Modeling Ultraviolet Radiation at the Earth's Surface. Part II: Model and Instrument Comparison

Abstract

High-resolution measurements in the spectral region of 280?400 nm using a double monochromator are compared with detailed radiative transfer calculations at Reading, United Kingdom (52°N, 0°), for clear and totally overcast days, using aerosol and cloud information deduced from empirical methods. For clear skies, instrument and model agree well in the UVA (320?400 nm), but agreement is worse in the UVB (280?320 nm). A number of possible reasons for the discrepancies are explored. Volcanic aerosols in the stratosphere of the model are found to improve agreement between the model and the instrument for high solar zenith angles by increasing the model UVB irradiances by as much as 6%. Convolving the model surface irradiances with the bandpass of the instrument leads to smaller differences between instrument and model at short wavelengths and also reduces the noisiness of the difference. When the model included stratospheric aerosol and the instrument's bandpass function, UVB irradiances within 10% of the measured irradiances could be produced by the model for clear skies. For cloudy conditions, differences between instrument and model are larger, reaching 20%, integrated over the UVB.