Estimates of Historical UV Indices

Based on historical ozone data and model calculation, the climatology of the UV Index for years preceding the development of the ozone hole was estimated for all sites with the exception of San Diego and Summit. The procedure is explained using South Pole as an example. Historical total ozone data at South Pole are available from observations of Dobson spectrophotometers performed by the Global Monitoring Division (GMD) of NOAA's Earth System Research Laboratory (ESRL) (Climate Monitoring and Diagnostics Laboratory, 2004). Measurements started in 1963. Data from 1963 - 1980 were used for estimating past UV levels.

Year-to-year variability of UV at South Pole is mainly influenced by total ozone. Variations of surface albedo in the UV and visible wavelength range are smaller than ± 1% (Grenfell et al., 1994). Except at times following volcanic eruptions, the aerosol optical depth at 500 nm is typically only 0.012 (Shaw,

Albedo Measurement Spectrophotometer

D J Month

Figure 3.2 Climatology of UV Index at South Pole. Measurements at 00:00 UT of the years 1991 - 2007 are indicated by black dots. Average and median are plotted as red and green lines. 10% of the measurements are below (above) the lower (upper) blue line. The daily maximum is indicated by a thin grey line. The plot starts at the winter solstice. Summer solstice, and vernal and autumnal equinoxes are indicated by broken lines. Panel (a): Unsmoothed data. Panel (b): Same data as in (a), but with an 11-day running-average applied to the average, median, 10% and 90% lines

1982). Attenuation by clouds is small due to the low atmospheric water vapor content and the moderation of cloud effects by the high (> 0.96) albedo (Nichol et al., 2003). By comparing measurements with the clear-sky model, it was determined that the average attenuation of spectral irradiance at 345 nm is only 6%. Cloud transmission at South Pole has also been determined from measurements of total irradiance (0.3 (am - 3.0 ^m) using pyranometers (Dutton et al., 2004). Although this study did not indicate a significant linear trend of cloud transmission between 1976 and 2001, an oscillation on a decadal timescale was observed with a small downward trend in the late 1970s, followed by an upward trend between 1982 and 1995, and a downward trend thereafter. The small relative changes in cloud transmission of about + 2% reported by Dutton et al., (2004) have a very small effect (< +1%) on the UV Index due to the diminished cloud influence at shorter wavelengths (Bernhard et al., 2004). Based on these considerations we assumed that all parameters affecting the radiative transfer did not change during the last 40 years, with the exception of atmospheric ozone concentrations. The historical clear-sky UV Index for the years 1963 -1980 was consequently modeled based on the average parameters used for processing South Pole Version 2 data of the years 1991 - 2007, with the exception of total ozone which was taken from the GMD Dobson data set.

The range of historical UV levels for a given day is mostly controlled by year-to-year changes in total ozone and cloudiness. To account for the variability introduced by clouds, we integrated measured and modeled spectra from the years 1991 - 2007 over the wavelength interval 337.5 nm - 342.5 nm, calculated the ratio of measurement to model, and used the results for estimating cloud-induced variability in UV intensities. This result was also applied to historical measurements. This approach assumes that year-to-year cloud-variability did not change during the last four decades, which is justified considering the small effect of clouds on UV discussed earlier. We also note that the wavelength interval of 337.5 nm - 342.5 nm is virtually unaffected by atmospheric ozone concentrations, and attenuation by clouds has only a weak wavelength dependence between 300 nm and 340 nm (Seckmeyer et al., 1996). The ratio of measurement to model for the 337.5 nm - 342.5 nm interval is therefore also appropriate for quantifying the effect of clouds on the UV Index. We estimate that the overall uncertainty in calculated historical UV Indices due to cloud effects is ± 2%.

