Results

1.3.1 Satellite-Derived UV Climatologies

The geographical distributions of daily UV radiation doses at earth's surface, averaged over the entire time period of (Nov. 1, 1978 - June 30, 2000) are shown in Figs. 1.2 -1.6. The upper panel in each figure shows values calculated by considering the effects of both ozone and clouds, as estimated from TOMS data, and are thus assessed to be nearest to the actual values experienced over this time period. The lower panels show climatological distributions estimated for hypothetical cloud-free skies (i.e., estimated from the ozone distributions without correcting for the presence of clouds).

UV-A: annual mean:kj/m-/day:ozone+clouds case

UV-A: annual mean:kj/m-/day:ozone+clouds case

UV-A: annual mean :kJ/m2/day;ozone-only ease

UV-A: annual mean :kJ/m2/day;ozone-only ease

1200 - 1200

Figure 1.2 Climatological daily doses of UV-A at earth's surface, derived from satellite (TOMS) observations of the atmospheric ozone column and cloud reflectivity at 380 nm and averaged annually over Nov 1, 1978 - June 30, 2000, with (upper) and without (lower) correcting for the presence of clouds

1200 - 1200

Figure 1.2 Climatological daily doses of UV-A at earth's surface, derived from satellite (TOMS) observations of the atmospheric ozone column and cloud reflectivity at 380 nm and averaged annually over Nov 1, 1978 - June 30, 2000, with (upper) and without (lower) correcting for the presence of clouds

UV-B: annual mean:kl/m2/day: r^one+clouds case

UV-B: annual mean:kl/m2/day: r^one+clouds case

Figure 1.3 Climatological daily doses of UV-B at earth's surface, as Fig. 1.2

I■ annual mean; kJ/m '/day :o.:o ne - c .11 y'

Figure 1.4 Climatological daily doses of erythemal UV at earth's surface, as Fig. 1.2

Figure 1.4 Climatological daily doses of erythemal UV at earth's surface, as Fig. 1.2

VtD: annual mean: kj/m2/day:ozone + clouds case

VtD: annual mean: kj/m2/day:ozone + clouds case

VtI3: annual mean: kJ/m2/day:ozone-only case
Figure 1.5 Climatological daily doses of UV weighted for pre-vitamin D3 synthesis at earth's surface, as Fig. 1.2

The zonally homogeneous distribution of UV-A calculated for cloud-free conditions shows almost exclusive dependence on solar position, with only small variations due to surface topography. Ozone column variations induce additional zonal variations in the distributions of cloud-free UV-B and UV weighted for either erythema or other biological response functions. However, the strongest longitudinal variations in the surface UV dose rate distributions are caused by climatological cloud distributions.

As expected, the highest doses are generally seen in the tropics, up to ca. 6 kJ m" -2 day-1 (60 SED day-1) for erythemal UV in the eastern Pacific and eastern Africa, but with substantial cloud-related reductions over western South America, parts of West Africa, and just north of the equator in the eastern and central Pacific. Middle latitudes of both hemispheres show a general pole-ward decrease from about 5 to 1 kJ m- day- , with some regional highs associated with higher elevations, smaller ozone columns, and infrequent cloudiness (e.g., the Andes

Figure 1.6 Climatological daily doses of UV weighted for non-melanoma carcinogenesis at earth's surface, as Fig. 1.2

LIV-13 : annua] mean: transmission thru doud

LIV-13 : annua] mean: transmission thru doud

Figure 1.7 Climatological annual mean cloud-related UV reduction factors for daily doses of UV-A derived from satellite (TOMS) observations of the atmospheric ozone column and cloud reflectivity at 380 nm for Nov 1, 1978 - June 30, 2000. Cloud-related UV reduction factors for the other UV functions discussed in this chapter are similar

Figure 1.7 Climatological annual mean cloud-related UV reduction factors for daily doses of UV-A derived from satellite (TOMS) observations of the atmospheric ozone column and cloud reflectivity at 380 nm for Nov 1, 1978 - June 30, 2000. Cloud-related UV reduction factors for the other UV functions discussed in this chapter are similar

Mountains, the Tibetan Plateau, central Mexico, and the southwestern U.S.). Lower values for those latitudes are noted for East Asia and the coastal eastern Pacific, associated with more frequent cloud cover. Figure 1.7 shows the cloud-related UV reduction factor, calculated as the ratio of the cloud-corrected climatological daily UV dose (upper panels of Figs. 1.2 - 1.6) to the climatological daily dose before cloud-correction (lower panels of Figs. 1.2 - 1.6).

