Coupled Climate and Environmental Modeling for the Baltic Sea Region

1
Coupled Climate and Environmental Modeling for
the Baltic Sea Region

• H.E. Markus Meier
• Swedish Meteorological and
• Hydrological Institute, Norrköping
• and
• Stockholm University, Stockholm
• Markus.Meier_at_smhi.se
• Collaborators Kari Eilola, Elin Almroth, Ralf
  Döscher, Anders Höglund, Robinson Hordoir

2
Dynamical downscaling

• Add regional details (land-sea mask)
• For the ocean consistent scenarios (with
  realistic SSTs) are necessary using coupled
  models

2009-11-15
3
The coupled system RCAO
RCA 44 km, 30 min RCO 11 km, 10 min Coupling
timestep 3 h
RCO
RCA
Model domain, covering most of Europe and parts
of the North Atlantic Ocean and Nordic Seas. Only
the Baltic Sea is interactively coupled.
The coupling scheme of RCAO. Atmosphere and
ocean/ice run in parallel.
Döscher et al. (2002)
2009-11-15
4
Regionalization is done for time-slices
from GCMs
Regional simulations
Results archived from a GCM-run
CO2
1800
1900
2000
2100
Present-day or a control climate
Climate scenario
(1961-1990)
(2071-2100)
4
2009-11-15
5
Mean maximum ice cover in control (blue) and
scenario (red)
(Meier et al., 2004)
5
2009-11-15
6
Figure 5. Seasonal mean SST differences between
the ensemble average scenario and simulated
present climate (in C) DJF (upper left), MAM
(upper right), JJA (lower left), and SON (lower
right). The figure is taken from Meier (2006,
Fig.13) with kind permission of Springer Science
and Business Media.
Sea surface temperature annual 1.9 3.9C
winter
spring
summer
autumn
2009-11-15
7
Time slice approach

• For the deep water response there is
• a spin-up problem due to the long
• time scales.
• Solution delta approach

2009-11-15
8
Model hierachy
2009-11-15
9
2009-11-15
2009-11-15
10
(Source Phil Graham, SMHI)
2009-11-15
2009-11-15
11
Salinity at Gotland Deep
Figure 1. Median profiles of salinity at
monitoring station BY15 for present climate
1961-1990 (black solid line, shaded areas
indicate the /- 2 standard deviation band
calculated from two-daily values for 1903-1998)
and in projections for 2071-2100 (colored lines).
In (a) only effects from wind changes are
considered whereas in (b) projections based upon
wind and freshwater inflow changes are shown.
Numbers in the legend correspond to the different
scenario runs (see Tab.1). The figure is taken
from Meier et al. (2006, Fig.2).
2009-11-15
2009-11-15
12
Sea surface salinity
Projection with the largest change RCAO-ECHAM4/A2
Present climate
5 psu
2009-11-15
13
Scenario simulations of the marine ecosystem

• Important to assess uncertainties due
• natural variability,
• model biases from GCMs and RCMs,
• greenhouse gas emission scenarios,
• land use and sewage treatment scenarios,
• biogeochemical processes)
• To get robust information ensemble simulations
• should be performed

2009-11-15
14
RCO-SCOBIHigh-resolution 3-D coupled
physical-biogeochemical model for climate and
process studies

• Pelagic variables
• nitrate (NO3)
• ammonium (NH4)
• phosphate (PO4)
• autotrophs (A1,A2,A3)
• (diatoms, flagellates, cyanobacteria)
• zooplankton (ZOO)
• detritus (DET)
• oxygen (O2)
• Hydrogen sulfide (H2S) is
  included as negative oxygen.

• The sediment contains nutrients in the form of
  benthic nitrogen (NBT) and phosphorus (PBT).
• Aggregated process descriptions for oxygen
  dependent nutrient regeneration, denitrification
  and adsorption of ammonium to sediment particles
  as well as re-suspension and permanent burial of
  organic matter.

