Correction Method for Infrared Detector Confirmed; Error in Clear Sky Bias Condition Remains Unresolved
Turner, D. D., National Oceanic and Atmospheric Administration
To obtain accurate, high-spectral resolution ground-based observations of infrared radiation, the Department of Energy's Atmospheric Radiation Measurement (ARM) Program uses a sensor called an Atmospheric Emitted Radiance Interferometer (AERI). The AERI instrument uses a mercury cadmium telluride (MCT) detector, which provides excellent sensitivity to infrared radiance from 5-25 μm. The MCT detector, however, has a non-linear response, which means that its output is not linearly proportional in wavelength to the measured radiant energy. A set of experiments conducted by ARM researchers in November 2003 and January 2004 confirmed the accuracy of a nonlinearity correction for calibrating the MCT detector, and also ruled out this procedure as the source of error found in clear, dry conditions calculated by a line-by-line radiative transfer model (LBLRTM).
To evaluate the nonlinearity correction procedure developed for AERI MCT detectors, ARM researchers used AERI data collected in November 2003 from the ARM Program's Southern Great Plains site in Oklahoma. Using uncorrected data, they found only a 0.3% error in the measured radiation (relative to the ambient radiance) for cloudy scenes, but for a clear sky case the error jumped to 0.75%. Therefore, uncertainties associated with the specifications of the nonlinearity correction were deemed most important for clear sky AERI applications, particularly in the 8-13 μm window. Within the ARM Program, an effort to compare downwelling radiance observed by the AERI with calculations from the LBLRTM was identified as the main application which could be affected by an inadequate nonlinearity correction. Results of this comparison effort at several sites showed that calculated residuals contained a significant and consistent negative bias for clear sky, low water vapor conditions.
To determine whether the nonlinearity correction was the source of this error, the researchers conducted a side-by-side test of two extended range AERIs at ARM's North Slope of Alaska locale in January 2004. A comparison of clear sky data demonstrated that the two instruments agreed to within measurement uncertainties. The researchers therefore concluded that the nonlinearity corrections applied to the two systems were correct, leaving the secondary question of the AERI LBLRTM clear sky bias unresolved.
The nonlinearity of the MCT detector, if not corrected for or inadequately corrected for, can lead to significant errors in the 8-13 μm window radiance data observed by the AERI in clear sky conditions. These errors are significant when evaluating clear sky radiative transfer algorithms—such as the LBLRTM—with AERI data, as well as studies involving aerosol or other optically thin layers using AERI data.
REF: Turner, D.D., H.E. Revercomb, R.O. Knuteson, R.G. Dedecker, and W.F. Feltz, "An Evaluation of the Nonlinearity Correction Applied to Atmospheric Emitted Radiance Interferometer (AERI) Data Collected by the Atmospheric Radiation Measurement Program," ARM TR-013, September 2004.