Calculations of the human vitamin D exposure from UV spectral measurements at three European stations

Abstract

The health benefits of solar UVB and vitamin D in reducing the risk of cancer and several other diseases have been well documented in recent years. In this study, quality-checked spectral UV irradiance measurements from three European stations (Jokioinen, Finland; Bilthoven, The Netherlands; and Thessaloniki, Greece) are used and the vitamin D effective dose (VDED) is calculated. The maximum average daily VDED is measured during the second half of June and it is up to 250 times higher than the corresponding winter minimum value. At each site, a polynomial fit between the VDED and the erythemal dose rates is proposed. The average VDED rates at local noon exceed a detection threshold value for the cutaneous production of vitamin D at Thessaloniki and Bilthoven throughout the year. The proposed standard vitamin dose cannot be attained, even for skin types I-III and exposure time of 60 minutes around local noon, under physiological atmospheric conditions at Bilthoven and Jokioinen during 3 and 4 months respectively. The daily VDED values, using the CIE action spectrum, are higher from 2% and 8% during summer and winter respectively at all sites, compared with those derived by the action spectrum proposed by MacLaughlin et al. (Science, 1982, 216, 1001–1003). These differences are comparable with the uncertainty of spectral measurements.

---

Introduction

The international countermeasures taken in the context of the Montreal protocol and its amendments are expected to lead to a slow total ozone recovery in the coming decades, and the onset of the ozone recovery has been detected.¹ However, many uncertainties still exist regarding the future UV radiation levels, because ongoing climate change interacts with ozone changes and in addition climate change can also have important direct implications for the future UV climate, through changes in the cloudiness and atmospheric dynamics. Up to 2006 at least, the surface ultraviolet (UV) irradiance continues to increase at rates of a few percent per decade at most midlatitude stations in the Northern Hemisphere.² Excess UV irradiance levels now, are, for a stationary population, expected to lead to a peak of global incidences of skin cancers around the year 2050. Excess cases of skin cancer are then of the order of 200 per million per year, if compliance with Montreal protocol is fulfilled.³ It is also estimated, that up to 60 000 deaths per year worldwide were caused by the excessive exposure to UV radiation,⁴ which is related to skin cancer, DNA damages etc.^5,6 In recent years however, the beneficial health effects due to UV exposure are being evaluated and discussed extensively.^7–11

Skin pigmentation and age evidently affect the cutaneous production of vitamin D₃.^12,13 The influence of season and latitude on the cutaneous synthesis of vitamin D₃ was initially described by Webb et al.¹⁴ They revealed that a minimum biologically effective UV dose rate is needed for the detection of provitamin D photoconversion to previtamin D. Based on their method results, a detection threshold value of 3.46 mW m⁻² was proposed by Engelsen et al.¹⁵ They applied a UV simulation toll, called FASTR,¹⁶ to compute the extent and the duration of cutaneous vitamin D production worldwide and throughout a year for different surface and atmospheric conditions. According to their findings and in dependence with atmospheric conditions, the vitamin D production can be absent for latitudes higher than 50 degrees (“vitamin D winter”). Webb and Engelsen¹⁷ defined the time required to obtain the recommended UV dose for adequate vitamin D₃ synthesis in human skin (1 SDD) for different skin types. In addition, they provided guidelines on UV exposure duration, taking into account all major variables, like atmospheric conditions, time of the day, percent body exposure and dietary vitamin D intake.

Kimlin et al.,¹⁸ used data from the U.S. Brewer network from year 2000 and theoretical calculations to highlight that the erythemally weighted UV data cannot be directly related with levels and fluctuations in vitamin D-synthesizing UV radiation, due to the stronger dependence of vitamin D action spectrum on the shorter UV wavelengths. McKenzie¹⁹ stated that the vitamin D weighted UV has an even stronger seasonal and geographic variability than erythemally weighted UV. According to that study, it appears that there is no region on the planet where there is no risk of sunburn in summer, yet ample UV for vitamin D production in the winter. More recently, McKenzie et al.,²⁰ based on 100 000 spectra measured at Lauder, New Zealand, developed a simple algorithm to estimate vitamin D production from UV index and determine the optimum conditions of UV exposure. Based on the results, they implied that, when a large area of the body is exposed to UV radiation, sufficient vitamin D could be produced in the mid-latitude winter.

In our study, we calculate the vitamin D effective dose (VDED) from UV spectral measurements at three European stations, Jokioinen (Finland), Bilthoven (The Netherlands) and Thessaloniki (Greece), representing different geographical and environmental conditions. All spectra are weighted with the previtamin D²¹ and the erythemal action spectra,²² in order to establish an empirical relationship between the two biological quantities. The days of year are examined where the VDED for local noon, 9 and 16 local times is higher than the detection threshold value of 3.46 mW/m². We also indicate the minimum recommended exposure time to achieve a standard VDED for different skin types. In many published studies on this subject the action spectrum proposed by MacLaughlin et al.²³ was used, therefore we examine the differences in VDED levels due to the two different action spectra, as derived from measurements.

