mathematical modeling of natural and anthropogenic change of regional climate and environment
DESCRIPTION
International Conference on Environmental Observations, Modeling and Information Systems ENVIROMIS-2004 17-25 July 2004, Tomsk, Russia. Mathematical modeling of natural and anthropogenic change of regional climate and environment V.N. Lykosov - PowerPoint PPT PresentationTRANSCRIPT
-
International Conference on Environmental Observations, Modeling and Information Systems ENVIROMIS-200417-25 July 2004, Tomsk, Russia
Mathematical modeling of natural and anthropogenic change of regional climate and environment V.N. Lykosov Russian Academy of Sciences Institute for Numerical Mathematics, Moscow E-mail: [email protected]
-
Climate System
1. ATMOSPHERE the gas envelope of the Earth (oxygen, nitrogen, carbon dioxide, water vapor, ozone, etc.), which controls the solar radiation transport from space towards the Earth surface.
2. OCEAN the major water reservoir in the system, containing salted waters of the World ocean and its seas and absorbing the basic part of the incoming solar radiation (a powerful accumulator of energy).
3. LAND surface of continents with hydrological system (inland waters, wetlands and rivers), soil (e.g. with groundwater) and cryolithozone (permafrost).
4. CRYOSPHERE continental and see ice, snow cover and mountain glaciers.
5. BIOTA vegetation on the land and ocean, alive organisms in the air, water and soil, mankind.
-
The Climate System(T. Slingo, 2002)
-
3
-1.2
-1.17
-1.12
-1.2
-1.09
-0.78
-0.87
-0.98
-1.17
-1.13
-1.13
-1.08
-1.2
-1.2
-1.38
-1.39
-1.08
-1.2
-1.42
-1.46
-1.68
-1.48
-1.68
-2.02
-2.07
-2.02
-2.46
-2.38
-2.29
-2.23
-2.25
-2.32
-2.16
-1.94
-1.83
-1.61
-1.37
-1.22
-1.2
-1.28
-1.29
-1.28
-1.04
-0.87
-0.85
-0.91
-1.01
-1.1
-1
-0.63
-0.57
-0.6
-0.87
-0.99
Decades
C degrees
Annually mean air temperature in Khanty-Mansiisk for the time period from 1937 to 1999 (Khanty-Mansiisk Hydrometeocenter).
1
193719381939194019411942194319441945194619471948194919501951195219531954195519561957195819591960196119621963196419651966196719681969197019711972197319741975197619771978197919801981198219831984198519861987198819891990199119921993199419951996199719981999200020012002
3.7-1.5-0.3-8.2-3.8-2.5-3.35.8-0.72.40.52.49-8.80.34.42-0.26.70.3-0.43.83.1-3.36.9-1.5-3.7-1.20.2-0.6-2.1-10.2-4.91.8-11-5.9-5.133.6-4.83.9-4.1-2.66.2-2.45.46.6-0.51.2-33.9-3.3-4.93-0.76.747.1-1.4-6.92.1-1.2-1.2-4.33.9
19371936193719381939194019411942194319441945194619471948194919501951195219531954195519561957195819591960196119621963196419651966196719681969197019711972197319741975197619771978197919801981198219831984198519861987198819891990199119921993199419951996199719981999200020012002
-1.5-3.42.12.70.7-2-41.43.9-1.9-0.51.6-0.20.83.93.2-2.9-0.61.3-2.72.1-5.5-52.42.5-5.5-5.64.6-1.55.75.6-2.53.6-1.4-31.92.14.2-0.21.31.92.1-0.5-0.9-4.73.15.33.64.32.15.97.83.5-4.71.32.83.45.83.25.73.4-5.562.14.4
1937-461938-471939-481940-491941-501942-511943-521944-531945-541946-551947-561948-571949-581950-591951-601952-611953-621954-631955-641956-651957-661958-661959-681960-691961-701962-711963-721964-731965-741966-751967-761968-771969-781970-791971-801972-811973-821974-831975-841976-851977-861978-871979-881980-891981-901982-911983-921984-931985-941986-951987-961988-971989-981990-99
189191192192193195193192191191190188188188186186189188187186189188186185187183184183184183183182182182184183188189188188192193191191190190188186187190189190190190
