OK, I admit that I am a certifiable radiometer geek, so I think this data from November 25 is really cool! The Storm Peak Lab (SPL) radiation plot and Total Sky Imager (TSI) movie both show a mostly clear-sky day, but the radiation plot that displays the total shortwave radiation (SW) measured by the Eppley precision spectral pyranometer (PSP) is much greater than the SPN-1 pyranometer total SW for the entire morning. Both plots exhibit smooth curves, so the first thought would be that there was something blocking the SPN-1 dome. But not so fast! This is a phenomenon unique to cold climates and low solar elevation angles.
What happened is that the PSP had frost on the top of its dome. The maximum solar elevation is just over 28 degrees at SPL on the day the measurements were taken. So the direct solar beam was reaching the detector at a low angle below the frost contamination, but the frost on the very top of the dome was also being illuminated by the direct sun as well.
The clear sky is blue, with the longer wavelengths not being scattered very much, but the frost is white and scatters all SW wavelengths about equally. Thus, greater irradiance is measured from that portion of the instrument view. The effect is the same as if a white cloud had parked right above the PSP all morning, giving an enhanced positive downwelling radiation cloud effect (i.e. SW greater than the equivalent clear-sky amount). In this case there was no cloud, and this is a PSP measurement error caused by frost on the dome. But fortunately the SPN-1 with its on-board internal heaters did not have frost on its dome and was not affected.
How do I know this is the case? Two things: experience, and the extra plot I added to today’s upload. The extra plot shows the SW irradiances plotted versus the cosine of the solar zenith angle. Note again that this was a mostly clear-sky day, so the afternoon total SW curve should almost match the morning curve, showing only small changes due to changing water vapor and aerosol loadings. But as the plot shows, the PSP total SW has a large “loop-de-loop” pattern. If this pattern persists day after day, it indicates that the instrument is not level. But in this case, the plot shows the PSP values being significantly too high for the morning. The frost melted away during the day, so the later afternoon readings are not affected, and agree well with the SPN-1 total SW measurements.
This case was easy to spot because it is a mostly clear-sky day. For overcast cases, the effect of a small amount of frost on the top of the dome is minimal because there are clouds directly above the instrument, and both situations behave about the same as far as the measurements are concerned.
The difficult case is for partly cloudy skies, where visually this is not so obvious: there are differences between the PSP and SPN-1 total SW, but it’s harder to tell which is the “good” measurement and which is the “bad.” It will be an interesting challenge for me to produce the “best estimate” radiation data set at the end of the STORMVEX campaign. — Cheers, Chuck