Estimating UV Trends Ground Based

An excellent example of UV trend detection is from ground-based measured solar irradiances at 305 nm and 324 nm at Thessaloniki, Greece. The irradiances shown in Fig. 5.13 are for cloud-free skies at a constant SZA of 63° (WMO, 2007, an extension of Bias et al., 1993). These data are obtained from a carefully maintained Brewer spectrometer located in an industrial area that is subjected to moderate amounts of pollution generated locally and also reaching Greece from other countries in Europe. There are also occasional dust episodes originating in northern Africa.

Figure 5.13 Combined effects of ozone, aerosols, and other absorbing components on UV radiation. Long-term variability in monthly mean solar spectral irradiances at 324 nm (upper panel) and at 305 nm (middle panel) measured at Thessaloniki, Greece, under clear skies at 63° SZA, shown as departures from the long-term (1990 - 2006) averages. The lower panel shows the corresponding departures in the ozone column of 375 DU (from WMO, 2007)

Figure 5.13 Combined effects of ozone, aerosols, and other absorbing components on UV radiation. Long-term variability in monthly mean solar spectral irradiances at 324 nm (upper panel) and at 305 nm (middle panel) measured at Thessaloniki, Greece, under clear skies at 63° SZA, shown as departures from the long-term (1990 - 2006) averages. The lower panel shows the corresponding departures in the ozone column of 375 DU (from WMO, 2007)

The radiation at 324 nm should not be significantly affected by ozone so that the cause of the upward trend at 324 nm (11.3% per decade) is almost certainly due to aerosol and pollution decreases. Decreasing amounts of aerosol and pollution that cause the upward trend at 324 nm will also affect 305 nm by approximately the same amount. Combining the changes seen for 324 nm with those observed for 305 nm (8.1% per decade) implies that the effect of increasing ozone amounts (0.9% per decade) on 305 nm irradiance is a statistically significant decrease of —(11.3 - 8.1)% = 3.2% per decade.

An easy way to check this conclusion is through the RAF defined as part of Eq. (5.1); RAF = -aQ sec(<9) = -4 for Q= 375 DU, and 0= 63°, the average measured values for Thessaloniki. Based on the RAF and the observed ozone change of 0.9% per decade, the change in 305 nm UV irradiance dF/F=RAF dO/O should be — 4(0.9)% = -3.6% per decade, consistent with the measurements of -3.2% and -3.4% per decade discussed above. In addition to the smaller ozone effects, Fig. 5.13 shows that a decline in air pollutants can cause increases in surface UV irradiance of 11.3% per decade in a local industrial site, such as Thessaloniki, Greece.

When data from cloudy and clear days are present in the UV time series, the measured trends in UV radiation at individual stations can have sufficient variation (typically 0% - 50%, and occasionally larger due to cloud cover) to make estimated long-term trends lose statistical significance. As shown in a report by WMO (2007), trend estimates for Toronto for the period of 1998 through 2005 were (1.5 ± 5)% per decade (1 standard deviation, 1 a) (WMO, 2007) during a period in which the total ozone amount was relatively constant. Even using unfiltered Toronto UV radiation data going back to 1990, no statistically significant trend is observable in the extended Toronto UV data despite ozone decreases that took place during the 1990s, due to the variability introduced by clouds. To relate the estimated trends to ozone changes requires knowledge of changes in aerosol and cloud amounts, which can be obtained from a wavelength not affected by ozone.

Fioletov et al. (2001) have made ground-based estimates of erythemal irradiance changes from two Brewer spectrometer stations (Montreal and Edmonton), and found that the UV-B trends were similar to those expected from changes in ozone alone, but with much larger uncertainty because of clouds and aerosols.

Was this article helpful?

0 0

Post a comment