NWSRFS Snow Modeling

1
NWSRFS Snow Modeling

• Cold Regions Workshop
• November 2004
• Andrea Holz
• NCRFC

2
North Dakota, 1966
Historic NWS Photo
3
SNOW-17 Model Overview

• Input Variables
• Air Temperature
• 10 days of maximum/minimum point (CCF) forecasts
• May use gridded temperature forecasts soon
  (maybe next spring)
• Precipitation
• 24 Hours of QPF
• 72 Hour QPF Contingency Forecasts
• Output
• Rain Melt
• Simulated Water Equivalent
• Simulated Areal Extent of Snow Cover
• Simulated Snow Depth

4
Other SNOW-17 Model calculations

• Snow Pack Information
• Accumulation of snow cover
• Surface Energy Exchange
• Heat storage in snow pack
• Rain on snow event
• Snow pack ripening
• Liquid water retention
• Ground Heat Exchange

5
Excavating Connelly Ditch, Minnesota, 1997
Photo Courtesy of St.Paul Corps of Engineers
6
SNOW-17

• Calculated on 6 Hour Time Intervals
• Air temperature determines form of precipitation
  input
• Can specify precipitation type during real time
  river forecasting
• Can adjust melt rate based upon observed
  information
• Dewpoints
• Windspeeds
• Observations from the field
• Can change snow water equivalent values based on
  observed data

7
South Dakota, 1966
Historic NWS Photo
8
Example of SNOW-17 Output with observed values
listed for comparison
Observed Snow Depth
Observed SWE
Observed AESC
9
(No Transcript)
10
Can create Snow Depth or SWE Areal Observations

• More than one observer in the basin
• 2 or more observers outside of the basin, yet
  nearby
• If one observation is missing for the day, no
  value is calculated considered missing

Historic NWS Photo
11
Tolerances used in Updating

• Automatically update SWE or AESC by setting
  tolerances
• SWE Updating occurs if sim
  ob gt tolerance x ob
• AESC Updating occurs if sim
  ob gt tolerance

12
NCRFC Tolerances

• High tolerances specified in the model prevent
  forecasters from making run time SWE or AESC
  modifications
• To prevent automatic updates and to allow
  forecasters to make real time modifications,
  NCRFC does not display observed WE
• Having ability to view observed WE would be
  operationally beneficial

13
Historic NWS Photo
14
River Forecasting using the Snow Model

• Sturgeon River at Alston
• Upper Peninsula of Michigan
• Snow Melt Rain on Snow Event
• March, 2004
• For purposes of illustrating mods, observed AESC
  WE are displayed

15
Sturgeon River at Alston
No modifications
6 hour RAIM around 0.2 inches, March 28 - 29
FS
? Observations
Simulation ?
16
SNOW-17 Output
Precipitation on March 29 is 0.28 inches of snow
0.06 inches of rain
Keep in mind Energy Exchange value for future
example
? Energy Exchange 0.5
17
Rain Snow Modification

• Field observations of rain instead of snow
• Rain/snow typing might be incorrect if there is
  an unknown temperature bias
• Timing might be off for an event with rain
  turning to snow in the evening
• Change precipitation typing in 6 hour increments
• Difficult to determine which 6 hour time period
  to change since output is in 24 hour time steps

18
Change precipitation to Rain March 28 12Z 29 12Z
19
Rain/snow Modification 24 hours
6 hour RAIM over 0.5 inches, March 28 - 29
20
UADJ Mod During Rain on Snow Events

• Rain on Snow must exceed 2.5mm per 6 hours for
  model to simulate an increase in melt (tenth of
  an inch)
• Modification can increase or decrease energy
  exchange
• .UADJ March 28 12Z 29 12Z
  ALSM4 5.0

21
Rain/snow UADJ mods
RAIM near 0.8 inches
22
UADJ 5.0March 28 - 29
? Energy Exchange 0.85
23
Snow Water Equivalent Change Mod
Observed values are listed in this table. Notice
3.9 inches on the 29th.
3.9 Observed Value ?
24
Water Equivalent Change Mod
Changing WE during melt automatically changes
AESC . . .
Best approach have WE values set before melt
begins.
25
Near Little South Pembina River, North Dakota,
1997
Photo courtesy of St.Paul Corps of Engineers
26
Updating Snow Model
No Modifications
Same Example moving forward in time
Too Much Peak ?
FS
? Missed peak
27
SNOW-17 Model Output
Melt already started . . .
Watton SWE
NOHRSC AESC
Watton SD
28
Start run further back in carryover
Melting begins . . .
29
SNOW-17 Output
No Modifications
Notice difference between Sim WE Obs WE
Energy Exchange .17 - .71
30
Low Energy Exchange
Low Energy Exchange
About 15 mm of melt modeled in any 24 hours
About 15mm of melt modeled in any 24 hours
31
Melt Factor Correction ModMarch 26 30, 2004

• .MFC 03260412Z 03300412Z
  ALSM4 4.0
• During time of active melt
• Meteorological conditions causing abnormal
  snowmelt
• High dewpoints
• High windspeeds
• Difficult to forecast MFC into future
• Watch meteorological forecasts closely
• Possible temperature bias
• Rain on snow

Can also use MFC to slow down melt!
32
MFC Mod Applied
33
MFC 4.0 March 26 - 30
Energy Exchange Values Increase . . .
Stimulating increase in melt
Energy Exchange Values Increase .36 1.54
34
Increased Energy Exchange
Increased Energy Exchange
45mm of melt ?
45mm of melt ?
35
How does Late April Peak Perform Now?
Within few tenths ?
? Caught peak
36
Current SNOW-17 Model

• Current output in 24 hour time steps
• Java display in 6 hour time steps?
• Possible to display observed WE without
  automatically updating yet still have ability to
  make WE mods?
• More Water Equivalent Data!
• Other observations from the field
• Water in ditches
• Runoff under snowpack

37
Photo Courtesy of NDSU