Biologically Active UV Radiation
The solar radiation at the top of the Earth's atmosphere
contains a significant amount of radiation of wavelength (l)
shorter - and therefore more energetic - than visible light (400-700 nm).
Wavelengths in the range 100-400 nm constitute the ultraviolet (UV) spectral
region. The shortest of these wavelengths (UV-C, 100-280 nm) are blocked
(absorbed) essentially completely by atmospheric oxygen (O2)
and ozone (O3). Wavelengths
in the UV-B range (280-315 nm* ) are absorbed
efficiently though not completely by O3,
while UV-A wavelengths (315-400 nm) are absorbed only weakly by O3
and are therefore more easily transmitted to the Earth's surface.
|Fig. 1.1. Biologically active UV radiation. The overlap between the spectral irradiance F(l) and the erythemal action spectrum B(l) given by McKinlay and Diffey (1987) shows the spectrum of biologically active radiation, F(l) x B(l). The area under the product function F(l) x B(l) is the biologically active dose rate. Thick lines are for a total ozone column of 348 DU, thin lines for 250 DU (one Dobson Unit, or DU, is defined as the height in milli-centimeters that pure gaseous ozone would occupy if compressed to 1013 hPa at 0°C, and thus equals 2.69x1016 molecules cm-2) (from Madronich and Flocke, 1997).|
Figure 1.1 illustrates the wavelength dependence of UV
spectral irradiance at the Earth's surface. The strong decrease below about
330 nm is due to absorption by atmospheric ozone. Reductions in ozone lead
to an increase at these wavelengths, as shown in the figure, with the largest
fractional (percentage) increase occurring at progressively shorter wavelengths
in the UV-A and UV-B ranges. Although UV-B irradiances are much smaller
than those in the UV-A region, many biological responses to UV exposure
are far greater at the shorter wavelengths. Thus even relatively small
increments of UV-B radiation can lead to substantial biological effects.
|Fig. 1.2. Dependence of erythemal ultraviolet (UV) radiation at the Earth's surface on atmospheric ozone, measured on cloud-free days at various locations, at fixed solar zenith angles. Legend: South Pole (Booth and Madronich, 1994); Mauna Loa, Hawaii (Bodhaine et al., 1997); Lauder, New Zealand (McKenzie et al., 1998); Thessaloniki, Greece (updated from Zerefos et al., 1997); Garmisch, Germany (Mayer et al., 1997); and Toronto, Canada (updated from Fioletov et al., 1997). Solid curve shows model prediction with a power rule using RAF = 1.10.|
To estimate the biological impacts of ozone-related UV increases, the wavelength dependence of the sensitivity to UV exposure must be known at least approximately. Spectral sensitivity functions (action spectra) have been determined in laboratory and field studies for a number of biological endpoints. Such action spectra allow the estimation of the effect of simultaneously changing radiation at different wavelengths by different amounts, as happens when ozone reductions occur. Figure 1.1 shows the action spectrum for erythema (skin-reddening) induction by UV radiation, and the spectral overlap between significant sensitivity (at shorter wavelengths) and significant spectral irradiance (at longer wavelengths). For this particular action, the overlap is greatest in the range 300-320 nm, and is quite sensitive to ozone amounts as shown in Figure 1.1. A useful measure of this overlap is the biologically active UV irradiance, or exposure UVbio, defined as the area under the spectral overlap function,
UVbio = ò F(l) B(l) dl
where F(l) is the spectral irradiance, B(l) is the action spectrum for a particular biological effect, and the integral is carried out over all UV wavelengths.
The sensitivity of UVbio to atmospheric ozone is frequently expressed with the Radiation Amplification Factor (RAF), defined as the percentage increase in UVbio that would result from a 1% decrease in the column amount of atmospheric ozone. The radiation amplification factors are given in Table 1.1 for a number of different known effects. The RAFs can generally be used only to estimate effects of small ozone changes, e.g. of a few percent, because the relationship between ozone and UVbio becomes non-linear for larger ozone changes. For action spectra that decrease approximately exponentially with increasing wavelength over 300-330 nm, the biologically active irradiances scale with larger ozone changes according to a power relationship (Madronich, 1993a, b; Booth and Madronich, 1994):
UVbio ~ (Ozone)-RAF
The RAFs presented in Table 1.1 have been computed with a model of the propagation of spectral UV radiation through the atmosphere, combined with the appropriate action spectra for the different effects. RAFs may also be derived from spectral UV measurements made at the Earth's surface, when these are combined (numerically) with the appropriate action spectra. Generally good agreement is found between these two methods, within the combined uncertainties of the measurements and the models (McKenzie et al., 1991; Booth and Madronich, 1994; Blumthaler et al., 1995; Bodhaine et al., 1997).
RAFs are useful indicators of the sensitivity of a particular effect (i.e., a particular action spectrum) to ozone changes. Large RAF values indicate that the radiation associated with a particular effect is strongly sensitive to changes in atmospheric ozone, while small RAF values indicate that the relevant UVbiois less sensitive to ozone changes. Values of RAF ~ 0 mean that the UVbio for that particular effect is not dependent on ozone, as occurs in cases when an action spectrum shows strong sensitivity to longer UV-A and visible wavelengths, but not to UV-B radiation.
In many cases the full spectral sensitivity is not well known, and only estimates of the RAF value can be made. A particularly important consideration is the potential role of longer (UV-A and visible) wavelengths, where even relatively low sensitivity (per photon) may be of importance because the ambient radiation increases strongly with increasing wavelength (see curve marked F(l) in Figure 1.1). To show this sensitivity to longer wavelengths, the RAFs given in Table 1.1. were also calculated by extrapolating the measured action spectra to 400 nm, and, for cases where such extrapolation lead to significant changes in the RAF, the re-calculated values are shown in brackets. The RAFs calculated from extrapolated action spectra are not necessarily more accurate than those without extrapolation, but rather show that such RAFs are quite uncertain, and more detailed measurements of the action spectra are need to assess the sensitivity to ozone.
It should be cautioned however that (i) neither the action
spectra nor the resulting weighted irradiances (UVbio)
give a measure of the absolute damage to any particular organism; (ii)
weighted irradiances computed from different action spectra cannot be compared
directly to one another, because the action spectra usually specify only
the spectral shape of the sensitivity, not the absolute value; (iii) even
for a single action spectrum, increases in UVbio
do not necessarily imply a proportionate increase in effect, if dose-response
relations for that effect are non-linear; (iv) any damage to a specific
organism must be viewed in the context of its entire ecosystem including
consideration of other stresses (e.g. nutrient availability, temperature)
and interactions with other organisms (e.g. species competition); (v) action
spectra are usually determined from short-term laboratory or field experiments,
while the effects of environmental UV increases may be felt on longer time
scales; and (vi) considerable uncertainties are inherent in the experimental
determinations of action spectra.