Widespread phenological shifts with temperature in Alaska’s marine fishes
Data files
Jan 22, 2026 version files 59.38 MB
-
EBS_spring_d98to162_11spp_Apr2024_larval.csv
13.44 MB
-
GOA_spring_d117to160_32spp_Jul2024_larval.csv
45.93 MB
-
README.md
4.83 KB
Abstract
Changes in the timing of fish spawning and early life stage development can affect the temporal match or mismatch of larvae with production of preferred prey as well as their availability to predators, with potential consequences for recruitment success, food-web dynamics, and fisheries. Using >370,000 observations from over four decades of spring ichthyoplankton surveys in the Gulf of Alaska and Bering Sea, we investigated long-term changes in the phenology of 29 fish species, including commercially important taxa such as Pacific cod, walleye pollock, and Pacific halibut. Larval size on a standardized date (size-at-date) was used as a proxy for larval developmental timing in spring, and reflects a combination of hatch timing (larval age), growth, and mortality. Spatiotemporal generalized linear mixed models were used to account for variable sampling effort in space and time in order to isolate long-term trends and thermal effects on larval size. For a majority of species, interannual variation in mean size-at-date was significantly and positively related to temperature, demonstrating widespread thermal effects on the phenology of fish early life stages. Despite the wide diversity of life history traits exhibited by the 29 species examined, patterns in size-at-date over time were similar across most species within each ecosystem, reflecting the common effect of temperature on phenology. While temperature affected size-at-date, there was little evidence of long-term trends, likely due to the lack of a linear trend in winter-spring temperatures observed in recent decades. We demonstrate a novel analytical method to assess changes in phenology from larval size observations sampled at variable locations and times, and detect phenological shifts that were not necessarily identifiable from larval abundance data alone. Our results suggest that earlier spring phenology due to warming will be a common response among fishes to projected future climate change in high-latitude ecosystems.
Dataset DOI: 10.5061/dryad.2280gb66b
Description of the data and file structure
Survey Data
Ichthyoplankton surveys in Alaska’s waters have been conducted by the National Oceanic and Atmospheric Administration’s Alaska Fisheries Science Center (NOAA AFSC) and partners since the 1970s, becoming more frequent in the 1980s (GOA) and 1990s (EBS). At each survey station, paired 60-cm bongo nets were towed obliquely from a depth of at least 100 m, or 10 m off the bottom in shallower waters. Net mesh sizes were either 333-µm or 505-µm. For each tow, the contents of one net were preserved in 5% formalin and later sorted and identified to the lowest taxonomic level possible at the Plankton Sorting and Identification Center in Szczecin, Poland. Up to 50 larvae of each taxa from each tow were measured for standard length after preservation. Species identifications were verified by expert taxonomists at NOAA AFSC. Survey coverage and timing has varied over the years due to differing survey objectives and ship time allocations, although many earlier surveys were specifically targeted for studying pollock eggs and larvae. For this analysis, we selected study areas that represented the most consistently sampled areas within each region and used only samples collected within the most consistently sampled temporal window (days 98-162 in the EBS and 117-160 in the GOA). This spring (April-June) sampling window captures a high diversity and abundance of larval fishes, reflecting common patterns of winter-spring spawning in these systems.
Species Selection
Only taxa identified to the species level were included for analysis. From those, we chose species that were present in at least 10 different years, with at least 10 observations in each of those years.
Files and variables
File: EBS_spring_d98to162_11spp_Apr2024_larval.csv
Description: Larval fish catch data and length measurements from NOAA EcoFOCI and partner surveys in the Bering Sea, collected between day of year 98 and 162. Data queried in April 2024.
Variables
- SPECIES_NAME: Scientific name of specimen.
- HAUL_ID: Unique identifier for haul (bongo net sampling event from a particular station and survey).
- CRUISE: Survey identifier.
- HAUL_NAME: Sequential haul number within a station.
- NET: 1 or 2, indicating which of two sampling nets was preserved and used for analysis.
