QuikSCAT Satellite

Introduction
The SeaWinds instrument on QuikSCAT is a Ku-band radar with ancestry in the Seasat scatterometer [e.g., Jones et al., 1982; Katsaros and Brown, 1991] and the NASA scatterometer, NSCAT [Naderi et al., 1991], on the Advanced Earth Observing Satellite I (ADEOS 1).Ê The QuikSCAT mission [Graf et al., 1998] was a quick recovery mission to make up for the loss of NSCAT, due to the failure of the power supply of ADEOS I.Ê NSCAT provided global surface winds from September 1996 through June 1997 but, unfortunately, the instrumentâs life span missed the Atlantic hurricane season. However, NSCAT did observe a major typhoon in the Pacific Ocean [Liu et al., 1997].

A scatterometer works on the principle that electromagnetic radiation transmitted toward the sea surface is scattered back towards the emitting antenna. The intensity of the backscatter at Ku-band is largely dependent on the centimeter-scale roughness at the sea surface, where the scale of the roughness elements is commensurate with the emitted radar wavelength. These small-scale waves are generated by the local wind stress on the sea. For a given wind speed, the backscattered power varies as a function of cos 2M, where M is the azimuth angle between the look direction and the wind direction, often with some ambiguity in the solution. Thus, by measuring the backscattered power from the ocean surface at several azimuth angles, the wind speed and direction may be retrieved.

Scatterometer data from the European Space Agency's C-band instrument on the European Remote Sensing (ERS) satellites 1 and 2 (launched in 1991 and 1995, respectively) have 50-km resolution and 500-km swath widths, which rarely cover an entire tropical cyclone. ERS-2 is still operating. Previous U.S. scatterometers were side-looking fixed beam sensors covering 500-km swaths on either side of the satellite subtrack, but with a 400-km wide gap between them. SeaWinds, on the other hand, has two conically scanning radar pencil-beams at incidence angles of 47E (H-pol) and 52E (V-pol) and provides 25-km resolution. It has an 1800-km wide continuous swath and can, therefore, completely cover most tropical cyclones.

After a comprehensive calibration/validation phase of the mission, QuikSCAT data were released to both the science and operational communities on February 1, 2000. The data set used in this study had not been officially released at the time of this demonstration project. The National Oceanic and Atmospheric Administration (NOAA)/National Environmental Satellite, Data and Information Service (NESDIS) has the responsibility for the near real-time (NRT) processing and distribution of QuikSCAT data for the operational community, and graphics ofthe NRT data are available to the general public at http://manati.wwb.noaa. gov/quikscat/.
Two approaches were taken to evaluate the usefulness of QuikSCAT data for studying tropical disturbances. The first approach was based on the hypothesis that the QuikSCAT wind vectors would give an early indication of rotation in the surface wind field. Closed circulation in the surface wind differentiates a developing tropical depression from an open tropical wave. Historically, hurricane forecasters typically identified potentially developing tropical cyclones by the rotation seen in the cloud top imagery obtained from geostationary satellites (GOES East over the tropical Atlantic Ocean). Some previous work indicates that tropical cyclone vorticity first develops at mid-levels and is transmitted downwards [Bosart and Sanders, 1981; Ritchie and Holland, 1997]. The principles involved in the release of latent heat as an energy source of hurricanes is discussed in detail by Malkus and Riehl [1960].

A second approach in evaluating the usefulness of QuikSCAT data in the study of tropical cyclones was to incorporate its scatterometer-derived wind field into the real time surface analyses provided by the Atlantic Oceanographic and Meteorological Laboratory's (AOML) Hurricane Research Division (HRD). These analyzed wind fields are primarily derived from data obtained by reconnaissance aircraft during penetrations into hurricanes [Powell et al., 1998] and also utilize surface winds from dropsondes, winds from an airborne microwave radiometer, and other available surface wind data to produce the wind field analyses.

A weather system must have a substantial region of convection and a definite organized surface circulation to be classified as a tropical depression [NOAA Office of the Federal Coordinator for Meteorological Services and Supporting Research, 1997]. Due to the westward translation of the air in the tradewind region, approximately a 10 m/s westerly (eastward) wind south of the depression's center is required to confirm officially that there is a bona fide tropical depression embedded in the tradewinds. We relax the criteria for this study and look simply for a consistent pattern of closed circulation.

Surface observations to assess wind speed and direction in the tropical Atlantic are rarely available from ships, buoys, or land stations and this has resulted in the dependence on observations of rotation in the satellite-observed cloud mass. Since clouds at higher levels may obscure the low-level cloud motions and delay detection of rotation, the working hypothesis for the 1999 hurricane season's demonstration project was that the vorticity might first be detected in the surface winds seen by QuikSCAT.