Toward a New UV Index Diagnostic in the Met Office's Forecast Model

The United Kingdom sporadically experiences low ozone events in the spring which can increase UV to harmful levels and is particularly dangerous as sunburn is not expected by the public at this time of year. This study investigates the benefits to the UV Index diagnostic produced by the UM at the Met Office of including either, or both of, a more highly resolved spectrum, and forecasted ozone profiles from the ECMWF CAMS database. Two new configurations of the spectral parameters governing the radiative transfer calculation over the UV region are formulated using the correlated‐k method to give surface fluxes that are within 0.1 UV Index of an accurate reference scheme. Clear‐sky comparisons of modeled fluxes with ground‐based spectral observations at two UK sites (Reading and Chilton) between 2011 and 2015 show that when raw CAMS ozone profiles are included noontime UV indices are always overestimated, by up to 3 UV indices at a low ozone event and up to 1.5 on a clear summer day, suggesting CAMS ozone concentrations are too low. The new spectral parameterizations reduce UV Index biases, apart from when combined with ozone profiles that are significantly underestimated. When the same biases are examined spectrally across the UV region some low biases on low ozone days are found to be the result of compensating errors in different parts of the spectrum. Aerosols are postulated to be an additional source of error if their actual concentrations are higher than those modeled. Plain Language Summary Providing the public with accurate forecasts of the risk of increased exposure to UV allows them to take protective measures to prevent sun damage. The UK sporadically experiences low ozone events in the spring which can increase UV and is particularly dangerous as sunburn is not expected by the public at this time of year. This study investigates the benefits to the UV Index produced by the Met Office of including either, or both of, more detail in the calculation of the relevant energy at UV wavelengths, and real‐time estimates of the vertical ozone distribution. A comparison of modelled UV with ground‐based measurements on cloud‐free days at two UK sites (Reading and Chilton) between 2011 and 2015 show that when ozone profiles are included noontime UV indices are always overestimated, by up to 3 UV indices when ozone is unusually low and up to 1.5 on a typical clear summer day, suggesting the ozone concentrations used are too low. A more detailed calculation of