Distribution of Megafaunal Taxa at the Pythia's Oasis Methane Seep

Abstract

Methane seeps are critical sources of productivity and habitat in the deep sea, potentially supporting major fisheries as well as contributing to nutrient cycling and carbon sequestration. Although thousands of methane seep sites have been detected, many remain unmapped and ecologically unexamined due to expense or time constraints. This study is the first survey of benthic megafauna at Pythia's Oasis, a seep with unusually high rates of venting fluid and fluid chemistry never before documented in the oceans. The fluids are thought to be sourced from overpressure in the Cascadia Subduction Zone and funneled to the site along the Alvin Canyon strike slip fault. Working from a 2019 dataset of georeferenced ROV images, this study used machine learning to increase the efficiency of image annotation and evaluated the relationship between faunal communities and substrate type. Over 22 taxa were present. The most abundant taxa included anemones, sponges, sea cucumbers, snails, corallimorphs, rockfish, and soft corals. Anemones and black corals at Pythia's Oasis may represent undescribed species. Hard and soft substrate communities are likely present, although substrate composition only explained ~16% of the variation in fauna distribution. This study demonstrates how machine learning enables scalability of image processing allowing establishment of an ecological baseline in methane seep habitats. As anthropogenic impacts on the deep sea escalate, these data are increasingly important, because they have the capacity to improve ecosystem modeling, provide the groundwork for conservation, and support sustainable fisheries management.