Since 1992 solar ultraviolet (UV) spectral irradiance (290-325 nm) has been measured at two Italian stations of Rome (urban site) and Ispra (semirural site) using Brewer spectrophotometry. The data collected under all sky conditions, are compared with the output of a sophisticated radiative transfer model (System for Transfer of Atmospheric Radiation-STAR model). The STAR multiple scattering scheme is able to cope with all physical processes relevant to the UV transfer through the atmosphere. The experience so far acquired indicates that, in spite of the unavoidable uncertainties in the input parameters (ozone, aerosol, surface albedo, pressure, temperature, relative humidity, cloud cover), measured and computed clear sky irradiances are in reasonable agreement. The STAR model is applied to build up the solar UV geographic patterns in Italy: the daily dose in the range 290-325 um is computed at about 70 sites where a thorough and homogeneous climatology is available. For each month the concept of an idealized "standard day" is introduced and the surface distribution of solar UV field determined. The map of solar UV patterns for Italy, available for the first time, meets the study requirements in the field of skin and eye epidemiology, as well as in other investigations dealing with the impact of UV on the biosphere. The results are interpreted in terms of atmospheric and meteorological parameters modulating UV radiation reaching the ground.

INTRODUCTION

The amount of solar ultraviolet (UV)t radiation at the earth's surface depends on the incoming solar energy and on the transmission properties of the atmosphere. LTV radiation is strongly absorbed by ozone in the spectral range of 200-310 nm, while the attenuation is increasingly weaker at higher wavelengths. After the discovery of the Antarctic ozone hole in 1985, the risk of a possible UV increase at ground, as a consequence of the observed stratospheric ozone depletion, has awakened the interest of the scientific community in view of potentially harmful effects on terrestrial and aquatic ecosystems (1-3). While the downward trend in stratospheric ozone is assessed, beyond any doubt, on the basis of observation supported by theory (4,5), a corresponding upward trend of UV irradiance is not easily detectable (6) even if several studies show increasing levels of UV irradiances in response to ozone reductions (7-11).

Reliable UV time series are much shorter than the total ozone series, and moreover the UV field is affected by complex absorption and scattering processes in the atmosphere (12) in which both astronomic and atmospheric factors (solar zenith angle, sun-earth distance, cloud cover, turbidity, albedo, pressure, temperature, humidity) play an important role. Some of these factors, in turn, may have a time behavior characterized by trends or long-period fluctuations.

Models are able to quantify UV irradiances at any time and location, provided the extraterrestrial solar radiation and the state of the atmosphere are known (6). Simulation tests allow to improve the knowledge on the UV-radiation sensitivity to atmospheric factors (such as ozone, aerosol load and cloudiness) while the coupling of Chemistry Transport Models (CTM) (13) and radiation transfer schemes can provide valuable forecasts of the ultraviolet index (UVI) (14,15), whose main purpose is to develop a public awareness about the effects of prolonged exposure to sun's rays. It is also possible to combine radiative transfer computations with satellite remote sensing data for mapping the surface UV fields over large regions ( 16,17).

In this paper, we have compared Rome and Ispra UV (290-325 nm) series (1992-1998) with multiple scattering model results.

An analysis of the observed and computed (spectral and integrated) irradiances has been carried out during an experimental field campaign to assess the agreement of the two sets of data when the most important atmospheric input parameters are known (total O3, NO2 and SO2 column, aerosol extinction coefficient, vertical pressure, temperature, relative humidity and cloud cover).

Our final aim is to derive the geographic patterns of solar UV fluxes in terms of mean climatological characters for a large number of sites, taking into account topographic factors as well. This work is stimulated by the interest and requirements expressed by scientists of disciplines, like medicine and biology, engaged in the investigation of the impact of UV radiation on the biosphere and, in particular, on epidemiological studies on human's skin and eye pathologies.

THE DATA SET

Measurements of UV irradiance and total ozone, by means of Brewer spectrophotometry, have been started at Rome, urban site (latitude, 41.9 deg N; longitude, 12.52 deg E; altitude, 60 m), and at Ispra, semirural site (latitude, 45.8N; longitude, 8.63 deg E; altitude, 240 m), in 1992. These are the only stations regularly measuring UV irradiance in Italy (18).

Both Brewers are single-monochromator instruments (version MKIV); measurements of global (direct plus diffuse) radiation are made in the spectral range from 290 to 325 nm at 0.5 nm stepwidth with a full bandwidth at half maximum of 0.5 to 0.6 nm. Values of damaging UV radiation, computed using the erythemal action spectrum (19) are also available for both stations. UV scans are made at every 50 zenith angle and near local solar noon; about 20 UV scans per day are available for both sites.