The seasonal variations of the 22-year UV dose climatologies are shown in Figs. 1.8 - 1.11. (The seasonal variability of UV weighted for non-melanoma carcinogenesis is similar in magnitude and distribution to that of UV weighted for pre-vitamin D3 synthesis, so it is not shown.) The latitudinal distributions are generally consistent with the annual variation of the subsolar point in the tropics, giving strong seasonal variations at temperate latitudes (out of phase by six months between the two hemispheres).

UV-A: kJ/m2/day:ozotie + clouds case : Dec
Figure 1.8 Seasonal variability of daily doses of UV-A. The figure shows the daily doses averaged over the period Dec. 1, 1978 - June 30, 2000 for the months of June (upper) and December (lower)
Figure 1.9 Seasonal variability of daily doses of UV-B, as Fig. 1.8
Figure 1.10 Seasonal variability of daily doses of erythemal UV, as Fig. 1.8
VtD: kJ/m-Vday : ozotic + clouds cast;: Dec
Figure 1.11 Seasonal variability of daily doses of UV weighted for pre-vitamin D3 synthesis, as Fig. 1.8

Detailed analysis of temporal trends is beyond the scope of this work, but some indications may be obtained by comparing the climatological values averaged over 1990 - 2000 with those averaged over 1979 - 1989. Figure 1.12 shows the changes in erythemal doses between these two 11-year periods. The total dose rate changes combine the effects of changes in the ozone column (Fig. 1.12, top panel) and changes in cloudiness (Fig. 1.12, center panel). The upper panel clearly shows the increase in surface UV resulting from stratospheric ozone reductions, not only in the Antarctic region, but also at mid-latitudes. The center panel shows a reduction in cloudiness, maximizing over north-central Europe, possibly as a result of the introduction of cleaner fuel-burning technologies in the west, combined with a contraction of heavy industry in the east. Increases in cloudiness are seen over the Bay of Bengal and over the Humboldt Current off the west coast of South America, possibly owing to the enhanced El Niño conditions prevalent during the 1990s. The apparent reductions in cloudiness around the coasts of southern Alaska and Hudson Bay may partly be artifacts related to decadal-scale changes in snow and ice cover. The lower panel shows the net change in the annual mean global distribution of erythemal UV dose rates. Significant changes are seen to have occurred, with annual mean dose rate increases of 8% or more in some regions, including north-central Europe, the eastern seaboard of the US, Siberia, and the Antarctic Ocean. Other regions experienced annual mean dose rate reductions of up to 6%.

Figure 1.12 Changes in average daily doses of erythemal UV between the periods of 1979 - 1989 and 1990 - 2000. Change in values calculated from the TOMS ozone data only (upper); change in cloud reduction factor (center); change in annual mean net daily erythemal UV dose (lower)

1.3.2 Comparison with Ground-Based Measurements

Comparisons with ground-based measurements are shown in Fig. 1.13 for annual averages of the daily UVeiy dose. Temporal overlap between the satellite observations and the ground-based observations is only available at some stations (see Table 1.1), and then only for a few years. The satellite-derived estimates show a long-term trend due to stratospheric ozone depletion, as has been reported previously (Herman et al., 1996a), whether or not cloud cover is considered. Note that the satellite-derived values apply to an extended region of typically 104 km2, while the ground-based observations pertain to a single, often urban, location. Figure 1.13 shows that the satellite-derived annual averages tend to overestimate measurements by averages of 5% for remote northern locations, 11% for Canadian cities, and 30% for polluted mid-latitude urban regions (San Diego and cities in Japan and cities in the Taiwan region). Herman et al. (1999) found similar differences for the measurements at Toronto, showing also that these were not seasonally dependent, except for snow-covered periods. These discrepancies remain under investigation and may stem from both measurements and satellite-derived values. Because of the imperfect cosine response of the entrance optics, Brewer instruments may underestimate true irradiances by 2% - 7% according to Bais et al. (1998), or by 6% ± 2% according to Herman et al. (1999). Network for the

