model development towards incorporating 3-d subsurface ...introduction models coupling results...
TRANSCRIPT
Introduction Models Coupling Results Future work
Model Development Towards Incorporating 3-DSubsurface Hydrologic Processes within theCommunity Land Surface Model (CLM):
Coupling PFLOTRAN and CLM
Gautam Bisht1, Richard T. Mills1, Forrest M. Hoffman1, andPeter E. Thornton1
1Oak Ridge National Laboratory
CCSM Land Model Working Group15, March, 2011
Gautam Bisht PFLOTRAN - CLM Coupling 1
Introduction Models Coupling Results Future work
Motivation
Soil moisture plays an important role in Land-Atmosphereinteraction
FIG. 10. (top) Average of !P(S) " !P(W ) across all 12 models and (bottom) average of !P(S) " !P(W ) across the 8 models that,at a given grid cell, reproduce most closely the observed mean June–August (JJA) precipitation.
AUGUST 2006 K O S T E R E T A L . 605
Fig 10 live 4/C
Koster et al., 2006, J. HydroMetero.
Gautam Bisht PFLOTRAN - CLM Coupling 2
Introduction Models Coupling Results Future work
Motivation
GRACE satellite: Groundwater depletion over India
Rodell et al. (2009), Nature
Time period: Aug, 2002 to Oct, 2008Depletion rate: 1 m per 3 years
Gautam Bisht PFLOTRAN - CLM Coupling 3
Introduction Models Coupling Results Future work
Objective
Improve subsurface hydrologic processes:
Lateral transport of water
Unified treatment of unsaturated and saturated zone dynamics
Gautam Bisht PFLOTRAN - CLM Coupling 4
Introduction Models Coupling Results Future work
CLM
1) Hydrologic cycle 3) Biogeochemistry2) Biogeophysics 4) Dynamic Vegetation
PFLOTRAN will be replacing this part within CLM
Gautam Bisht PFLOTRAN - CLM Coupling 5
Introduction Models Coupling Results Future work
PFLOTRAN
Models Multiscale-Multiphase-Multicomponent Subsurface ReactiveFlows
Funded by SciDAC-2
PI: Peter C. Lichtner (LANL)
Development team: LANL, PNNL, ORNL, ANL, UIUC, Univ. ofUtah
http://www.nccs.gov/wp-content/uploads/2009/06/300yrs.pnghttp://www.pnl.gov/science/images/highlights/computing/uvi.jpg
Gautam Bisht PFLOTRAN - CLM Coupling 6
Introduction Models Coupling Results Future work
PFLOTRAN
Various Modes:
Richards
Multiphase
Multiple immiscible phases
Two phase liquid water-gas flow
Thermal-Hydrologic-Chemical
Reactive Transport
Written in Fortran 90, with a modular object-orientedapproach
Fully implicit time-stepping
Uses Portable, Extensible Toolkit for Scientific Computation(PETSc)
Gautam Bisht PFLOTRAN - CLM Coupling 7
Introduction Models Coupling Results Future work
PFLOTRAN
0
500
1000
1500
2000
2500
4092 8184 16380 32760 65532 131064
Tim
e in
sec
onds
Number of processor cores
PFLOTRAN on Cray XT5 and IBM BlueGene/P: 1 billion DoF problem, Transport stage
On Cray XT5On IBM BlueGene/P
Adaptive Mesh RefinementMimetic Finite Difference
PFLOTRAN is one of the six Early Science Applications for OLCF-3
Gautam Bisht PFLOTRAN - CLM Coupling 8
Introduction Models Coupling Results Future work
Governing Equation
Porosity
Poorly Sorted Fracture
IntragranularWell Sorted
Decreased Porosity
Dissolution
by Diogenesis
Diogenesis – The formation of rock; pores fill up with precipitations of mineral and reduce porosity
1.72, Groundwater Hydrology Lecture 2 Prof. Charles Harvey Page 4 of 10
Φ - Porositys - Saturation
q(x+δx)q(x)
q(z)
q(z+δz)
