Climatology of UVA Irradiance

Figure 3.9 shows the climatology of UV-A irradiance (spectral irradiance integrated between 315 nm and 400 nm) for all UVSIMN sites but Summit. All plots show individual measurements, as well as the average, median, and 10th and 90th percentiles, and overall daily maximum. Since long-term changes in cloud cover and albedo are not considered in this study, historical estimates are virtually identical with recent measurements and are therefore not included in Fig. 3.9. Changes in total ozone have practically no influence on UV-A irradiance. Patterns in Fig. 3.9 show mostly seasonal variations in cloudiness, surface albedo, and to a lesser extent, atmospheric aerosol loading.

UV-A irradiance at the South Pole is very symmetric about the solstice due to constant high albedo year-round, and little influence by clouds. Daily maximum UV-levels are considerably above the 90th percentile, indicating that enhancement of UV radiation by scattered clouds—which is a well-known effect at mid-latitude sites (Mims and Frederick, 1994)—can also occur at the South Pole. The maximum enhancement is about 30%. Enhancement of the spectral integral 400 nm - 600 nm can be as high as 70%. An extreme example is shown in Fig. 3.10, which displays three spectra measured at 19:00 UT, 19:15 UT, and 19:30 UT on 17 December 2000 at South Pole. During the first spectrum starting at 19:00 UT, the sun was

Figure 3.9 UV-A irradiance measured at South Pole (Panel (a)), McMurdo (Panel (b)), Palmer (Panel (c)), Ushuaia (Panel (d)), San Diego (Panel (e)), and Barrow (Panel (f)). Individual measurements, average, median, and 10th and 90th percentiles, as well as daily maximum, are shown as in Fig. 3.2(b)

hidden by a stable cloud, leading to a reduction of UV and visible irradiance of about 10% - 20% compared to the clear-sky model (Fig. 3.10(b), red line). Measurements were also compared against a second model spectrum where a wavelength-independent cloud optical depth of 1.83 was used as an additional model input parameter. The ratio of the measured spectrum with this model spectrum (Fig. 3.10(b), orange line) is close to one and virtually independent of wavelength, confirming that the radiation field during the period of the scan (approximately 13 minutes) was very stable. Measurements of total irradiance with a pyranometer also indicate constant conditions (Fig. 3.10(c)). During the second spectrum, starting at 19:15 UT (Fig. 3.10, green lines), total irradiance increased sharply during the first part of the scan; spectral irradiance increased up to 60% relative to the clear-sky model. Total irradiance increased by up to 72%. As reflections from nearby obstacles can be excluded, this pattern can only be explained by enhancement due to scattered clouds surrounding the (unoccluded) disk of the sun. Photons passing through a hole in the cloud are scattered multiple times between the snow-covered surface and the cloud-ceiling. This effect leads to a large enhancement of downwelling radiation and cannot be observed at locations with small surface albedo. The third spectrum starting at 19:30 UT (Fig. 3.10, blue lines) agrees well with the clear-sky model. Pyranometer measurements were close to the value expected for clear-sky.

UV-A irradiance at McMurdo (Fig. 3.9(b)) is generally symmetric about the solstice. Radiation levels are somewhat smaller in January than December, probably

Figure 3.10 Enhancement of global irradiance by a broken cloud at South Pole. Panel (a): Three spectra of global irradiance measured on 17 December 2000 at 19:00 UT, 19:15 UT, and 19:30 UT. The clear-sky model spectrum for 19:00 is also shown. Panel (b): Ratios of measured and modeled spectra. Panel (c): Total irradiance measured by a pyranometer during the recording of the three spectra and plotted against the wavelength being sampled by the spectroradiometer as a surrogate of time

Figure 3.10 Enhancement of global irradiance by a broken cloud at South Pole. Panel (a): Three spectra of global irradiance measured on 17 December 2000 at 19:00 UT, 19:15 UT, and 19:30 UT. The clear-sky model spectrum for 19:00 is also shown. Panel (b): Ratios of measured and modeled spectra. Panel (c): Total irradiance measured by a pyranometer during the recording of the three spectra and plotted against the wavelength being sampled by the spectroradiometer as a surrogate of time due to smaller albedo in summer when compared with spring. UV-A irradiance at Palmer (Fig. 3.9(c)) shows a much larger variability than observed at South Pole and McMurdo due to frequent cloud cover with optical depths typically ranging between 20 and 50 (Ricchiazzi et al., 1995). The ocean surrounding Palmer freezes over during the winter. Terrain and glaciers at Palmer are typically covered by snow up to mid-December. Surface albedo is therefore larger in winter and spring than in summer. This leads to the small asymmetry in the annual cycle UV-A irradiance discernable in Fig. 3.9(c). Ushuaia, like Palmer, is affected by persistent cloudiness, leading to a large difference of the 10th and 90th percentiles (Fig. 3.9(d)). The Beagle Channel adjacent to Ushuaia does not freeze, but snow typically enhances the effective surface albedo to approximately 0.2 to 0.3 between June and October, leading to some enhancement of UV-A during the winter. UV-A irradiance during spring and summer is almost symmetric about the solstice. The clear-sky limit of UV-A irradiance at San Diego (Fig. 3.9(e)) is symmetric about the solstice, but the average and 10th percentile are affected by seasonal patterns in cloudiness. Cloud attenuation is largest during May and June, whereas most days in August are cloud-free at solar noon. Enhancement of UV-A irradiance by scattered clouds beyond the clear-sky limit is remarkably small, and less pronounced than at the South Pole. This is attributable to low surface albedo (< 0.05) and the near absence of broken cumulus clouds, which can enhance UV radiation at other mid-latitude locations by up to 25% (WMO, 2007). UV-A irradiance at Barrow displays a strong annual cycle due to seasonal differences in cloudiness (more prevalent in summer) and surface albedo (0.83 + 0.08 between November and May; smaller than 0.1 during summer). The effect of the two factors has been quantitatively described by Bernhard et al. (2007).

Was this article helpful?

0 0

Post a comment