Download Atmosphere-Biosphere-Hydrosphere Modeling

Regional Atmosphere-BiosphereHydrosphere Modeling:
Current Potentials & Future Requirements
H. Kunstmann, E. Haas, R. Knoche,
R. Forkel, R. Grote, K. Butterbach-Bahl,
Institute for Meteorology and Climate Research
Forschungszentrum Karlsruhe/Garmisch-Partenkirchen
Regional Atmosphere-Biosphere-Hydrosphere Modeling
HGF-Workshop Regional Modeling 2./3.11.2006
Interdependent Systems: Brief Review
⇒ Met./climate variables,
air chemistry
⇐ Biogenic/microbial
exchange, transpiration
Terrestrial
Biosphere
Atmosphere
⇒ Dynamic vegetation
properties, leaching
⇐ Water availability, land
surface & physical &
chemical soil properties
⇒ Met./climate variables
⇐ Latent & sensible
heat fluxes,
land surface & soil
properties
Terrestrial
Hydrosphere
& Land Surface
+ anthropogenic impact/management
Regional Atmosphere-Biosphere-Hydrosphere Modeling
HGF-Workshop Regional Modeling 2./3.11.2006
Current Potentials
Regional Atmosphere-Biosphere-Hydrosphere Modeling
HGF-Workshop Regional Modeling 2./3.11.2006
Regional Atmospheric Modeling
Atmosphere
Concept of dynamical
downscaling
Regional Atmosphere-Biosphere-Hydrosphere Modeling
HGF-Workshop Regional Modeling 2./3.11.2006
Regional Atmospheric Modeling
Basic facts
• Scales: 1 km (non-hydrostatic) < ∆x < 50 km
10-60 vertical layers for troposphere
3 sec < ∆t < 150 sec
• Boundary & initial cond. by global atmospheric model
Atmosphere
• Nesting approach (successive refinement)
• Grid scale: general physical laws
(conservation of energy, momentum, mass)
• Subgrid scale: parameterisations
(turbulence, convection, cloud & precipitation physics)
• Requires land surface-, soil-, vegetationproperties, DEM, etc.
Regional Atmosphere-Biosphere-Hydrosphere Modeling
HGF-Workshop Regional Modeling 2./3.11.2006
Regional Atmospheric Modeling
Performance of Atmospheric Models
Difference MM5-ECMWF 19 km
vs. DWD gridded stations
1979-1993
Regional Atmosphere-Biosphere-Hydrosphere Modeling
HGF-Workshop Regional Modeling 2./3.11.2006
Regional Atmospheric Modeling
Example: Regional Climate Modeling
Impact of Climate Change on Precipitation
Expected precipitation change
MM5-ECHAM4, A2
Regional Atmosphere-Biosphere-Hydrosphere Modeling
HGF-Workshop Regional Modeling 2./3.11.2006
Regional Atmospheric Modeling
Air chemistry models
• Solving both for atmospheric transport & chemical kinetics
• Conservation laws (energy & mass) for all trace species (up to 200 species)
• Chemistry mechanisms defined by species & respective reaction paths
• Reducing complexity by lumping of similar species into groups
• Surface interaction both at lower boundary and particle surface
• Numeric:
time splitting schemes
≈ factor 10 increased complexity compared to pure meteorological model
• Linking terrestrial biosphere: e.g. through canopy chemistry
Requires met. variables at all grid points
• Online vs. offline coupled codes
Regional Atmosphere-Biosphere-Hydrosphere Modeling
HGF-Workshop Regional Modeling 2./3.11.2006
Regional Atmospheric Modeling
Example: Regional Meteorology-Chemistry Simulation
Air Pollution in Metropolitan Areas
11.4
11.6
48.4
11.6
11.8
48.4
48.4
Airport-Munich
NO2
Airport-Munich
A92
A92
Neufahrn
Neufahrn
48.3
48.3
Latitude (°N)
Dachau
Garching
A99
Ismaning
48.2
Allach
Johanneskirchen
Pasing
Stachus
A96
5 ppbv
48.1
A94
A95
A9
A92
A99
48.2
A99
Stachus
48.1
Taufkirchen
A95
11.6
-30 %
A94
Riem
Fürstenried
48.1
A99
-40 %
-50 %
Taufkirchen
A8
11.4
48.2 -20 %
Johanneskirchen
A96
48.1
-10 %
Ismaning
Allach
Pasing
10 %
0 %
Garching
A8
Riem
Fürstenried
0 ppbv
48.3
Dachau
A8
48.2
48.3
A9
A92
Latitude (°N)
15 ppbv
10 ppbv
11.4
11.8
48.4
A8
11.8
Longitude (°E)
Air pollution distribution for the Munich
area in summer 2000
Regional Atmosphere-Biosphere-Hydrosphere Modeling
11.4
11.6
11.8
Longitude (°E)
Air pollution difference due to emission
projection in 2010
HGF-Workshop Regional Modeling 2./3.11.2006
Regional Atmospheric Modeling
Example: Regional Climate-Chemistry Simulation
Impact of Climate Change on Air Chemistry
3,5
ECHAM4, IS92a
3
1991-2000
2031-2039
Frequency
2,5
2
1,5
1
0,5
ECHAM4, IS92a
0
0
20
40
60
80
100 120 140 160 180 200 220 240
Ozone (mg/m3)
Increased occurrence of extreme value conditions
Regional Atmosphere-Biosphere-Hydrosphere Modeling
HGF-Workshop Regional Modeling 2./3.11.2006
Terrestrial Hydrological Modeling
How does a standard atmospheric model account for
terrestrial hydrology and vegetation?