Saturation
Dep
th (z
)
1.00.0Unsaturatedzone
Saturated zone
q(y)
q(y+δy)Mass conservation
CLM PFLOTRAN
∂∂t (θ) = −∇ · (q) + qw
∂∂t (ρφs) = −∇ · (ρu) + Qw
q = −K (ψ)∇(ψ + z) u = −(κκr (s)µ
)∇(P + ρgz)
θ = φs K = ρgκµ P = ρgψ
Gautam Bisht PFLOTRAN - CLM Coupling 9
Introduction Models Coupling Results Future work
CLM : Conceptual Domain
Gautam Bisht PFLOTRAN - CLM Coupling 10
Introduction Models Coupling Results Future work
PFLOTRAN - CLM: Conceptual Domain
Gautam Bisht PFLOTRAN - CLM Coupling 11
Introduction Models Coupling Results Future work
PFLOTRAN - CLM: Soil properties initialized by CLM
Organic Percolating
OrganicNon-
PercolatingMineral
x
z
Computation of Kz
Gautam Bisht PFLOTRAN - CLM Coupling 12
Introduction Models Coupling Results Future work
PFLOTRAN - CLM: Soil properties initialized by CLM
Organic Percolating
OrganicNon-
PercolatingMineral
x
z
Computation of Kz
Organic Percolating
OrganicNon-
PercolatingMineral
Computation of Kx & Ky
Gautam Bisht PFLOTRAN - CLM Coupling 12
Introduction Models Coupling Results Future work
PFLOTRAN - CLM: Initial conditions
0.3Saturation Pressure
1 m
Dep
th
Dep
th
Gautam Bisht PFLOTRAN - CLM Coupling 13
Introduction Models Coupling Results Future work
PFLOTRAN - CLM: Time-stepping
CLM
PFLOTRAN
T T+δt
Gautam Bisht PFLOTRAN - CLM Coupling 14
Introduction Models Coupling Results Future work
PFLOTRAN - CLM: Time-stepping
PFLOTRAN
T T+δt
CLMCLM
Gautam Bisht PFLOTRAN - CLM Coupling 14
Introduction Models Coupling Results Future work
PFLOTRAN - CLM: Time-stepping
PFLOTRAN
T T+δt
CLMCLM
Gautam Bisht PFLOTRAN - CLM Coupling 14
Introduction Models Coupling Results Future work
PFLOTRAN - CLM: Time-stepping
PFLOTRAN
T T+δt
PFLOTRAN
Implicit time stepping
CLMCLM
Gautam Bisht PFLOTRAN - CLM Coupling 14
Introduction Models Coupling Results Future work
PFLOTRAN - CLM: Time-stepping
T T+δt
PFLOTRAN
CLMCLM
Gautam Bisht PFLOTRAN - CLM Coupling 14
Introduction Models Coupling Results Future work
PFLOTRAN - CLM: Time-stepping
T T+δt
PFLOTRAN
CLMCLM
Gautam Bisht PFLOTRAN - CLM Coupling 14
Introduction Models Coupling Results Future work
Experiment-1: Sanity check of coupling
CLM
PFLOTRAN PFLOTRAN
One-way forced PFLOTRAN
Offline PFLOTRAN
Gautam Bisht PFLOTRAN - CLM Coupling 15
Introduction Models Coupling Results Future work
Experiment-1: Sanity check of coupling
0.5 0.6 0.7 0.8 0.9
0
0.5
1
1.5
2
2.5
3
3.5
4
4.5
5
Saturation
Soi
l Dep
th [m
]
Time: t0
PFCLM−PF
0.5 0.6 0.7 0.8 0.9
0
0.5
1
1.5
2
2.5
3
3.5
4
4.5
5
Saturation
Soi
l Dep
th [m
]
Time: t0
PFCLM−PF
Gautam Bisht PFLOTRAN - CLM Coupling 16
Introduction Models Coupling Results Future work
Experiment-1: Sanity check of coupling
0.5 0.6 0.7 0.8 0.9
0
0.5
1
1.5
2
2.5
3
3.5
4
4.5
5
Saturation
Soi
l Dep
th [m
]
Time: t1
0.5 0.6 0.7 0.8 0.9
0
0.5
1
1.5
2
2.5
3
3.5
4
4.5
5
Saturation
Soi
l Dep
th [m
]
Time: t1
PFCLM−PF
PFCLM−PF
Gautam Bisht PFLOTRAN - CLM Coupling 16
Introduction Models Coupling Results Future work
Experiment-1: Sanity check of coupling
0.5 0.6 0.7 0.8 0.9
0
0.5
1
1.5
2
2.5
3
3.5
4
4.5
5
Saturation
Soi
l Dep
th [m
]
Time: t2
0.5 0.6 0.7 0.8 0.9
0
0.5
1
1.5
2
2.5
3
3.5
4
4.5
5
Saturation
Soi
l Dep
th [m
]
Time: t2
PFCLM−PF
PFCLM−PF
Gautam Bisht PFLOTRAN - CLM Coupling 16
Introduction Models Coupling Results Future work
