| Abstract: |
The accuracy of a calibration target or blackbody directly impacts the resulting brightness temperature measurement of a radiometer system. Typical space-borne radiometer systems scan across an internal calibration target as part of their operational procedure, and use a cold space look as the second calibration point. The dependence of brightness temperature on view angle of the calibration target is often overlooked but is an important contributor to overall calibration uncertainty.
Radiometer calibration targets can have significant scan-angle dependence, but past studies have only considered normal incidence performance. Brightness temperature is dependent on physical temperature and emissivity. Emissivity is often approximated through a measurement of normal incidence monostatic reflectivity, and in this paper we investigate this approximation. We present measured monostatic reflectivity of two different calibration target geometries at incidence angles up to 15 degrees off of normal. We investigate a pyramidal-array and a conical-cavity blackbody calibration target at two frequencies in G-band, 165. GHz and 183.3 GHz. The reflectivity is measured by stepping the targets through space and measuring the complex S11 using a network analyzer, G-band frequency extender head, and standard gain horn antenna. Stepping through separation distance increments of less than a sixteenth of a wavelength allows the interference standing wave to be clearly resolved. A flat metal plate is also measured with the same technique at normal incidence allowing a one-port calibration, similar to that of a network analyzer using a short and multiple offset-shorts. The angle of the frequency extender head and antenna is varied on a rotating stage at 1 degree increments between 0 degrees and 15 degrees off of normal. The distance stepping procedure is repeated for each incidence angle and both calibration targets for a total of 32 calibrated reflectivity measurements.
The pyramidal array exhibits a maximum in monostatic reflectivity at or very near normal incidence, while the monostatic reflectivity of the cone increases with incidence angle. The magnitude of measured reflectivity from the cone is below that of the pyramidal array at both frequencies across the range of view angles. The maximum reflectivity of the pyramidal array is -34.5 dB and -36.1 dB at 165.5 GHz and 183.3 GHz respectively. Reflectivity of this magnitude corresponds to an emissivity as low as 0.981, a poor approximation of a blackbody, causing calibration bias errors of more than 5 K at brightness temperatures of 300 K. The maximum reflectivity of the conical cavity is -58 dB and -62.2 dB at 165.5 GHz and 183.3 GHz respectively suggesting much higher emissivity than the pyramidal array.
We have also modeled the two targets using finite element modeling software to gain understanding of the validity of the monostatic measurements for approximating emissivity. The finite element software allows a full wave analysis of the geometries and allows us to calculate the full hemispherical bi-static scattering field as opposed to only the monostatic or backscatter case. The monostatic reflectivity is often used as an approximator for the full hemispherically integrated reflectance, but this assumption is often unsupported. We show that for a pyramidal array, the monostatic reflectivity is a reasonable approximator of the total reflectance, whereas for the conical cavity the monostatic reflectivity significantly underestimates the total reflectance. The reason for this is the large amount of non-backscatter reflection for the conical geometry compared to the pyramidal array. The simulation results suggest that the monostatic reflectivity may be a better approximator of emissivity for pyramidal array calibration targets than for conical cavities.
|