- MESH: Mesh size of sampling net in microns.
- LAT: Latitude, degrees.
- LON: Longitude, degrees.
- GEOGRAPHIC_AREA: Region of sampling, either GOA (Gulf of Alaska) or BS (Bering Sea).
- YEAR: Year of collection.
- GMT_DATE_TIME_TXT: Date and time of sampling.
- GMT_DATE_TIME: Date of sampling.
- CORRECTED_LENGTH: Standard length of formalin-preserved larvae (mm).
- LARVALCATCHPER10M2: Calculated catch of larvae of this species per 10 m2 from this haul.
- LARVALCATCHPER1000M3: Calculated catch of larvae of this species per 1000 m3 from this haul.
- NUMBER_MEASURED_COUNTED: Number of larvae of this species measured for length at this station (not necessarily equal to total number caught).
- COMMON_NAME: Common name of species.
- YDAY: Day of year of sampling.
- nyrs: Total number of years this species is present in dataset (by region).
File: GOA_spring_d117to160_32spp_Jul2024_larval.csv
Description: Larval fish catch data and length measurements from NOAA EcoFOCI and partner surveys in the Gulf of Alaska, collected between day of year 117 and 160. Data queried in July 2024.
Variables
- SPECIES_NAME: Scientific name of specimen.
- HAUL_ID: Unique identifier for haul (bongo net sampling event from a particular station and survey).
- CRUISE: Survey identifier.
- HAUL_NAME: Sequential haul number within a station.
- NET: 1 or 2, indicating which of two sampling nets was preserved and used for analysis.
- MESH: Mesh size of sampling net in microns.
- LAT: Latitude, degrees.
- LON: Longitude, degrees.
- GEOGRAPHIC_AREA: Region of sampling, either GOA (Gulf of Alaska) or BS (Bering Sea).
- YEAR: Year of collection.
- GMT_DATE_TIME_TXT: Date and time of sampling.
- GMT_DATE_TIME: Date of sampling.
- CORRECTED_LENGTH: Standard length of formalin-preserved larvae (mm).
- LARVALCATCHPER10M2: Calculated catch of larvae of this species per 10 m2 from this haul.
- LARVALCATCHPER1000M3: Calculated catch of larvae of this species per 1000 m3 from this haul.
- NUMBER_MEASURED_COUNTED: Number of larvae of this species measured for length at this station (not necessarily equal to total number caught).
- COMMON_NAME: Common name of species.
- YDAY: Day of year of sampling.
- nyrs: Total number of years this species is present in dataset (by region).
Survey Data
Ichthyoplankton surveys in Alaska’s waters have been conducted by the National Oceanic and Atmospheric Administration’s Alaska Fisheries Science Center (NOAA AFSC) and partners since the 1970s, becoming more frequent in the 1980s (GOA) and 1990s (EBS). At each survey station, paired 60-cm bongo nets were towed obliquely from a depth of at least 100 m, or 10 m off the bottom in shallower waters. Net mesh sizes were either 333-µm or 505-µm. For each tow, the contents of one net were preserved in 5% formalin and later sorted and identified to the lowest taxonomic level possible at the Plankton Sorting and Identification Center in Szczecin, Poland. Up to 50 larvae of each taxa from each tow were measured for standard length after preservation. Species identifications were verified by expert taxonomists at NOAA AFSC. Survey coverage and timing has varied over the years due to differing survey objectives and ship time allocations, although many earlier surveys were specifically targeted for studying pollock eggs and larvae. For this analysis, we selected study areas that represented the most consistently sampled areas within each region and used only samples collected within the most consistently sampled temporal window (days 98-162 in the EBS and 117-160 in the GOA). This spring (April-June) sampling window captures a high diversity and abundance of larval fishes, reflecting common patterns of winter-spring spawning in these systems.
Species Selection
Only taxa identified to the species level were included for analysis. From those, we chose species that were present in at least 10 different years, with at least 10 observations in each of those years.