L. R. Thorne and J. Vitko, Jr.
Sandia National Laboratories
The Department of Energy (DOE) has recently instituted the Atmospheric Radiation Measurement (ARM) program to further its long-term goal of improving the accuracy of general circulation models (GCM). Among other applications, GCMs are used to predict climate changes that may result from increased atmospheric CO2 produced by human activities (e.g., burning of fossil fuels for energy production). Improving the subgrid-scale parameterization of clouds and radiation is viewed as critical to increasing GCM accuracy (Atmospheric Radiation Measurement Program 1990). As a part of improving GCM parame-terizations, the ARM program, plans to make detailed ground-level radiation budget measurements over a 30- x 30-km area which will be expanded to a 200- x 200-km area using supplementary data from extended sites.
Satellite measurements will play a significant role in the ARM program by providing measurements of the radiation budget at the top of the atmosphere (TOA) for the region associated with the ARM site. The most important radiation measurements include incident solar radiation, reflected solar radiation, and emitted terrestrial infrared radiation. This report focuses on operational satellite radiometers that measure reflected solar radiation in the 0.4-1.0-mm spectral region (visible and near infrared wavelengths). The satellite-based radiometers of greatest value to the ARM program are scanning radiometers (Advanced Very High Resolution Radiometer [AVHRR], Satellites Probatoires d'Observations de la Terre/High Resolution Visible [SPOT/HRV], Landsat Thematic Mapper [TM], and Visible and Infrared Spin-Scan Radiometer/VISSR Atmospheric Sounder [VISSR/VAS]), because they can spatially resolve the radiance of the atmospheric and surface regions near the ARM site. Selected characteristics of these radiometers are briefly summarized in Table 1. Non-scanning satellite radiometers exist (e.g., Earth Radiation Budget Experiment [ERBE]), but their fields of view are much wider than the 200- x 200-km extended ARM site and, thus, are not compatible with the ARM measurement objectives.
Absolute radiation measurements must be made in order to determine the TOA radiation budget and, as such, require a calibration reference. Calibration is critically important to 1) making absolute measurements, 2) merging one data set with other data sets (either from the same instrument over a period of time or from one instrument to another instrument), and 3) monitoring long-term trends in the radiation field (climate). Unfortunately, recent studies (Staylor 1990; Abel 1990a, 1990b; Price 1987) have shown that calibration of the visible channels generally changes significantly from the prelaunch calibration (by up to 25% for AVHRR on NOAA-9,-10 [Teillet et al. 1990; 1988]) and that the calibration can drift over time (~7% per year for AVHRR on NOAA-9 [Teillet et al. 1988]). Moreover, unlike the infrared channels, there is no on-board absolute calibration for the visible-wavelength channels of the current operational satellites (Price 1987) other than a space view which provides a practical zero radiance.
One solution to this calibration problem is to use targets on the Earth's surface as reference radiance sources. In this way, an indirect (i.e., not intrinsic to the instrument) calibration can be made. The purpose of this report is to evaluate the accuracy of various indirect calibration methods reported in the literature to determine what limits the accuracy of these methods, to evaluate the possibility for improving their accuracy, and to recommend which methods should be used in the ARM program.