Terrestrial
Hydrosphere
& Land Surface
Soil-Vegetation-Atmosphere-Transfer models
Regional Atmosphere-Biosphere-Hydrosphere Modeling
HGF-Workshop Regional Modeling 2./3.11.2006
Regional Atmospheric Modeling
How does a standard atmospheric model account for
terrestrial hydrology and vegetation?
• Vertical discretization: e.g. 4 soil layers till 2 m depth
• Solving heat diff. equation (soil temp.) & Richards equation (soil moisture)
• Canopy transpiration via Penman-Monteith & resistance approach
• Stability dependent exchange coefficient for sensible and latent heat fluxes
• Lower boundary condition for deep soil temperature & moisture
• fixed vegetation properties: e.g. vegetation type
• seasonally fixed: e.g. LAI, albedo, roughness length, water stress parameters
Regional Atmosphere-Biosphere-Hydrosphere Modeling
HGF-Workshop Regional Modeling 2./3.11.2006
Terrestrial Hydrological Modeling
Example: Regional Climate Modeling
Impact of Climate Change on Infiltration Excess
DJF
JJA
Change in infiltration excess [%] 2070-2099 vs. 1960-89, ECHAM4 & MM5, B2
Regional Atmosphere-Biosphere-Hydrosphere Modeling
HGF-Workshop Regional Modeling 2./3.11.2006
Terrestrial Hydrological Modeling
Basic differences between SVAT-based hydrological models
and “traditional” hydrological models
• SVAT-Hydro Models (designed for atmospheric feedback purposes):
full energy balance (soil heat & sensible heat fluxes)
2-way interaction with PBL
• “Traditional”-Hydro models (designed for pure hydrol. applications):
lateral water fluxes, surface runoff routing
deeper soils considered
finer vertical & horizontal resolutions
often groundwater interaction
often extensions for reactive flow & transport, erosion, etc.
but: depending on specific model choice
Regional Atmosphere-Biosphere-Hydrosphere Modeling
HGF-Workshop Regional Modeling 2./3.11.2006
Biosphere Modeling
Biosphere
Macro-scale ecosystem
Micro-scale ecosystem
Regional Atmosphere-Biosphere-Hydrosphere Modeling
HGF-Workshop Regional Modeling 2./3.11.2006
Biosphere Modeling
Basic facts (DNDC; MOBILE)
• 1-d column approach (soil & vegetation layers)
• 10 sec. < ∆t < 1 day
• Driven by met./climate & air chemistry data
Biosphere
• Macroscale ecosystem
1) Vegetation dynamics: e.g. stand level, foliage, LAI
2) Soil-hydrology approaches: a) bucket b) SVAT
• Microscale ecosystem
1) Physiological (vegetation) & biogeochemical (soil)
processes (C & N- fluxes & balances): mass fluxes,
dissolution, decomposition, oxidation, adsorption,
complexation, assimilation & reverse processes
2) optionally (celluar level): biochemical processes
C-metabolism (e.g. VOC emission)
Regional Atmosphere-Biosphere-Hydrosphere Modeling
HGF-Workshop Regional Modeling 2./3.11.2006
Biosphere Modeling
Process based vs. empirical approaches
Example: VOC modeling
relative emission rate
1.2
1
0.8
0.6
0.4
0.2
0
0
10
20
30
40
50
temperature
1.2
relative emission rate
1
0.8
0.6
0.4
Guenther
0.2
0
25
275
525
radiation
775
1000
Empirical
Regional Atmosphere-Biosphere-Hydrosphere Modeling
Process based
HGF-Workshop Regional Modeling 2./3.11.2006
Biosphere Modeling
Process based vs. parameterized approaches
Example: VOC modeling
Empirical
Process-Based
• Transparent relations
• Better representation of variability
• Reliable within the
parameterized range