1937-461938-471939-481940-491941-501942-511943-521944-531945-541946-551947-561948-571949-581950-591951-601952-611953-621954-631955-641956-651957-661958-661959-681960-691961-701962-711963-721964-731965-741966-751967-761968-771969-781970-791971-801972-811973-821974-831975-841976-851977-861978-871979-881980-891981-901982-911983-921984-931985-941986-951987-961988-971989-981990-99
-1.2-1.17-1.12-1.2-1.09-0.78-0.87-0.98-1.17-1.13-1.13-1.08-1.2-1.2-1.38-1.39-1.08-1.2-1.42-1.46-1.68-1.48-1.68-2.02-2.07-2.02-2.46-2.38-2.29-2.23-2.25-2.32-2.16-1.94-1.83-1.61-1.37-1.22-1.2-1.28-1.29-1.28-1.04-0.87-0.85-0.91-1.01-1.1-1-0.63-0.57-0.6-0.87-0.99
1
00
00
00
00
00
00
00
00
00
00
00
00
00
00
00
00
00
00
00
00
00
00
00
00
00
00
00
00
00
00
00
00
00
00
00
00
00
00
00
00
00
00
00
00
00
00
00
00
00
00
00
00
00
00
00
00
00
00
00
00
00
00
00
00
00
00
00
00
1936-2002 ( -10,8)
2
00
00
00
00
00
00
00
00
00
00
00
00
00
00
00
00
00
00
00
00
00
00
00
00
00
00
00
00
00
00
00
00
00
00
00
00
00
00
00
00
00
00
00
00
00
00
00
00
00
00
00
00
00
00
00
3
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
-
1
3.912.810.46.4
7.812.310.46.4
6.711.310.46.4
7.211.910.46.4
3.711.710.46.4
3.77.510.46.4
7.811.810.46.4
14.521.710.46.4
16.323.210.46.4
17.221.810.46.4
6.712.810.46.4
10.719.110.46.4
14.620.510.46.4
9.817.710.46.4
5.4910.46.4
8.714.910.46.4
13.620.210.46.4
4.123.810.46.4
9.622.110.46.4
8.613.110.46.4
10.615.810.46.4
13.220.710.46.4
18.724.310.46.4
16.320.210.46.4
11.815.610.46.4
7.814.610.46.4
5.110.710.46.4
13.921.910.46.4
15.620.710.46.4
9.414.310.46.4
5.911.310.46.4
C degrees
Air temperature. May 2003 (Khanty-Mansiisk Hydrometeocenter)
2
0.4
3.9
6.3
5.4
2.2
0
0
7.3
0
0
4.3
1.2
4.9
27.9
64 . 47
(). 2003.
1
2003.
-
13.912.810.40-6.36.400
27.812.310.401.86.4000.4
36.711.310.4036.4003.9
47.211.910.4046.4006.3
53.711.710.402.36.4005.4
63.77.510.401.46.4002.2
77.811.810.402.56.400
814.521.710.406.96.400
916.323.210.4011.76.400
1017.221.810.4012.86.4000
116.712.810.4036.4000
1210.719.110.401.56.400
1314.620.510.407.46.400
149.817.710.406.56.4007.3
155.4910.402.16.400
168.714.910.402.26.4000
1713.620.210.405.86.400
184.123.810.408.96.400
199.622.110.4014.56.400
208.613.110.4036.4000
2110.615.810.403.36.400
2213.220.710.403.66.400
2318.724.310.4013.86.400
2416.320.210.4013.96.4004.3
2511.815.610.407.16.4001.2
267.814.610.404.66.4004.9
275.110.710.4016.400
2813.921.910.406.96.400
2915.620.710.4010.16.400
309.414.310.4066.400
315.911.310.4036.40027.9
509.3168.363.8
-
Features of the climate system as physical object-IBasic components of the climate system atmosphere and ocean are thin films with the ratio of vertical scale to horizontal scale about 0.01-0.001.
On global and also regional spatial scales, the system can be considered as quasi-twodimensional one. However, its density vertical stratification is very important for correct description of energy cycle. Characteristic time scales of energetically important physical processes cover the interval from 1 second (turbulence) to tens and hundreds years (climate and environment variability).
Laboratory modelling of such system is very difficult.
-
Features of the climate system as physical object-II
It is practically impossible to carry out specialized physical experiments with the climate system. For example, we have no possibility to pump the atmosphere by the carbon dioxide and, keeping other conditions, to measure the system response. We have shirtterm series of observational data for some of components of the climate system.
Conclusion: the basic (but not single) tool to study the climate system dynamics is mathematical (numerical) modeling.
Hydrodynamical climate models are based on global models of the atmosphere and ocean circulation.
