**4. Variations of UV irradiance and spectral composition in cloudy conditions**

prolonged episodes reaching a week or even more in length were recorded when column ozone

Aerosol-size distribution is characterized by the fine mode fraction, e.g., how much the submicronic particles contribute to the AOD at 500 nm. Smoke often contributes more than 90% of small particles' influence in AOD, reducing radiation more strongly at shorter wave‐ lengths [28-31]. Smoke was a major reason for a large AOD in years when there were prolonged dry periods in summer. The major season of forest fires in the region is in July-August and

In most of the years, the dryness and frequency of fire outbreak are moderate. In 2002-2012 the AERONET system recorded AOD values above the threefold and twofold median were met at Tartu-Tõravere meteorological station in about 6% and 17%, respectively, out of all 1500 days of the AERONET measurement data. The influence of the landscape fire episode in April-May 2006 on the UV spectrum is described in [32]. In conditions of seasonal normal ozone between 379 and 391 DU three almost clear days with AOD values at 340 nm between 0.979

**Figure 6.** Comparison of mean daily spectral doses for the group of days manifesting smoke induced large AOD and a

The cloud influence is often described by the cloud modification factor (CMF). CMF is defined as a ratio of measured cloudy irradiance to that of normal conditions of clear sky. Here we use a similar aerosol modification factor (AMF) to compare total influence on the UVR. The AMF in daily irradiance totals was found to be 0.84 in UVA and 0.75 in UVB ranges. It is comparable to the CMF of middle-level clouds. The major difference is that the AMF decreases with

clear day with almost equal noon SZA, similar column ozone 384 DU and moderate AOD (0.16 at 340 nm)

was even more than 50 DU lower of its normal seasonal level.

that of landscape burnings in late April to early May.

and 1.299 were found.

128 Solar Radiation Applications

Clear skies are not common at the study site or in the neighbouring Northern European area. The clear conditions are met most frequently in March. In April to September the number of cloudless days in a month is only one to two on average [33]. Most of the solar radiation energy is received in May to August. In broadband solar radiation the relative contribution of direct sunshine energy is highest on average in May, reaching 45% to its assumed clear value. By August it steadily decreases to 40.5% [15].

Similarly a change of the global radiation energy takes place, from 70% of the assumed clear in May to 66% in August, on average. The lowest direct irradiance energy relative to the assumed clear, 13%, is received in November when the average global irradiance relative to clear is the lowest, at 20%. In low sun conditions during October to February, the overcast conditions prevail and the received UVR energies are small. The aerosol and total ozone contributions in their variability are much less than those from cloudiness. Overcast skies in the summer half-year are also met less frequently than the partly cloudy days, which are typical for the climatic region of study. Clouds may significantly reduce the ground-level irradiance but also enhance it by reflections from bright clouds. A key question is in what conditions the UVR level is reduced or enhanced [34-37].

The Cloud Modification Factor (CMF) is defined as a ratio of measured cloudy irradiance to that of normal conditions of clear sky. When the Sun is shaded by clouds the irradiance is reduced and CMF is below 1. When the Sun is not shaded the reflection from clouds near the Sun may enhance the irradiance, and CMF may exceed 1. In Figure 7 an example of both effects is presented in the case of SZA 36˚ when fortunately both situations occurred during the same day in conditions of variable Cu, Cb and Ac cloudiness.

The data for the cloudless background were taken from spectral dataset of another day. One can see that both enhancement and attenuation increase the relative contribution of UVA radiation in the UVR spectrum. Daily total cloud effect depends much on the cloudiness characteristics during noon hours when the received energy is the largest. In dry summers there are fewer clouds in that time period and less influence on the UVR spectra. In wetter summers, both the attenuation and enhancement are stronger due to the convection of moist air. Total effect depends on the cloud amount and on their vertical extent. Most of the daily variations in UV irradiance in May to August occur during three hours before and after the midday. Statistical relationships between broadband solar irradiance and dose, and those of UVA are stronger than between broadband solar irradiance and UVB radiation.

