Engineering topics
Publications and source records attributed to Abrams, R. B..
Descriptive and procedural background material for understanding results from radiometric calibration of the reflective bands and digital image products of the thermal mapper (TM) are discussed. The radiometric subsystems of the TM are described, with emphasis on the internal calibrator pulse shapes and timing cycle.
LANDSAT-4 Thematic Mapper (TM) digital image products recorded onto computer compatible tapes (CCTs), which were available for internal research purposes prior to August, 1983, are reviewed. The SCROUNGE image processing system at Goddard Space Flight Centr generated in tape formats: (1) raw band-sequential data (CCT-BT), generally used for internal transportation of digital data from one ground processing system to another; (2) calibrated data (CCT-AT), useful for reseachers doing radiometric characterization; and (3) geometrically resampled data (CCT-PT), the final product. The formats represent different steps in the process of producing fully-corrected TM data. The CCT-BT images are re-sequenced from telemetry format to image format, but are uncorrected radiometrically and geometrically. The CCT-AT images had data from two faulty data channels replaced and all data radiometrically calibrated. The CCT-PT images were resampled by cubic convolution procedures to provide a geometrically corrected image using satellite ephemeris and altitude data and scan-mirror correction data. The final product, the CCT-PT, is the one to which all of the radiometric and geometric corrections were applied.
The radiometric subsystem of NASA's LANDSAT-4 Thematic Mapper (TM) sensor is described. Special emphasis is placed on the internal calibrator (IC) pulse shapes and timing cycle. The procedures for the absolute radiometric calibration of the TM channels with a 122-centimeter integrating sphere and the transfer of radiometric calibration from the channels to the IC are reviewed. The use of the IC to calibrate TM data in the ground processing system consists of pulse integration, pulse averaging, IC state identification, linear regression analysis, and histogram equalization. An overview of the SCROUNGE-era (before August 1983) method is presented. Procedural differences between SCROUNGE and the TIPS-era (after July 1983) and the implications of these differences are discussed.
Computer compatible tapes (CCTs) of LANDSAT 4 thematic mapper (TM) digital image products are compared and reviewed. The following tape formats are discussed: (1) raw band-sequential data (CCT-BT); (2) calibrated data (CCT-AT); and (3) geometrically resampled data (CCT-PT). Each format represents different steps in the process of producing fully corrected TM data. The CCT-BT images are uncorrected radiometrically or geometrically, CCT-AT data are radiometrically calibrated, and CCT-PT images are both radiometrically and geometrically corrected.
Prelaunch and postlaunch internal calibrator, image, and background data is to characterize the radiometric performance of the LANDSAT-4 TM and to recommend improved procedures for radiometric calibration. All but two channels (band 2, channel 4; band 5, channel 3) behave normally. Gain changes relative to a postlaunch reference for channels within a band vary within 0.5 percent as a group. Instrument gain for channels in the cold focal plane oscillates. Noise in background and image data ranges from 0.5 to 1.7 counts. Average differences in forward and reverse image data indicate a need for separate calibration processing of forward and reverse scans. Precision is improved by increasing the pulse integration width from 31 to 41 minor frames, depending on the band.
The image characteristics of digital data generated by LANDSAT 4 thematic mapper (TM) are discussed. Digital data from the TM resides in tape files at various stages of image processing. Within each image data file, the image lines are blocked by a factor of either 5 for a computer compatible tape CCT-BT, or 4 for a CCT-AT and CCT-PT; in each format, the image file has a different format. Nominal geometric corrections which provide proper geodetic relationships between different parts of the image are available only for the CCT-PT. It is concluded that detector 3 of band 5 on the TM does not respond; this channel of data needs replacement. The empty bin phenomenon in CCT-AT images results from integer truncations of mixed-mode arithmetric operations.
