Modeling Investigation of Water Partitioning at a Semiarid Hillslope

1
Modeling Investigation of Water Partitioning at a
Semi-arid Hillslope

• Huade Guan, John L. Wilson
• Dept. of Earth and Environmental Science, NMT
• Brent D. Newman
• Earth and Environmental Sciences Division , LANL
• Jirka Simunek
• Department of Environmental Sciences, UCR
• AGU Fall, 2003

2
Acknowledgements

• The analysis in this presentation was supported
  by SAHRA
• the NSF Science and Technology Center
• for Sustainability of semi-Arid Hydrology and
  Riparian Areas
• Site data was collected as part of the Los Alamos
  Environmental Restoration Project
• Modifications to the numerical code were funded
  by NSF grant, SAHRA, and Swedish Research Council

3
Motivation Mountain Front Recharge
Is distributed mountain block recharge
significant?
4
Hillslope scale
Preliminary (generic) simulations
5
Field site ponderosa pine hillslope at a
semi-arid area
Figures from Wilcox et al. (1997)
6
Objectives of this study

• Use numerical modeling to synthesize the
  observations and previous generic simulations
• Is the percolation into the bedrock really
  negligible?
• It wasn't directly observed, just inferred.
• If it is negligible, why?
• What impedes downward movement of water into the
  highly permeable tuff?
• For what situations will percolation to
  bedrock
• become significant for this climate?
• and with this permeable volcanic bedrock?

7
What we know and dont know

• We know
• Soil horizons and hydraulic parameters
• Root density profile
• Precipitation and other meteoric parameters
• Soil moisture
• Surface runoff and interflow
• Root-derived macropore flow
• We dont know
• ET
• Percolation

8
Modeling challenges

• Modeling ET
• System-dependent ET model
• Appropriate root-water-uptake model
• Modeling macropores
• Root-derived macropores
• Sub-parallel to the slope
• Numerical issues
• Highly non-linear, coupled processes
• Dual permeability

We used a modified version of HYDRUS-2D
9
Hillslope setting
Moisture profiles at three seasons, 1993
Figure from Wilcox et al. (1997)
10
ET modeling

• ET accounts for 95 of the annual water budget
  (Brandes and Wilcox, 2000)
• ET modeling

11
Calibration of ET modelillustrated using
measured moisture profiles for 4 of 19 sampled
days
PE70, PT30, h4-50m Root density ABw2.0,
Bt0.3
PE50, PT50, h4-50m Root density ABw0.59,
Bt0.4
PE70, PT30, h4-50m Root density ABw0.65,
Bt0.35
PE70, PT30, h4-15m Root density ABw2.0,
Bt0.3
PE70, PT30, h4-15m Root density ABw0.65,
Bt0.35

12
Representing root-derived macropores
1. Annular root macropore aperture
3. Equivalent root dip angle
2. Radial root distribution
13
Conceptual models for macropore flow

• Control Model without macropores
• Single continuum (sc)
• Models with macropores
• Single continuum with anisotropic K with three
  root dips (x11, x215, x330)
• Composite continuum (cc)
• Dual permeability model (dp)

14
Simulated1994 water balance
15
Simulated and observed runoff
No macropore (sc)
Composite continuum (cc)
simulated
Observation
observed
Macropore, ß1 (x1)
Macropore, ß15 (x2)
Macropore, ß30(x3)
16
Results of best-fit simulation(x2)
Infiltration (cm) ET 48.5 46.0 Runoff
Interflow Percolation 3.0 0
0.38 (0.7P)
17
What happens if root-zone directly contacts the
tuff?
simulated
observed
Infiltration (cm) ET 48.5 (x2 48.5)
39.3 (46.0) Runoff Interflow
Percolation 3.0 (3.0) 0 (0) 5.0,
10.0P (0.38, 0.7P)
18
Conclusions

• The simulated percolation across the soil-bedrock
  interface at this site is less than 1 of annual
  precipitation, in good agreement with previously
  inferred.
• The simulation results are consistence with
  Wilcox et als (1997) alternative hypothesis that
  the CB horizon, without roots, behaves as a
  barrier to downward movement of water into the
  bedrock.
• The results also indicates that sub-horizontal
  root-derived macropore flow increases the
  infiltration capacity and decreases surface
  runoff at this site.
• In this climate, at a location with a shallower
  soil layer where the root zone contacts the
  highly permeable tuff, percolation can be as
  large as 10 of the annual precipitation.

19
The End
20
Implication about the ET model
Feddes model overestimate ET loss based on the
observed wilting point (h4). S-shape model is
better if the numerical instability can be
avoided.