Surface Measurements of Solar Irradiance: A Study of the Spatial Correlation between Simultaneous Measurements at Separated Sites

Abstract

Pyranometers have been used for many years to measure broadband surface incoming solar irradiance, data that is necessary for surface energy budget, cloud forcing, and satellite validation research. Because such measurements are made at a specific location, it is unclear how representative they may be of a larger area. This study attempts to determine a reasonable spacing between measurement sites for such research by computing the correlation, and standard deviation from perfect correlation, between simultaneous measurements of incoming solar irradiance for a network of surface measurement sites covering a 75 km ? 75 km area. Using 1-min data collected from this network of 11 sites during the NASA First ISSCP Radiation Experiment/Surface Radiation Budget Project temporal averages were calculated. The correlation between any two of these sites was determined by comparing simultaneous measurement averages for the 55 possible combinations of site pairs, along with the distances between them. In an attempt to remove the effect of the diurnal cycle, thus leaving clouds as the primary influence on correlation of the radiation field, model results for a clear day were used to normalize measured irradiances and correlations were again calculated. For individual days, the correlation between sites varied widely, depending primarily on the type of cloud cover the region experienced that day. Removal of the diurnal cycle, as expected, significantly decreased these correlation values. Comparisons using the continuous experiment records from 13 October through 2 November 1986, however, show that a relatively high degree of correlation existed with or without the diurnal cycle removed. Plotting these correlation coefficients versus the distance between sites, the expected trend for a decrease in correlation with increasing distance is observed. Results also confirm that, whether using the complete record for the duration of the experiment or by individual day, the correlation between site station pairs increases with increasing averaging times. Finally, the standard deviation from perfect correlation suggests a predictive relationship within about 6% of clear-sky irradiance for daily averages at a distance of 75 km. Thus, a spacing of 150 km between measurement sites seems reasonable for studies of midlatitude frontal weather regimes using daily averages over periods of weeks or more.