Depletion of solar radiation at discrete wavelengths by nitric oxide has a significant impact on the dissociation rate of this gas in the delta(0, 0) and delta(1, 0) bands. Inclusion of the opacity provided by a typical NO profile reduces the dissociation rate in the upper stratosphere to 50-75 pct of that predicted when the optical depth calculation omits this contribution. A substantial fraction of the NO column abundance as measured near the stratopause resides in the lower thermosphere where correlations of NO with solar and magnetic activity are well documented. Variations in the thermospheric NO abundance therefore modulate the radiation field at the precise wavelengths absorbed by this molecule in the mesosphere and upper stratosphere. Predicted changes in the dissociation rate arising from a varying thermospheric opacity exceed 8 pct throughout the mesosphere and reach 15 pct between 65 km and 95 km.
It is pointed out that numerous fluorescent emissions from the Herzberg bands of molecular oxygen lie in the spectral region 242-300 nm. This coincides with the wavelength range used by orbiting spectrometers that observe the Rayleigh backscattered spectrum of the earth for the purpose of monitoring the vertical distribution of stratospheric ozone. Model calculations suggest that Herzberg band emissions in the dayglow could provide significant contamination of the ozone measurements if the quenching rate of O2(A3Sigma) is sufficiently small. It is noted that this is especially true near 255 nm, where the most intense fluorescent emissions relative to the Rayleigh scattered signal are located and where past satellite measurements have shown a persistent excess radiance above that expected for a pure ozone absorbing and molecular scattering atmosphere. Very small quenching rates, however, are adequate to reduce the dayglow emission to negligible levels. Noting that available laboratory data have not definitely established the quenching on the rate of O2(A3Sigma) as a function of vibration level, it is emphasized that such information is required before the Herzberg band contributions can be evaluated with confidence.
Fluorescent emissions from nitric oxide appear imposed upon the Rayleigh backscattered spectrum of the earth's atmosphere between 250 and 300 nm in wavelength. Satellite instruments designed to monitor the global ozone distribution can routinely observe these signals although techniques for exploiting the data are not yet available. Application of a radiative transfer equation developed for an atmosphere including absorption by ozone, molecular scattering, and nitric oxide fluorescence shows the three most prominent NO emissions relative to the 250-300 nm backscattered sunlight to be the (1,4), (1,6), and (0,3) gamma bands. Analysis of the contribution function for each emission band indicates that the fluorescent signals can provide useful information on the magnitude and variability of nitric oxide between 40 and 140 km in altitude.
Satellite-borne measurements of solar radiation backscattered by the earth and atmosphere allow a determination of the surface albedo in the near ultraviolet. The data yielded albedos at tropical latitudes during a 15-month period in 1979-1980. The wavelengths studied are 331.2 and 339.8 nm. Sixty-nine percent of the measurements imply albedos less than 0.3. Higher values include a contribution due to reflection from clouds and are consistent with previous estimates of the fractional cloudover in the tropics. An albedo histogram based on a bin width of 0.1 shows results in the range 0.1-0.2 to be the most frequent, appearing in 29% of the cases, although values which span the entire range 0.0-1.0 are present. The derived albedos show no correlation with solar zenith angle. This result is consistent with the fact that, at the wavelengths considered, the diffuse integrated intensity is larger than the attenuated direct solar beam near the ground.
Analysis of backscattered ultraviolet radiances observed at tropical latitudes by the Atmosphere Explorer-E satellite reveals both annual and semiannual cycles in upper stratospheric ozone. The annual variation dominates the signal at wavelengths which sense ozone primarily above 45 km while below this, to the lowest altitude sensed, 35 km, the semiannual component has comparable amplitude. Comparison of radiance measurements taken with the same instrument at solar minimum during 1976 and solar maximum in 1979 show no significant differences. This suggests that variations in upper stratospheric ozone over the solar cycle are small, although the data presently available do not allow a definite conclusion.
Analysis of more than one year of backscattered ultraviolet radiances obtained from an equatorial orbit by the Atmosphere Explorer-E satellite reveals an annual cycle in tropical, high altitude ozone which shows no latitude dependence between 20 degrees south and 20 degrees north. The amplitude of the variation increases with altitude. In the 35 to 45 km altitude region statistics of the data suggest no variation whatsoever, while in the lower mesosphere the radiances indicate an ozone increase of 25 to 30 percent between January and July with a decrease thereafter provided the solar irradiance remained constant except for the variation with earth-sun distance.