Direct sun and radiance sphere calibration values are measured at distributed calibration sites. The NASA GSFC calibration facility manages direct solar calibrations and radiance sphere calibrations. In addition, NASA GSFC is responsible for maintaining "master" instruments that meet high operating standards and determining the apparent extraterrestrial constants at Mauna Loa Observatory in Hawaii. Other distributed calibration sites include direct sun measurements in Izana, Spain and radiance sphere measurements in Lille, France (PHOTONS), El Arenosillo, Spain (RIMA), and Canberra, Australia (CSIRO).

Instruments located initially at a calibration facility are considered in "pre-deployment" status. Pre-deployment direct sun and radiance calibration data values are obtained at the calibration facility. The instrument is considered as "field-deployed" once the instrument leaves the calibration facility. After a determined measurement period, the instrument is returned from its measurement location to a calibration facility. At this stage, "post-field deployment" calibration direct sun and radiance values are measured and applied to the data. Some window cleaning or filter maintenance may be necessary. After the lab maintenance, the instruments are considered in a "pre-deployment" status again.

Note: Field-deployed instruments should normally collect data from 6-12 months. Data collected after twelve months may be susceptible to environmental conditions (e.g., spider webs or dust), which may jeopardize the post-calibration values necessary for raising data to the highest quality (i.e., Level 2).

Direct Sun Calibration

Calibration refers to the determination of the calibration coefficients needed to convert the instrument output digital number (DN) to a desired output, in this case aerosol optical depth (AOD), precipitable water (cm), and radiance (W/m2/sr/um). Field instruments are generally returned to GSFC (Goddard Space Flight Center) for intercomparison with reference instruments approximately every 6 to 12 months in order to maintain accurate calibration. The GSFC reference Cimels are calibrated by the Langley technique at Mauna Loa Observatory in Hawaii on a frequent basis. The Langley Plot is a logarithm of the DN taken during these times plotted against the optical air mass between a range of 5 and 2 (between 3.5 and 2 for 340 nm), where the intercept is the calibration coefficient (zero air mass DN) and the slope is the optical depth. Langley plots from NOAA's Mauna Loa Observatory have been made to determine the spectral extraterrestrial voltage for these instruments since 1994. The observatory's high altitude and isolation from most local and regional sources of aerosols provides a very stable irradiance regime in the mornings, and is ideally suited to our purposes.

AERONET reference instruments are typically recalibrated at MLO every 2-3 months using the Langley plot technique. The zero air mass voltages [Vo, instrument voltage for direct normal solar flux extrapolated to the top of the atmosphere (Shaw, 1983)] are inferred to an accuracy of approximately 0.2 to 0.5% for the MLO calibrated reference instruments (Holben et al., 1998). Therefore the uncertainty in AOD due to the uncertainty in zero air mass voltages for the reference instruments is better than 0.002 to 0.005.

Radiance Calibration

For the sky radiance measurements, calibration is performed at the NASA Goddard Calibration Facility using a calibrated integrating sphere to an accuracy of +/- 5%. For the 940 nm channel that includes water absorption, calibration is performed using a variant of the modified Langley method. With respect to the long term stability of the calibration coefficients, the optical interference filters are the limiting factors. On average, there has been a decrease from 1 to 10% per year. Therefore, instruments are calibrated on a 6- to 12-month rotation and filters are changed when needed.

Interpolation and Uncertainty

The Sun-sky radiometers at sites other than GSFC are intercalibrated against a MLO calibrated AERONET reference instrument both before deployment in the field and post- deployment. A linear rate of change in time of the zero air mass voltages is then assumed in the processing of the data from field sites. Our analysis suggests that this results in an uncertainty of approximately 0.01 - 0.02 in AOD (wavelength dependent) due to calibration uncertainty for the field instruments.