Note: (10/13/00): All 1996-2000 datasets now include a
change in method of calculating optical depth. The other years of
data, processed using the new method, will be made available soon.
Note: (2/6/01) The March 2000 dataset has been reprocessed
using a GOES-8 calibration against VIRS.
We believe VIRS provides the
most accurate calibration reference available. All other GOES8 datasets
will likely be reprocessed soon using the calibration against VIRS.
Note: (5/4/01) Datasets from December 1996 - November 1999
have been reprocessed using the latest GOES-8 calibration against
VIRS . We believe VIRS provides the
most accurate calibration reference available. The March 2000 dataset
uses an earlier VIRS calibration, which will soon be updated to the latest
available.
Note: (5/4/01) The GMT times recorded in the netcdf
files of recently replaced datasets (based on AVHRR calibration) from
December 1996, April-May, August, and October 1999 were 15 minutes ahead of
the actual processing times. The times have been corrected in the latest
netcdf files.
Note: (5/12/01) The 0.3 degree Jul98 dataset
has been updated; erroneous nighttime values were corrected.
Note: (6/5/01) Datasets from December 1996 - November 1999
and August-November 2000
have been reprocessed using the latest GOES-8 calibration against
VIRS . We believe VIRS provides the
most accurate calibration reference available. The March 2000 dataset
uses an earlier VIRS calibration, which will soon be updated to the latest
available.
Note: (6/6/01) LBTM has a tendency to underpredict
low cloud amounts especially during twilight (SZA > 80 degrees)
and night. VISST (Visible Infrared Solar-Infrared Split-Window
Technique) should help increase accuracy of cloud detection and
cloud height determination; datasets using this technique will
be made available soon.
It should be noted that VISST's visible optical depth
for high solar zenith angles (approximately 50-90 degrees)
will be significantly smaller than that derived by LBTM.
Note: (7/30/01) The datasets from April 1-6, 2000,
formerly processed with the March 2000 data, have been reprocessed
with the rest of April 2000. This yields some differences
in the products.
Note: (8/17/01) LBTM datasets are produced without
a snow mask, so snow may be misinterpreted as cloud. This effect is
especially marked with the months of Nov 2000 through March 2001,
although any month with snow cover is affected. Datasets processed using
VISST, when they become available, should be more accurate in cases
of snow cover.
(12/19/01) As of 8/30/01, the LBTM dataset
from ARESE-I
(25 Sep - 1 Nov 1995) was reprocessed using the updated GOES-VIRS
calibration (see 6/5/01 note) and change in method of calculating optical
depth (see 10/13/00 note). All available datasets from 1995-2001 have
now been reprocessed in this manner.
Note: (2/13/02) The values of total (level=4)
Visible_Optical_Depth_SD, total (level=4) Cloud_Center_Temperature_SD,
Clear_Temperature_SD and Narrowband_VIS_Albedo_SD are
erroneous and should be disregarded. This is true for all
datasets between 1995-2001. The error will be corrected
as the datasets are re-processed.
(4/15/02,updated 11/9/04) The last 12 hours of some
monthly datasets have used a suspect sounding. The datasets may
include Dec96-1997,Feb98,Dec98-Mar99,Dec99,Feb00,Dec00,
Feb01-Mar01,Aug01,Nov01-Feb02.
(9/6/02) Datasets from 1996 are being reprocessed
by month, rather than by IOP. This will result in some
minor differences in the results. This also means previous
issues (i.e. notes from 2/13/02 and 4/15/02)
will be corrected in these new datasets (currently
Apr96, May96, Jun96).
(1/3/03) Dec02 dataset may have some
snow misinterpreted as cloud (see 8/17/01 note).
(8/27/03) Mar03 is the last GOES-8 monthly dataset,
and Apr03 represents the first GOES-12 dataset.
