model development towards incorporating 3-d subsurface ...introduction models coupling results...

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


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.



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



Objective

Improve subsurface hydrologic processes:

Lateral transport of water

Unified treatment of unsaturated and saturated zone dynamics



CLM

1) Hydrologic cycle 3) Biogeochemistry2) Biogeophysics 4) Dynamic Vegetation

PFLOTRAN will be replacing this part within CLM



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



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)



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



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



CLM : Conceptual Domain



PFLOTRAN - CLM: Conceptual Domain



PFLOTRAN - CLM: Soil properties initialized by CLM

Organic Percolating

OrganicNon-

PercolatingMineral

x

z

Computation of Kz



PFLOTRAN - CLM: Soil properties initialized by CLM

Organic Percolating

OrganicNon-

PercolatingMineral

x

z

Computation of Kz

Organic Percolating

OrganicNon-

PercolatingMineral

Computation of Kx & Ky



PFLOTRAN - CLM: Initial conditions

0.3Saturation Pressure

1 m

Dep

th

Dep

th



PFLOTRAN - CLM: Time-stepping

CLM

PFLOTRAN

T T+δt




PFLOTRAN

T T+δt

CLMCLM




PFLOTRAN

T T+δt

PFLOTRAN

Implicit time stepping

CLMCLM




T T+δt

PFLOTRAN

CLMCLM



Experiment-1: Sanity check of coupling

CLM

PFLOTRAN PFLOTRAN

One-way forced PFLOTRAN

Offline PFLOTRAN




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




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




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




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




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



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




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]




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



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



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



Next steps

Restart capability

Map projection

Application of the coupled model to a study site

Nutrient transport

Thermal Hydrologic transport



Acknowledgements

DOE ASCR

PFLOTRAN team

Daniel Ricciuto



Questions?

Gautam Bisht : [email protected]

Richard Mills : [email protected]

Forrest Hoffman : [email protected]

Peter Thornton : [email protected]


model development towards incorporating 3-d subsurface ...introduction models coupling results...

Documents