Snow Processes and
Parameterisations
John Pomeroy
Canada Research Chair in Water Resources & Climate Change,
Centre for Hydrology, Univ Saskatchewan, Saskatoon
and collaborators
Dan Bewley (UBC), Sean Carey (Carleton), Richard Essery (Edinburgh),
Masaki Hayashi (Calgary), Rick Janowicz (Yukon Env), Tim Link (Univ Idaho),
Danny Marks (USDA ARS), Al Pietroniro (Env Canada), Nick Rutter (Sheffield),
Diana Verseghy (Env Canada)
and Centre for Hydrology Researchers
Chris DeBeer, Pablo Dornes, Chad Ellis, Warren Helgason, Edgar Herrera,
Jimmy MacDonald, Matt MacDonald, Nicholas Kinar, Michael Solohub
Study Elements
z
Processes (also see Marks talk)
z
z
z
z
Parameterisations (also see Ellis, Essery talks)
z
z
z
z
z
z
Snow accumulation, structure and observation
Radiation effects on snowmelt under vegetation
Turbulent transfer to snow
Blowing snow over complex terrain
Sub-canopy snowmelt
Spatial associations between accumulation and melt
Contributing area for snowmelt
Areal infiltration and runoff generation
Prediction (see Pietroniro talk)
z
z
Level of spatial complexity necessary in models
Snow modelling contribution to MESH
Marmot Creek
Marmot Creek, Alberta
-montane, subalpine forest
-alpine tundra
-steep alpine topography
-eastern slopes climate
-10 meteorological stations
-WSC streamflow gauge
-3 U of S streamflow gauges
Wolf Creek
Wolf Creek, Yukon
-boreal forest
-shrub and sparse tundra
-sub-arctic to arctic climate
-4 meteorological stations
-Snow pillow
-4 Yukon Environment streamflow
gauges
Marmot Creek Research Basin
x x
x
xx
x
x
Snow Accumulation, Structure, and
Observation
Acoustic observation of depth, porosity
and tortuosity from snow layers
z Interception and unloading of snow from
mountain forests
z
Acoustic Determination of SWE
1
2
•Digital signal processing adapted from Frequency-Modulated ContinuousWave (FMCW) Radar
•Continuous sound pulse in the audible frequency range (20 Hz to 20 kHz)
•Solution based on Biot’s theory and Berryman’s relationship
•Layer and total depth and porosity estimated from returned acoustic signal
•Non destructive sampling
SWE Estimates Possible where Snow Tubes Fail
Comparing Measured and Acoustic SWE
(Granger Basin Site)
500
Acoustic Device
Snowpit
400
SWE(mm)
Average Difference = 27%
300
Low Shrub
200
100
0
1
2
3
4
5
6
Sample Point
7
8
9
10
Low Shrub
Snow Interception & Unloading
-Greater interception than observed elsewhere
-Frequent unloading events in mountains
-Current model might be wrong…….
-Unloading during wind gusts
Alpine Blowing Snow: Flow Separation
GRU-to-GRU Blowing Snow Transport
Over Complex Terrain for MESH
QSUBLIM
WINDWARD HILL,
BARE GROUND,
GRASS
QTRANSP
LEEWARD HILL,
FOREST
SHRUB,
DEPRESSION
BLOWING SNOW
IF CAPACITY/THRESHOLD
IS EXCEEDED
WINDWARD
HILL
GRU
LEEWARD
HILL
GRU
QSUBLIM
QSUBLIM
QTRANSP
Research Basins:
• Wolf Creek
• Marmot Creek
TOPOGRAPHIC
DEPRESSION
GRU
BARE
GROUND
GRU
QTRANSP
SHRUB
GRU
GRASS
GRU
FOREST
GRU
Radiation Effects on Snowmelt
under Vegetation
Longwave radiation from shrubs, tree
trunks and needles
z Shortwave extinction by shrubs, and
forests on slopes
z Radiation in sparse canopies and gaps
z Associated sub-canopy turbulent transfer
z
Radiation to Snowmelt
on 25o Forest Slopes,
Marmot Creek Research
Basin, Alberta
Net Shortwave Radiation
Net Allwave Radiation
Net Longwave Radiation
Slope Forest Transmissivity
τ a function of
LAI,
Foliage inclination
Crown coverage
Slope,
Aspect,
Solar azimuth,
Solar elevation
South Face Forest
50
100
150
200
250
azimuth angle
300
350
25
40
55
70
0.0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
10
Model of solar
radiation
transmission
through
continuous
evergreen
canopy on
slopes
elevation angle
70