**Figure 7.** Example of enhancement and attenuation of UV spectral irradiance by clouds at SZA 36˚

**Figure 8.** Diurnal cycles of spectral irradiance in overcast by thick *Frnb* cloudiness at selected UVB (left) and UVA (right) wavelengths, 21 June 2010

Some examples of variable diurnal cycles of UV irradiance spectral density at selected wavelengths are presented in this section. In our work [38] it was noted that in seemingly uniform overcast conditions the recorded ground-level UV irradiance varies across a wide range.

In Figure 8 diurnal cycles of the UVB irradiance at wavelengths 300 and 306 nm and of the UVA irradiance at wavelengths 340 and 400 nm are presented in the case of thick *fractusnim‐ bus* (*Frnb*) cloudiness. One can see strong variations over the relatively low irradiance level.

In Figure 9 diurnal cycles at the same wavelengths are presented in conditions of almost no direct sunshine but close to three times more global irradiance. The cloud cover on that day (5 July 2012) consisted of bright and more transparent *altocumuli* (*Ac*), *stratocumuli* (*Sc*) and *cumulonimbus* (*Cb*) clouds.

**Figure 9.** Diurnal cycles of spectral irradiance in overcast by thinner *Ac, Sc, Cb* cloudiness at selected UVB (left) and UVA (right) wavelengths, 5 July 2012

The UV irradiance levels were significantly larger than in the previous thick cloudiness case. In Figure 10 to Figure 12 the daily cycles of spectral irradiance at the same selected wavelengths were presented in partly cloudy conditions. In Figure 10 the daily broadband direct irradiance relative to clear was 0.29; in Figure 11 it was 0.49 and in Figure 12 the largest 0.82.

The relative global irradiances in these selected days were 0.67, 0.75 and 0.92, respectively. The selected days were 8 July, 9 July in 2012 and 17 July in 2009. Clouds were typical combinations of *cirrus* (*Ci*), *Cb*, *Cu*, and *Ac* in two first cases. Only two cloud types, *Ci* and *Cu,* were met in the third case.

**Figure 8.** Diurnal cycles of spectral irradiance in overcast by thick *Frnb* cloudiness at selected UVB (left) and UVA

**Figure 7.** Example of enhancement and attenuation of UV spectral irradiance by clouds at SZA 36˚

(right) wavelengths, 21 June 2010

130 Solar Radiation Applications

**Figure 10.** Diurnal cycles of spectral irradiance at selected UVB (left) and UVA (right) wavelengths at relatively large amounts of partial cloudiness, 8 July 2012

**Figure 11.** Diurnal cycles of spectral irradiance at selected UVB (left) and UVA (right) wavelengths at mean amounts of partial cloudiness, 9 July 2012

All of these cases might be considered as typical good weather cloud situations in summer months. A common feature is that in morning and evening the change of UVR with time is relatively smooth.

Instrumentation and Measurement of Ground-Level Ultraviolet Irradiance and Spectral Composition in Estonia http://dx.doi.org/10.5772/59615 133

**Figure 12.** Diurnal cycles of spectral irradiance at selected UVB (left) and UVA (right) wavelengths at small amounts of partial cloudiness, 17 July 2009

Contrasting variations begin after heating the surface to the ability of initiating convection, producing *Cu* and *Cb* clouds. Locally produced convective clouds are the major factor causing variability in ground-level solar irradiance. There also exists background cloudiness, often consisting of *Ci*, *Ac*, *Sc* clouds and contributing to the variability of ground-level spectral irradiance. Frontal clouds are also frequent in some years, manifesting high cyclonic activity. Then full days or parts of days are overcast. Usually the attenuation of solar radiation in those conditions is larger [39]. Generally, the attenuation by clouds in the UV range is somewhat less than in the whole incoming solar radiation. In the UVA range the attenuation decreases with the decrease in wavelength and is the smallest reaching UVB spectral range where the increasing absorption by tropospheric ozone leads to an increase of attenuation.

