Growth Regimes in the NW Atlantic Following Temperature Regime Shift
Size at Age Analysis of the Northeast US Groundfish Community
Abstract
Physical observations of the region indicate that the NW Atlantic has experienced a regime shift that has accelerated warming in the area. Temperature has a profound impact on the growth, behavior, production, and mortality in ectotherms. For these reasons species are expected to follow their thermal preferences, otherwise risking consequences in the forms of metabolic costs and/or other forms of stress. Using trawl survey data we monitored we tracked the individual responses of 17 species during the ten years prior, and ten years following the regime shift in 2010. Species response to warming was studied in terms of their behavioral shifts via shifting geographic center of biomass,and their biologic response via their size-at-age relationships through time. The majority of species in the southern half of the study area shifted their center of biomass Northward, with no northern species shifting anywhere south of Cape Cod. A concurrent pattern emerged in growth patterns, with the majority of species experiencing faster early-stage growth and smaller adult size. These patterns suggest that even among species that are able to follow thermal preferences, there are additional metabolic costs associated with the shifting environment. Changes in growth rates across such a wide range of species have implications for ecologists and fisheries managers seeking to understand climate impacts on fisheries.
Introduction
Literature on TSR and Growth Expectations
Introduce Study Area and Existing Research
Recent changes in Gulf stream positioning have altered the relative influences of both the Gulf stream and Labrador current on temperature and salinity regimes for the region. Understanding that this new temperature and salinity regime should alter both the energetics of the region, as well as food-web dynamics. We seek to identify whether these changes have had a measurable impact on the growth of several groundfish species.
Despite many of these species living at or near the bottom, research suggests that the forces driving these thermal regimes should have far-reaching impacts biological impacts. These impacts may directly change the near-bottom environment through mixing patterns, and/or indirectly through food-web impacts.
Through the lens of an environmental regime change we will assess changes to body size, size-at-age, and growth characteristics derived using the Von-Bertalanffy growth equation.
Methods
Sea Surface Temperature
Global Sea surface temperature data was obtained via NOAA’s optimally interpolated SST analysis (OISSTv2), providing daily temperature values at a 0.25° latitude x 0.25° longitude resolution (Reynolds et al. 2007). A daily climatology for every 0.25° pixel in the global data set was created using average daily temperatures spanning the period of 1982-2011. Daily anomalies were then computed as the difference between observed temperatures and the daily climatological average. OISSTv2 data used in these analyses were provided by the NOAA PSL, Boulder, Colorado, USA from their website at https://psl.noaa.gov.
Temperature Study Region
Temperature data was regionally averaged to match the sampling area for the fisheries independent survey program providing the age-at-length data. SST anomalies were averaged over the entire sampling region, consisting of continental shelf habitats from Cape Hatteras to Nova Scotia, to produce a daily time series. This time series was then processed into an annual timeseries of surface temperatures and anomalies. All area-averaging was done with area-weighting to account for differences in the relative areas of the latitude/longitude grid of the OISSTv2 data.
Temperature Regime Shift
Annual temperature anomalies were checked for a regime shift using the bayesian change points package (Erdman and Emerson 2007) following the approach of (“A Bayesian Analysis for Change Point Problems: Journal of the American Statistical Association: Vol 88, No 421,” n.d.) implemented in the R statistical programming language.
Warming Trends
We computed the linear trend (\(°C/Decade\)) in the daily Gulf of Maine SST time series using standard linear regression. We considered the two time periods: all years before, and all years following the regime shift. Warming rates and long-term averages were estimated independently for each period.
| Area | rate_1982_2021_C | rate_1982_2021_F | mean_temp_1982_2009_C | mean_anom_1982_2009_C | mean_temp_2010_2021_C | mean_anom_2010_2021_C | mean_temp_1982_2009_F | mean_anom_1982_2009_F | mean_temp_2010_2021_F | mean_anom_2010_2021_F |
|---|---|---|---|---|---|---|---|---|---|---|
| northeast shelf | 0.038 | 0.069 | 11.830 | -0.043 | 12.928 | 1.055 | 53.294 | -0.078 | 55.270 | 1.899 |
| Gulf of Maine | 0.048 | 0.087 | 9.818 | -0.067 | 11.197 | 1.312 | 49.672 | -0.120 | 52.154 | 2.362 |
| Global Oceans | 0.015 | 0.027 | 18.063 | -0.012 | 18.375 | 0.301 | 64.513 | -0.021 | 65.075 | 0.542 |
Data Collection
Fishery Independent data on groundfish size-at-age was collected as part of the NEFSC’s northeast trawl survey. This survey is conducted each year in the spring and in the fall, with sample locations determined following a stratified-random survey design with effort allocated in proportion to stratum area. Add an Explanation of the Strata we Exclude. Correction factors were applied for changes in vessels, gear, and doors when appropriate. Prior to sampling, a CTD cast is performed to collect environmental conditions for the water column at/near the start of the trawl sample providing information on bottom habitat conditions like bottom temperatures.
