NASA Remote Sensing Validation Data: Data Processing
The NASA Remote Sensing Validation Data were processed as described below.
Post-Processed Saudi Network and Baseline Surface Radiation Network Data
Until April 18, 2000, the data available for the Saudi Arabia network stations was quality assessed and flagged based on the use of a single composite calibration factor for the pyranometer deployed at each station. As of April 18, 2000, the global horizontal data posted was corrected for the cosine response of the individual pyranometer deployed at each station. During individual radiometer calibration and characterization, a calibration factor for each of ten 9-degree-wide zenith angle bins was constructed. To obtain the corrected data, the team applied a calibration factor interpolated from the two nearest zenith angle bin responsivities. This interpolated zenith angle correction factor was scaled according to the estimated direct beam contribution to the global measurement.
The post-processed data also contains a global horizontal irradiance (derived global) computed from the direct normal irradiance and the diffuse sky radiation. The global is computed as Idn * cos(z) + D, where Idn is the measured direct normal irradiance measured with a normal incidence pyrheliometer (NIP) (within 5 degree field of view of the sun), z is the solar zenith angle (complement of the solar elevation angle), and D is the diffuse sky radiation measured by a pyranometer shaded with a tracking disk subtending a solid angle of five degrees and obscuring the sun as seen from the pyranometer detector. The Baseline Surface Radiation Network data set also includes the derived global horizontal irradiance, calculated using direct normal measurements from an all-weather cavity radiometer.
Diffuse Pyranometer Thermal Offsets
The Eppley Laboratory Precision Spectral Pyranometer used to measure the diffuse sky radiation is known to have a thermal offset voltage that is a function of the net infrared radiation (or infrared radiation balance between the radiometer and the sky). Study of the correlations between the offsets seen in the Saudi network radiometers and the net infrared show that offsets range from 0 to -15 watts per square meter, with a mean value of -4 watts per square meter and two-sigma standard deviation of 7 watts per square meter. However, the standard error of an estimate based on a linear correlation with net infrared is on the order of +/- 8 watts per square meter. Thus, the team chose not to apply any correction for Saudi diffuse thermal offsets. Adding an additional 4 watts per square meter to the derived global horizontal is just as effective a correction and much simpler. However, the uncertainty in applying the 4- watt correction is itself +/-4 watt per square meter.