Skip to main content
Dryad

Northeast Atlantic species distribution shifts over the last two decades

Cite this dataset

Le Luherne, Emilie (2024). Northeast Atlantic species distribution shifts over the last two decades [Dataset]. Dryad. https://doi.org/10.5061/dryad.cc2fqz6fd

Abstract

Marine species are widely shifting their distributions in response to global changes, and it is commonly expected they will move northward and to greater depths to reach cooler, less disturbed habitats. However, local manifestations of global changes, anthropogenic pressures, and species characteristics may lead to unanticipated and varied responses by individual species. In this regard, the Celtic-Biscay Shelf is a particularly interesting study system because it has historically been heavily fished and occurs at the interface between two distinct biogeographic provinces, its community is thus comprised of species with diverse thermal preferenda. In the context of rapidly warming temperatures and intense fishery exploitation, we investigated the distribution shifts of 93 taxa (65 Actinopteri, 10 Elasmobranchii, 11 Cephalopoda, 5 Malacostraca, and 2 Bivalvia), which were sampled annually from 1997 to 2020 during a scientific bottom trawl survey. We used a set of 11 complementary spatial indices to quantify taxon distribution shifts over time. Then, we explored the relative effect of taxon abundance, fishing pressure, and climatic conditions on taxon’s distribution shift when a significant shift was detected. We observed that 56% of the taxa significantly shifted. Not all taxa will necessarily shift northward and to deeper areas, as is often expected. Two opposite patterns were identified: taxa either moving deeper and to the southeast, or moving closer to the surface and to the northwest. The main explanatory factors were climate change (short- and long-term temperatures) and taxon abundance. Fishing pressure was the third, but still significant, explanatory factor of taxa of greater commercial importance. Our research highlights that taxa are displaying complex distribution shifts in response to the combined anthropogenic disturbances and underscores the need to conduct regional studies to better understand these responses at the ecosystem scale to develop more suitable management plans and policies.

README: Northeast Atlantic species distribution shifts over the last two decades

https://doi.org/10.5061/dryad.cc2fqz6fd

Raw data of this study comes from the scientific bottom trawl survey EVHOE (Evaluation of fishing resources of Western Europe; EVHOE cruise, R/V Thalassa, IFREMER, DOI: 10.18142/8) carried out between October and December from 1997 to 2020 along the Celtic-Biscay Shelf (CBS). The latter occurs at the interface between the Celtic Sea (CS) and the Bay of Biscay (BoB). EVHOE relies on a stratified random sampling design in which 119 to 158 stations were sampled annually. The EVHOE survey is funded through the European Union Data Collection Framework.

At each station, specimens were collected using a 36/47 GOV bottom trawl (opening width: 20 m; height: 4 m; mesh size: 20 mm) operated for 30 min at a towing speed of 4 knots. Once on board, species were identified to the finest possible taxonomic scale. Due to variation in taxonomic resolution, some species were assigned to a coarser taxonomic level, hereafter named “taxon” and “taxa”. We excluded the 2017 dataset, as an incident of boat engine failure prevented completion of the sampling campaign. To reduce potential bias due to the sampling method, we only retained the Taxon with a reliable temporal and spatial coverage. We selected Taxon collected during at least 80% of the studied years and at 5% or more of the sampling stations within a given year. This selection was adapted for the analysis of temporal shifts in taxon distribution but not for the exploration of invasive species income.

The dataset contains tables summarizing the Taxon characteristics, Taxon populations annual spatial indices, explanatory variables, and models results of Taxon distribution shift analysis and the explanatory variables influencing distribution shifts.

Please read the paper or contact the authors for more details.