Distribution Shift Analyses
Analyses focusing on the shifts in distribution among species were limited to data from the spring survey season. Previous work has shown no significant changes in timing of sampling, or the mean annual latitude and longitude of the spring survey sampling (Nye et al. 2009). Our analyses followed the continued movements of 30 species consistent with Nye et al. 2009. These species were originally selected for being consistently sampled across all years, and as representatives of a wide range of taxonomic groups. Our analysis looks at both the annual (combined spring and fall) centers of biomass for latitude, longitude, depth, and bottom temperature distributions.
Calculating Spatial Metrics
Movements through time were characterized for each species using the following metrics: center of biomass, area-occupied, mean depth of occurrence, and the mean temperature of occurrence. Each metric was weighted by the biomass of a species sampled at each station such that.
\[X_j = \frac{ \sum_{}^{} w_iX_{ij} }{ \sum~w_i }\] Where \(X\) is the spatial metric, \(j\) is the year, & \(w\) is the biomass (in kg) for each station \(i\).
Size at Age Analyses
The available age data is a subset of the overall catch data from the NMFS trawl survey. This biological data subset contains additional information on individual fishes such as otolith-derived aging that require additional workup to ascertain. Not all fish species have age available for the full time period, with special attention being given to aid in the management of commercially valuable species. The data used for size-at-age analyses comes from both the Spring and Fall surveys.
Size at age analyses were performed on 14 species. Species were omitted from analyses if age data was not available across both temperature regimes (bluefish & ocean pout) or if their physiology prevented accurate aging (elasmobranch species). The following 14 species had sufficient age data across both temperature regimes to be included in the analysis: american plaice, atlantic cod, atlantic herring, atlantic mackerel, black sea bass, butterfish, haddock, pollock, scup, silver hake, white hake, winter flounder, witch flounder and yellowtail flounder.
Growth by Temperature Regime
To assess the impact of an elevated temperature regime, the size-at-age data was grouped into decadal periods relating to the shift in surface temperature regimes (2000-2009 & 2010-2019). Growth in fishes is commonly modeled using the “Von-Bertalanffy” growth function (VBGF) to capture how a fishes size (length or weight) changes with age. For each species the growth was modeled by fitting the VGBF to the length distribution-at-age.
\[L_t~=~L_{\infty}(1-e^{-K(t-t_0)})\] Where \(L_t\) is the length in cm of a species at age \(t\). \(L_\infty\) is the asymptotic maximum length, \(K\) is _____, and \(t_0\) is x intercept, or the hypothetical age at 0 cm length.
Growth Impacts on Productivity
Yield-per-recruit analysis similar to Baudron et al. 2014?
Results
Temperature Regime Shift
SST anomaly time series of the region indicate a jump in annual temperature anomalies indicative of a regime shift centered around 2010. A bayesian change point analysis showed highest support for a regime shift among 3 the years: 2009-2011. Change point probabilities for these years were 31.66, 32.01, 36.17% respectively.
Warming Trends
| Period | Decadal Rate |
|---|---|
| 1982-2019 | 0.10 |
| 2010-2019 | -0.07 |
Temperatures prior to the 2010 regime shift rose at a rate of 0.7 \(°C/Decade\). Regional temperatures then peaked in 2012 and have remained high while decreasing slowly at a rate of 0.02 \(°C/Decade\). Average temperatures following the regime shift (2011-2021), were on average 1.09°C warmer than the than the average over the previous 28 years (1982-2009). Temperatures within each regime were relatively stable, indicative of a jump in mean temperatures rather than a rate change between them. While temperatures in the current regime are falling, the rate is such that
Distribution Shifts
Center of Biomass Trends
Taking the annual weighted center of biomass locations from each species we can see that there is a prevailing direction to the majority of these shifts.
Changes to Growth
Changes in the size-at-age growth relationship varied across species, with the majority (14/18) of the species exhibiting smaller L-infinity during the warmer temperature regime.
The species that showed larger asymptotic lengths include the three hake species: red hake, silver hake, & white hake as well as Atlantic herring. These species are all among the majority group of species that is found primarily in the Northern part of the survey area, in and around the Gulf of Maine.