Figure 1.13 Differences between ground-based and satellite-derived annual mean daily erythemal UV doses. Points show the mean of differences for all years where both ground-based and satellite-derived values are available. Lines show the standard deviations at each station. Filled diamonds = satellite-derived UV values including adjustment for the presence of clouds, and mean ground-based values; open circles = satellite-derived UV values without cloud adjustment and maximum ground-based values. Stations are listed in order of decreasing latitude. See Table 1.1 for station details

Figure 1.13 Differences between ground-based and satellite-derived annual mean daily erythemal UV doses. Points show the mean of differences for all years where both ground-based and satellite-derived values are available. Lines show the standard deviations at each station. Filled diamonds = satellite-derived UV values including adjustment for the presence of clouds, and mean ground-based values; open circles = satellite-derived UV values without cloud adjustment and maximum ground-based values. Stations are listed in order of decreasing latitude. See Table 1.1 for station details

Table 1.1 Details of ground-based Brewer spectrophotometer stations referenced in Fig. 1.13

Station

Agency*

Station I.D.

Latitude (°)

Longitude (°)

Altitude (m)

Churchill

MSC

77

58.8

-94.1

35

Obninsk

IEM-SPA

307

55.1

36.6

0

Edmonton

MSC

21

53.6

-114.1

766

Saskatoon

MSC

241

52.1

-106.7

550

Regina

MSC

338

50.2

-104.7

592

Winnipeg

MSC

320

49.9

-97.2

239

Poprad-Ganovce

SHMI

331

49.0

20.3

706

Saturna

MSC

290

48.8

-123.1

178

Montreal

MSC

319

45.5

-73.8

24

Halifax

MSC

321

44.7

-63.6

31

Toronto

MSC

65

43.8

-79.5

198

Sapporo

JMA

12

43.1

141.3

19

Tateno

JMA

14

36.1

140.1

31

San Diego

NSF

239

32.8

-117.1

0

Kagoshima

JMA

7

31.6

130.6

283

Naha

JMA

190

26.2

127.7

29

Taipei

"CWBT"

95

25.0

121.5

30

Chenkung

"CWBT"

306

23.1

121.4

10

Mauna Loa

MSC

31

19.5

-155.6

3397

Ushuaia

NSF

339

-54.5

-68.2

7

Palmer

NSF

292

-64.5

-64.0

0

* Agency abbreviations: Meteorological Service, Canada (MSC); Institute of Experimental Meteorology-Scientific Production Association (IEM-SPA), Russia; Swedish Meteorological and Hydrological Institute (SHMI); Japan Meteorological Agency (JMA); National Science Foundation (NSF), USA; "Central Weather Bureau of Taiwan (CWBT)." The World Ultraviolet Radiation Data Centre (WOUDC), where the data is archived, has a stated goal of annually intercalibrating the instruments.

* Agency abbreviations: Meteorological Service, Canada (MSC); Institute of Experimental Meteorology-Scientific Production Association (IEM-SPA), Russia; Swedish Meteorological and Hydrological Institute (SHMI); Japan Meteorological Agency (JMA); National Science Foundation (NSF), USA; "Central Weather Bureau of Taiwan (CWBT)." The World Ultraviolet Radiation Data Centre (WOUDC), where the data is archived, has a stated goal of annually intercalibrating the instruments.

