SEDAC Socioeconomic Data and Applications Center
Environmental Effects of Ozone Depletion 1998 Assessment


Measurements of UV Radiation

The last decade has seen a great increase in the number and general quality of solar UV measurements. Many new commercial and research-grade UV detectors have been developed, calibration procedures have been improved, and several national and international intercomparisons have been carried out (e.g. McKenzie et al., 1993; Gardiner et al., 1993; Seckmayer et al., 1994; Thompson et al., 1997; Kjeldstad et al., 1997; Webb, 1997; Leszczynski et al. 1998). Agreement among similarly calibrated spectro-radiometers is typically 5% or better in the UV-A range, and 5-10% in the UV-B range. Comparisons between different types of instruments (e.g. spectro-radiometers, broad band meters, filter radiometers) are more difficult, because of the need to put the different measured quantities on a similar basis, for example through the use of model interpolations (e.g. Slaper et al., 1995; Bernhard et al., 1996; Mayer and Seckmeyer, 1996; Booth, 1997).

    Direct measurements of surface UV radiation confirm to a large extent the theoretical expectations, if allowances are made for local conditions (e.g. Booth and Madronich, 1994; Forster et al., 1995; Geogdzhaev et al., 1996; Gardiner and Martin, 1997; Mayer et al., 1997; Pachart et al., 1997; Pfister et al., 1997; Weihs and Webb, 1997). However the analysis, interpretation, and utilization of the measurements still lag behind the growing data archives. Some general patterns of temporal and geographical variations are also being identified (e.g. Seckmeyer et al., 1995; Bais et al., 1997; Bodhaine et al., 1997; Lu and Li, 1997; Orce et al., 1997; Qu et al., 1997; Sasaki et al., 1997; Zerefos et al., 1997; Bigelow et al. 1998). For example, ground-based measurements show that summertime erythemal UV irradiances in the Southern Hemisphere exceed those at comparable latitudes of the Northern Hemisphere by up to 40% (Seckmeyer et al., 1995), whereas corresponding satellite-based estimates yield only 10 to 15% differences (WMO, 1998). Atmospheric pollution may be a factor in this discrepancy between ground-based measurements and satellite-derived estimates. UV-B measurements at more sites are required to determine whether the larger observed differences are globally representative.

Ozone-Related UV Radiation Changes

The evidence is overwhelming that under cloud-free skies UV-B radiation is controlled largely by ozone (Bais et al., 1993; Kerr and McElroy, 1993; Booth and Madronich, 1994; Diaz et al., 1994; Frederick et al., 1994; Bojkov et al., 1995; Huber et al., 1995; Mims et al., 1995; Varatsos and Kondratyev, 1995; Björn and Holmgren, 1996; Hofmann et al., 1996; Bernhard et al., 1997; Bodhaine et al., 1997; Chubarova and Nezval, 1997; Kirchoff et al., 1997; Koskela et al. 1997; Mayer et al., 1997; Seckmeyer et al., 1997; Taalas et al., 1997; and references therein). The response of UV-B radiation to ozone changes is strongly dependent on wavelength because of the rapid increase of the ozone absorption cross section toward shorter wavelengths, with greater sensitivity at short wavelengths and low sun, where the slant ozone optical depth is greater [see for example the review by Madronich et al. (1995), the more recent measurements by Fioletov and Evans (1997), and references therein]. For biologically weighted radiation, measurements under cloud-free skies also show the theoretically expected dependence on ozone. Figure 1.2 shows the sensitivity of erythemal (skin-reddening) UV radiation to the ozone column amount, as measured at a number of different locations and for different solar zenith angles. When expressed on a relative (percentage) basis, the increases in erythemal UV radiation are seen to correlate closely with ozone reductions, whether the latter stem from natural fluctuations and seasonal cycles, or from systematic long-term depletion. The largest percent UV increases are associated with the largest percent reductions in atmospheric ozone.

    Current losses of stratospheric ozone are discussed in the Scientific Assessment of Ozone – 1998 (WMO, 1998). Relative to the values in the 1970s, these are estimated to be about 50% in the Antarctic spring (the ozone hole), about 15% in the Arctic spring, about 6% at Northern Hemisphere mid latitudes in winter/spring, about 3% at Northern Hemisphere mid latitudes in summer/fall, and about 5% at Southern Hemisphere mid latitudes year-round. The corresponding increases in erythemal UV radiation are estimated to be 130%, 22%, 7%, 4%, and 6%, respectively. No significant ozone trend has been found in equatorial regions. The geographical extent and severity of the Antarctic ozone hole have remained essentially unchanged since the early 1990s. Relatively little change in the mid latitude ozone losses has been observed in the last half-decade.

