3 Role of SAR in Surface Energy Flux Measurements Over Sea Ice
59
These equations were encoded in a computer program and confidence intervals were
generated for predicted albedo values. The generated confidence intervals were used
as thresholds for image classification. For each 0.1 dB change in the independent variable, the corresponding albedo was calculated with 90% upper and lower confidence
limits. Classes were formed whenever the upper limit of the upper confidence interval
Generate change image from repeat-pass ERS- l data
Apply land mask
....
+
Map MYI albedo classes
using regre sion confidence
interval interpolation
~
Isolate pure MYI from
conglomerated MYI and ---+determine mean albedo
cia s
1
A ign FYI an albedo
4% greater than that of
prre MYI
Produce map of MYI and FYI albedo
Fig.14. Flowchart of the procedure used to map albedo from MYI t1cr"
59
These equations were encoded in a computer program and confidence intervals were
generated for predicted albedo values. The generated confidence intervals were used
as thresholds for image classification. For each 0.1 dB change in the independent variable, the corresponding albedo was calculated with 90% upper and lower confidence
limits. Classes were formed whenever the upper limit of the upper confidence interval
Generate change image from repeat-pass ERS- l data
Apply land mask
....
+
Map MYI albedo classes
using regre sion confidence
interval interpolation
~
Isolate pure MYI from
conglomerated MYI and ---+determine mean albedo
cia s
1
A ign FYI an albedo
4% greater than that of
prre MYI
Produce map of MYI and FYI albedo
Fig.14. Flowchart of the procedure used to map albedo from MYI t1cr"