Detection of Atmospheric Composition Change (NDACC) quality ground based instruments, which have superior cosine responses and include corrections for these errors, also show much lower UV irradiances than satellite derived values in polluted locations, but good agreement in pristine locations (McKenzie et al., 2001). The TOMS-based method does not account for UV-absorbing aerosols, which according to Herman et al. (1999) lead to a systematic overestimate of ca. 8% ± 2% at Toronto. Aerosols are likely to be significant at other locations during pollution episodes (e.g., Wenny et al., 1998), or if an area suffers routinely from pollution (e.g., the Southeast Asian sites). Smaller errors (e.g., 5% or less) are associated with several other factors (e.g., instrument calibrations, extraterrestrial irradiances used in the model), but at present there is no basis for estimating the sign of any resulting bias. The strong disagreement at Palmer Station, where we underestimate the irradiance by 50%, illustrates the consequence of using

TOMS-observed reflectivity to infer cloudiness at high-latitude locations. The high reflectivity recorded by the satellite may be due to snow and ice rather than to clouds, thus leading to a significant underestimate of the UVery reaching the surface.

Using a similar TOMS-based technique, Frederick and Erlick (1995) computed noon-time erythemal irradiances, as well as their trends and interannual variability, for regions of 6° latitude by 10° longitude centered over New Zealand, Malaysia, Sweden and the eastern U.S. Lubin et al. (1998) used monthly mean ozone from TOMS and monthly mean cloud cover data derived from the Earth Radiation Budget Experiment (ERBE) to compute global UVery distributions for several months during 1988 and 1989. Sabziparvar et al. (1999) presented a global climatology of daily doses for January, April, July, and October, computed from monthly-averaged climatological ozone (TOMS, 1985 - 1989) and cloud data (International Satellite Cloud Climatology Project (ISCCP), 1983 - 1991). Global distributions of UVery for 1988 (January, March, July, and September) are presented by Herman et al. (1999), who also carried out detailed comparisons to observations obtained with the Toronto Brewer instrument. Our results are generally consistent with these studies, for example, predicting the strong latitudinal gradients as well as most of the regional anomalies. However, detailed values are not directly comparable because of our use of daily rather than monthly ozone and cloud data, and our integration over an extended time period (1978 - 2000).

1.3.3 Discussion of Uncertanties

We recognize a number of limitations in our study that can hopefully be addressed by future work. The parameterization of cloud effects via the 380 nm reflectivity is obviously crude compared to the complexity of real cloudiness, and its failure in the presence of high albedo surfaces has been discussed. Other sources of cloud information exist and show promise in extending the climatology to higher latitudes (e.g., Mayer and Madronich, 1998). Additionally, pollutants present in the lower atmosphere can attenuate surface UV irradiances. Regional-scale absorbing aerosols, probably associated with plumes of biomass burning, have been detected by the TOMS instrument (Krotkov et al., 1998), although quantification remains a challenging area of research. On smaller scales, such as highly polluted urban areas, substantial absorption of UV radiation is possible from smog-generated ozone, SO2, NO2, and absorbing aerosols (e.g., soot) in the lower atmosphere. These absorbers are not easily detected from satellite platforms, so that a climatology based on direct ground-based UV radiation measurements, if available, is preferable for such locations.

The presence of snow, whether seasonal or year-round, creates another challenge. The misinterpretation of snow cover as cloud in the satellite data leads to underestimates of ambient UV. This underestimate stems not only from the inappropriate application of the cloud-related UV reduction factor, but also from the use of a surface albedo of 5%, whereas the reflectivity of snow is usually much higher and causes stronger surface-atmosphere radiative coupling. Hence, even our "cloud-free" climatology may be an underestimate for UV dose in snowy regions. If snow cover is interspersed with lower albedo surfaces such as forest, the bias will be reduced, but not eliminated. Another factor is the increase in effective UV dose received just above a snow surface by reflection from that surface. At higher latitudes and low altitudes, persistent widespread snow is likely to be present only during winter when UV irradiance is already low due to large SZAs. In mountainous regions, snow cover may also persist at lower latitudes, and for longer seasons. In each case, the climatologies presented here underestimate the actual ambient UV dose.

The results shown here do not give the short-term variations in erythemal UV, although some of the inter-annual variability may be inferred from Fig. 1.12. Daily data (1979 - 1994 and 1996 - 2000) are available and were used to compute the long-term averages; however, space limitations preclude their presentation here.

Was this article helpful?

0 0

Post a comment