MODICE using other.pic data

These images show MODICE results before and after adding the "other.pic" data to the MODICE algorithm. The other.pic data are derived using the modsort changes Tom gave me--they basically remove the category for lake ice from the sorting algorithm at the end of each set of MODSCAG model runs.

Data are for 2001 season, Juneau scene. I'm using other.pic data, greater than 50%, that's deemed "adequate" by Painter's .signal mask. If these data are considered "usable"--there are no clouds, and zenith angle is good, then they're set to a permanent ice value, and never reset. These results are a lot cleaner than the first time we tried this (compare to all the noise in the last image in this page).

N.B. there's a *little* bit of noise left, see the speckles on the zoom around 58.8N and 133 W).

MODICE 2001, Juneau

MODICE + other.pic, 2001, Juneau, using new modsort that removes lake ice category

MODICE + other.pic, 2001, Juneau (this was the one with the original lkice configuration in modsort)

Notes about original way: The first time we tried this, images show MODICE results before and after adding the "other.pic" data to the MODICE algorithm. Data are for 2001 season. I'm using other.pic data, greater than 15%, that's deemed "adequate" by Painter's .signal mask. If these data are considered "usable"--there are no clouds, and zenith angle is good, then they're set to a permanent ice value, and never reset. I thought we needed to be more discriminating in our algorithm. This approach puts a lot of confidence into a single hit in this file, and never gives us the opportunity to turn it off again.