LDRD 04

Report
A dynamically-coupled groundwater, land surface
and regional climate model to predict seasonal
watershed flow and groundwater response
Reed M. Maxwell1, Stefan J. Kollet1, Qingyun Duan1
and Fotini K. Chow2
1Atmospheric,
Earth, and Energy Sciences Dept, Lawrence
Livermore National Lab
2Civil and Environmental Engineering Dept, University of
California, Berkeley
This work was performed under the auspices of the U.S. Department of Energy by University of California,
Lawrence Livermore National Laboratory under contract No. W-7405-Eng-48.
UCRL-PRES-XXXXXX
Three talks for the price of one
• Brief overview on RC/LS/OF/GW coupled model project
underway at LLNL
• Some details of a new coupled overland flow-groundwater
work (Kollet and Maxwell, 2005)
• Example problem that illustrates distributed LS/OF/GW
modeling
Coupled Model
provides much
more accurate
predictions of:
Runoff (mm/month)
Results of first coupled* model study show important
gains resulting from process feedback
250
200
150
100
50
0
J-68
D-68
Observations
D-69
Coupled
Uncoupled
• Runoff
Observations
• Soil moisture
• Water
balance
-1
Coupled
Uncoupled
*Maxwell and Miller, J. Hydromet,6(3), 2005. Others have shown similar results.
D-70
This project integrates four models in a
unique way
Land surface model
This project integrates four models in a
unique way
Land surface model
This project integrates four models in a
unique way
Land surface model
Overland flow
This project integrates four models in a
unique way
Regional climate model
Land surface model
Overland flow
This project integrates four models in a
unique way
Explicitly incorporates
fluxes at air/landsurface/subsurface
interfaces
Regional climate model
Precipitation/Advection
Land surface model
Runoff/Routing
Moisture/heat flux
Evapotranspiration
Infiltration/Seepage
Overland flow
Project tasks, details
• Run RCM over central US w/ a detailed study area over Little
Washita watershed
– SGP/ARM site
– Data to validate all models (need lots)
• P1 run in a nested mode, RCM and GW/LS/OF models
uncoupled (control run)
– RCM (ARPS) passes LS forcing to coupled model
– Coupled model spun up w/ obs, forced by RCM
• P2 fully couple models, re-run
– Verify that models are coupled and balancing mass and energy
– Look at water, heat fluxes across the LS and at weather and
weather generating processes
Surface contours for 1 km resolution grid,
07/08/99 12 UTC
RCM/ARPS Results
• Norman Oklahoma sounding comparison 07/08/99, 12 UTC
θ [K]
U [m/s]
φ[degrees]
q [g/kg]
Sacramento Model Calibration for Little Washita
Borehole data used to create 3D geostatistical
realization of the subsurface
Free-surface overland-flow boundary condition,
coupled groundwater overland flow
• Lots of motivation for coupled model
– Watershed modeling
– Climate
– Water quality
• Most (all?) coupled models rely on interface between SW and
GW
– “conductance concept”
– Hard to find field data to support this
– Need for a more general formulation
• Desire for parallel model w/ a robust non-linear solver
– Integrate into ParFlow
– Take advantage of infrastructure
Kollet and Maxwell, Advances in Water Resources, in press, 2005.
The Conductance Concept
 s

 v  s  q r ( x )  q e  x 
t
qe x  x s  p 
Ss Sw
 p
t

S w  p 
t
   q  q s  m' qe
New Overland Flow Boundary in ParFlow
 k x k r   z   
Ss Sw
  ,0
t

 v  ,0  q( x)
S  

 w
   q  qs
t
t
Kollet & Maxwell, 2005
Verification Examples
1D Slope:
200 min rain, 100 min recession
Jabar & Mohtar, 2004
2D Tilted V-Catchment:
90 min rain, 90 min recession
Panday & Huyakorn, 2004
Simulation Examples II
Low-K slab
Simulation Examples III
Random (Gaussian) Heterogeneity
Five Realizations
Kgeo = qrain
Scaled Parallel Efficiency
E(np,p) = T(n,1) / T(np,p)
Integrating land surface processes into ParFlow
an Example
• We add in LS processes (parts of CLM) into ParFlow
– Hydro, Runoff handled by PF as detailed earlier
– Fully distributed
– Fully parallel
• Use a well-resolved, large scale 2D “Classic” Example
problem to investigate coupled model performance and
behavior
– Dx=100m; Dz=2m; 40km x 0.54km domain
– Toth Problem: sinusoidal topography in a large basin
– J. Toth, 1963. A theoretical analysis of groundwater flow in small
drainage basins. J. Geophys. Research 68:4795-4842.
• Forced uniformly with PILPS midlatitude for one year
Toth problem uses an sinusoidal topography
Initial Pressure
July Pressure
Coupled model forced by PILPS midlatitude, produces
realistic looking hydrograph
Averaged, Cumulative ET
Distributed Ground Surface Temperature
Non-uniform
freezing
Distribution of
temps, non-uniform
water content
Non-uniform thaw
Initialization
Distributed Water and Heat Fluxes
Summary
• We are working on lots of stuff, but have a lot yet to do
– Coupled RC/LS/OF/GW project in Y1/control run phase, soon will
start dynamic coupling
– Coupled overland flow and groundwater method looks very
promising
– Integrating LS processes into GW provides interesting
distributed results need to compare to field site (Little Washita,
Valdai)
– Still overall question regarding quantifying impacts and scale of
coupled processes

similar documents