1. Koichi Sakaguchi and Xubin Zeng
Effects of soil dryness, plant litter and under-canopy atmospheric stability on ground evaporation in NCAR CLM3.5
Koichi Sakaguchi and Xubin Zeng
Department of Atmospheric Sciences, The University of Arizona
For CCSM Land Model Working Group Meeting

2. Improvement to ET components by soil resistance
(Oleson et al., 2008)
LH flux
(W/m2)
Monthly mean diurnal cycle
From Stöckli et al., 2008
ET in other spatial & temporal scales were also improved

3. Issues with soil resistance
Physically inconsistent over wet soil
(rsoil= 52 s/m at saturation)
Harvard forest, 2003 simulation
Blue:
Green:
Explain that the soil resistance was derived by Sellers et al., 1992 using the 1987 FIFE data set, which was also used by Song et al., 1997.
2. Implicitly includes the effect of plant litter
Mesoscale model study by Song et al., 1997 :
high litter content reduced evaporation at FIFE site

4. Our approach Revisited surface resistance / moisture limitation
b) New formula for the explicit effect of plant litter: rlitter
c) Developed a formula for under-canopy atmospheric stability
Kondo et al., 1990; Mahfouf & Niolhanm, 1991; Lee & Pielke 1992
Bristow et al.,1986; Bussiére & Cellier, 1994; Schaap and Bouten, 1997 +
Baldocchi et al., 2000; Niu and Yang, 2004; Miller et al., 2007;

5. a) Revisited surface resistance
CLM3.5
Rsoil ≈ 300 s/m
Reduce evaporation only when the soil is dry
Based on Lee & Pielke, 1992 or
New β
Field capacity
θ(volumetric soil moisture content)
Mention the value of the soil resistance for the field capacity = 300

6. b) Plant litter resistance
rlitter with Lllitter =1
10,000
1,000
(s/m)
100
Converge to zero when Llitter = 0
- Consider only dry plant litter
Snow cover effect is considered
Assumed to be isothermal with soil
For this study, Llitter = 1.0
u* (m/s)
Magnitude is comparable to soil resistance and observed litter resistance (Schaap and Bouten,1997)
Applied only over vegetated surface

7. c) Stability effect of under-canopy air
CLM3.5
with
Ts > Tg (stable)
Ts < Tg (unstable)
Functional form
maximum reduction of Cs,dense is 1/6 for stable condition, to be conservative
The model was not sensitive to modification for unstable condition

8. S8
Average over global land from
Eg % of total ET: [New formulation] - [CLM3.5] global mean difference = -0.4%
(%)

9. Monthly Eg in Western U.S.
Eg % of total ET
100 50
Eg fraction is excluded when the evaporation components are very small or negative, because Eg fraction becomes unreasonable values such as -10, +5, etc.
Lat: 30 ~ 50°N
Lon: 105 ~ 125°W
LAI+SAI = 1 ~ 1.7
(Months with negative or very small E are excluded) β
Litter
Stability
Sparse canopy
Small difference

10. Litter S10 Monthly Eg in Europe Eg % of total ET
100 50
Barbour et al., 2005: 20~40%
Williams et al., 2004: 20~30%
Similar result was found over China.
Barbour et al., 2005
Low LAI (<3), low-productivity, temperate rain forest with dominant conifer trees, New Zealand: Eg is 20~40% of total ET
Williams et al., 2004.
10-day study of semi-arid ecosystem (Olive orchard) . Eg is about 20~30%.
Eg is a large component
of total ET.
Lat: 45 ~ 60°N
Lon: 0 ~ 32°E
LAI+SAI = 1.2 ~ 2.8
Intermediate canopy
Thickness β
Litter
Stability
Larger improvement on ET partitioning

11. Monthly Eg in Amazon forest
Eg % of total ET
100 50
(Eg% = 4% in Choudhury 1998)
But Eg is small fraction of total ET.
Mention that the stability correction is as important as the soil resistance, especially in eastern part of Amazon forest where there is a strong dry season.
Choudhury 1998’s Amazon ET components shows Eg of 4%
Lat: 11°S ~ 11°N
Lon: 50 ~ 72°W
LAI+SAI = 4.5~5 β
Litter
Stability
Thick canopy
New formulation reduced Eg more.

12. S12
Summary
Soil resistance in CLM3.5 significantly improved Eg compared to CLM3.0., but has two issues:
Physically inconsistent in wet soil condition
Implicitly includes litter effect
Developed more physically-based formulations for both wet and dry soil, plant litter resistance, and also under-canopy atmospheric stability.
These revisions did not change global average compared to CLM3.5, but improved ET partitioning over certain regions.
They are simple to implement. Litter resistance will also facilitate the direct use of CN-DGVM litter output in CLM Eg computation in the future.
We are lack of reliable global ET partitioning data to prove our modifications are better, but our modifications are more physically robust.

13. Thank you!