• Principally suitable for scenario studies
(e.g. climate change & land use change)
• Requires data on time
scale of prediction
(often not available)
Regional Atmosphere-Biosphere-Hydrosphere Modeling
HGF-Workshop Regional Modeling 2./3.11.2006
Biosphere Modeling
]
Field
Fieldscale
scale
Bellenden Ker
Pin Gin Hill
80
Kauri Creek
60
14
16
18
20
22
24
26
28
30
32
-2
12
Temperature [°C]
T_A
T_MAX - temp[l] T_A
)
*e
fact_temp [l] = (
T_MAX - T_OPT
*
temp[l]- T_OPT
T_MAX
60
40
-1
1+ e
water[l]- W_CRIT
W_DELTA
350
litter layer
80
h]
20
0
1
fact_moist [l] = 1 -
40
0
50
100
150
30
28
26
24
22
20
Precipitation [mm h
WFPS [%] Temperature [°C]
100
N2O emission [µg N m
N 2O-Emission rate [µg N kg -1SDW d -1]
-1
Laboratory scale
- T_OPT
d nitr_akt[l ] =
temp_moist _fact
MUEMAX * (
- nitr_akt[l ])
KM_DOC + DOC[l] + KM_N + n[l]
dt
DOC[l]
n[l]
600
400
250
200
150
100
50
0
27 Jan
10 Feb
Kauri Creek
24 Feb
Calibration
Parametrisation
9 Mar
Bellenden Ker
23 Mar
Pin Gin Hill
Validation
Process oriented biogeochemical model [PnET-N-DNDC]
C lim at e
Ve g e tatio n
S o il prop e rtie s
G e o g ra p h ic
In fo r m a tio n
S y s te m
Regional scale
Regional Atmosphere-Biosphere-Hydrosphere Modeling
N 2 O Inventory
HGF-Workshop Regional Modeling 2./3.11.2006
40
0
20
20
40
0
80
-1
Precipitation
Field observations
PnET-N-DNDC model results
-1
CO2-emission [Kg C ha d ]
|
Temperature
60
60
80
Precipitation [mm]
Temperature [°C]
Biosphere Modeling
40
20
0
Höglwald forest
Jan
Apr
Jul
1995
Okt
Jan
Apr
Jul
1996
Okt
Jan
Apr
Jul
Okt
1997
Validation
Regional Atmosphere-Biosphere-Hydrosphere Modeling
HGF-Workshop Regional Modeling 2./3.11.2006
Biosphere Modeling
25
NO emission
N2O emission
30
Solling
gap (1991)
Schottenwald, beech (1997)
Speuld, douglas fir (1997)
Schottenwald
beech (1996)
Höglwald, beech (1995)
Klausen-Leopoldsdorf, beech (1997)
Wildbahn, pine (1996)
Höglwald spruce (1997)
Speuld, beech (1997)
20
Höglwald, beech (1995)
-1
Höglwald
beech (1996)
10
Höglwald, spruce (1996)
Höglwald, spruce (1997)
Höglwald, beech (1997)
Höglwald, beech (1996)
Kopenhagen, spruce (1992)
Höglwald, spruce (1995)
Wildbahn, pine (1997)
Harvard Forest, hardwood (1989)
0
Höglwald, spruce (1995)
-1
NO simulated [g N ha day ]
5
Höglwald, beech (1997)
10
Hubertusstock, pine (1997)
Hubertusstock, pine (1996)
Höglwald spruce (1997)
Solling, beech (1991)
-1
15
Klausen-Leopoldsdorf, beech (1996)
-1
N2O Simulated [g N ha day ]
20
0
Villingen spruce (1994)
Oak ridge, oak (1988)
Villingen spruce (1995)
0
5
10
15
20
-1
-1
N2O Measured [g N ha Tag ]
25
0
10
20
30
-1
-1
NO Measured [g N ha day ]
Validation for different European climate and vegetation regions
Regional Atmosphere-Biosphere-Hydrosphere Modeling
HGF-Workshop Regional Modeling 2./3.11.2006
General Coupling Concepts
external
A
external
A
B
Stand alone
No coupling
B
One-way
coupling
B
Two-way
coupling
external
A
Regional Atmosphere-Biosphere-Hydrosphere Modeling
HGF-Workshop Regional Modeling 2./3.11.2006
General Coupling Considerations
∆ta
Atmosphere
∆ta→b
∆tb→a
∆tb
Terrestrial
Biosphere
∆th→a
∆ta→h
∆th→b
∆tb→h
Relation between
• Process time scales ∆tb, ∆ta, ∆th
• exchange time scales
Regional Atmosphere-Biosphere-Hydrosphere Modeling
∆th
Terrestrial
Hydrosphere
& Land Surface
determines coupling
strategy & data exchange
frequency
HGF-Workshop Regional Modeling 2./3.11.2006