-
Objectives of climate modelingTo reproduce both climatology (seasonal and monthly means) and statistics of variability: intra-seasonal (monsoon cycle, characteristics of storm-tracks, etc.) and climatic (dominated modes of inter-annual variability such as El-Nino phenomenon or Arctic Oscillation)
To estimate climate change due to anthropogenic activity
To reproduce with high degree of details regional climate: features of hydrological cycle, extreme events, impact of global climate change on regional climate, environment and socio-economic relationships
Fundamental question (V.P. Dymnikov): what climatic parameters and in what accuracy must by reproduced by a mathematical model of the climate system to make its sensitivity to small perturbations of external forcing close to the sensitivity of the actual climate system?
-
Computational technologies
Global climate model (e.g. model with improved spatial resolution in the region under consideration) implemented on computational system of parallel architecture (CSPA)
Methods of regionalization: 1) statistical approach (downscaling);
2) hydrodynamical mesoscale simulation ( e.g., mesoscale model 5) requires CSPA; 3) large-eddy simulation of geophysical boundary layers (requires CSPA)
Assessment of global climate change and technological impact on regional environment
-
Observational data to verify models1) ECMWF reanalysis ERA-15 (1979-1993 ..), ERA-40 (1957-2001) http://www.ecmwf.int/research/era
2) NCEP/NCAR (National Center of Environment Protection/National Center of Atmospheric Research, USA), 1958-1997, http://wesley.wwb.noaa.gov/reanalysis.htm
3) Precipitation from 1979 to present time http://www.cpc.ncep.noaa.gov/products/globalprecip 4) Archive NDP048, containing multiyear data of routine observations on 225 meteorological stations of the former USSR http://cdiac.esd.oml.gov/ftp/ndp048 Numerical experiments with modern global climate models produce a large amount of data (up to 1Gb for 1 month). This requires special efforts for its visualization, postprocessing and analysis.
-
Large-scale hydrothermodynamics of the atmosphere Subgrid-scaleprocessesparameterization
,
,
,
,
,
where
.
_1119850510.unknown
_1119850544.unknown
_1120137379.unknown
_1120137402.unknown
_1119850556.unknown
_1119850536.unknown
_1119850482.unknown
-
Parameterization of subgrid-scale processes Turbulence in the atmospheric boundary layer, upper ocean layer and bottom boundary layer Convection and orographic waves
Diabatic heat sources (radiative and phase changes, cloudiness, precipitation, etc.)
Carbon dioxide cycle and photochemical transformations Heat, moisture and solute transport in the vegetation and snow cover Production and transport of the soil methane
Etc.
-
T.J. Philips et al. (2002). Large-Scale Validation of AMIP II Land-Surface Simulations
-
Table 2. Model codes and features of the sixteen AMIP2 models analysed in Zhang et al. (2002)
Code
Resolution
Land-surface components
No. of layers in soil temp. calculations
No. of layers in soil moist. calculations
Model
Country
Soil model complexity
Canopy representation
A
B
C
D
E
F
G
H
I
J
K
L
M
N
O
P
T42L18
T63L45
4x5 L21
T159L50
T63L30
T42L18
T62L18
T42L18
3.75x2.5 L58
3.75x2.5 L19
T47L32
4x5 L20
T42L30
T42L18
4x5 L24
4x5 L15
bucket
force-restore
multi-layer diffusion
multi-layer diffusion
multi-layer diffusion
multi-layer diffusion
multi-layer diffusion
multi-layer diffusion
multi-layer diffusion
multi-layer diffusion
multi-layer diffusion
multi-layer diffusion
multi-layer diffusion
multi-layer diffusion
bucket
bucket
const. canopy resistance
intercept. + transpiration
intercept. + transpiration
intercept. + transpiration
intercept. + transpiration
intercept.+transpiration+CO2
intercept. + transpiration
intercept. + transpiration
intercept.+transpiration+CO2
intercept.+transpiration+CO2
intercept. + transpiration
intercept. + transpiration
intercept. + transpiration
intercept.+transpiration+CO2
no
no
3
2
24
4
4
6
3
2
4
4
3
2
3
6
1
1
1
2
24
4
3
6
2
3
4
4
3
3
3
6
1
1
CCSR, Japan
CNRM, France
INM, Russia
ECMWF, UK
JMA, Japan
NCAR, USA
NCEP, USA
PNNL, USA
UGAMP, UK
UKMO, UK
CCCMA, Can
GLA, USA
MRI, Japan
SUNYA, USA
UIUC, USA
YONU, Korea
PAGE
35
-
The Taylor diagram for the variability of the latent heat flux at the land surface as follows from results of AMIP-II experiments (Irannejad et al., 2002).
