The growing threats to deep-sea ecosystems (e.g., offshore wind, seabed mining, aquaculture) have the potential to damage these understudied systems and to impact the fisheries that they support. However, to date, there has not been a uniform assessment identifying which commercially landed species are associated with the cold-water corals and other structure-forming invertebrates (SFI) found in the deep-sea. We conducted a meta-analysis to identify the species of fish and invertebrates landed by commercial fisheries in California, USA (2011-2024), and evaluate their potential associations with SFI. Captured species are assigned to Landing Categories, which range in specificity from individual species to broad taxonomic groups. We developed a referenced list of observations documenting associations, and use three metrics to identify which of the 313 Landing Categories are (or are not) associated with SFI: (1) Adjacent: species observed within one body length of SFI (target fish) or touching SFI (target invertebrates); (2) General Proximity: species observed within the general area of SFI or based on modeling; (3) Habitat: species’ depth ranges and habitat requirements suggest associations. Approximately 30% of Landing Categories were associated with cold-water SFI, but the specific percentage varied across metrics (Adjacent: 23%; General Proximity: 27%; Habitat: 36%). Commercially targeted fish were more likely to be associated with SFI than invertebrates, but many invertebrates had undocumented life histories and/or non-specific Landing Categories. This study provides a step toward understanding the importance of SFI for commercial fisheries and demonstrates the broad co-occurrence between commercially landed marine species and living deep-water habitats.

Study Site

California’s ocean is characterized by two Marine Ecoregions: a temperate, upwelling driven system north of Point Conception (i.e., Montereyan Pacific Transition) and a subtropical, current driven system south of Point Conception (i.e., Southern Californian Pacific) (Wilkinson et al., 2009). A transitional zone is located at the Point Conception geographic break which centers around the northern Channel Islands and Santa Barbara Channel. However, the range of the warm-water species can extend northward, facilitated by warming effects of El Niños, marine heat waves, and the Pacific Decadal Oscillation (Micheli et al., 2020). Due to the confluence of these distinct marine regions, California’s marine environment is characterized by high biodiversity relative to other Western states. The deep sea in California encompasses most state and Federal waters, including seamounts, submarine canyons, and relatively featureless plains of soft sediment. Such deep-sea areas are targeted by commercial and recreational fisheries. In rocky areas, cold-water SFI cover cobble, boulders, and bedrock. These deep-sea corals and sponges provide habitat for fishes and invertebrates, including those of commercial importance.

Commercial fishing in California takes place in 14,937 km² of the ocean (Figure 1). Operations include small-scale boats with a single individual or a few crew members to large-scale boats with large crews. California’s commercial fisheries use a variety of fishing gears and fishing methods, ranging from traps, to purse seines, to bottom trawls (CDFW, 2021). Fishing operations catch many species of fish and invertebrates that live on or near the seafloor, and provide significant revenue to the state (CDFW, 2024a). Some areas of the deep ocean are protected as part of Federal National Marine Sanctuaries (which allow fishing in most areas), California's network of marine protected areas (MPAs), including some MPAs that are closed to commercial and recreational fishing such as State Marine Reserves (SMR), and Groundfish Exclusion Areas (which are intended to mitigate potential impacts to sensitive deep water coral habitat and prohibit all commercial and recreational groundfish fishing activity) (Figure 1). 

Fisheries Dataset

Ex-vessel fisheries landings data (Fish Ticket Data) from CDFW document the species of marine life that have been caught by commercial fisheries and landed in California. In the Fish Ticket dataset, fishers assign landed catches a category name (‘Species Name’) and corresponding number (‘Species ID’). In practice, the ‘Species Name’ and ‘Species ID’ reporting categories include a non-consistent range of species (from single species to whole taxonomic groups – see below). To distinguish between species-specific information and information about the species that are grouped within a ‘Species Name’ category, we use the term ‘Landing Categories’ to refer to catch data.

Many Landing Categories are resolved to a single species (e.g., Greenspotted rockfish (Sebastes chlorostictus), Market squid (Doryteuthis opalescens), Pacific angel shark (Squatina californica)). However, other Landing Categories are assigned to groups that range in levels of breadth. For example, some multi-species categories are based on: (a) broad taxonomic groupings (e.g., unidentified molluscs (Phylum Mollusca)); (b) taxonomy and habitat preferences (e.g., slope rockfish (Sebastes spp.) as defined by CDFW (CDFW, 2021, 2024b); and (c) groups of species that are difficult to differentiate in the field (e.g., Canary (Sebastes pinniger) and Vermillion (Sebastes miniatus) Rockfish Group).

The accuracy of reporting for landings has been evaluated for a subset of species, and varies among groups and through time. For example, some single species categories of rockfish have high accuracy and contain over 95% of a single species (e.g., Greenspotted, Yellowtail (S. flavidus), and Copper rockfishes (S. caurinus)) (Pearson & Almany, 1995). Other categories of rockfishes have been recorded with lower accuracy (e.g., Blackgill rockfish (S*. melanostomus)*, Cowcod (S. levis), and Splitnose rockfish (S. diploproa)) (Pearson & Almany, 1995). Furthermore, the precision of records has changed through time - some species are currently placed in single species categories, but were lumped into broad categories historically (e.g., Greenspotted rockfish vs. Rockfish - unspecified). CDFW acquires data from its own fisheries management activities and from mandatory reporting requirements on the commercial and recreational fishery pursuant to the Fish and Game Code and the California Code of Regulations. These data are constantly being updated, and data sets are constantly modified. CDFW may provide data upon request, but, unless otherwise stated, does not endorse any particular analytical methods, interpretations, or conclusions based upon the data it provides.

