Sub-grid scale processes: the role of snow
and lakes at the arctic forest/tundra transition
Philip Marsh1,2, Stefano Endrizzi1, Cuyler Onclin2,
Mark Russell2, Newell Hedstrom2, and Chris Marsh1
2.
1.
University of Saskatchewan
National Hydrology Research Centre
Environment Canada, Saskatoon
Outline
1. Study sites
2. 2006/07 data collection
3. Snow processes
– Compare the effect of shrub and tundra surfaces on snow
accumulation and melt
4. Process modelling
5. Lakes (WILL NOT DISCUSS LAKES IN THIS TALK)
6. Plans for the remainder of year 2 and year 3
Field Sites - Inuvik, NWT area
Big Lake
Beaufort Sea
Denis Lagoon
Trail Valley Creek
(TVC)
TVC
HPC
TVC Lake
0
50
K ilo m e tre s
Mackenzie River
@ Arctic Red R.
Havikpak Creek
(HPC)
TVC vegetation cover and station location
NWRI and
MSC station
WSC TVC
discharge
Tundra station
TVC Lake
Lake basin
discharge
Shrub station
Similar vegetation maps are available for Havikpak Creek
TVC Tundra
TVC Tundra
HPC Forest
TVC Shrub
TVC Shrub
TVC Lake
Lake instrumentation
TVC Lake
Big Lake
Water Survey Canada lake water level gauge
Terrestrial sites: Instrumentation – 2006-07
Terrestrial sites (annual)
Instrument
Air temp/RH
HMP35CF
Barometric Pressure
Setra
Wind speed/direction
RM Young
Rainfall
Tipping bucket
Precipitation
Geonor weighing alter
Incoming/outgoing short/long
CNR1
Surface temperature
Everest IR
Ground temp + heat flux
Campbell themistors and HFP
Snow on ground
Campbell SR50
Sensible and Latent Fluxes
Campbell CSAT + Krypton
Hygrometer + 21x
Discharge
Water Survey Canada
Stream velocity
Price + acoustic doppler
7 stations including:
1. HPC main met,
2. TVC MSC,
3. TVC main met,
4. TVC shrub,
5. TVC tundra,
6. Denis Lagoon and
7. Big Lake
Sensible and latent flux measurements are seasonal
Lake sites: Instrumentation 2007
Lake sites (summer only)
Instrument
2 sites including:
Same as terrestrial sites as
appropriate, plus
Water level
RBR pressure recorder
Water temperature
RBR temperature recorder
Lake discharge
Acoustic doppler
1. TVC Lake
2. Denis Lagoon
Typical problems with unattended winter measurements
Process and Parameterization
•
Observations at all study sites during 2006/07, and analysis of these data are
ongoing
•
Process modelling to date include:
• Cold Regions Hydrologic Model (CRHM)
• Glen Liston’s Snow model
• GEOTOP
• TOPOFLOW
– All of these are small scale hydrologic models that will allow testing of, and
comparison of, parameterizations on snow accumulation and melt, and allow
comparison of methods to scale to the correct scale needed for prediction
Role of shrubs in controlling snow
accumulation and melt
• Previous work has shown the expansion of shrub areas
into the tundra north of Inuvik
• Previous modelling of snow accumulation and melt at
shrub sites at TVC was shown to have lower accuracy than
for tundra sites or for drift locations (Pohl et al.).
• This uncertainty is due to difficulties in modelling:
• shrub behaviour
• effect of shrubs on radiation and
• effects of shrubs on turbulent fluxes.
During spring 2003, most of the shrubs
were ”bent” over, and covered by snow
May 2003
Snow depth between 30 to 60 cm
Aug 2003
Shrub heights between 1.5 m to 3 m
2003
2004
2005
Range of conditions during the 5 study years, with shrubs almost completely covered
in 2003, mostly standing above the snow in 2005, and the other years being between
these extremes.