Description of the data and file structure

  • 1_Taxa_population_guilds
Column Explanation
taxa_code EVHOE survey Taxon code
common_name Taxon common name
name_latin Taxon latin name
distri Taxon population distribution according to presence-absence data
pop_nb Taxon population number identified in the area according to presence-absence data
class Taxonomic class
position_guild Taxon vertical distribution guild (pelagic, demersal, benthic swimmer, or benthic)
biogeographic_guild Taxon biogeographic guild (Atlantic, boreal, or Lusitanian)
rel_com_imp Taxon relative commercial importance (lesser or greater commercial importance)
lat_min_biblio Minimum latitude recorded for this Taxon in the literature
lat_max_biblio Maximum latitude recorded for this Taxon in the literature
ref_range_lat_data Reference for the minimum and maximum latitudes

distribution_final : CBS = 1 population distributed in the CBS area, CS = 1 population distributed in the CS area, BoB = 1 population distributed in the BoB area,
CS_BoB = 2 populations (1 in CS and 1 in BoB)

Based on information from the scientific literature, FishBase, and ICES, each taxon was assigned to a biogeographic guild (Atlantic, boreal, or Lusitanian) and a vertical distribution guild (pelagic, demersal, benthic swimmer, or benthic). Two benthic guilds were defined according to Taxon swimming capacity: the benthic swimmer Taxon were Actinopteri, Elasmobranchii, and Cephalopoda and the benthic Taxon were Malacostraca and Bivalvia.

We calculated each taxon’s relative commercial importance using ICES mean catches in FAO divisions 27.7 g-h and 27.8 a-b, corresponding to the Celtic Sea and the Bay of Biscay delineated in our study, respectively, between 2006 and 2019 (source: Eurostat/ICES compilation of catch data and statistics, ICES 2021, Copenhagen). On the basis of this information, Taxon were considered to be of lesser commercial importance (mean catch for taxon < 0.25 quantile of the distribution of mean catches for all taxa) or of greater commercial importance (mean catch for taxon ≥ 0.25 quantile of the distribution of mean catches for all taxa).

  • 2_Spatial_index_abun_biomass Annual spatial index calculated for each Taxon population.
Column Explanation
year year
taxa_code EVHOE survey Taxon code
distri Taxon population distribution according to presence-absence data
sub_pop Taxon population when two populations were identified in the studied area
long_CG Taxon population longitude center of gravity
long_q5 Taxon population longitude quantile 0.05 = western limit
long_q95 Taxon population longitude quantile 0.95 = eastern limit
lat_CG Taxon population latitude center of gravity
lat_q5 Taxon population latitude quantile 0.05 = southern limit
lat_q95 Taxon population latitude quantile 0.95 = northern limit
depth_CG Taxon population depth center of gravity
depth_q5 Taxon population depth quantile 0.05 = lower limit
depth_q95 Taxon population depth quantile 0.95 = upper limit
I Taxon population inertia
PA Taxon population positive area
abun_IA Taxon population relative abundance index
biomass_tons_IA Taxon population biomass index

A relative abundance index was used as a descriptor of population status (Hutchings & Baum, 2005). To account for the randomly stratified sampling design, we calculated taxon relative abundance index as the number of individuals of a given taxon elevated to the surface area (in square nautical miles) of the CS , BoB or CBS depending on the delineation of the populations within the distribution zones (Mahé & Poulard, 2005).

The biomass index estimated using EVHOE data is used to calculate a harvest rate and to quantofy the influence of fishing pressure on Taxon distribution shifts.

  • 3_Model_results_temporal_change_analysis Results of GLM and GLS models for the analysis of Taxon significant distribution shift during the studied period.
Column Explanation
taxa_code EVHOE survey Taxon code
distri Taxon population distribution according to presence-absence data
sub_pop Taxon population when two populations were identified in the studied area
"index"_p P value of the statistical analysis
"index"_slope Slope estimated by the model
"index"_test Model used for the analysis of temporal changes of the index (GLM or GLS)

"index" are either lat, long, depth (CG, q5, and q95), I or PA.

