Modelling of regional CO2 balance

1
Modelling of regional CO2 balance
Tiina Markkanen
with Tuula Aalto, Tea Thum, Jouni Susiluoto and
Niina Puttonen
2
Contents

• Modeling framework in Snowcarbo
• Models
• REMO regional climate model
• JSBACH land surface scheme
• Demonstration of model performance
• Regional
• Local
• Conclusions and future perspectives

3
Modelling framework in Snowcarbo

• Models applied
• REgional climate MOdel of MPI -M, Hamburg REMO
• Produce regional climatic forcing
• Land surface scheme (LSS) of GCM ECHAM JSBACH
• Produce regional CO2 balance consisting of
  assimilation and emissions in ecosystems

4
Modelling framework in Snowcarbo

• Relatively high spatial resolution
• 18km
• High time resolution
• 1 hour
• Stays close to actual weather of target years
  2001-2011
• Climate initialised once per day
• REMO and JSBACH are offline coupled
• REMO does not get feedback from JSBACH but
  interacts with its own surface scheme

5
Vegetation type classification from remote
sensing
Land cover classification from remote sensing
Anthropogenic and ocean CO2 sources, fires
JSBACH
CO2 flux
Detailed meteorology
REMO2008 tracer
REMO2008
Indicators of regional CO2 balance
Snow data, Phenology, etc.
Observed climate (ECWMF)
CO2 flux (NEE) field 2D
CO2 concentration field 3D
Snow data
Validation against various data
Snow cover and phenology related variables
Flux and concentration data
6
Vegetation type classification from remote
sensing
Land cover classification from remote sensing
Anthropogenic and ocean CO2 sources, fires
JSBACH
CO2 flux
Detailed meteorology
REMO2008 tracer
REMO2008
Indicators of regional CO2 balance
Snow data, Phenology, etc.
Observed climate (ECWMF)
CO2 flux (NEE) field 2D
CO2 concentration field 3D
Snow data
Validation against various data
Snow cover and phenology related variables
Flux and concentration data
7
REMO forcing and initialisation

• Regional climate model requires as boundary data
• Atmospheric conditions
• Wind speeds, Temperature, Humidity
• Sea surface temperature, ice cover
• Surface parameter fields
• As initial data in addition to those above
• Soil temperature and moisture
• Sources of initial and boundary meteorological
  data
• General circulation models
• Re-analysis data products, here ERA-Interim,
  ECMWF

8
REMO basic characteristics

• Dynamic core of DWD operational model
• Physics and surface model from ECHAM
• Surface parameter maps for
• Surface background albedo, roughness length,
  vegetation ratio, leaf area index, forest
  fraction, soil field capacity
• Rotated spherical grid
• Close to rectangular
• Resolution applied in Snowcarbo 0.1667

9
REMO surface parameter maps

• Standard land cover of USGS classified according
  to Olson cover types (n100)
• Parameter values allocated to each Olson type
• Parameters aggregated from 1km USGS map to maps
  of resolution of the model
• In Snowcarbo the USGS map is replaced by National
  Corine Land Cover (CLC), European CLC and
  Globcover datasets
• Allocations from new land cover classes to Olson
  ones needed

10
REMO influence of land cover forest fraction
Standard USGS land cover
National Corine land cover
11
REMO

• In this project runs in forecast mode
• Model initialised daily at 6pm
• Spun-up until midnight in order to let the
  flowfield to develop into a reasonable state
• Run for 24 consequent hours with hourly output
• Weather stays close to observed
• Sensitive to meteorological boundary
• Not very sensitive to surface parameterisation

12
Vegetation type classification from remote
sensing
Land cover classification from remote sensing
Anthropogenic and ocean CO2 sources, fires
JSBACH
CO2 flux
Detailed meteorology
REMO2008 tracer
REMO2008
Indicators of regional CO2 balance
Snow data, Phenology, etc.
Observed climate (ECWMF)
CO2 flux (NEE) field 2D
CO2 concentration field 3D
Snow data
Validation against various data
Snow cover and phenology related variables
Flux and concentration data
13
Modelling framework in Snowcarbo JSBACH