2006
2007
Effects of shrubs on albedo during melt
Incoming shortwave
300
TUP and TTS
TTS air temperature
30
20
200
10
0
100
Air Temperature (oC)
40
400
Shortwave Radiation at
a shrub and tundra site
-10
0
Apr/1
Apr/11
Apr/21
May/1
1.2
May/11
May/21
May/31
Jun/10
Jun/20
2003
Albedo
1.0
0.8
0.6
0.4
TUP
TTS
0.2
0.0
Apr/1
Apr/11
Apr/21
May/1
Outgoing Shortwave
300
250
May/21
May/31
Jun/10
Jun/20
May/21
May/31
Jun/10
Jun/20
May/21
May/31
Jun/10
Jun/20
2003
TUP
TTS
200
150
100
50
0
Apr/1
Apr/11
Apr/21
May/1
350
250
200
May/11
2003
300
Net Solar
May/11
TUP
TTS
150
100
50
0
Apr/1
Apr/11
Apr/21
May/1
May/11
2003
Lower albedo at the
shrub site, results in
less outgoing
shortwave, and
therefore increased
absorbed shortwave
Outgoing longwave radiation at a shrub and tundra
2005
site
450
TUP
TTS (above "canopy"
W/m
2
400
350
300
250
450
120
140
160
180
Day number
400
TUP
TTS (above "canopy"
2
W/m
2006
350
300
250
450
120
140
160
Day number
180
2007
TUP
TTS (above "canopy"
W/m
2
400
350
300
250
120
140
160
Day number
180
Increased longwave
emission at the shrub site
due to the occurrence of
exposed shrubs with
warmer stems
Sensible Heat Flux at a Tundra site (TUP)
Positive flux is upward
Sensible Heat
Air Temperature
Wind Speed
350
Only first 2 days of melt
Is the flux towards the surface
TUP
300
250
2
Sensible heat flux (W/m ) or
Temperature (C) or Wind Speed (m/s)
400
7 day melt period
200
150
100
50
0
2003
May/25
May/30
Jun/04
Sensible Heat Flux at a Shrub site (TTS)
Positive flux is upward
Sensible Heat
Air Temperature
Wind Speed
350
300
250
2
Sensible heat flux (W/m ) or
Temperature (C) or Wind Speed (m/s)
400
5 day melt period
Only 1 day + brief
periods is the flux
towards the
surface
TTS
200
150
100
50
0
May/25
May/30
2003
Jun/04
Change in snow covered area and snow water equivalent
with increased melt at the shrub site, and snow being
removed from first at the shrub site
SCA
1 .0
14
SWE
0 .8
12
10
0 .6
8
6
4
2
0
1 6 /M a y
0 .4
T u n d ra S W E
S h ru b S W E
S h ru b s n o w c o v e r a re a
T u n d ra s n o w c o v e r a re a
2 0 /M a y
2 4 /M a y
D a te (2 0 0 3 )
0 .2
2 8 /M a y
0 .0
0 1 /J u n
Snow covered area
Snow Water Equivalent (cm)
16
Process modelling to date include:
• Cold Regions Hydrologic Model (CRHM)
• Glen Liston’s Snow model
• GEOTOP (see poster by Stefano Endrizzi)
• TOPOFLOW
Cold Regions Hydrological Model (CRHM
•
Preliminary test as applied to Trail Valley Creek for 1998/99
•
Initial run set up as follows:
– 4 HRUs
• Tundra (69.8% of basin)
• Shrub (21.5%)
• Forest (0.5%)
• Drift (0.2%)
– Evaporation – used the Granger method
– Discharge
• Soil component set to have minimal effect
• Did not use snow from PBSM
• Instead used 97 mm SWE from end of winter snow survey
CRHM – snow accumulation and melt
TVC 1998/99
140
Pohl et al.
CRHM Tundra SWE
CRHM Shrub SWE
120
Drift = 457 mm
CRHM Forest SWE
CRHM Drift SWE
100
Forest = 110 mm
Measured SWE
80
mm
Tundra = 114 mm
60
Shrub = 197 mm
40
20
0
27-Sep
1-Nov
6-Dec
10-Jan
14-Feb
21-Mar
25-Apr
30-May
4-Jul
Estimated SWE is underestimated for all sites, but it seems likely that this is due to
underestimating snowfall. However, it appears that the drifts are severely under
estimated.
