Estimation of fractional sky cover from broadband shortwave radiometer measurements

We outline a methodology for estimating fractional sky cover for an effective 160° field of view from an analysis of surface measurements of downwelling total and diffuse shortwave (SW) irradiance. The data are screened for optically thicker overcast cases, after which an empirically derived formulation is used to estimate the fractional sky cover for the remaining data. The retrieved fractional sky cover time series is then evaluated to mitigate times of anomalous behavior caused by the thick overcast screening. The resultant sky cover estimates show a high degree of repeatability given nominally well maintained and operated radiometer systems and the use of the Long and Ackerman (2000) methodology for estimating the clear‐sky total and diffuse SW. Thus the resultant fractional sky‐cover estimates appear to be fairly independent of the particular climate regime and model of radiometers used, at least for the climate regimes we have tested so far. The sky‐cover estimates agree to better than 10% root mean square sky cover amount with sky imager retrievals and human observations, which is as good as the agreement between sky imaging systems and observers themselves. As such, this methodology becomes a powerful tool for satellite and model validations and climatological analyses including the study of trends in cloud amount. Analysis shows that the technique also produces realistic frequency distributions, showing that the continental midlatitude regimes included in the study are typified by clear‐sky occurring about 1/3 of the time, overcast about 1/3 of the time, and partly cloudy skies to varying extent occurring the remaining 1/3 of the time. By contrast, the tropical western Pacific oceanic regime during the Nauru99 field experiment exhibits far more frequent occurrence of partly cloudy skies, with sky cover amounts of 20% to 50% occurring about half the time.