SGP/RSS/C1 - Stability and nonlinearity corrections

From the very beginning after deployment in May 2003 we observed several effects that we 
did not anticipate during the pre-deployment lab test.   They were: (1) Rapid drop in 
responsivity; (2) Responsivity having 1/lambda wavelike periodicity; (3) Appearance of 
nonlinearity within 3 months; (4) Larger than expected diurnal wavelength changes and long term 
wavelength drift. 

Right away we responded with development of correction algorithms and methods.  (1) and 
(2) are corrected to within 2-3% with biweekly lamp calibrations; (3) Nonlinearity factor 
is derived from calibration data and applied in correction; and (4) is efficiently 
corrected with Fourier correlation algorithm applied to every spectrum (once a minute).

Furthermore we perform Langley regression processing  (not a part of official data stream) 
that corrects longer term drifts in responsivity and  should correct calibration lamp 
biases.  The Vo's from Langley regressions are available via RSS  web page.

The overhaul of the instrument is planned to eliminate or at least reduce  nonlinerarity 
stability problems.  We hope that after the overhaul data processing would  be simplified 
and the frequency of calibrations would be reduced.

• Raw data stream - documentation not supported(raw)

• Wavelength(wavelength)
• scan pixel shift in red(red_shift)
• Radiation, Shortwave, Downwelling, Global(totalhorizontal)
• 1 sigma std. dev, fraction of direct normal(sdf_dn)
• scan pixel shift in blue(blue_shift)
• 1 sigma std. dev, fraction of total horizontal(sdf_th)
• Radiation, Shortwave, Direct normal(directnormal)
• 1 sigma std. dev, fraction of diffuse(sdf_diff)
• Radiation, Shortwave, Downwelling, Diffuse(diffuse)