Summary Tables of Growth Change
?(caption)
| Changes in L-infinity | ||
| Common Name | Change in Linf (cm) | Percent Change |
|---|---|---|
| Body Size Decrease | ||
| scup | -25.96 | -38.00 |
| american plaice | -12.98 | -22.63 |
| witch flounder | -5.81 | -11.22 |
| atlantic cod | -12.46 | -10.15 |
| winter flounder | -4.99 | -9.57 |
| yellowtail flounder | -4.22 | -8.84 |
| haddock | -4.48 | -6.98 |
| butterfish | -1.50 | -5.71 |
| pollock | -5.38 | -4.18 |
| black sea bass | -1.30 | -1.65 |
| Body Size Increase | ||
| atlantic mackerel | 0.17 | 0.37 |
| atlantic herring | 0.44 | 1.50 |
| white hake | 22.16 | 13.44 |
| silver hake | 9.73 | 19.51 |
?(caption)
| Changes in Condition Factor (Fulton's K) | |||
| Common Name | Early Regime | Warm Regime | Change in Condition (K) |
|---|---|---|---|
| Condition Stable | |||
| black sea bass | 1.33 | 1.33 | 0.00 |
| pollock | 1.02 | 1.02 | 0.00 |
| haddock | 0.96 | 0.96 | 0.00 |
| Condition Decrease | |||
| winter flounder | 1.20 | 1.19 | -0.01 |
| yellowtail flounder | 0.84 | 0.82 | -0.02 |
| Condition Increase | |||
| scup | 2.00 | 2.04 | 0.05 |
| butterfish | 1.83 | 1.87 | 0.03 |
| atlantic herring | 0.84 | 1.01 | 0.17 |
| atlantic cod | 0.95 | 0.97 | 0.01 |
| atlantic mackerel | 0.85 | 0.88 | 0.03 |
| american plaice | 0.73 | 0.78 | 0.05 |
| white hake | 0.74 | 0.75 | 0.01 |
| silver hake | 0.62 | 0.67 | 0.05 |
| witch flounder | 0.56 | 0.57 | 0.01 |
Growth with Changes in The Environment
Temperature
Depth
VB Fits to Data:
american plaice
atlantic cod
atlantic herring
atlantic mackerel
black sea bass
butterfish
haddock
pollock
scup
silver hake
white hake
winter flounder
witch flounder
yellowtail flounder
Discussion
Bonus: Data Explorations
Age Data Availability
The availability of age data varied by species and through time, with older ages becoming less frequently sampled in recent years across many species.
Strong age classes were particularly prominent in the following species, showing lasting numbers in subsequent years.
| Species | Strong Year Classes |
|---|---|
| Atlantic Cod | 2003, 2013 |
| Haddock | 2003, 2010, 2013 |
| Atlantic Herring | 2008, 2011 |
| Pollock | 2000, 2012 |
Growth Increment Changes
Annual Growth increments for each species were computed to track the change in growth of specific age-groups through time.
Age specific growth (annual growth increments) help detail where in a species life cycle it may be experiencing favorable or less-favorable conditions for growth. These can be a result of differences in habitat use or prey resources at different ages, and/or a change in exposure to temperature or other stressors that inflict a metabolic cost.
Temperature and Age Increments
I’m also trying to think about how to look at lagged temperature effects on size, with a particular interest in MHWs. A starting point might be to look at cross-correlation (ccf) between temperature and the age increments. That would enable us to see the time period over which potential lagged effects operate.
This will produce lots of plots (one for each age increment), so let’s start with just a few species…maybe plaice, cod, silver hake, and summer flounder based on some differences I see in the VB curves.
Parameters & Environment
The following figures plot the change in Linf and K against the center of biomass values of depth, temp, lat, and lon.
Retired Assets:
This section exists to isolate code/analyses that were removed
Length at Age by Regime
Weight at Age by Regime
Shifts in Growth
Growth Change Horizons
References
“A Bayesian Analysis for Change Point Problems: Journal of the American Statistical Association: Vol 88, No 421.” n.d. https://www.tandfonline.com/doi/abs/10.1080/01621459.1993.10594323.
Erdman, Chandra, and John Emerson. 2007. “Bcp: An r Package for Performing a Bayesian Analysis of Change Point Problems.” Journal of Statistical Software 23 (December). https://doi.org/10.18637/jss.v023.i03.
Nye, Ja, Js Link, Ja Hare, and Wj Overholtz. 2009. “Changing Spatial Distribution of Fish Stocks in Relation to Climate and Population Size on the Northeast United States Continental Shelf.” Marine Ecology Progress Series 393 (October): 111–29. https://doi.org/10.3354/meps08220.
Reynolds, Richard W., Thomas M. Smith, Chunying Liu, Dudley B. Chelton, Kenneth S. Casey, and Michael G. Schlax. 2007. “Daily High-Resolution-Blended Analyses for Sea Surface Temperature.” Journal of Climate 20 (22): 5473–96. https://doi.org/10.1175/2007JCLI1824.1.