#### **5. Summary and conclusions**

All of these cases might be considered as typical good weather cloud situations in summer months. A common feature is that in morning and evening the change of UVR with time is

**Figure 11.** Diurnal cycles of spectral irradiance at selected UVB (left) and UVA (right) wavelengths at mean amounts of

**Figure 10.** Diurnal cycles of spectral irradiance at selected UVB (left) and UVA (right) wavelengths at relatively large

relatively smooth.

partial cloudiness, 9 July 2012

amounts of partial cloudiness, 8 July 2012

132 Solar Radiation Applications

Environmental effects of UV irradiance at any site depend on the received radiation energy and its spectral composition. The availability of the most efficient UVB irradiance depends on the presence of direct sunshine and on solar elevation. Geographical site and time-dependent regular daily cycle of solar elevation reaches its highest value at noon in real solar time. Noon solar elevation manifests a regular annual cycle, being the smallest at winter solstice and the largest at summer solstice. The highest levels of solar irradiance including UVR in clear weather conditions are recorded at noon.

At summer solstice the shortwave threshold at the study site in the UVB range is 294 nm in normal column ozone and atmospheric transparency conditions. In early morning and late evening the shortwave threshold drops to approximately 310 nm. The UVA spectral range of recorded irradiance is always equal to 316 to 400 nm. In the UVB range it is much shorter, and varies during the day from 5 nm to 20 nm. Due to absorption by tropospheric ozone and partly by aerosols the UVA/UVB ratio of irradiances in the UVA and UVB spectral ranges changes during a year and during a day in wide range.

Immediately after sunrise and before sunset it may reach about 500-600 nm, and decreases to less than 50 in summer noon hours. In cloudless conditions the major modulators of the ground-level UV irradiance are column ozone and spectral aerosol optical depth. Variations in both may result in comparable effects in irradiance levels and spectral composition. The influence of these factors is important to consider in the presence of sunshine. Cloudless UV irradiance may vary more than twice at the same SZA in the UVB spectral range. In most of the UVA range the influence of column ozone is negligible and variations in irradiance result from AOD variations. At solar elevations below 10˚ (SZA above 80˚) the accuracy of spectral irradiance measurements is lower than at higher sun conditions and the spectral influence of column ozone and AOD is less clear.

The regular annual cycle of the atmospheric column ozone with the maximum (380-390 DU) in March-April and minimum (270-280 DU) in October-November is a reason for the differ‐ ences in UVB irradiance levels in spring and autumn. These differences become obvious during sunshine episodes, which tend to be less frequent in autumn and in some years are extended in spring. On the contrary, in spring episodes when column ozone drops to close to usual levels, autumnal ones are met.

Snow conditions at the study site contributing significantly to the surface albedo have been variable in recent decades. In spring in some years snow persists until the middle of April (noon SZA around 48˚) but in other years the presence of snow remains episodic during the whole winter. In autumn there have been separate years when snow cover appeared for some time in October (noon SZA around 69˚). In other cases there had been almost no snow by January. Albedo of snow is high, sometimes close to 0.90, for fresh snow and much smaller for wet and dirty snow. Reflection from the fresh clean snow increases the recorded irradiance significantly. In autumnal period the irradiance level is low before reflecting from surface and its UVB part very small. Since the beginning of November, snow does not increase vitamin D synthesizing irradiance to necessary level. Without snow the surface albedo is only a few percent.

A major contributor to variations of both UVB and UVA irradiance in Northern Europe and Estonia is cloudiness. In cloudy weather the irradiance levels tend to be somewhat lower and the threshold of sensitivity several nanometres larger than in cloudless conditions. Cloud influence on the UVR spectra is related to attenuation and enhancement of irradiance.

Attenuation of solar irradiance happens when the Sun is shaded by cloud. When the Sun shines and bright clouds are close to the visible solar disk, reflection and scattering by these clouds often leads to larger irradiances than in cloudless conditions. A cloudy sky background is a significant irradiance modulating factor. Daily total cloud effect depends much on the cloudiness characteristics during noon hours when the received energy is the largest. In dry summers there are less clouds in that time period and less influence on the UVR spectra. In wetter summers, both the attenuation and enhancement are stronger due to the convection of moist air.

