utilize Observations from many passes to gener- 
ate time series and to do statistical analyses. This 
implies that the locational accuracy of data from 
different passes must be sufficiently precise to al- 
low compositing. For these purposes, an abso- 
lute navigational accuracy of 1 km for high- 
resolution data and 5 km for low-resolution global 
data are required. 
Maximum Scan Angle and 
Revisit Interval 
The topic of revisit interval engendered much 
discussion during the workshop. In fact, having a 
daily revisit interval emerged as the sine qua non 
for most commercial and operational users. The 
baseline SeaWiFS instrument design presented 
during the initial workshop had a +45° scan an- 
gle, resulting in a scan swath of 1500 km and a 2- 
day revisit interval. This design was selected 
based on the understanding that data from scan 
angles greater than about +30° could not be in- 
terpreted quantitatively. However, as demon- 
strated in Section 2, commercial and operational 
users find great qualitative value in imagery ac- 
quired at scan angles in excess of 40°. With the 
Landsat-6 orbit, a scan angle of +58.3° would be 
required to provide a daily revisit interval (2800 
km swath), and this angle can be accommodated 
in the instrument design without difficulty. (This 
scan angle will give complete coverage at the 
equator with increasing overlap at higher and 
lower latitudes.) The positions of both panels on 
this issue and its resolution are presented below. 
Commercial and Operational Users' 
Requirements for a Daily Revisit Interval 
A daily revisit cycle was strongly recommend- 
ed by commercial and operational users since the 
temporal variability of ocean-color and tempera- 
ture features are of prime interest to them. A 
knowledge of the daily changes in these features 
is critical in understanding water-mass move- 
ments and in observing the movements of ocean- 
circulation features. For example, the movements 
of waves along the north wall of the Gulf Stream 
change significantly within a 1-day period. In ad- 
SYNTHESIS OF REQUIREMENTS 
dition, navigating ship movements for research 
station locations, positioning frontal boundaries 
for naval operations, and locating fish populations 
all require a daily revisit interval. 
This revisit-interval requirement is based on 
experience. Usually cloud cover hampers rou- 
tine coverage of areas of interest. With a revisit 
interval of every other day, if clouds hamper the 
image for 1 day (as is often the case in tropical or 
subtropical regions during Summer and in polar 
regions during winter), then no imagery is ob- 
tained for a period of 4 days. This is not ade- 
quate, since many ocean frontal boundaries are 
moving too rapidly to be able to infer the temporal 
variability from data that differs by 4 days. Ex- 
perience has also shown that the image- 
processing technique of "movie looping" succes- 
sive images for flow visualization requires at least 
one image a day in frontal regions in order for the 
human eye to interpolate or see and understand 
the motions of the ocean features. Circulation 
patterns are readily discernable in movie loops 
using twice-daily AVHRR images. Once-daily im- 
ages provide poor, but usable, results; however, 
circulation patterns are not recognizable with 
every-other-day coverage. 
An important and attractive use of the Sea- 
WIFS data will be in gaining an understanding of 
the temporal variability of the ocean surface. If 
the data are undersampled in time so that the dai- 
ly scales, or bidaily scales, of variability are not 
obtained, the utility of the sensor for any type of 
near-real-time commercial or operational applica- 
tion would be greatly limited. 
Research Users' Requirements for Revisit 
Interval 
Research requirements for world-wide chlo- 
rophyll observations dictate global measurements 
at least once every 3 or 4 days. However, be- 
cause imagery is lost due to cloud cover and sun 
glint, this requirement results in a need for global 
coverage on a 1- or 2-day cycle. 
Currently, algorithms for correcting ocean- 
color data for the effects of atmospheric degra- 
dation are viable to a scan angle of approximately 
41 