25
40
55
x
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
10
elevation angle
85
85
North Face Forest
50
100
150
200
250
azimuth angle
300
350
Hot Trees
z
z
z
z
z
Longwave radiation is
an important component
of snowmelt energy
Longwave from trees
dominates longwave in
forests
High melt rates are
observed close to tree
trunks in sparse forests
Many assume that the
tree temperature is
equal to the air
temperature during
snowmelt
This assumption has not
been well tested……
Longwave Exitance Pine Stand
Estimating Canopy Contribution to
Downward Longwave
Air Temperature Model: all temperatures are air temperature, segregate by
sky view Vf and emissivity ε
Air Temperature & Shortwave Extinction Model: common temperature at air
temperature with shortwave extinction term B K*H that accounts for
shortwave conversion to longwave
Radiation Paradox for Snowmelt: Albedo and Slope Effects
-Longwave radiation model (Sicart et al 2004)
-Shortwave Radiation Model (Pomeroy & Dion, 1996)
-Driven by IP3 Basin data
By Topography: radiation to snow also strongly controlled by By Location: radiation to snow strongly controlled by
albedo; little difference in total radiation between sites of
differing elevation and/or latitude.
albedo but also by topography. Increased forest cover results in grater total radiation to steep north‐facing forests regardless of albedo.
More typical snow albedo at melt
Slight gains in radiation at higher forest cover densities
Turbulent Exchange:
Mountain generation mechanisms
upper level winds
strong shear zone
transported turbulence
valley winds
tributary valley
winds
surface winds (internal B-L)
flux tower in clearing
Roughness Length (z0m)
Alpine
Meadow
Prairie
Lake
Ridge
0
10
⎤
U 1⎡ ⎛ z ⎞
= ⎢ln⎜
⎟ −ψ m (ζ )⎥
u * k ⎣ ⎝ z0m ⎠
⎦
-1
10
0 < ζ < 0.1
-2
z0m (m)
10
Enhanced roughness in
mountain valleys due to
horizontal component
of turbulence from
surrounding terrain
-3
10
expected range
for snow
-4
10
-5
10
-6
10
0
10
20
U u -1
*
30
40
Flow has intermittent
turbulence at open
prairie site typical of
tundra plains
Typical Bulk Transfer Estimation
Results
Sensible heat flux in meadow site (mountain valley)
0
Horizontal component of
turbulence adds
momentum without
contributing to vertical
transport, causing failure
of normal stability
parameterisations of
turbulent transfer in
mountain valleys
Measured Flux (W m -2)
-10
-20
-30
-40
-50
-60
-60
-50
-40
-30
-20
Modeled Flux (W m -2)
-10
0
Basin Snowmelt and Areal
Depletion
Variable SWE
z Variable Melt Rate
z Spatial Associations between the two
z Modelling Implications
z
Snowcover Depletion in Alpine Basins
08-May
02-Jun
04-Jul
¾ Temporal snow cover depletion patterns
differ considerably between slopes within ~1
km2 cirque basin
¾ Single SCD curve cannot be applied even
for relatively ‘small’ scales in mountain terrain
• Spatial association between SWE,
melt rates, and melt timing.
•Greater end-of-winter SWE at
higher elevations,
•Earlier melt onset at lower
elevations,
•Higher melt rates later in spring
when deep snow at higher
elevations is melting.
Snowmelt Runoff
North Face
Valley Bottom
South Face
Next Steps
z
z
z
z
z
z
Multi-layer snow physics model
Snow Interception and Unloading in Mountains
Blowing snow parameterisation for CLASS/MESH
3-D Blowing Snow – small, medium and MEC scale
Snow-atmosphere exchange parameterisation
Complex Terrain Forest Melt Algorithms
z
z
z
z
z
Sparse canopy
Full integration of warm vegetation canopy effects
Sub-grid distributions of accumulation, energy, melt for
complex terrain
Quantify and Correct for the Effect of Spatial
Associations on Runoff Contributing Area from Melt in
Complex Basins
CRHM basin modelling and interface to MESH