The largest daily doses of UVR spectral density are recorded in the presence of medium cloud amounts. At small cloud amounts, clouds are less frequently located close to the Sun. At large cloud amounts shading of the Sun by clouds and attenuation of irradiance dominates. In moderate cloud amounts sunshine episodes are relatively frequent and related to the en‐ hancement of irradiance. Both enhancement and attenuation increase the relative contribution of UVA radiation in the UVR spectrum. Daily total cloud effect depends much on the cloudi‐ ness characteristics during noon hours when the received energy is the largest. Total effect depends on the cloud amount and on their vertical extent. Deep convective clouds attenuate the UVB range of incoming irradiance significantly more than shallow stratiformed clouds. The CMF of stratiform clouds like *Sc* does not necessarily decrease with decreasing wave‐ length, as is commonly considered for cloudy atmospheres.

## **Acknowledgements**

evening the shortwave threshold drops to approximately 310 nm. The UVA spectral range of recorded irradiance is always equal to 316 to 400 nm. In the UVB range it is much shorter, and varies during the day from 5 nm to 20 nm. Due to absorption by tropospheric ozone and partly by aerosols the UVA/UVB ratio of irradiances in the UVA and UVB spectral ranges changes

Immediately after sunrise and before sunset it may reach about 500-600 nm, and decreases to less than 50 in summer noon hours. In cloudless conditions the major modulators of the ground-level UV irradiance are column ozone and spectral aerosol optical depth. Variations in both may result in comparable effects in irradiance levels and spectral composition. The influence of these factors is important to consider in the presence of sunshine. Cloudless UV irradiance may vary more than twice at the same SZA in the UVB spectral range. In most of the UVA range the influence of column ozone is negligible and variations in irradiance result from AOD variations. At solar elevations below 10˚ (SZA above 80˚) the accuracy of spectral irradiance measurements is lower than at higher sun conditions and the spectral influence of

The regular annual cycle of the atmospheric column ozone with the maximum (380-390 DU) in March-April and minimum (270-280 DU) in October-November is a reason for the differ‐ ences in UVB irradiance levels in spring and autumn. These differences become obvious during sunshine episodes, which tend to be less frequent in autumn and in some years are extended in spring. On the contrary, in spring episodes when column ozone drops to close to usual

Snow conditions at the study site contributing significantly to the surface albedo have been variable in recent decades. In spring in some years snow persists until the middle of April (noon SZA around 48˚) but in other years the presence of snow remains episodic during the whole winter. In autumn there have been separate years when snow cover appeared for some time in October (noon SZA around 69˚). In other cases there had been almost no snow by January. Albedo of snow is high, sometimes close to 0.90, for fresh snow and much smaller for wet and dirty snow. Reflection from the fresh clean snow increases the recorded irradiance significantly. In autumnal period the irradiance level is low before reflecting from surface and its UVB part very small. Since the beginning of November, snow does not increase vitamin D synthesizing irradiance to necessary level. Without snow the surface albedo is only a few

A major contributor to variations of both UVB and UVA irradiance in Northern Europe and Estonia is cloudiness. In cloudy weather the irradiance levels tend to be somewhat lower and the threshold of sensitivity several nanometres larger than in cloudless conditions. Cloud influence on the UVR spectra is related to attenuation and enhancement of irradiance.

Attenuation of solar irradiance happens when the Sun is shaded by cloud. When the Sun shines and bright clouds are close to the visible solar disk, reflection and scattering by these clouds often leads to larger irradiances than in cloudless conditions. A cloudy sky background is a significant irradiance modulating factor. Daily total cloud effect depends much on the cloudiness characteristics during noon hours when the received energy is the largest. In dry

during a year and during a day in wide range.

134 Solar Radiation Applications

column ozone and AOD is less clear.

levels, autumnal ones are met.

percent.