Results of these calculations are shown in Fig. 3.3. The red line is the historical average UV Index estimated from the average total ozone column of the years 1963 -1980 measured with the Dobson, and the average attenuation by clouds of about 6% derived from Version 2 data. Figure 3.3 also includes the estimated 10th and 90th percentiles from the historical variability of total ozone (thin black lines in Fig. 3.3) and year-to-year differences in cloud transmission estimated from recent measurements (broken black lines). The two ranges were combined in quadrature to estimate the 10th and 90th percentiles for both effects (blue lines). During summer, variations induced by ozone and clouds are of similar magnitude. Variations during spring are dominated by ozone.

Winter

Spring

_ Summer

Autumn

South Pole

— Average

— ± Ozone

±CloLids ■

-

fi

— ± Al!

Figure 3.3 Estimate of the historical UV Index at South Pole. The red line is the average. The ranges defined by the 10th and 90th percentiles caused by variability of total ozone and cloud transmission are indicated by thin black and broken black lines respectively. The two ranges were combined in quadrature resulting in the span indicated by blue lines

J ASONDJ FMAMJ Month

Figure 3.3 Estimate of the historical UV Index at South Pole. The red line is the average. The ranges defined by the 10th and 90th percentiles caused by variability of total ozone and cloud transmission are indicated by thin black and broken black lines respectively. The two ranges were combined in quadrature resulting in the span indicated by blue lines

A comparison of UV Indices measured during the years 1991 - 2007 and UV Indices modeled for the years 1963 -1980 is shown in Fig. 3.4(a). An 11-day moving average was applied to all lines, except the line of the daily maximum. To better emphasize differences between the two periods, the average, and 10th and 90th percentiles from the current measurements were ratioed against the respective data from the historical period. The results are shown in Fig. 3.4(b). The difference of measurements performed before and after the solstice (22 December) is further highlighted in Fig. 3.5. Data from 21 December were ratioed against data from 23 December, data from 20 December were ratioed against data from 24 December, and so forth. If atmospheric conditions had been the same before and after the solstice, the ratio would be close to unity and only slightly (<±1%) affected by the different earth-sun distance before and after the midsummer mark. Figure 3.5 shows that the actual situation is very different. Results shown in Figs. 3.4 and 3.5 are further discussed below.

" : Spring

Summer

" South Pole :

_ ■ If '*'

i i i . i i South Pole j

\\ ' Maximum

JA SOND-I FMAMJ Month (b)

JA SOND-I FMAMJ Month (b)

Figure 3.4 Comparison of UV Index measured during the last 17 years with historical estimate for South Pole. Panel (a): Individual measurements, average, median, range of 10th - 90th percentiles, and daily maximum. Recent measurements are indicated by thick lines; historical data are indicated by thin lines and grey-shading. Panel (b): Ratio of recent-to-historical data for average, and 10th and 90th percentiles

0 10 20 30 40 50 60 70 80 90 100 Day be I ore or after summer solstice

Figure 3.5 Comparison of the UV Index at the South Pole for periods before and after the summer solstice (22 December). Data from 21 December were ratioed against data from 23 December, data from 20 December were ratioed against data from 24 December, and so forth. Ratios of recent measurements are indicated by thick lines; historical data are indicated by thin lines and grey-shading. Red lines refer to the average noontime UV Index, broken and solid blue lines to the 10th and 90th percentiles, respectively. The thin grey line is the ratio for the daily maximum of recent data. The broken vertical line indicates the equinox

0 10 20 30 40 50 60 70 80 90 100 Day be I ore or after summer solstice

Figure 3.5 Comparison of the UV Index at the South Pole for periods before and after the summer solstice (22 December). Data from 21 December were ratioed against data from 23 December, data from 20 December were ratioed against data from 24 December, and so forth. Ratios of recent measurements are indicated by thick lines; historical data are indicated by thin lines and grey-shading. Red lines refer to the average noontime UV Index, broken and solid blue lines to the 10th and 90th percentiles, respectively. The thin grey line is the ratio for the daily maximum of recent data. The broken vertical line indicates the equinox

Was this article helpful?

0 0

Post a comment