Retrieval of Snow covered area with SAR SCA Helmut Rott Contributions from: IMGI, HUT, Norut and IFA

1
Retrieval of Snow covered area with SAR
(SCA)Helmut RottContributions from IMGI, HUT,
Norut and IFAc
2
SCA mapping by SAR

• Parameter Snow covered area (SCA)
• Definition Percentage of unit area covered with
  snow
• Sensors Synthetic aperture radar (SAR), C-band
  (? ? 5 cm)
• Background/Method Detection of backscatter
  changes
• Advantage Weather and illumination independent
• Limitation-1 C-band SAR is insensitive to dry
  snow
• Limitation-2 SAR geometry obscures part of area
  in mountains
• Achievement Robust methods for SCA of wet snow

3
EnviSnow developments for SAR SCA retrievals

• IMGI Change detectiondry snow algorithm, after
  Nagler and Rott (2000), using ASAR and Radarsat,
  any IM, APS, WS mode, used for real time runoff
  forecasting demo etc.
• Norut Change detectiondry snow algorithm. Using
  ASAR WS to obtain coverage, used for
  pre-operational SCA mapping demo
• HUT based on change detection, suitable for
  boreal forest
• IfAC based on change detection, demonstration
  products

4
IMGI SCA Processing Line
5
IS2
SCA mapping data base - ASAR
IS6
6
ASAR Snow Map Example
Track 172, IS2, VV
Red Wet Snow 24 April 2003 (without
postprocessingfor dry snow)
Ötztal Basin Area 784
km2 Layover/shadow 43.4 Wet Snow
48.6 ( of observed area)
7
Multitemporal Snow Map IS6
Snow extent 4 May 2004 8 June 2004 14 July
2004 Yellow ? lt 17 (6.8 of area)
8
ASAR Wide Swath 23 May 2005
9
ASAR Wide Swath Mode Snow Map
Wet snow 4-May-10 cyanblue 2004-Jun-14 -
blue
10
Norut SCA product
20050524

• Method
• Wet snow detected with change detection against
  dry snow reference.
• Dry snow inferred above wet snow line.
• Dry snow estimates may be improved with
  temperature input from met.stations.
• Forests, lakes and urban areas masked.

11
NRT SCA Production Line
DEM
ftp
Current SAR
ref.image
Geocode
Resample 250m
SCA-map 100m
SCA-map 250m
Confidence
mask
Confidence
SCA alg.
Web-gui to database
Temp mask
12
Norut 2005 demonstration
13
Multitemporal SCA Analysis 2004 IfAC

• Wet snow detected with a threshold set to 3 dB
• Dry snow in April Pixels classified as wet snow
  in May but not in April.
• Snow free in May Pixels classified as wet in
  April but not in May
• Dry snow at both dates Pixels not classified as
  wet snow at both dates, but located at a height
  above the median altitude of the wet snow in
  April.

5 April 2004
10 may 2004
Light blue dry-snow Blue wet snow Green
forests Brown bare soil Red layover and shadow
areas (rocks)
14
Satellite Data Processing Chain for HUT SCA Method
Satellite Data (Raw format)
Geocoding (ErMapper)
Land-use Stem volume Information
Forest Compensation (Matlab)
Linear Interpolation (Matlab)
2 x Reference Images
SCA map
15
SCA Estimates from Radarsat data - HUT
SCA from Radarsat data, 5 May -04
SCA from AVHRR data, 6 May -04
16
Conclusion

• Improved, reliable methods for SCA retrieval from
  C-band SAR have been developed for mountainous,
  forested and Alpine areas
• Automated procedures are available for
  terrain-corrected geocoding and snow
  classification
• WS data are useful to give wide coverage.
• Any available polarization and imaging mode can
  be used
• The data flow and processing methods were
  successfully tested in semi-operational
  environment, to produce SCA time series in
  Scandinavia and in the Alps