This work has been supported by the project 3.2.0801.11-0041 'Estonian radiation climate' funded by the European Regional Development Fund. In previous years the work has been supported by several ETF grants. The authors thank the AERONET team and Weather Service of the Estonian Environment Agency for the data collected at the Tartu-Tõravere meteorolog‐ ical station.

## **Author details**

Kalju Eerme\* , Margit Aun\* and Uno Veismann

\*Address all correspondence to: kalju.eerme@to.ee

Tartu Observatory, Tõravere, Estonia

## **References**

[1] Caldwell M M, Bormann J F, Ballare C L, Flint S D, and Kulandaveilu G. Terrestrial ecosystems, increased ultraviolet radiation, and interactions with other climate change factors. Photochemical & Photobiological Sciences 2007; 6 252-266.


[16] Eerme K, Veismann U, Ansko I, and Lätt S. Year-to-year variations of the vitamin D synthesis related UV-B radiation in Estonia. In: Slusser J R, Schäfer K, Comerón A. Remote Sensing of Clouds and the Atmosphere XI, proceedings of SPIE 6362, 11-14 September 2006, Stockholm, Sweden. Doi:10.1117/12.688976.

[2] den Outer P N, Slaper H, and Tax R B. UV radiation in the Netherlands: Assessing long-term variability and trends in relation to ozone and clouds. Journal of Geophys‐

[3] Häder D-P, Kumar H D, Smith R C, and Worrest R C. Effects of solar UV radiation on aquatic ecosystems and interactions with climate change. Photochemical and Pho‐

[4] Boa H, Xinghua Z, and Yuesi W. Variability in UVB radiation in Beijing, China. Pho‐

[5] Jackson S P and Bartek J. The DNA-damage response in human biology and disease.

[6] Kakani V G, Reddy K R, Zhao D, and Sailaja K. Field crop responses to ultraviolet-B radiation: A review. Agricultural and Forest Meteorology 2003; 120 191-218.

[7] Martinez-Lozano J A, Utrillas M P, Nunez J A, Esteve A R, Gomez-Amo J L, Estelles V and Pedros R. Measurement and analysis of broadband UVB solar radiation in

[8] Meleti, C, Bais A F, Kazadzis S, Kouremeti N, Garane K, Zerefos C. Factors affecting solar ultraviolet irradiance measured since 1990 at Thessaloniki, Greece. Internation‐

[9] Sabburg J M, Parisi A V, and Kimlin M G. Enhanced spectral UV irradiance: a 1 year

[10] Sandmann H, and Stick C. Spectral and spatial UV sky radiance measurements at a seaside resort under clear sky and slightly overcast conditions. Photochemistry and

[11] Lindfors A, Heikkilä A, Kaurola J, Koskela T, Lakkala K. Reconstruction of solar spectral surface UV irradiance using radiative transfer simulations. Photochemistry

[12] Veismann U, Eerme K, and Koppel R. Solar erythemal ultraviolet radiation in Estonia in 1998. Proceedings of the Estonian Academy of Sciences. Physics. Mathematics

[13] Eerme K, Veismann U, and Koppel R. Ultraviolet irradiance in meteorologically con‐ trasting summers of 1998 and 1999 in Estonia. Proceedings of the Estonian Academy

[14] Eerme K, Veismann U, and Koppel R. Variations of erythemal ultraviolet radiation

[15] Eerme, K. Interannual and intraseasonal variations of the available solar radiation.

and dose at Tartu-Tõravere, Estonia. Climate Research 2002; 22 245-253.

ical Research 2005; 110 D02203, doi:10.1029/2004JD004824.

Spain. Photochemistry and Photobiology 2012; 88 1489-1496.

preliminary study. Atmospheric Research 2003; 66(4) 261-272

al Journal of Remote Sensing 2009; 30(15-16) 4167-4179.

tobiological Sciences 2007; 6 267-285.

136 Solar Radiation Applications

Nature 2009; 461(7267) 1071-1078.

Photobiology 2014; 90 225-232.

2000; 49 122-132.

and Photobiology 2009; 85(5) 1233-1239.

