P5.55
Application of MODIS data for deriving ocean surface currents
PAPER WITHDRAWN
R. L. Bernstein, SeaSpace Corp., Poway, CA; and K. S. Prasad
Geostationary satellite imaging sensors have been used for over 25 years to derive atmospheric wind vectors, based on the displacement of visible and/or thermal infrared features in time-sequential imagery. Typically, a two- or three-image sequence, separated by 30 minutes, using 5 km resolution sensor data, is the basic input. If the sensor data is perfectly earth-located, the "least count" of this process is 5000 m / 1800 s or approximately 3 m/s, and this determines the limiting accuracy of the derived wind speeds and directions.
In principle, a similar approach for deriving ocean surface currents could be employed, using for example the NOAA polar orbiting satellites AVHRR sensor. While a number of published papers have described results from such an approach, in practice it has seen limited operational usage, for a variety of reasons including: restriction to thermal infrared (sea surface temperature) features; earth-location inaccuracies in AVHRR due to spacecraft attitude uncertainties; six-hour time separation between overpasses, which increases likelihood of cloud cover moving over features that were successfully imaged in the first overpass, thus eliminating the required minimum two-image time sequence.
The advent of the MODIS sensor, now flying on the NASA EOS Terra satellite (overpass time 10:30AM/PM), and also on its sister satellite Aqua, scheduled for launch in late 2001 (overpass time 1:30AM/PM), opens up new opportunities to apply the displacement technique for deriving ocean currents. MODIS has both thermal infrared and ocean color channels. In addition, the Orbview-2 SeaWiFS ocean color sensor provides data at local noon, exactly midway between Terra and Aqua.
The earth location accuracy of MODIS is better than 200 m, i.e. considerably smaller than the 1000 m resolution of the ocean color channels of MODIS and SeaWiFS. The "least count" for surface currents using these sensor inputs is determined by the sensor spatial resolution and the time separation. Focusing for the moment only on Terra/Aqua, which will be separated by three hours (10,000 s), the least count will be 1000 m / 10 000 s or 0.1 m/s (10 cm/s). This should be sufficient for determining surface flow of major ocean currents with useful accuracy.
Examples of surface currents derived with a combination of MODIS and SeaWiFS imagery, in the Gulf of California and the upwelling region along the U.S. west coast, demonstrates that this approach does provide useful levels of current mapping capability, when flow speeds are in excess of 50 cm/s. This limitation is based on the 5000 s time separation between MODIS and SeaWiFS overpass times. The availability of Aqua/MODIS will provide an additional factor of two improvement in "least count"; additionally, as is also the case with geostationary winds, a three-image sequence will permit significantly better quality control checks on displacement-derived surface currents
Poster Session 5, New Technology and Methods (Continued)
Thursday, 18 October 2001, 9:15 AM-11:00 AM
Previous paper Next paper