• LSS of ECHAM to account for
• Surface energy partitioning e.g. water balance
• Carbon cycle
• In offline coupled mode JSBACH is used to account
  for ecosystem carbon balance - CO2 exchange
• Process model
• Processes described down to as small scale as
  possible
• Limited by computational resources

14
JSBACH characteristics

• 4 tiles, i.e. 4 PFTs (plant functional type) for
  each grid cell
• Photosynthesis of C3 and C4 plants
• Radiation in canopy
• Carbon storages in soil and vegetation
• Q10 approach for soil decomposition
• LAI (leaf area index) dynamics described with
  four phenology models

15
JSBACH parameterisations for PFTs

• Tropical broadleaf evergreen trees
• Tropical broadleaf deciduous trees
• Temperate broadleaf evergreen trees
• Temperate broadleaf deciduous trees
• Coniferous evergreen trees
• Coniferous deciduous trees
• Raingreen shrubs
• Deciduous shrubs
• C3 grass
• C4 grass
• Tundra
• Swamp (not used)
• Crops
• Glacier

• Phenology
• Photosynthesis
• Carbon storage sizes
• Decomposition rates of carbon storages
• Albedo for NIR and VIS
• Roughness length, LAImax, etc.
• Dynamic vegetation
• Nitrogen cycle

PFT distribution is revised with more detailed
land cover products
16

17
Offline coupling in Snowcarbo

18
Vegetation type classification from remote
sensing
Land cover classification from remote sensing
Anthropogenic and ocean CO2 sources, fires
JSBACH
CO2 flux
Detailed meteorology
REMO2008 tracer
REMO2008
Indicators of regional CO2 balance
Snow data, Phenology, etc.
Observed climate (ECWMF)
CO2 concentration field 3D
CO2 flux (NEE) field 2D
Snow data
Validation against various data
Snow cover and phenology related variables
Flux and concentration data
19
JSBACH
REMO in tracer mode
radiation

fAPAR
Thermal and hydrological conditions
gc unstressed
CO2 concentration
gc stressed
H LE
Gross assimilation, Rd
NPP
CO2 flux
Albedo roughness length
LAI
20
Observed climate (ECWMF)
Anthropogenic and ocean CO2 sources, fires
JSBACH
REMO in tracer mode
radiation

fAPAR
Thermal and hydrological conditions
gc unstressed
CO2 concentration
gc stressed
H LE
Gross assimilation, Rd
NPP
CO2 flux
Albedo roughness length
LAI
21
REMO tracer run

• Uses the same meteorological initial and boundary
  data as the first REMO run
• Additionally
• Gets CO2 flux estimates of vegetation from JSBACH
• Utilizes prescribed anthropogenic and land fire
  emissions from a database
• Ocean sources from a database
• Requires background CO2 concentrations
• Produces 3D CO2 concentration fields

22
Demonstration of model performance

• JSBACH was forced
• regionally with REMO derived climatic forcing
• Forecast mode for climate model
• Standard land cover in both models
• Default carbon storages in JSBACH
• for flux measurement sites Sodankylä and Hyytiälä
• About a decade of measurements as climatic
  forcing
• 1000 years spin up for soil carbon storages with
  present climate

23
Demonstration of model performance

• Regional daily average NEE ltstart animationgt

24
Demonstration of model performance

• Daily NEE at Sodankylä Scots pine site (gm-2s-1)

25
Demonstration of model performance

• Daily NEE at Hyytiälä Scots pine site (gm-2s-1)

26
Conclusions and future perspectives

• Assimilation and emissions of CO2 in ecosystems
  explicitly modeled
• Offline coupled REMO JSBACH framework produces
  CO2 balance in high temporal and in relatively
  high regional resolution
• National level estimates of CO2 balance can be
  extracted from the regional maps
• To be done
• Evaluation of the results against CO2 flux and
  concentration data
• Adjustment of the relevant parameters in order to
  produce better regional estimates