Methodology:

Occurrence records and environmental data were downloaded from online and open access sources (see section Reference). After identifying overlapping datasets between repositories, remaining occurrence records were thinned to retain only one occurrence per season per grid cell (resolution: 0.25° x 0.25°) using the "spThin" R package (Aiello-Lammens et al. 2015). This removes the fewest records necessary to substantially reduce the effects of sampling bias, while simultaneously retaining the greatest amount of useful information. This left a total of 27,727 occurrence records and this cleaned dataset is version provided here.

Occurrence and background points were matched to the conditions from the season and era from which they were collected. This was done using an adapted version of the "depth_extract" R function available from: https://figshare.com/articles/Duffy_Chown_MEPS12011_model_virtualspecies/4315052. The adapted version we use is available in the R Markdown file we provide.

The steps that followed included optimising Maxent model settings, using the final model to predict habitat suitability across the study region, and running several sensitivity analyses. These processes are documented in the R Markdown file provided.

- Folder: env_dat_rasters, Volume: 193 megabytes

Contents: This folder contains the gridded environmental data. Each subfolder contains the raster files (ascii format) for a particular time period (era) and season. E.g. subfolder "era1_amj" contains the raster data during April-May-June of era 1 (1955-1984).

Data files and units:

bathy.asc = bathymetry (metres)

chla.asc = sea surface chlorophyll a concentration (mg.m-3)

par.asc = sea surface photosynthetically active radiation (einstein.m-2.day-1)

ph.asc = sea surface ph (total scale)

salinity.asc = sea surface salinity (psu)

seaicecon.asc = sea-ice concentration (%)

silicate.asc = sea surface molar concentration of silicate (µmol.kg-1)

slope.asc = bathymetric slope (degrees)

temp.asc = sea surface temperature (degrees celsius)

velocity.asc = sea surface current velocity (m.s-1)

- File: occurrence_dat_cleaned.csv, Volume: 4.9 megabytes

Contents: This file contains the cleaned and processed occurrence records of Calanus finmarchicus and the season-and-era-specific environmental conditions assigned to each presence locality.

Data columns and units:

NAME = identifier for species "Calanus_finmarchicus"

LON = longitude (decimal degrees)

LAT = latitude (decimal degrees)

MONTH_COLLECTED = month of sample collection (1=Jan, 2=Feb etc.)

SEASON = season assigned to each point (1=Jan-Feb-Mar, 2=Apr-May-Jun, 3=Jul-Aug-Sep, 4=Oct-Nov-Dec)

MINIMUM_DEPTH = minimum vertical depth of sample (metres from sea surface)

MAXIMUM_DEPTH = maximum vertical depth of sample (metres from sea surface)

MIN_MAX_DEPTH_RANGE = vertical range of sample (metres)

YEAR_COLLECTED = year of sample collection

TIME_PERIOD = era assigned to each point (1=1955-1984, 2=1985-2017)

LIFESTAGE = copepodite life stage assigned to the record at collection (CI-Adult)

DATASET = dataset/cruise from which sample was collected (see file "occurrence_dat_citations.txt" for full list)

DATABASE = database from which sample was archived (see file "occurrence_dat_citations.txt" for full list)

BATHY = bathymetry (metres)

CHLA = sea surface chlorophyll a concentration (mg.m-3)

PAR = sea surface photosynthetically active radiation (einstein.m-2.day-1)

PH = sea surface pH (total scale)

SALINITY = sea surface salinity (psu)

SEAICECON = Sea-ice concentration (%)

SILICATE = sea surface molar concentration of silicate (µmol.kg-1)

SLOPE = Bathymetric slope (degrees)

TEMP = sea surface temperature (degrees Celsius)

VELOCITY = sea surface current velocity (m.s-1)

Abbreviations:

CI - CVI = lifestage Copepodite I - VI, w...(42)

Data collection:

All analyses were carried out in R version 3.4.1.

Data quality:

Occurrence records represent those remaining after undertaking quality assurance and control processes as detailed in section 6. Further actions to limit the potential for sampling bias issues were carried out and are included in the R Markdown file we provide.

Due to limited data availability, seasonal environmental data during the older time period (1955-1984) are limited to: bathymetry, slope, temperature, salinity, sea-ice concentration and current velocity.

Sensitivity analyses were carried out to (a) account for potential misidentification of occurrences between Calanus finmarchicus and its congener, C. glacialis, (b) account for potential bias stemming from the seasonal assignment process of occurrences to environmental, and (c) test the sensitivity of model outputs to the inclusion of sea-ice concentration. Code for these analyses is within the R Markdown file we provide.