Snowpack Radiative Heating: Influence on Tibetan Plateau

Snowpack Radiative Heating:
Influence on Tibetan Plateau
Climate
Mark Flanner
Charlie Zender
University of California at Irvine
Earth System Science Department
Thanks to:
Richard Armstrong
Siri Jodha Singh Khalsa
Steve Warren
Zong-Liang Yang
Keith Oleson
Mariana Vertenstein
Chao Luo
Flanner and Zender, Geophys. Res. Lett., 32, L06501, 2005
This study demonstrates:
• Powerful snow-albedo feedback in environments
of intense insolation.
• The importance of subsurface snowmelt in
triggering this feedback.
• 80% reduction in winter/spring snowmass on
Tibetan Plateau, attributed entirely to realistic
representation of vertically-resolved radiative
absorption.
– Significant improvements in model predictions of
snow mass and air temperature.
Koh and Jordan, “Sub-surface
melting in a seasonal snow
cover”, J. Glaciol, 41, 1995:
Detected meltwater in low
density snow at depths between
2 and 5 cm
• Daytime subsurface
temperature maximum
because of:
– Strong longwave cooling from
surface
– Subsurface absorption of
solar radiation
– Good insulating properties of
low density snow, allowing
sustained temperature
gradients to persist
SNow, ICe, and Aerosol Radiative Model (SNICAR)
Two-stream, multi-layer radiative transfer model
[Wiscombe and Warren, 1980; Toon et al, 1989]
• Fraction of total solar absorption occuring more than 2cm
beneath surface ranges from 5-60% of total absorption
• Subsurface absorption depends on:
– Snow grain size
– Zenith Angle
– Snow Density
Climate Model Experiments
Off-line
Simulation
Coupled
Simulation
Snow RT
Model
Median
Radius
Shortwave
Absorption
CLM-B
CAM-B
SNICAR
200 µm
Top Layer
CLM-C
CAM-C
SNICAR
200 µm
All Layers
(Identical Snow Albedo in all Experiments)
CLM: NCAR Community Land Model 3 [Oleson et al., 2004]:
- Represents snow with up to 5 layers, based on Jordan [1991].
- All solar radiative absorption is in top snow layer (2 cm)
CAM: NCAR Community Atmosphere Model 3 [Collins et al., 2004]
without snow-albedo feedback
… and with
Cause: Sub-surface melt, followed
by meltwater seepage into soil,
reducing snow column depth and
areal snow coverage, triggering
snow-albedo feedback.
Climate Model Experiments
Off-line
Simulation
Coupled
Simulation
Snow RT
Model
Median
Radius
Shortwave
Absorption
CLM-A
CAM-A
CLM
n/a
Top Layer
CLM-B
CAM-B
SNICAR
200 µm
Top Layer
CLM-C
CAM-C
SNICAR
200 µm
All Layers
CLM-D
CAM-D
SNICAR
100-1000 µm All Layers
- High spatial variability in TP snow depth (terrain) and large observational
uncertainty (remoteness), but clear high winter bias predicted by vanilla CAM.
- Microwave high-bias because of thin TP atmosphere.
- High precipitation bias in model
- Unique situation: Strong snow-albedo feedback in winter!
Vanilla CAM severely
over-predicts winter
snow mass on Tibetan
Plateau.
Changes in 2-meter air temperature tracks changes
in snow coverage
Conclusions
• Subsurface snow melt occurs regularly in low
density, mid-latitude snowpack
• Snow mass reduction from subsurface melt
triggers strong snow-albedo feedback on the
Tibetan Plateau.
• How do these changes impact soil moisture?
• How important is this mechanism in times of
greater snow coverage and greater NH winter
solar insolation?
Soil Moisture: Preliminary Results
• Offline model comparison, with realistic
precipitation.
• Vanilla CLM shows no winter soil moisture
because ground is frozen.
Snow Depth Changes:
Offline Land Model