    High levels of UV-B radiation have been observed directly in association with the Antarctic ozone hole (Frederick and Alberts, 1991; Booth et al., 1994; Seckmeyer et al., 1995, Frederick et al., 1998), and on occasion the measured DNA-damaging radiation at Palmer Station, Antarctica (64° S) has been found to exceed maximum summer values at San Diego, USA (32° N) (Booth et al., 1994). It should be noted that monitoring of UV-B irradiances in Antarctica began only in 1989, well after the appearance of the ozone hole, so that the UV-B levels in pre-ozone hole years can be only estimated.

    The smaller increases of UV-B radiation at mid latitudes, while expected, have not yet been detected unambiguously. The record of mid-latitude UV-B measurements is not sufficient for the derivation of statistically significant trends. Little or no reliable historical information on the climatology of UV radiation is available from pre-ozone depletion days (e.g., pre-1980). The few available long-term UV measurement records have been hampered by the difficulty in maintaining stability of UV-measuring outdoor instruments over periods of decades, and by changes in atmospheric turbidity associated with local pollution. For example, measurements obtained with Robertson-Berger meters over 1974-85 suggested a decrease in UV radiation at 14 US locations (Scotto et al., 1988); however a recent re-analysis of these data has identified calibration shifts which, when removed, indicate that no significant trend can be derived from the data record (Weatherhead et al., 1997). Furthermore, increases in UV due to stratospheric ozone reductions may have been masked in some urban areas experiencing increasing levels of local air pollution (e.g. Garadzha and Nezval, 1987). Pronounced ozone losses have occurred for shorter periods of time, e.g. in the few years after the 1991 eruption of Mt. Pinatubo (Gleason et al., 1993) and over the Arctic during six of the past nine winters (Müller et al., 1997; Rex et al., 1997; Stolarski, 1997), with correspondingly higher measured UV-B radiation levels (e.g. Kerr and McElroy, 1993; Fioletov and Evans, 1997).

    Tropospheric ozone is also an effective absorber of UV-B radiation (Brühl and Crutzen, 1989). In urban and industrialized regions, tropospheric ozone is formed by the photo-chemical reactions of some pollutants (nitrogen oxides and hydrocarbons), while in remote regions it stems from both downward transport from the stratosphere, and from in-situ photochemical production by both natural and anthropogenic precursor compounds transported from source regions (WMO, 1994a). Model-based estimates suggest that for industrialized regions of the Northern Hemisphere, the increases in tropospheric ozone since pre-industrial times may have reduced DNA-damaging UV radiation by 3-15% (Brühl and Crutzen, 1989; Madronich et al., 1991; Frederick et al., 1993; Blumthaler et al., 1997; Ma and Guicherit, 1997). Comparisons between spectral UV measurements in Germany and New Zealand also suggest that the lower UV radiation levels observed in Germany may be explained partly by higher tropospheric ozone levels (Seckmeyer and McKenzie, 1992). Recent changes in tropospheric ozone are estimated to be much smaller those since pre-industrial days, with both positive and negative trends reported for different geographic locations (WMO, 1994a, 1998). Their contributions to the trend in the total ozone column are much smaller than those from changes in stratospheric ozone over the same time period (e.g., 1980 to present). Other gases such as sulfur dioxide (SO2) and nitrogen dioxide (NO2) can also reduce atmospheric UV transmission, however significant effects are limited to some extremely polluted urban environments.

Cloud-Related UV Radiation Changes

Clouds generally reduce surface UV irradiances, although the magnitude of this effect is highly variable depending on cloud amount and coverage, cloud cell morphology, particle size distributions and phase (water droplets and ice crystals), and possible in-cloud absorbers (esp. tropospheric ozone). It is useful to note that under some conditions, UV irradiances can be higher than for clear sky, as for example when both direct sunlight and light scattered by clouds (e.g. the sides of bright broken clouds) reaches the observer (Nack and Green, 1974; Mims and Frederick, 1994).

    Numerous statistical correlations between UV transmission and cloud cover have been carried out (e.g. Paltridge and Barton, 1978; Cutchis, 1980; Josefsson, 1986; Ilyas, 1987; Ito, 1993; Björn and Holmgren, 1996; Estupinian et al., 1996; Schafer et al., 1996; Gillotay et al., 1997), but because of the high spatial and temporal variability of clouds, no single value can be given for their effects on surface UV levels. For example, analysis of the Robertson-Berger meter data record shows that monthly average UV levels are reduced by 10-50 percent, depending on season and location in the US (Frederick et al., 1989, 1993). An important aspect of clouds is that, by introducing strong variability in the UV intensities reaching the Earth’s surface, they complicate the detection of long-term trends (Frederick and Erlick, 1995; Lubin and Jensen, 1995; Nunez et al., 1997).

