Research database

Project information
Keywords
pelagic ecosystem sensitivity feedbacks Arctic Ocean acidification
Project title
ECOAN WP3-OA9: Investigate pelagic ecosystem sensitivity and feedbacks to Arctic ocean acidification
Year
2015
Project leader
Phil Wallhead (NIVA) and Morten Skogen (IMR)
Participants

Andre Staalstrom (NIVA)

Cecilie Hansen (IMR)

Flagship
Ocean Acidification
Funding Source

Framsenter Flagship

Summary of Results

SINMOD results

During 2015 we analysed projections from the SINMOD ocean biogeochemical model under an SRES A1B emissions scenario.  This run used atmospheric forcings the REMO climate model1, riverine inputs from a hydrological model2, and open ocean boundary conditions from the Bergen Climate Model3 with bias correction to the CARINA dataset4.  CO2 system variables such as pH and aragonite saturation state Ωar were calculated using CO2SYS.m5 and SINMOD output for dissolved inorganic carbon (DIC), total alkalinity (Alk), temperature (T), salinity (S), silicate (Si), phosphate concentrations estimated by multiple linear regression on (T, NO3, Si) (fitted to the CARINA data), pressure estimated using a quadratic regression function of depth, dissociation constants from (6, 7) and total borate concentrations from (8).  Note that CO2SYS assumes that the calcium ion concentration is proportional to the salinity in calculating the calcite and aragonite saturation states9.

As a first step in analysing the projections, we consider the output averaged over the surface 50 m, where impacts of climate change and acidification are expected to be strong, and averaged over all seasons and years within two decades, thus focusing on the baseline environmental conditions (or oceanic climate).   These baseline conditions are expected to be important for determining the long-term viability of populations (e.g. copepods) in a given region, although changes in magnitude and phenology of the seasonal cycle may also have strong impacts and will be examined in future work.  To assist future comparison with NORWECOM projections10 we consider bidecadal averages over the periods 1980-2000 and 2045-2065 as reference and future climate projections separated by 65 years.

SINMOD projects warming of the 0-50m layer by up to 5˚C over the 65 years, with the largest changes occurring in the N. Barents Sea and Greenland Sea (Fig. 1, upper).  This is likely driven in part by increasing penetration of Atlantic water, because the salinity projections show increased salinity in most of the warming hotspots (Fig. 1, lower).  However, this may also reflect loss of ice cover (Fig. 2, upper) which acts to increase albedo and atmospheric heat absorption, and is in turn driven by increasing air temperature and heat fluxes from the Atlantic water (hence a positive feedback).  An exception here is the Kara Sea, where strong salinity increases are collocated with only mild temperature increases.  We hypothesize that these increases might reflect a weakening of the cross-shelf salinity gradient due to loss of ice cover (Fig. 2, upper) and associated increases in wind stress and horizontal mixing.

Projected DIC changes (Fig. 3, upper) reflect a broadscale increase due to vertical penetration of anthropogenic CO2, but also apparently reflect the increasing penetration of Atlantic water and loss of ice cover, with strongest increases in the Barents Sea and Greenland shelf.  Although the losses of ice cover are generally associated with an increased annual primary production (Fig. 2, lower), this apparently only weakly offsets the DIC increases.

SINMOD projects pH drops of up to 0.25 units in the Barents Sea, Kara Sea, E. Greenland Shelf and Eurasian Arctic Basin (Fig. 4, upper).  Aragonite saturation state Ωar drops to corrosive levels (~1) in the Kara Sea and Arctic Basin (Fig. 4, lower).  This is mainly driven by increasing DIC and freshening/reduced alkalinity due to increasing river discharge and ice melt (cf. Figs. 1, 3).  Again, this acidification is apparently only weakly offset by the increased primary production in areas losing ice cover (cf. Fig. 2).

Figure 1.  SINMOD bidecadal average temperature (upper) and salinity (lower) averaged over 0–50m and 1980-2000 (left), 2045-2065 under A1B scenario (middle), and the differences (right).  

Figure 2.  SINMOD bidecadal average ice cover (i.e. number of days per year with >0 ice thickness, upper) and primary productivity integrated over the euphotic zone (lower), averaged over 1980-2000 (left), 2045-2065 under A1B scenario (middle), and the differences (right).  

 

 Figure 3.  SINMOD bidecadal average dissolved inorganic carbon (upper) and total alkalinity (lower) averaged over 0–50m and 1980-2000 (left), 2045-2065 under A1B scenario (middle), and the differences (right). 