15
Long period experimentForcing and model set-up
of the 100 years model run

• RCO 6 nm run on the period 1902-1998 with
  reconstructed atmospheric forcing and river
  discharge data.
• Nutrient loading from land (rivers and coastal
  runoff) is based on climatological mean
  concentrations from the period 1970-1993.
• Point sources are based on HELCOM estimates from
  the 1990s.
• Atmospheric nitrogen deposition is based on
  HELCOM estimates from the 1980s and 1990s.

16
To estimate uncertainties an ensemble of 20
simulations has been performed

• four climate scenarios using RCAO forced
  with two emission scenarios (A2, B2) and two
  GCMs (SMHI)
• 1) ECHAM4/A2 SST 3.7ºC, SSS -3.2 psu,
  increased mixing
• 2) ECHAM4/B2 SST 2.9ºC, SSS -3.0 psu,
  increased mixing
• 3) HADAM3H/A2 SST 3.2ºC
• 4) HADAM3H/B2 SST 2.1ºC

17
Scenario simulationsMedian, first and third
quartiles at Gotland Deep

• Present climate (black)
• ECHAM4/A2 (red) and ECHAM4/B2 (orange)
  (2071-2100)
• HADAM3H/A2 (green) and HADAM3H/B2 (blue)
  (2071-2100)

18
Annual mean phytoplankton concentration and
changes (0-10 m) mgChl/m3
Reference (1969-1998)
Present climate
ECHAM4/A2
HADAM3H/A2
ECHAM4/B2
HADAM3H/B2
2009-11-15
19
Summary of the scenario simulations

• Changing climate will affect the Baltic Sea
  ecosystem because the impact of changing physical
  conditions on the marine ecosystem isimportant,
  especially when wind and precipitation are
  changing significantly
• The sensitivity of the response to climate change
  depends on key processes that are not well
  understood
• The response is highly non-linear (the impact of
  climate change depends on the nutrient level)

20
New scenarios are planned

• New components, e.g. wave model to improve
  resuspension

21
Coupled wave model
Mean effect of wind waves on circulation Mean
effect of wind waves on vertical stress and
turbulence
22
New scenarios are planned

• New components, e.g. wave model to improve
  resuspension
• Higher resolution in the ocean

23
New scenarios are planned

• New components, e.g. wave model to improve
  resuspension
• Higher resolution in the ocean
• Direct coupling with improved wind extremes
  (mixing)

24
temperature
salinity
Arkona Basin (BY2)
Bornholm Basin (BY5)
1980-2004
Median, first and third quartile Observations
(solid), Model results (dashed)
Gotland Basin (BY15)
Gulf of Finland (LL07)
Bothnian Sea (SR5)
25
Annual mean halocline depth for 1981-2004
26
Surface wind correction(measured mean wind is
the maximum 10 minute mean wind over the last 3
hours,measured wind gust is the maximum 2 sec
mean wind over the last 10 minute period)

• Wind gust estimate method (Brasseur, 2001)
  assumes that wind gusts develop when air parcels
  higher up in the boundary layer deflects down to
  the surface by turbulent eddies (Nordström, 2005)
• Empirical linear relationship between peak gusts
  and mean wind speeds (Davis and Newstein, 1968),
  1.6 in 10 m height
• Wind speed correction U10new max(Ugust/1.6,
  U10)

27
Observations (red), RCA model wind
(green), Modified model wind using gustiness
parameterization (blue)
Gotska Sandön
28
Results with corrected winds taking gustiness
into account
temperature
salinity
Arkona Basin (BY2)
Bornholm Basin (BY5)
Median, first and third quartile Observations
(solid), Model results (dashed)
Gotland Basin (BY15)
Gulf of Finland (LL07)
Bothnian Sea (SR5)
29
New scenarios are planned

• New components, e.g. wave model to improve
  resuspension
• Higher resolution in the ocean
• Direct coupling with improved wind extremes
  (mixing)
• Transient simulations 1960-2100

30
Model hierarchy in ECOSUPPORT (Advanced modeling
tool for scenarios of the Baltic Sea ECOsystem to
SUPPORT decision making)