  • 4_Significant_temporal_change_interpreted Significant results of GLM and GLS models for the analysis of Taxon distribution shift during the studied period and their shifting direction.
Column Explanation
taxa_code EVHOE survey Taxon code
distri Taxon population distribution according to presence-absence data
sub_pop Taxon population when two populations were identified in the studied area
spatial index Spatial index
p_value Significant p value (alpha ≤ 0.05)
slope Slope estimated by the model
test Model used for the analysis of temporal changes of the index (GLM or GLS)
moving Shifting direction
  • 5_Climate_explanatory_variables
Column Explanation
year Year
annual_AMO Annual Atlantic Multidecadal Oscillation index
winter_NAO Winter North Atlantic Oscillation (December of the previous year to March of the focal year)
BT_ano_CS Bottom temperature anomalies in the Celtic Sea
BT_ano_BoB Bottom temperature anomalies in the Bay of Biscay
BT_ano_CBS Bottom temperature anomalies in the Celtic-Biscay Shelf
  • 6_Taxon_population_relative_F
Column Explanation
taxa_code EVHOE survey Taxon code
distri Taxon population distribution according to presence-absence data
sub_pop Taxon population when two populations were identified in the studied area
year Year
biomass_tons_IA Taxon population biomass index
catches_fao Catches reported by the FAO
relative_F Fishing pressure index

To quantify fishing pressure, we estimated a harvest rate as the ratio between nominal catches (sources: Eurostat/ICES compilation of catch data and statistics - ICES 2011 & 2021, Copenhagen) and the biomass estimated during EVHOE (within the CS, BoB, and CBS). We decided against using fishing mortality (e.g., ICES analytical assessment) as a proxy for fishing pressure because both analytical stock assessments are lacking for most of the studied taxa and EVHOE’s survey area did not entirely match up with ICES fish stock delineations.

  • 7_Relative_importance_of_explanatory_variables

The relative importance (i.e., the cumulative AICc weight) of each explanatory variable was calculated by summing Akaike weights across all models that included the explanatory variable (Table S2; Burnham & Anderson, 2002; Arnold, 2010).

Column Explanation
taxa_code EVHOE survey Taxon code
distri Taxon population distribution according to presence-absence data
sub_pop Taxon population when two populations were identified in the studied area
spatial index Spatial index
moving Shifting direction
abun_IA Relative importance of abundance
relative_F Relative importance of fishing pressure
annual_AMO Relative importance of annual AMO
winter_NAO Relative importance of winter NAO
BT_ano Relative importance of bottom temperature anomalies
  • 8_Model_average_estimates_of_explanatory_variables

The variables’ estimates and 80% confidence intervals were obtained by computing the average model using all the models (Arnold, 2010).

Column Explanation
taxa_code EVHOE survey Taxon code
distri Taxon population distribution according to presence-absence data
sub_pop Taxon population when two populations were identified in the studied area
spatial index Spatial index
moving Shifting direction
abun_IA Abundance estimate
relative_F Fishing pressure estimate
annual_AMO Annual AMO estimate
winter_NAO Winter NAO estimate
BT_ano Bottom temperature anomalies estimate

Sharing/Access information

Data was derived from the following sources:

  • EVHOE survey data are available at https://www.seanoe.org/data/00688/80041/
  • NAO data are available at https://www.ncdc.noaa.gov/teleconnections/nao/
  • AMO data are available at https://psl.noaa.gov/data/timeseries/AMO/
  • The monthly aggregated bottom temperatures were provided by the Atlantic Margin Model FOAM (Forecasting Ocean Assimilation Model). The data had a horizontal resolution of 7 km (0.111 × 0.067°) and covered the period from 1997 to 2020 (UK Met Office Operational Suite, http://marine.copernicus.eu/).
  • Taxon’s relative commercial importance was estimated using ICES mean catches in FAO divisions 27.7 g-h and 27.8 a-b, corresponding to the Celtic Sea and the Bay of Biscay delineated in our study, respectively, between 2006 and 2019 (source: Eurostat/ICES compilation of catch data and statistics, ICES 2021, Copenhagen).
  • Fishing pressure was estimated with a harvest rate as the ratio between nominal catches (sources: Eurostat/ICES compilation of catch data and statistics - ICES 2011 & 2021, Copenhagen) and the biomass estimated during EVHOE (within the CS, BoB, and CBS).

Methods

This dataset was collected during the scientific bottom trawl survey EVHOE (Evaluation of fishing resources of Western Europe; EVHOE cruise, R/V Thalassa, IFREMER, DOI: 10.18142/8) carried out between October and December from 1997 to 2020 along the Celtic-Biscay Shelf.

Funding

Ifremer

European Union Data Collection Framework