution hydroacoustic mapping with oceanographic context, habitat distribution, and behavioral ecology.

Pockmarks are known from diverse marine and lacustrine settings and often correlate with methane seepage and groundwater discharge. Their dimensions range from meters to hundreds of meters in diameter and depth up to hundreds of meters for giant examples. However, non-fluid-related mechanisms have also been proposed, including scouring around obstacles, and biogenic excavation by gray whales and fishes. In the German Bight, a large field of shallow depressions north of Heligoland emerged and waned over months and was attributed to episodic methane degassing and storm infilling. Yet documented active seepage in the German Bight is sparse and mainly linked to localized features (Dogger Bank, Palaeo-Elbe, and anthropogenic events). Prior reports connected depressions to areas with thin Holocene cover, but evidence for shallow gas (e.g., acoustic blanking, flares, gas migration pathways) is limited or ambiguous in the study area. These observations leave room for alternative, biology-driven explanations for shallow, morphologically atypical depressions.

The team conducted hydroacoustic surveys across the Sylt Outer Reef (SOR) in the German Bight, focusing on 20–40 m water depths with fine to medium sands. Data acquisition: multibeam echosounders (MBES) included KONGSBERG EM710 (2013, 2017, 2021; 70–100 kHz FM chirp, 0.5×1° RX/TX, 400 soundings/ping; DGPS positioning), KONGSBERG EM712 (0.5×0.5°; 70–100 kHz; dual-ping; Seapath INS), and a NORBIT iWBMS (400 kHz chirp, 80 kHz bandwidth, theoretical range resolution ~0.009 m) with Applanix Wavemaster II INS and RTK GNSS (0.02–0.05 m accuracy). Sound velocity was measured via AML keel probe and vertical casts. Data processing: multibeam data were calibrated for roll, pitch, yaw (Qimera), corrected for refraction and tides via modeled water level time series (BSH), spline-filtered to remove outliers, and gridded to 1×1 m for visualization (QGIS 3.16); high-quality 400 kHz point clouds were also analyzed directly (GMT). Time-lapse mapping: repeat MBES surveys of the same area in May 2021 and November 2021 enabled assessment of pit evolution. Automated depression mapping: within ArcMap, circular depressions were filled to their pour points; the filled grid was subtracted from originals to obtain differential heights; polygons were generated for regions with ≥0.05 m height difference; small holes were filled if <50% of polygon area; outlines smoothed using PEAK; features <10 m² removed; final polygons were visually quality-checked and mean/maximum depth and area calculated. Complementary datasets: - Sidescan sonar (Edgetech 4200MP, ~300 kHz; towed 10 m above bottom; 140–230 m range) to resolve backscatter patterns. - Sub-bottom profiling (R/V Merian: PARASOUND DS3 P70 at 19.3 kHz; ~4.5° beam; up to ~13 m penetration; R/V Heincke: INNOMAR SES-2000 medium, 6–15 kHz) to detect shallow gas, stratigraphy, and Holocene thickness. - Groundtruthing: 45 HELCOM Van Veen grab samples (top 2 cm; carbonates and organics removed; grain-size analyzed with CILAS 1180; Gradistat statistics) and drop-camera video transects with laser scaling (~0.1 m spacing). - Oceanographic context: satellite-derived sea-surface patterns to locate tidal mixing fronts; review of regional hydrography. - Ecological context: harbor porpoise seasonal density modeling (generalized additive models) from 2005–2013 line-transect surveys using predictors such as depth, distance to shore, proximity to sandeel grounds, SST, frontal proxies, and day length; sandeel habitat maps to assess spatial overlap with pits.

