People frequently ask us questions related to the uncertainties of pyranometer measurements. As the sales & service engineer at the head office I have answered these questions coming from all over the world, especially from customers in renewable energy. I have summed it up in a short overview.
During each calibration process we transfer the sensitivity of a reference pyranometer (traceable to the world standard in Davos, Switzerland) to the instrument that needs calibration. The calibration process of the reference pyranometer and the conditions under which the calibration transfer is performed, plus all the specifications of the instruments, determine the actual accuracy of the sensitivity in which this transfer results. This is defined as the calibration uncertainly.
Each Pyranometer from Kipp & Zonen is supplied with an individual calibration certificate that shows the traceability of the reference pyranometer and the calibration uncertainty. An example can be seen on the Certificate under the header “Justification of total instrument calibration uncertainty”. Click here to see an example.
When a pyranometer is in operation all the measurements include the calibration uncertainty, but the performance of it is correlated to a number of parameters; such as temperature, level of irradiance, angle of incidence, etc. The supplied sensitivity figure, or a defined output scale, is used to calculate the irradiances. But, if the operating conditions differ significantly from the calibration conditions, additional uncertainty in the calculated values must be expected. The measurement uncertainty is often described as the maximum expected uncertainty in the daily total of irradiance with respect to the ‘absolute truth’. The confidence level of Kipp & Zonen uncertainties is 95%, which means that 95% of the data-points lie within the given uncertainty interval representing the absolute value.
It does not mean that throughout each moment of the day you can expect the same uncertainty, because the conditions are changing. At sunset and sunrise the directional response causes the biggest error at those particular moments, while the same effect causes very little error at noon. When you are not using the instrument at high solar zenith angles the daily uncertainty will be much better. At noon and with temperatures corresponding closely to the calibration environment the extra error of the measurement can should be small. So each hourly uncertainty can be different than the daily uncertainty.
Daily and hourly uncertainty
Kipp & Zonen determines daily uncertainty figures for a range of conditions based on the measured pyranometer characteristics and many years of experience in field measurements. Individual uncertainties can be determined for the effects of the parameters of the measuring environment and can be combined using the RSS (Root Sum Square) method. The total daily uncertainty of the instrument includes the calibration process and by RSS (Root Sum Square) calculation of the instruments daily uncertainty and the calibration uncertainty a RSS calculation is Daily Uncertainty = √ (measurement uncertainty² + calibration uncertainty²).
For a pyranometer meeting the ISO 9060 Secondary Standard classification (the highest quality) an uncertainty in the hourly irradiance total of about 3%, and in daily total an error of 2%, is expected for most non-extreme climates. Some of the response variations cancel each other out if the integration period is long.
The daily and hourly uncertainties assume that the pyranometer has a valid calibration, is correctly installed and maintained, and that the data logger does not degrade the instrument performance. Most importantly, the dome must be clean and dry! Dirt and water droplets will have a much greater effect than all the other variables.
Many years of experience has shown that pyranometer performance can be improved concerning zero offsets type A and B by using a well-designed ventilation system. The Kipp & Zonen CVF4 ventilation unit is recommended and can reduce these errors by a half. It will also help to keep the dome clean and dry (particularly if heated) and maximise the up-time of good data.
Each CMP21 and CMP22 model come with temperature response and direction response profiles that allow you to correct the measurement data for the influence of temperature and the angle of the sun, resulting in a smaller uncertainty. SMP21 and SMP22 Smart models have individually optimized internal digital correction for the influences of temperature.
Another way to improve daily uncertainties is to exclude data from measurements when the sun is very low in the sky; for example, if it is less than 5° above the horizon.
Remember, the lower your uncertainty is, the more reliable and trustworthy your data will be!
Download our latest whitepaper where the uncertainty of pyranometers is compared to satellite irradiance data here.