    Cloud transmission depends somewhat on wavelength. In the UV-A region, transmission increases slightly toward shorter wavelengths due to increased multiple reflections between cloud and the surrounding air molecules (Nack and Green 1974; Seckmeyer et al., 1996; Chubarova et al., 1997). At shorter wavelengths, in the UV-B range, long photon path lengths in clouds can increase absorption by tropospheric ozone, resulting in a sharp decrease in effective transmission (Frederick and Lubin, 1988; Mayer et al., 1998).

Aerosol-Related UV Radiation Changes

Small particles suspended in air (aerosols) can have a significant effect on the transmission of UV-B radiation to the surface. The magnitude of the effect is highly variable, depending on the number of particles and their physical and chemical make-up (e.g. sulfate haze, soot, dust, sea-salt aerosols). Such particles are frequently found in the lowest part of the troposphere (the boundary layer), and are often associated with pollution.

    Liu et al. (1991) estimated that anthropogenic sulfate aerosols (associated primarily with fossil fuel combustion) have decreased surface UV-B irradiances by 5-18% in industrialized regions of the Northern Hemisphere. Additional evidence for the role of aerosols comes from simultaneous monitoring of UV irradiances and atmospheric turbidity in relatively polluted environments (Garadzha and Nezval, 1987; Varotsos et al., 1995; Estupinian et al., 1996; Mims, 1997), from differences between locations in the Northern (more polluted) and Southern (less polluted) hemispheres (Seckmeyer and McKenzie, 1992; Seckmeyer et al., 1995), and from the increases in UV irradiances with increasing surface elevation, in excess of those expected from pollution-free conditions (Cabrera et al., 1995; Madronich et al., 1995; Piazena 1996; Blumthaler et al., 1997). The measured effects on UV radiation are highly variable and specific to the various locations (e.g.,Wenny et al., 1998).

    An important consideration is whether the aerosol particles are highly absorbing (e.g. soot) or simply scatter (re-direct) the incident radiation (e.g. sulfate aerosols). All particles tend to reduce the UV irradiance (defined as the radiation incident on a horizontal surface). However, scattering by non-absorbing aerosols can actually increase the UV exposure on non-horizontal surfaces due to the additional radiation incident from low angles (e.g. Blumthaler et al. 1997; Dickerson et al., 1997; Loxsom and Kunkel, 1997). The net effects on biota from such changes in direction of incidence are not well understood.

    Stratospheric aerosols are usually too sparse to have any effect on atmospheric UV transmission. An exception arises following a major volcanic eruption, such as that of Mt. Pinatubo (Philippines) in June 1991 which injected large amounts of ash and sulfur dioxide (SO2) into the stratosphere. The heavier ash sedimented out of the stratosphere relatively quickly and its optical effects were of limited geographical extent. Gaseous SO2, on the other hand, was removed from the stratosphere mainly by chemical reactions to form H2SO4 molecules, which then readily nucleated into sulfate aerosol particles. Higher stratospheric sulfate aerosol loadings were observed for several years after the eruption, during which time these particles were distributed on global scales. Calculations indicate that the effects on biologically weighted UV irradiances were quite small, of order of a few percent (Madronich et al., 1991; Vogelmann et al., 1992; Tsitas and Yung, 1996), with even some possible enhancements at very short wavelengths and low sun when aerosols scatter some photons directly downward thus allowing a shorter crossing of the stratospheric ozone layer (Michelangeli et al., 1992; Davies, 1993). Ground-based measurements of UV irradiance after the Mt. Pinatubo eruption confirm the small decreases and also show a strong increase in diffuse/direct radiation at all wavelengths, in good agreement with theoretical models (Blumthaler and Ambach, 1994; Zeng et al., 1994, Lantz et al., 1996). A less direct but more important UV-related consequence of stratospheric aerosols is their effect on stratospheric ozone itself. Significant destruction of stratospheric ozone by heterogeneous chemical processes involving the aerosols was predicted (Hoffman and Solomon, 1989; Brasseur et al., 1990) and observed for several years after the Mt. Pinatubo eruption (Gleason et al., 1993; WMO, 1994a).


CIESIN Copyright © 1997 For more information about CIESIN and our activities contact CIESIN User Services. E-mail: Tel.: (914) 365-8988