Tuesday, 11 January 2000
Thomas J. Kleespies, NOAA/NESDIS, Camp Springs, MD; and L. M. McMillin
The recent launch of NOAA-15 and the accompanying on-orbit verification efforts caused us to take a critical look at the radiometric characteristics of the new Advanced Microwave Sounding Unit (AMSU) instruments. It was discovered that certain instrument count positions are populated preferentially in histograms of the scene data. These preferential positions are separated by powers of two. This behavior has been seen in all of the radiometers examined thus far (AMSU-A, AMSU-B, High Resolution InfraRed Sounder (HIRS) on TIROS-N and NOAA-6 through 15). The cause of this anomaly is small imperfections in the resistors that make up the analog to digital converter in each instrument. The errors associated with this phenomenon can be as much as several tenths of a degree in the colder sensing infrared channels.
Another oddity has been discovered in the level 1B HIRS data. The raw data flowing from the HIRS is signed binary, in which zero can be either positive or negative.
In most of the HIRS channels that sense both positive and negative counts, the negative zero is overpopulated by approximately a factor of three, and the positive zero is unpopulated. This problem apparently causes the entire range of counts to be misrepresented for the affected channels. This second phenomenon seems to be endemic in the design or fabrication of the instrument, since it is evident in all instruments examined, from TIRON-N, and NOAA-6 through 15.
This paper will describe the nature of these problems, the impact on the radiances, and present a potential method for mitigating their effect.
- Indicates paper has been withdrawn from meeting
- Indicates an Award Winner