-
Sensitivity of the climate system to small perturbations of external forcing
(invited lecture at the World Climate Conference, Moscow, 29 September 3 October, 2003)
V.P. Dymnikov, E.M. Volodin, V.Ya. Galin, A.S. Gritsoun, A.V. Glazunov, N.A. Diansky, V.N. LykosovInstitute of Numerical Mathematics RAS, Moscow
-
INM coupled atmosphere - ocean general circulation model
AGCM
- Finite difference model with spatial resolution 5x4 and 21 levels in sigma-coordinates from the surface up to 10 hPa.
- In radiation absorption of water vapour, clouds, CO2, O3, CH4, N2O, O2 and aerosol are taken into account. Solar spectrum is divided by 18 intervals, while infrared spectrum is divided by 10 intervals.
- Deep convection, orographic and non-orographic gravity wave drag are considered in the model. Soil and vegetation processes are taken into account.
||
Non-flux-adjusted coupling
||
OGCM
-The model is based on the primitive equations of the ocean dynamics in spherical sigma-coordinate system. It uses the splitting-up method in physical processes and spatial coordinates. Model horizontal resolution is 2.5x2, it has 33 unequal levels in the vertical with an exponential distribution.
-
The climate model sensitivity to the increasing of CO2
CMIP - Coupled Model Intercomparison Projecthttp://www-pcmdi.llnl.gov/cmipCMIP collects output from global coupled ocean-atmosphere general circulation models (about 30 coupled GCMs). Among other usage, such models are employed both to detect anthropogenic effects in the climate record of the past century and to project future climatic changes due to human production of greenhouse gases and aerosols.
-
Response to the increasing of CO2CMIP models (averaged)INM model
-
Global warming in CMIP models in CO2 run and parameterization of lower inversion clouds
T - global warming (K), LC - parameterization of lower inversion clouds (+ parameterization was included, - no parameterization, ? - model description is not available). Models are ordered by reduction of global warming.
Model
T
LC
NCAR-WM
3.77
?
GFDL
2.06
-
LMD
1.97
-
CCC
1.93
-
UKM03
1.86
-
CERF
1.83
-
CCSR
1.75
-
CSIRO
1.73
+
GISS
1.70
-
UKMO
1.59
-
BMRC
1.54
+
ECHAM3
1.54
-
MRI
1.50
-
IAP
1.48
+
NCAR-CSM
1.26
+
PCM
1.14
+
INM
0.99
+
NRL
0.75
+
-
Mesoscale non-hydrostatic modelingMM5 - Penn State/NCAR Mesoscale Modeling System http://www.mmm.ucar.edu/mm5Program code: Fortran77, Fortran90, C. Hybrid parallelization (shared and distributed memory) + vectorizationDocumentationImplemented by V. Gloukhov (SRCC/MSU) on MVS-1000M
-
International Conference on Computational Mathematics ICCM-2004, 21-25 June, 2004, Novosibirsk, Russia
Large-Eddy Simulation of Geophysical Boundary Layers on Parallel Computational Systems V.N. Lykosov, A.V. Glazunov Russian Academy of Sciences Institute for Numerical Mathematics, Moscow E-mail: [email protected], [email protected]
-
Geophysical Boundary Layers (GBLs) as elements of the Earth climate systemAtmospheric Boundary Layer HABL ~ 102 - 103 mOceanic Upper Layer HUOL ~ 101 - 102 mOceanic Bottom Layer HOBL ~ 100 - 101 m
GBL processes control:
1) transformation of the solar radiation energy at the atmosphere-Earth interface into energy of atmospheric and oceanic motions
2) dissipation of the whole Earth climate system kinetic energy
3) heat- and moisture transport between atmosphere and soil (e.g. permafrost), sea and underlying ground (e.g. frozen one).