CHRM – Evaporation using Granger method
TVC 1998/99
160
CRHM Tundra evap
140
120
mm
100
CRHM Shrub evap
CRHM Forest evap
CRHM Drift evap
Measured LE
80
60
40
20
0
1-Jun
11-Jun
21-Jun
1-Jul
11-Jul
21-Jul
31-Jul
10-Aug 20-Aug
Measured LE is from Eddy Correlation
30-Aug
CRHM - Discharge
TVC 1998/99
5
5.5
5
4.5
Measured Discharge
4.5
4
Precipitation
4
3.5
CRHM Discharge
3
3
2.5
2.5
mm
m3/s
3.5
2
2
1.5
1.5
1
1
0.5
0.5
0
0
20May
30- 9-Jun 19May
Jun
29Jun
9-Jul
19Jul
29Jul
8-Aug
18Aug
28- 7-Sep 17Aug
Sep
27Sep
First test shows that CHRM underestimates total spring discharge and
overestimates response to rain events
750
650
Liston Snow model, end
of winter SWE
550
450
350
250
150
Liston Snowmodel
50
S. Pohl et al.
C. Marsh et al. 2007
Observed vs modelled snow covered area
SPOT
May 23
1999
May 28
1999
June 10
1999
Model
Oblique photos
Liston Model
Planned activities over the next year
NRC Twin Otter Aircraft
9 flight lines (in red), each approx.
16 km in length
aircraft observations - 1999
TVC
Tower
• aircraft eddy correlation measurements
provide an estimate of basin average
measurement of latent and sensible
fluxes, and gridded fluxes at 1 km
resolution. Currently working with Dr.
Ray Desjardins (Agriculture Canada) to
compare aircraft to modelled fluxes
Tower vs aircraft comparison
2
Flux (W/m )
100
80
Tower
Tower
120
150
150
140
100
Aircraft
50
100
50
0
Tower
0
0
50
100
Aircraft
150
0
50
100
Aircraft
150
60
Aircraft
40
Tower
20
Sensible
0
May 20
May 24
May 28
Date (1999)
Jun 1
May 20
Latent
May 24
May 28
Date (1999)
TVC Tower vs TVC transect #8
Jun 1
Jun 5
Tower vs aircraft fluxes
200
200
#11
#9
#8
150
150
Flux (W/m2)
#9
100
Flight
#5
50
100
#7
Flight #7
#5
50
#8
#11
0
0
Tower
Tower
-50
-50
-100
May 25
May 28
Snowcover =
39%
May 28
Latent
Sensible
May 31
Date (1999)
Jun 3
-100
Jun 6 May 25
May 28
May 31
Jun 3
Date (1999)
Jun 6
Lidar
1. DEM – is of sufficient accuracy for small scale hydrologic
modelling
2. Vegetation – unfortunately the Lidar was flown with minimal
overlap between flight paths. Although it is still useful, it has
significantly reduced our ability to estimate vegetation height
and density. As a result, we hope to acquire new Lidar data
during the summer of 2008
TVC Basin
TVC Lake
Collaborative Modelling Efforts
•
We are currently working with Murray Mackay (MSC) and Paul
Bartlet (MSC) to use CLASS to better understand factors controlling
melt at a shrub and tundra site at Trail Valley Creek
•
Will be providing data to the following IP3 groups
– Ric Soulis and Frank Seglenieks – data to initialize MESH for
TVC (done) and HPC
– Diana Verseghy – data to run CLASS for TVC
Lake studies
• We
are collaborating with Peter Blanken (U. of
Colorado Boulder) to analyze our lake flux data
• Will be working with Raoul Granger to consider
implementing his lake evaporation parameterization to
the Inuvik area study sites
• We are collaborating with Dr. Ray Desjardins and hope
to analyze the NRC aircraft data for the flight lines over
lake rich regions near Inuvik
Map vegetation cover over entire
domain using Lidar, Quickbird,
Landsat and ground truthing
Beaufort Sea
Quickbird
TVC
HPC
Lidar vegetation Mackenzie River
0
50
K ilo m e tre s
@ Arctic Red R.
THE END