Duration of vitamin D synthesis from weather model data for use in prospective epidemiological studies

In order to investigate the influence of solar radiation on vitamin D status and its association with different health outcomes in population based studies, appropriate estimates of the subjects' UV radiation exposure are needed. This unique study describes a method that estimates the daily number of vitamin D effective hours (VD-hours) at arbitrary ground locations throughout the period 1957-2002. The method is particularly suited for large-scale prospective epidemiological studies with questionnaire-based information on sun exposure, and where blood measures of vitamin D status are not available. The model takes total cloud cover fraction and total ozone column at noon as input from the ERA-40 data series (i.e. the 40 + year European Centre for Medium-Range Weather Forecasts Re-Analysis archive). By comparing the model results against high accuracy measurements at two different locations in Norway, we found the method for estimating the number of VD-hours to be accurate within 2.5 +/- 7% or better for moderate solar zenith angles (< 65 degrees). For higher solar zenith angles (> 65 degrees) the results are more variable, but the contribution to a population's vitamin D level from solar radiation when the sun is this low in the sky is rather small. The program code to compute VD-hours from ERA-40 files is written in Perl (v 5.8.7) and may be obtained free of charge by contacting corresponding author.