 

 Figure 4.  SINMOD bidecadal average dissolved inorganic carbon (upper) and total alkalinity (lower) averaged over 0–50m and 1980-2000 (left), 2045-2065 under A1B scenario (middle), and the differences (right).

 

 

 

NORWECOM results

 

 

From NORWECOM.E2E two different simulations  are now available forced by downscaled physics using ROMS for two different climate models and emission scenarios.  In the first one the ROMS model is used to downscale the GISS-AOM climate model under A1B for the period 1980-2000 and 2046-2065, while in the second one the NorESM model is downscaled for the period 2006-2070 under RCP4.5.  Focusing on the change on pH (Figure 1), both simulations give equal patterns with a mean decrase in most areas between 0.1 and 0.2, and with a higher reduction along the Norwegian coast.  The main differences is in the Arctic and in the Fram strait where the NorESM simulation give an increase in pH, while the GISS-AOM give a decrease.

 

A

 B

 

 Figure 1: Change in surface pH between control and future climate for the two different NORWECOM.E2E realisations. A: 2065-2000 forced by downscaled  the GISS-AOM. B: 2069-2006 forced by the downscaled NorESM model.

 

 

 

 

 

NoBa results

 

Running NoBa with impact of ocean acidification on large zooplankton (ZL)

 

To explore direct effects of ocean acidification on lower trophic levels and indirect effects on higher trophic levels and potential implications for management, an end-to-end model, NoBa, has been implemented and run for the Nordic and Barents Seas. Impact of ocean acidification is implemented in NoBa using a pH-scalar which is multiplied with the growth rate and consumption rate, whereas the mortality rate is divided by it. Here, we will only present results from the Barents Sea. NoBa was in these scenarios run for two periods using physical (salinity, temperature, volume fluxes) from ROMS (Table 1). These two periods represent the years 1980-2001 and 2046-2065. The first period (1980-2001) was run as a control run. For the second period, three simulations was performed; first a control run including no effects of pH, thereafter one run where the pH scalar was 1.36, and in the last one the pH scalar was 0.64. On the 24th November, a group of experts were gathered for one day in Tromsø to define a set of pH scalars that NoBa will use to run a range of realistic scenarios exploring potential effects of ocean acidifications.

 

For the two runs including an effect of ocean acidification, we chose to apply the pH scalar only on large zooplankton. For all simulations we present effects of ocean acidification or physical forcing itself on a few selected components in the system; the commercially important (cod, saithe, haddock and herring) and important prey items (polar cod and capelin). All simulations are 45 years long, with a spinup time of 25 years. In the spin-up period, the physical forcing is held constant at the first year of the period (1980 or 2045).

 

 

 

Identifying the “most important” zooplankton group in the system

 

Large zooplankton, which in NoBa represents all zooplankton with a length above 2 mm, was chosen based on a small sensitivity study, where the total effect on the ecosystem by decreasing one zooplankton group at a time was measured by exploring the total changes in the biomasses of the other components in the system. The simulations performed are described in Table 1.

 

 

 

Table 1: Overview of the runs performed exploring the sensitivity of the ecosystem to the four different classes of zooplankton; Large (ZL), medium (ZM), small (ZS) and gelatinous (ZG). The parameters tuned are growth rates (mum_XXX; mgN d-1), mortality rates (mQ; d-1) and consumption rates (C_XXX; mgN d-1). In the simulations where we determine which zooplankton group is most important for the flow in the ecosystem, we have decreased the growth rates of all the four zooplankton groups by 25%.

 

Run

Period

Species tuned

Parameter tuned

Parameter value, old

Parameter value, new

1

1981 (loop – 55 years)

ZL

mum_ZL

0.076

0.057

2

As in run 1

ZM

mum_ZM

0.1

0.075

3

As in run 1

ZS

mum_ZS

3.55

2.6625

4

As in run 1

ZG

mum_ZG

0.02

0.015

20C3M

1980-2001

-

-

-

-

SRESA1B_cntr

2045-2065

-

-

-

-

SRESA1B_pos

As in run 6

ZL

mum_ZL, C_ZL, mQ_ZL

7.6e-2,

0.2,

8e-10

1.03e-1,

0.272,

5.88e-10

SRESA1B_neg

As in run 6

ZL

mum_ZL, C_ZL, mQ_ZL

7.6e-2,

0.2,

8e-1

4.86e-2,

0.128,

1.25e-9

 

 

 

In the sensitivity runs, the changes in average biomass of the last 10 years were calculated for all components in the system.  We thus identified the large zooplankton group as the zooplankton group that the ecosystem was most vulnerable to.