Solar radiation. Croatia: InTech; 2012.

of Sciences. Physics. Mathematics 2000; 49 251-265.

tochemistry and Photobiology 2013; 89 745-750.


(in Russian). Tartu: Academy of Sciences of the Estonian SSR, Institute of Astrophy‐ sics and Atmospheric Physics; 1989. p54 -66.


[37] Mayer B, Kylling A, Madronich S, and Seckmeyer G. Enhanced absorption of UV ra‐ diation due to multiple scattering in clouds: Experimental evidence and theoretical explanation. Journal of Geophysical Research 1998; 103(D3) 31241-31254.

(in Russian). Tartu: Academy of Sciences of the Estonian SSR, Institute of Astrophy‐

[27] Veismann U, Pehk M, Kübarsepp T. Standards for radiometric calibrations of electrooptical devices in Estonia, In: Baltic Electronic Conference: Proceedings of the 4th Bi‐

[28] Barnaba F, Angelini F, Curci G, and Gobbi G P. An important fingerprint of wildfires on the European aerosol load. Atmospheric Chemistry and Physics 2011; 11

[29] Krzyscin J W, Eerme K, and Janouch M. Long-term variations of the UV-B radiation over Central Europe as derived from the reconstructed UV time series, Annales Geo‐

[30] Witte J C, Douglass A R, da Silva A, Torres O, Levy R, and Duncan B N. NASA A-Train and Terra observations of the 2010 Russian wildfires. Atmospheric Chemistry

[31] Jaroslavski J P and Krzyścin J W. Importance of aerosol variations for surface UV-B level. Analysis of ground based data taken at Belsk, Poland, 1992-2004. Journal of Ge‐

[32] Aun M, Eerme K, Ansko I, Veismann U, Lätt S. Modification of spectral ultraviolet doses by different types of overcast cloudiness and atmospheric aerosol. Photochem‐

[33] Schwander H, Koepke P, Kaifel A, and Seckmeyer G. Modification of spectral UV ir‐ radiance by clouds. Journal of Geophysical Research 2002; 107(D16) AAC 7-1–AAC

[34] Seckmeyer G, Glandorf M, Wickers C, McKenzie R, Henriques D, Carvalho F, Webb A, Siani A M, Bais A, Kjeldstad B, Brogniez C, Werle P, Koskela T, Lakkala K, Gröb‐ ner J, Slaper H, den Outer P, and Feister U. Europe's darker atmospheres in the UV-

[35] Staiger H, den Outer P N, Bais A F, Feister U, Johnsen B, and Vuilleumier L. Hourly resolved cloud modification factors in the ultraviolet. Atmospheric Chemistry and

[36] Thiel S, Ammannato L, Bais A, Bandy B, Blumthaler M, Bohn B, Engelsen O, Gobbi G P, Gröbner J, Jäkel E, Junkermann W, Kazadzis S, Kift R, Kjeldstad B, Kouremeti N, Kylling A, Mayer B, Monks P S, Reeves C E, Shallhart B, Scheirer R, Schmidt S, Schmitt R, Schreder J, Silbernagl R, Topaloglou C, Thorseth T M, Webb A R, Wen‐ disch M, and Werle P. Influence of clouds on the spectral actinic flux density in the lower troposphere (INSPECTRO): Overview of field campaigns. Atmospheric Chem‐

ennial Conference, 9-14 October, Tallinn, Estonia. Tallinn; 1994 p187-194.

physicae 2004; 22(5) 1473-1485, doi:10.5194/angeo-22-1473-2004.

ophysical Research 2005; 110 D16201, doi: 10.1029/2005JD005951.

B. Photochemical and Photobiological Sciences 2008; 7 925-930.

sics and Atmospheric Physics; 1989. p54 -66.

10487-10501.

138 Solar Radiation Applications

and Physics 2011; 11 9287-9301.

istry and Photobiology 2011; 87 461-469.

7-12, doi:10.1029/2001JD001297.

Physics 2008; 8 2493-2508.

istry and Physics 2008; 8 1789-1812.


**Solar Applications in Architecture**