• Mapping extent and morphology: In the northern study region, 4,965 km of MBES profiles spanning 2,128 km² were analyzed. An automated algorithm detected 42,456 depressions covering about 9% of the seafloor. Pits are truncated-cone depressions with a remarkably consistent shallow mean depth of 0.11 m (typical range 0.05–0.2 m; mode 0.08 m) and a mean surface area of 297 m² (circular-equivalent radius ~9.7 m). Virtually no pits exceed 0.2 m depth even when up to ~50 m in diameter. Shapes include furrows, concentric highs, and adjacent mounds; many pits are non-circular (e.g., half-moon) and exhibit linear, curvilinear, radial, or clustered alignments with equidistant spacing. - Temporal evolution: Time-lapse (May vs. November 2021) shows smoothing and widening of existing pits, commingling into larger irregular scour-depressions, and the appearance of numerous new smaller pits (2–6 m wide, ~0.2 m deep) predominantly in previously pitted areas. - Sedimentology and setting: Pits occur exclusively in fine to medium sands within 20–35 m water depth; no pits were observed in gravelly coarse sands (confirmed by grab samples). - Lack of gas-seep indicators: Across extensive sub-bottom profiles, no acoustic blanking, bright spots, gas flares (the single previously reported 1.8 m high anomaly is ambiguous and likely a fish school), or migration pathways (pipes/chimneys) were identified. Holocene sand thickness varies from 0–10 m; pit occurrence does not correlate with minimal Holocene cover only. Dissolved methane near Heligoland (~4 nM) is close to North Sea background (~3.5 nM); no chemosynthetic communities or MDAC were reported. - Biogenic mechanism supported: Spatial coincidence was found between pits, tidal mixing fronts (areas of enhanced productivity), known sandeel habitats, and modeled high spring densities of harbor porpoise. Harbor porpoises (estimated abundance 23,219 in the German Bight; ~4,078 in the study area in spring) are capture-suction feeders that perform benthic foraging, with stomach contents sometimes containing sand. Behavioral observations (captive/other cetaceans) indicate bottom grubbing 0.1–0.2 m deep, consistent with pit depths and morphologies (furrows, mounds). - Energetic/foraging plausibility: Average porpoise consumption is ~1.96 kg/day; benthic fraction ~1.14 kg/day; sandeel share ~413 g/day (~31 fish/day), implying ~125,743 sandeels consumed per day by porpoises in the study area, supporting the capacity to produce large numbers of feeding pits. - Evolution model: Initial feeding pits act as nuclei for tidal-current scouring to form larger pit-scours (≥10 m), with merging of pits; episodic winter storms can level features, resetting the seafloor. - Management relevance: High-resolution MBES can identify benthic feeding grounds, informing conservation and spatial planning amid offshore wind expansion.

The morphological consistency (shallow, decimeter-deep, non-circular pits with furrows and mounds), their restriction to fine–medium sands at 20–35 m depths, and their temporal dynamics are inconsistent with classic methane-driven pockmarks but align with repeated biogenic excavation from above. The absence of geophysical or geochemical evidence for shallow gas further undermines a seepage origin. The strong spatial association with tidal mixing fronts and sandeel habitats, together with high modeled harbor porpoise densities and known benthic foraging behavior, supports the hypothesis that porpoises create small feeding pits that then evolve through hydrodynamic scouring into larger depressions. This biogenic mechanism reframes interpretations of similar shallow depressions and highlights megafauna-driven macro-bioturbation as an important modulator of sediment transport, seabed roughness, and benthic habitat structure. Recognizing these features as biologically mediated also provides a novel, non-invasive means to map vertebrate feeding grounds using modern MBES, with implications for marine spatial planning and protection of productive shelf habitats.

High-resolution multibeam mapping and multidisciplinary contextual data show that tens of thousands of shallow seafloor depressions in the German Bight are best explained as harbor porpoise feeding pits that subsequently undergo scouring, rather than methane-related pockmarks. The decimeter-scale, morphologically distinctive pits occur in fine–medium sands, evolve over months, and correlate with sandeel habitats and porpoise density near tidal mixing fronts. This work underscores the underappreciated role of megafauna-driven macro-bioturbation in shaping continental shelf seafloors and suggests that millions of similar pits likely occur transiently in the North Sea and globally. Future research should seek direct in situ documentation of foraging events creating pits, quantify hydrodynamic thresholds for pit-to-scour transitions, assess contributions from other benthic-feeding vertebrates (e.g., seals), and integrate seabed habitat mapping into conservation and offshore development planning.

• The initial excavation events by harbor porpoises were not directly observed in situ; the biogenic origin is inferred from morphology, spatial correlations, behavioral literature, and time-lapse changes. - Time-lapse coverage spans only a six-month interval and a limited area; seasonal and interannual variability remain to be quantified. - Sidescan sonar proved unreliable for pit detection, potentially limiting cross-sensor validation. - Sub-bottom data penetration was ~13 m (Merian) and may miss deeper gas signatures, though the Holocene cover is thin and no shallow gas indicators were observed. - The contribution of other predators (e.g., seals) to pit formation was not quantified. - Consumption and abundance estimates rely on published models and assumptions and provide order-of-magnitude plausibility rather than direct rates of pit formation.