 

 

 

Sensitivity to ocean acidification and changes in physics

 

The expected changes in the runs would be that the simulation including a negative impact on ZL would yield lower fish biomass than the one with the positive effect by OA on ZL. However, this is not always the case.  Here, we find that decreasing the large zooplankton introduced increases in most of the species that we have explored, with the exception of haddock and saithe (Table 3 and Table 4). The changes are small though, for all species except capelin, which has an increase of 85% in the negative run. Looking at Figure 2, this is explained by the large peak that the population experiences around year 17.  The recruitment success of this particular period needs to be thoroughly explored before we draw any further conclusions about it.

 A

 

 

B

 

Figure 1: The development over the last 20 years of the simulations. The figures shows that there are differences introduced by the physical forcing alone, this should be kept in mind when looking at the results from the ocean acidification runs.

 

A

 

 B

 Figure 2: The development of the biomass of cod and capelin over the last 20 years of the simulation.

 

Table 3: Changes (%) in average biomass over the last 10 years of the simulation compared to the average biomass in the last 10 years of the 20C3M control run.

Species

SRESA1B_cntr

SRESA1B_pos

SRESA1B_neg

Cod

-4.6

-5.9

3.5

Capelin

13.3

-3.1

85.3

Polar cod

-3.1

-8.4

11.5

Herring

-1.1

-2.9

3.4

Haddock

-2.5

-9.6

-3.4

Saithe

1.9

0.8

-0.6

 

Table 4: Changes (%) in average biomass over the last 10 years of the simulation compared to the average biomass over the last 10 years of the SRESA1B control run.

Species

SRESA1B_pos

SRESA1B_neg

Cod

-1.4

8.5

Capelin

-14.4

63.6

Polar cod

-5.5

15.0

Herring

-1.8

4.5

Haddock

-7.3

-1.0

Saithe

-1.1

-2.5

 

Based on these preliminary results, we cannot conclude that impact of ocean acidification on large zooplankton will have any impact on management of the large commercial stocks in the Barents Sea. It also has to be noted that the largest introduction of changes in the biomass comes when changing the physical forcing, not by introducing effects of ocean acidification.  The buffering effect in the Barents Sea has also been shown in literature to be rather strong, and we might not see any effects before perturbing several parts of the ecosystem at the same time.

 

References

1Keup-Thiel, E., Gottel, H., Jacob, D. 2006. Boreal Environment Research. 11(5): 329-339.

2Dankers, R., Middelkoop, H. 2008. Climatic Change, 87: 131-153.

3Tjiputra, J. F., Assmann, K., Bentsen, M. et al. 2010. Geoscientific Model Development, 3: 123-141.

4Key, R. M., Tanhua, T., Olsen, A. et al. 2010. ESSD 2, 105-121.

5van Heuven, S., Pierrot, D., Rae, J. W. B. et al. 2011. In ORNL/CDIAC-105b. Carbon Dioxide Information Analysis Center, Oak Ridge  National Laboratory, U.S., Department of Energy, Oak Ridge, Tennessee

6Millero, F.J. 2010. Marine and Freshwater Research 61(2) 139-142.

7Dickson, A.G., 1990.  J. Chemical Thermodynamics, 22:113-127.

8Uppstrøm, L.R., 1974. Deep Sea Research I,  21(2) 161-162.

9Riley, J. P. and Tongudai, M., Chemical Geology 2:263-269, 1967.

10Skogen, M. D. et al. 2014. Journal of Marine Systems 131, 10–20.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

For the Management

The project has progressed as planned in the proposal and all milestones and deliverables have been met.

Published Results/Planned Publications

A peer-review paper on the SINMOD projections is planned for 2016. 

Communicated Results

SINMOD projections were presented in a poster at the FRAM Science Days conference.

Interdisciplinary Cooperation

This project has demanded collaboration between biogeochemical modellers (SINMOD, NORWECOM), ecosystem modellers (NoBa), the socioeconomic modellers in ECOAN-WP4 as well as the observationalists in ECOAN-WP1 and ECOAN-WP2.  A workshop on Calanus sensitivity to climate change and acidification took place in November to promote this collaboration and gather information needed for future work in WP3 and WP4.

Could results from the project be subject for any commercial utilization
No
Conclusions

The project has progressed in line with the proposal, and is providing important insights into projected climate change, acidification, and possible ecosystem responses.  Interdisciplinary collaborations with other workpackages have been established and utilized.