UV Index: Is it Validated?

Each year the National Weather Service performs a validation of the summer time UV Index forecasts. This effort incorporates the cooperation of several government agencies and private companies, hospitals, and colleges who provide observations of surface UV radiation. Once received, observations are scrutinized for obvious errors. Statistics of the differences between the observations taken during the solar noon hour and the UV Index forecast are then made. From the validation of the UV Index, the NWS learns what the shortcomings of the UV Index are.

Figure 1 shows a time series of UV Index forecasts and observations at Boston, Massachusetts. Generally, the UV Index forecast coincides with the observations. Notice though that on days when the observations are high that the forecast is smaller and on days when the observations are low, the forecast is greater. This inability to match the large dynamic range of the actual observations is a short coming of our current forecasting technique. However, for the majority of the forecasts at all sites, the probability of making a correct forecast is quite good.

Figure 2 conveys this in the form of a histogram of all the differences between the observation sites and the forecasts. The histogram shows that 26% of the time the UV Index is exactly correct. 65% of the time the UV Index forecast is within ± 1 UV Index unit. And 84% of the time the UV Index is within ± 2 UV Index units.