Electronic Supplementary Material for:
“Potential consequences of climate change for primary production and fish production in large marine ecosystems”
Julia L. Blanchard1,2*, Simon Jennings3,4, Robert Holmes 5, James Harle6, Gorka Merino 5, Icarus Allen5, Jason Holt6 , Nicholas K. Dulvy7 & Manuel Barange5
A. Model Equations and Parameters
Table S1. Equations for the dynamic coupled size-spectrum model. Subscripts: i {P = pelagic predators, B = benthic detritivores} and D=detritus
Equations / UnitsDynamical system:
/ m-3 yr-1g-1
/ m-3 yr-1g-1
Temperature effect:
Feeding rates result from allometric search rates, prey size preference and availability across the size spectra:
/ g yr-1
give growth rates:
/ g yr-1
Flux terms from death included:
Predation mortality
Intrinsinc and senescence mortality
/ yr-1
yr-1
Fishing mortality
Total death rates were:
Fishing mortality was only applied to the pelagic community where F was equal across all size > 1.25 g and the value of F was varied. / yr-1
yr-1
Table S2. Parameter definitions, values and units for the dynamic coupled size spectrum model. In cases where two values are given the first value is for pelagic predators and the second value is for benthic detritivores.
mmin / minimum body mass of plankton / 10-12 / g
mP,min / minimum body mass of pelagic predators (also of max body size of plankton) / 10-3 / g
mB,min / minimum body mass of benthic detritivores / 10-3.5 / g
mmax / maximum body size in the whole system. / 106 / g
k / Boltzmann constant / 8.62E-5 / eV K-1
E / activation energy / 0.63 / eV
c1 / constant / 25.55
w / preference for prey in either the pelagic or benthic spectrum / 0.5
β / preferred predator-prey mass ratio log10(PPMR) / 2.0
σ / measure of the width of the log10 PPMR distribution / 1.0
A / pre-factor of search volume / 64, 6.4 / m3 yr-1
a / exponent of search volume / 0.82, 0.75
K / net growth conversion efficiency / 0.2, 0.1
z0 / pre-factor for background mortality / 0.1 / yr-1
zi / exponent for intrinsic background mortality / -0.25 / g yr-1
ms / start size for senescence mortality / 1 / kg
zs / exponent for senescence background mortality / 0.3 / g yr-1
B. Model application and output
The plankton size spectrum is not modelled explicitly but has temporal dynamics that are driven by output from a coupled physical-biogeochemical model, POLCOMS-ERSEM. Temporal changes in the intercept of the plankton component of size-spectrum were estimated from phytoplankton and microzooplankton biomass density. The plankton size spectrum was described as log N(m) = a + blog(m) where N is abundance density per unit volume at mass, and mass is m (in grams). We estimate a from the total biomass density of phytoplankton and microzooplankton functional groups from the physical-biogeochemical model output by spreading the density across a realistic size range (10-14 to 10-4 g) and assuming a fixed slope b of -1, in keeping with global empirical studies of phytoplankton size spectra [38]. Near sea floor detritus biomass density estimates from the physical-biogeochemical model were used to force the detritus dynamics. Temperature estimates were taken form the mixed layer depth and near sea floor vertical layers for the pelagic predators and benthic detritivores, respectively.
The total biomass density (g m-3 ) across a size-range (m1=1 g to m2= 100 kg) was computed from the predicted numerical density at size N(m) at time across the continuum of body masses m as:
(1)
Similarly, the total fish production (g m-3 yr-1) was calculated from the growth rates at size G(m) and the numerical density at size:
(2)
The total annual catch (or ‘yield’, in g m-3 yr-1) across a size-range was calculated by combining (1) with the fishing mortality rate at size :
(3)
C. Details of the large regional marine ecosystem domains
Table S3. The large regional domains of coastal and shelf seas included in the analysis and the corresponding Large Marine Ecosystems (LME) that make up the domains, with information on the primary production, surface area and 2005 fish catches. These 28 LMEs contributed approximately 77% of the landed catch from all LMEs combined around the world, and 60% of the global marine landed catches.
PP / Area / Fish Catch 2005Domain / LME / (mgC.m-2.d-1) / (103.km2) / (t)
Bengal / Bay of Bengal / 729 / 3,657 / 3,818,679
Benguela / Benguela Current / 1,387 / 1,470 / 1,095,408
Canary / Canary Current / 1,196 / 1,120 / 2,077,314
Guinea / Guinea Current / 980 / 1,959 / 859,111
Humboldt / Humboldt Current / 876 / 2,619 / 9,855,464
Indo-Pacific / Yellow Sea / 1,613 / 439 / 2,063,739
Indo-Pacific / South China Sea / 477 / 5,662,985 / 6,606,620
Indo-Pacific / Sulu-Celebes Sea / 573 / 1,016 / 1,111,075
Indo-Pacific / Indonesian Sea / 772 / 2,289,597 / 1,609,263
Indo-Pacific / Gulf of Thailand / 780 / 391,665 / 862,066
NW Pacific / East China Sea / 891 / 1,008 / 3,329,876
NW Pacific / Kuroshio Current / 422 / 1,333,074 / 612,605
NW Pacific / Okhotsk Sea / 815 / 1,627 / 2,662,794
NW Pacific / Oyashio Region / 716 / 664 / 584,048
NW Pacific / Sea of Japan/ East Sea / 604 / 1,054 / 1,166,937
NE Atlantic / North Sea / 1,115 / 690,041 / 1,885,726
NE Atlantic / Celtic Biscay Shelf / 956 / 766,550 / 1,086,603
NE Atlantic / Faroe Plateau / 422 / 151 / 316,817
NE Atlantic / Iberian Coastal / 758 / 301 / 299,644
NW Atlantic / Newfoundland- Labrador / 809 / 675 / 354,768
NW Atlantic / NE US Continental Shelf / 1,536 / 308 / 741,834
NW Atlantic / Scotian Shelf / 1,395 / 413 / 351,017
Nordic / Norwegian Shelf / 491 / 1,109 / 1,341,406
Nordic / Icelandic Shelf / 551 / 521,237 / 1,312,248
Nordic / Greenland Sea / 477 / 1,176 / 127,504
Nordic / Baltic Sea / 1,910 / 397 / 702,404
California / California Current / 613 / 2,225 / 728,988
California / Gulf of California / 1,199 / 216 / 195,308
D. Model cross-validation
Figure S1. Comparison of predicted percentage changes (climate relative to control) in overall fish biomass density from the static metabolic scaling model and the dynamic coupled size spectrum model. Each point represents an EEZs group by colour according the large ecosystem domains (codes in legend). Solid line is 1:1 relationship.
E. Effect of fishing on fish production and resilience to climate change across ecosystems.
Figure S2. Climate-projected changes in (a) large fish production and (b) Coefficients of Variation (CVs) of large fish biomass density time series. Changes are relative to the unexploited control scenario when fishing mortality rates are = 0 (dark grey, climate only), 0.2 (light grey, low fishing) and 0.8 yr-1 (white, high fishing) for all economic zones grouped by their 11 regional ecosystems domains.