-
Dynamic structure of GBLs
Three types of motion:
totally organized mean flowcoherent semi-organized structures (large eddies and waves)chaotic three-dimensional turbulence
-
Turbulence in PBLs
Rough surface Large scales Stratification
-
Inertial range Dissipation range Energy range Synoptical variationsBoundary-Layer flows
-
Differential formulation of models
Models are based on Reynolds type equations obtained after spatial averaging of Navier-Stokes equations and added by equations of heat and moisture (or salt):
-
Turbulent ClosureEquation for turbulent kinetic energy (of subgrid-scale motions): Equation for turbulent kinetic energy dissipation:Constraint on maximal value of sub-grid turbulence length scale:
-
Finite-difference approximation on C grid
Explicit time scheme of predictor-corrector type (Matsuno scheme)
Numerical scheme
-
Calculation of tendenciesdiffusionCoriolis forcegravityadvection
pressure gradientdiffusionadvectiondiffusionadvectionBoundary conditionsVelocity components
Calculation of eddy viscosity and diffusioncoefficientsSumming up of tendencies, dissipationTemperature, moisture, salinitySolver for the Poisson equationCalculation of the Poisson equation R.H.S., dissipationproduction, dissipation, non-linear terms velocity componentstemperature, moisture, salinity
, dissipation
-
Parallel version of models is developed to be mainly used on supercomputers with distributed memoryProcesor-to-processor data exchange is realized with the use of MPI standard non-blocked functions of the data transfer-receive 3-D decomposition of computational domain on each time step, processes are co-exchanged only by data which belongs to boundary grid cells of decomposition domains The Random Access Memory (operative memory) is dynamically distributed between processors (the features of FORTRAN-90 are used) Debuging and testing of parallel versions of models is executed on supercomputer MVS1000- of Joint Supercomputer Center (768 processors, peak productivity - 1Tflops)
PARALLEL IMPLEMENTATION
-
Extreme case:
Domain 768 x 768 x 256 (= 150 994994) grid points
574 processors
3D decomposition (12 x 12x 4)
MGD Poisson solver
15 Gbytes of memory
1 step of scheme ~ 20 s of computer time.
EMBED MSGraph.Chart.8 \s
_1113840107.xls
-
Spectra of kinetic energy calculated using results of large-eddy simulation of the convective upper oceanic layer under different spatial resolution (m3)
-
von Karman votrex street behind a round cylinder, Re =200: top - from . (1986). , bottom - model results
-
Turbulent flow between buildingsWind 2.24 /s
-
Lagrangian transport of fine-dispersive particle tracer
Large number of particles in turbulent flow generated by the LES-model of the atmospheric boundary layer.
The calculation of trajectories is carried out simultaneously with numerical integration of the hydrodynamic part of the model.
It is assumed that each of particles has non-zero mass and size. It is possible to sort particles in few groups
accordingly to their size and density.
The code of tracer transport is parallelized with the use of MPI. The data exchange between processors is realized with the help of non-blocking transfers and takes place on the background of basic calculations.
Using results of calculations, the spatial distribution of the particle tracer concentration is calculated for each time step.
On parallel computational system this algorithm allows to simultaneously calculate trajectories of tens millions particles. The time which is needed to calculate the particles transport is significantly less then the time of the ABL model integration.
Equation of the particle motion in the air flow:
where xi (i=1,2,3) particle coordinates, m mass of particle, f drag force.
To calculate f , the Stokes formula is used:
r , , V , .
EMBED Equation.3
EMBED Equation.3
_1143542808.unknown
_1148897591.unknown
-
An example of the particle transport by the turbulent flow between buildings Wind Particle concentrationParticles are ejecting near the surface (along the dotted line). The maximal number of particles is about 10 000 000.
-
SnowUpper iceWaterGroundLower iceUH,LEEsEaSThermodynamics of shallow Reservoir (Stepanenko & Lykosov, 2003, 2004)
1) One-dimensional approximation.
2) On the upper boundary: fluxes of momentum, sensible and latent heat, solar and long-wave radiation are calculated On the lower boundary: fluxes are prescribed
3) Water and ice: heat transport Snow and ground: heat- and moisture transport
U wind velocityH sensible heat fluxLE latent heat fluxS shirt-wave radiationEa incoming long-wave radiation Es outgoing long-wave radiation
-
Mathematical formulation
- for water and ice:
, - heat conductivity - for snow:
- temperature
- liquid water
- for ground:
- temperature
- liquid water
- ice
-
Kolpashevo
-
Simulated snow surface temperature versus measured one on meteorologicalstation Kolpashevo (1961)Quality of snow surface temperature reproduction is an indicator of quality of heat transfer parameterization in the atmospheric surface layer snow system.
Accuracy of temperature measurementson meteorological station is about 0.5 .
-
Syrdakh Lake
-
Ground temperature under Syrdakh Lake (modeling results)
As follows from results of numerical experiments, the talik is stably existing during all of integration time (20 years, 1965-1984). Its depth varies from 1.2 to 2 meters under the lake bottom.
-
THANK YOU for YOUR ATTENTION