Mutual System Interaction & Interfaces
Atmosphere
Models
Biosphere
Models
Hydrological
Models
Vegetation:
type, LAI, z0, α
External forcing
(Seasonally fixed)
Dynamic
External forcing
(Seasonally fixed)
Emission:
BVOC, NOx
External forcing
Dynamic
-
Meteorology:
T, P, p, v, RH
Dynamic
External forcing
External forcing
Dynamic
(1-d)
Dynamic
(2-d, 3-d)
Water redist.: Dynamic
Vertical&lateral (3-d,1-d)
Model coupling allows to introduce state variables where originally
constant or externally provided parameters had to be applied
Regional Atmosphere-Biosphere-Hydrosphere Modeling
HGF-Workshop Regional Modeling 2./3.11.2006
1 Way Coupling
Example: Climate-Hydrology Modeling
Impact of Climate Change on Terrestrial Water Balance
Upper Jordan
Catchment
LEBANON
6x6 km²
18x18 km²
54x54 km²
SYRIA
Jordan
Catchment
High resolution dynamical
downscaling of global climate
scenarios
Regional Atmosphere-Biosphere-Hydrosphere Modeling
GOLAN
HEIGHTS
Ayun
Dan Saar
Snir Banyas
ISRAEL
Yoseph Bridge
Distributed hydrological modeling
of surface and subsurface
water balance in 90 m resolution
HGF-Workshop Regional Modeling 2./3.11.2006
1 Way Coupling
Example: Climate-Hydrology Modeling
Impact of Climate Change, Near East/Upper Jordan River
Regional Atmosphere-Biosphere-Hydrosphere Modeling
HGF-Workshop Regional Modeling 2./3.11.2006
1 Way Coupling
Example: Climate-Hydrology Modeling
Impact of Climate Change, Near East/Upper Jordan River
Recharge
Runoff
Precipitation
Precipitation [mm]
160
-35.2%
140
120
100
80
60
40
+4.9%
12
70
-22.5%
10
60
-35.6%
-22.0%
+16.7%
+116.7%
20
0
-3.4%
-80%
Jan Feb Mrz Apr Mai Jun Jul Aug Sep Okt Nov Dez
1961 - 1990
2070 - 2099
Regional Atmosphere-Biosphere-Hydrosphere Modeling
Total Runoff
[mm] [mm]
Groundwater
Recharge
180
8
50
6
40
4
30
2
20
0
10
-2
0
-4
-34%
-30%
-33%
-39%
-36%
-32%
-24%
-8%-28%
-22%
-33%
-14%
-41%
-25%
-92%
-18%
-243%
-33%
-15% -33%-35%
-110% -127%
-200%
Jan Feb Mrz Apr Mai
Mai Jun
Jun Jul
Jul Aug
Aug Sep
Sep Okt
Okt Nov
Nov Dez
Dez
1961
1960 -- 1990
1990
2070
2070 -- 2099
2099
HGF-Workshop Regional Modeling 2./3.11.2006
1 Way Coupling
Example: Climate-Biosphere Modeling
Impact of Climate Change on Biogeochemical Emissions
N2O emission change in % (future minus present)
Regional Atmosphere-Biosphere-Hydrosphere Modeling
HGF-Workshop Regional Modeling 2./3.11.2006
Towards fully dynamic 2-way coupled
atmosphere-biosphere-hydrosphere
modeling …
Regional Atmosphere-Biosphere-Hydrosphere Modeling
HGF-Workshop Regional Modeling 2./3.11.2006
2 Way Coupling
Example: Atmosphere-Biosphere-Hydrosphere Modeling
Initialization
Initialization
MCCM
Time-Loop
MOBILE
MM5 Meteorology
Vegetation
development
CO
CO22 (atmospheric
(atmospheric // endogene)
endogene)
sugars
sugars
from
from starch
starch degradation
degradation
or
or xylem
xylem transport
transport
Chloroplast
Chloroplast
Triose-P
Triose-P
Triose-P
Triose-P
dxs
dxs
Antropogenic
Emissions
Acetyl-CoA
Acetyl-CoA
Vegetation
C & N balance
Canopy air
chemistry
Transport
& Mixing
Canopy & soil
physics
Mevalonate
Mevalonate
Atmospheric
Atmospheric
N-deposition
N-deposition
Atmospheric
Atmospheric
N-deposition
N-deposition
Plant
Plantlitter
litter
production
production
Plant
Plant N
N
Soil C & N
balance
Organic
Organic
matter
matter
Plant
Plant
N-uptake
N-uptake
Model Coupling
