How accurate are long-term surface flux measurements for trend detection and variability analysis?

Recently, the calibration and usage of solar broadband (.3 to 4 micrometer response) total (direct+diffuse) pyranometer measurements have been called into question due to documented problems with a night-time offset. This issue of the night-time offset combined with the already known problems associated with the angular response of these instruments, renews the questions concerning the usefulness of these measurements. Yet, there remains a large and long time series of such measurements. So, the question remains as to whether or not these measurements are useful for climate questions and if so how to best use them.

In this presentation, we give a brief overview of the uncertainties and quality control issues relating to a large database of daily broadband total measurements collected for use by the Global Energy and Water Cycle Experiment (GEWEX) Surface Radiation Budget project. These observations are used to validate solar radiative flux estimates that pre-date more modern Baseline Surface Radiometer Network (BSRN) measurements. The measurements from many of these sites are used to form the monthly averages available through the Global Energy Balance Archive (Gilgen and Ohmura, 1998). For a limited number of sites we will determine the mean standard deviation and the seasonal-to- interannual/intraannual climate variability as a function of the characteristics of each site (e.g., latitude, water vapor, surface ecosystem type, and elevation). We apply the statistical techniques to the time series of measurements from these sites as developed by Weatherhead et al. (1998) to quantify the usefulness of these data for trend detection and variability analysis. These methods account for the uncertainties in the measurements and for gaps in the time series. We will attempt to match appropriate application strategies with the reliability of these data.