Landing Categories

Using the CDFW Fish Ticket dataset, we developed a list of specific marine species and multi-species groups that are landed across a recent 14-year period (2011-2024). We excluded some Landing Categories from our analysis: ten freshwater species, three algae, and herring and salmon eggs/roe. From the final list of Landing Categories, we reviewed a variety of sources to determine if the species contained within each Landing Category were likely associated with SFI during any point in their life history.

Metrics of Association with Structure-Forming Invertebrates

For the species contained within each Landing Category, we evaluated three metrics of association with SFI using three commonly applied metrics (Figure 2). These measures - used in previous studies of commercially targeted species within the U.S. West Coast Exclusive Economic Zone (Rooper et al., 2019; Salgado et al., 2018; Yoklavich et al., 2013) - vary in how conservatively they identify associations or co-occurrences:

1.     Adjacent (most conservative):
Observed within one body length of SFI (for fish) or observed touching SFI (for invertebrates). This measure scales with body size.

2.     General Proximity (moderately conservative):
Observed within the general area of SFI (e.g., within 2 m) or found to be associated with SFI using mathematical models.

3.     Habitat & Depth (least conservative):

Species’ habitat preferences and depth ranges suggest likely associations with cold-water SFI.

For Habitat & Depth Associations, each species was scored as:

0 = not associated

1 = probable association

2 = definite association.

Habitat & Depth-based associations (scores of 1 or 2) required that species:

(a) occur at depths >50 m, and

(b) are associated with rocky habitats or rocky ecotones (transitional areas
between habitats).

Several Landing Categories included more than one species. For these multi-species Landing Categories, we assessed associations for each species within the group. Within a Landing Category if all species were associated for a given metric, the entire Landing Category was classified as Associated. Within a Landing Category if association information conflicted or the category was broad (e.g., Rockfish – unspecified; Shrimp – unspecified), we categorized the Landing Category’s association as a multispecies group with mixed associations.

Our objective was to document associations based on our three metrics defined above, and not to evaluate the strength of those associations. Due to the widely varying levels of information about species within Landing Categories, we considered that species were associated if there was at least one documented association. If there were conflicts between observations (e.g., where a species in one study was found within one body length of SFI, but then other studies where they were found more than one body length of SFI) we defaulted to scoring these as ‘Associated’. Species observed as Adjacent to SFI were automatically considered General Proximity associates, since the latter represents a broader definition. Habitat & Depth associations were assessed separately. Species determined to be unlikely to associate with SFI (e.g., those occurring at depths <50 m or with pelagic life histories) were classified as ‘Not Associated’.

We evaluated associations across both juvenile and adult life history phases. Species exhibiting associations during either the juvenile or adult phase were classified as associated. We did not distinguish associations by life stage in our dataset because life stage–specific association data were lacking for many species.

Literature Review

We compiled information from multiple sources to document associations between target species and SFI. We conducted a literature search in Web of Science using the terms: (deep sea coral OR cold water coral OR deep water coral OR structure forming invertebrate) AND (habitat) AND (California OR Washington OR Oregon OR Alaska OR West Coast OR British Columbia) AND (associat OR co-occur). This search yielded 198 publications. We screened abstracts and reviewed relevant papers for evidence of associations involving target species. Using citation chaining and snowball methods, we identified additional references. Next, we consulted with regional experts to identify additional resources including technical memoranda, natural history books, and unpublished reports (e.g., NOAA Cruise Reports). Because peer-reviewed literature on associations was limited, gray literature was an important data source. To fill remaining gaps, we conducted targeted Google Scholar searches of the form: [species name] AND (deep sea coral OR cold water coral OR deep water coral OR structure forming invertebrate). Finally, we incorporated species life history information from online databases such as the CDFW Marine Species Portal and SeaLifeBase. This analysis focused exclusively on hard-bottom SFI, excluding associations with sea pens, which are characteristic of soft-sediment habitats. We focused on data sources from the eastern Pacific, except for a few cases for species with broad distributions (e.g., swordfish). From each reference, we recorded the study location, species, depth range, habitat associations, and type of association. When common and extreme depth ranges were provided (see, for example, FishBase) we included both estimates. Finally, we recorded if the reference used Adjacent, General Proximity, or Depth & Habitat measures of association, if relevant. Data sources for each species are summarized in Table S1, and the complete list of references reviewed appears in Table S2.

Associations by Taxon Group, Species Group, and Gear

To provide a baseline assessment of associations between cold water SFI and target species we summarized how the three standardized association metrics influenced association estimates for fish and invertebrate Landing Categories.  To evaluate if the three metrics differed in their distribution, we used a Pearson’s X² Test between associated and non-associated measures for all Landing Categories, and for Fish and Invertebrate Landing Categories. To evaluate differences in the distribution of associations between Fish and Invertebrate Landing Categories, we used a Pearson’s X² Test for all measures.  For this analysis and summary reporting, we grouped probable and definite habitat-based associations.

To evaluate how these associations varied by taxon (fish vs invertebrates) and species groups, we grouped Landing Categories into species groups (Leonard & Watson, 2011) and summarized the number of Landing Categories for each association metric and score. To evaluate how these associations varied by gear, we summarized gears into gear groups based on Leonard & Watson (2011) and summarized the number of Landing Categories for each association metric and score.