Interface
PEP
PEP
PEP
PEP
vegetation
vegetation developmentcanopy air
C- & Nchemistry
balance
Module
coupler
soil
physics
C- & N(canopy &
balance biosphere
soil)
water
balance
Ammonification
Ammonification
NH
NH44++
Nitrification
Nitrification
N
N22O
O
NO
NO33--→
→NO
NO22--→
→NO
NO→
→N
N22O
O→N
→N22
N
N22
NO
NO -- Denitrification
33
Denitrification
NH
NH44++→
→NH
NH22OH
OH→
→(NOH)
(NOH) →
→NO
NO22-- →
→NO
NO33-↑↓
↑↓
N
N22O
O++NO
NO
Death
Deathof
of
microbes
microbes
microbial
microbial
N-immobilisation
N-immobilisation
mikrobial
mikrobial biomass
biomass
NO
NO
NO
NO33---leaching
-leaching
Pyr
Pyr
DOXP
DOXP
dxr
dxr
NADPH
NADPH
MEP
MEP
HMG-CoA
HMG-CoA
Cytoplasm
Cytoplasm
Chemical
Transformation
Pyr
Pyr
IDP
IDP
DMADP
DMADP
??
NADP
NADP++
IsoS
IsoS
(3x
(3x IDP)
IDP) IDP
IDP ++ DMADP
DMADP
2x
2x IDP
IDP ++ DMADP
DMADP
FDP
FDP
Sesquiterpenes
Sesquiterpenes
GDP
GDP
GGDP
GGDP
Higher
Higher
isoprenoides
isoprenoides
Isoprene
Isoprene
climate/
air chemistry
Monoterpenes
Monoterpenes
MonoS
MonoS
Model
coupler
Data
regional
water
balance
N
N22-fixation
-fixation
Input / Output
Finalization
Water
balance
Termination
Fully coupled model system, under development
Regional Atmosphere-Biosphere-Hydrosphere Modeling
HGF-Workshop Regional Modeling 2./3.11.2006
2 Way Coupling
Example: Atmosphere-Biosphere-Hydrosphere Modeling
[∆O3 ppb]
Differences Atmo-Chemistry vs. Atmo-Bio-Hydro Model (August 2003)
Regional Atmosphere-Biosphere-Hydrosphere Modeling
HGF-Workshop Regional Modeling 2./3.11.2006
Future Requirements
Regional Atmosphere-Biosphere-Hydrosphere Modeling
HGF-Workshop Regional Modeling 2./3.11.2006
Challenges of Coupled Regional Model System(s)
1) Aggregation of stand alone 1d-modules to regional application:
problem of scales & problem of regional validation
usually only point measurements available
⇒ use of remote sensing techniques for areal validation necessary
- soil moisture dynamics
- sensible & latent heat fluxes (e.g. SEBAL method)
- areal vegetation dynamics, e.g. via LAI
but: coupling can provide new validation possibilities
2) Parameter estimation in coupled model system
- scale dependence of parameters
- non-uniqueness of solution due to large number of degrees of freedom
- multi-objective calibration of adjustable parameters & validation required
3) Specification of system parameters & initialization of state variables
(soil, vegetation, land surface, atmosphere, …)
4) Technical realization of data exchange between stand-alone modules!
Regional Atmosphere-Biosphere-Hydrosphere Modeling
HGF-Workshop Regional Modeling 2./3.11.2006
Final remarks
• Computational efficient realization of tight model coupling:
parallelization and efficient implementation on HPC platforms
• Application and validation to different vegetation and climate zones
• Scaling laws? Often pragmatically neglected!
laboratory scale ⇔ field scale ⇔ regional scale
Model coupling must be accompanied by scaling law investigation
• Open model architecture & data exchange
• Community based approach
• Robustness required to ensure usability by all disciplines involved
Regional Atmosphere-Biosphere-Hydrosphere Modeling
HGF-Workshop Regional Modeling 2./3.11.2006
Thank you for your attention
Regional Atmosphere-Biosphere-Hydrosphere Modeling
HGF-Workshop Regional Modeling 2./3.11.2006