Errors in Net Radiometry: Comparison and Evaluation of Six Radiometer Designs

Abstract

Net radiation is a fundamental variable in meteorology, but net radiometers have not been used extensively since the 1960s because of what is commonly considered as unreliable function. This study was conducted to determine whether this problem was caused by design, by calibration, or if some more fundamental problem was associated with net radiometry. Comparative measurements between six different types of net radiometers, calibrated in two wares were carried out under different climatic conditions. Calibration factors obtained from shadowing varied between 75% and 129% of the manufacturer's. The range was 85%?161% for calibrations relative to reference measurements of individual radiation components. Differences between the reference and individual meters were between 6% and 20%. After correction for different responsivities for longwave and shortwave radiation, the differences decreased markedly, especially in nighttime, except for two meters that had bad cosine responses. Three meters showed a marked hysteresis. Excess temperatures at the radiation absorbing surfaces could explain well the temperature dependence of all the instruments. Only one, externally ventilated, instrument gave reliable results independent of dew, snow, rain, and frost. One instrument always showed superior performance, and its output was almost entirely within the accuracy of the reference measurements. This design was based on principles established during the 1960s and earlier, primarily published in German. It could be considered as a de facto net radiation standard until further progress is made. Such progress primarily depends on the possibility of absolute calibration in the longwave range.