Experiment-1: Sanity check of coupling
0.5 0.6 0.7 0.8 0.9
0
0.5
1
1.5
2
2.5
3
3.5
4
4.5
5
Saturation
Soi
l Dep
th [m
]
Time: t3
0.5 0.6 0.7 0.8 0.9
0
0.5
1
1.5
2
2.5
3
3.5
4
4.5
5
Saturation
Soi
l Dep
th [m
]
Time: t3
PFCLM−PF
PFCLM−PF
Gautam Bisht PFLOTRAN - CLM Coupling 16
Introduction Models Coupling Results Future work
Experiment-1: Sanity check of coupling
0.5 0.6 0.7 0.8 0.9
0
0.5
1
1.5
2
2.5
3
3.5
4
4.5
5
Saturation
Soi
l Dep
th [m
]
Time: t4
0.5 0.6 0.7 0.8 0.9
0
0.5
1
1.5
2
2.5
3
3.5
4
4.5
5
Saturation
Soi
l Dep
th [m
]
Time: t4
PFCLM−PF
PFCLM−PF
Gautam Bisht PFLOTRAN - CLM Coupling 16
Introduction Models Coupling Results Future work
Experiment-2: Single soil column simulation
PFLOTRAN
CLMCLM
CLM alone Two-way coupledCLM-PFLOTRAN
Experiment setup similar to Maxwell and Miller (2005, J. ofHydroMeterology) : Rain for 14 days
Gautam Bisht PFLOTRAN - CLM Coupling 17
Introduction Models Coupling Results Future work
Experiment-2: Single soil column simulation
10−2
10−1
100
101
102
0
0.5
1
1.5
2
2.5x 10
−3 Rain[m
m/s
ec]
10−2
10−1
100
101
102
0
50
100
150
200Shortwave Radiation
[W/s
ec−2
]
Time [days]
Gautam Bisht PFLOTRAN - CLM Coupling 18
Introduction Models Coupling Results Future work
Experiment-2: Single soil column simulation
10−2
10−1
100
101
102
0
0.5
1
1.5
2
2.5x 10
−3 Overland[m
m/s
ec]
CLM CLM−PFLOTRAN
10−2
10−1
100
101
102
0
0.5
1
1.5
2
2.5
3x 10
−3 Infiltration
[mm
/sec
]
CLM CLM−PFLOTRAN
Gautam Bisht PFLOTRAN - CLM Coupling 18
Introduction Models Coupling Results Future work
Experiment-2: Single soil column simulation
Gautam Bisht PFLOTRAN - CLM Coupling 18
Introduction Models Coupling Results Future work
Development of coupling with HPC in mind
0
1
2
3
4
5
6
7
8
9
10
11
048
121620242832
159
131721252933
12
13
14
15
16
17
18
19
20
21
22
23
26
10141822263034
3711151923273135
36 37 38 3940 41 42 4344 45 46 47
CLM domain PFLOTRAN domain
Processor - 0Processor - 1Processor - 2
Gautam Bisht PFLOTRAN - CLM Coupling 19
Introduction Models Coupling Results Future work
Development of coupling with HPC in mind
0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
048
121620242832
159
131721252933
26
10141822263034
3711151923273135
36 37 38 3940 41 42 4344 45 46 47
CLM domain PFLOTRAN domain
Processor - 0Processor - 1Processor - 2
Gautam Bisht PFLOTRAN - CLM Coupling 19
Introduction Models Coupling Results Future work
Development of coupling with HPC in mind
0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
048
121620242832
159
131721252933
26
10141822263034
3711151923273135
36 37 38 3940 41 42 4344 45 46 47
CLM domain PFLOTRAN domain
Processor - 0Processor - 1Processor - 2
Gautam Bisht PFLOTRAN - CLM Coupling 19
Introduction Models Coupling Results Future work
Next steps
Restart capability
Map projection
Application of the coupled model to a study site
Nutrient transport
Thermal Hydrologic transport
Gautam Bisht PFLOTRAN - CLM Coupling 20
Introduction Models Coupling Results Future work
Acknowledgements
DOE ASCR
PFLOTRAN team
Daniel Ricciuto
Gautam Bisht PFLOTRAN - CLM Coupling 21
Introduction Models Coupling Results Future work
Questions?
Gautam Bisht : [email protected]
Richard Mills : [email protected]
Forrest Hoffman : [email protected]
Peter Thornton : [email protected]
Gautam Bisht PFLOTRAN - CLM Coupling 22