Data and methodology

The methodology that was used in this work aimed to derive erythemal and vitamin D weighted doses at three different European locations with different geographical conditions (Table 1). The spectral measurements that were used for the analysis were obtained from the European Ultraviolet Data Base (EUVDB) located at the Finnish Meteorological Institute (FMI), (http://uvdb.fmi.fi/uvdb/index.html).

Table 1 Information on measurement sites at the ground used in this study Caption

Location	Longitude (°E), Latitude (°N)	Measurement time-period	Number of spectra
Thessaloniki, Greece	22.95, 40.63	01/1990–12/2005	75733
Bilthoven, The Netherlands	5.19, 52.12	01/1996–12/2005	227055
Jokioinen, Finland	23.49, 60.81	01/1996–12/2005	49788

---

In the framework of the EC-funded project ‘‘Quality Assurance of Spectral Ultraviolet Measurements in Europe through the development of a transportable unit’’ (QASUME) (http://lap.physics.auth.gr/qasume/), a traveling reference spectroradiometer was developed to provide quality assurance of spectral solar UV measurements conducted by spectroradiometers operating in Europe. The three instruments agreed within 6% in the UVB spectral region with the traveling standard.²⁴

The spectral irradiance data were first standardized to 1 nm spectral resolution using the SHICrivm algorithm²⁵ The same algorithm was used to extend the spectral data up to 400 nm, since the instruments at Thessaloniki and Jokioinen perform measurements of the spectral irradiance up to 325 nm and 365 nm respectively. All spectra were weighted with the vitamin D and the erythemal action spectra and the momentary dose rates were calculated.

Daily integrals of effective UV irradiances were chosen as the most representative quantity for UV exposure. Significant uncertainties may be introduced in the calculation of daily integrals, if the frequency of measurements is too low or if to large gaps during a day occur. To reduce such uncertainties, we set limitations for the spacing of data within a day taking into account the usual schedules of measurements at each station: The following algorithm was tested and finally used for the daily data selection criteria:

• Two consecutive scans are less than 1.5 hours apart.

• A value at local noon is available (within 40 minutes from true local noon at each station).

• Data are available from at least 70% of the total day time duration (the time difference between the first and the last scan covers at least 70% of the total hours from sunrise to sunset).

Climatology of vitamin D dose

The daily VDED as a function of the day of the year at the three sites are presented in Fig. 1. Averaged values over the available 10 to 16 years are given. As expected, the maximum values are observed during the second half of June (between days 165 and 180) and reveal the dominant effects of solar zenith angle and cloud variability. At all sites, an excess of UV radiation in late summer compared with periods with similar solar zenith angles earlier in the year can be detected; this cannot be considered only as an ozone effect, since cloudiness and aerosols may not be symmetrical to the summer solstice.

Fig. 1 Climatological average daily VDED production as a function of the day of year at Thessaloniki, Bilthoven and Jokioinen. Ten years of measurements were used for the last two sites and 16 years for Thessaloniki.

The maximum values of VDED are observed at Thessaloniki (around 8 kJ m⁻² at the end of June). The daily doses at Bilthoven and Jokioinen are close to 4 kJ m⁻² and 3.5 kJ m⁻² respectively for the same time period. The range of variation (±1σ) is close to 10% from the average values during all year at Jokioinen and Bilthoven. The same variation is observed at Thessaloniki only from autumn to spring, while it is minimized to the almost 4% during summer, because there are plenty of cloud-free days.

The variability of solar zenith angle and cloudiness throughout the year affects the measured daily VDED at the three sites in different ways. The solar zenith angle at Thessaloniki during local noon is less than 63 degrees throughout the year and the cloudiness is relatively less compared with the other two sites. So, the winter average values at Thessaloniki are 5% of the summer ones. In contrast, the minimum daily VDEDs are 1.3% and 0.4% of the corresponding summer maximum values at Bilthoven and Jokioinen respectively.

Relationship between vitamin D and erythemal dose rates

Momentary values of erythemal dose rates are routinely measured at many sites around the world. These measurements could be probably used to estimate also the VDED rates. In this study, we propose an empirical relationship between the two biological quantities. The dose rates at each site, derived from all spectral measurements under different atmospheric conditions, are presented in Fig. 2. According to the results, the relationship between the two quantities could not be considered as linear. This is because of the differences of the action spectra, with the vitamin D action spectrum being more sensitive to solar irradiance at lower UV-B wavelengths. In wintertime, the solar irradiance below 305 nm is diminished because of the high values of solar zenith angle and the increased cloudiness. The increased optical path of solar irradiance through the atmosphere (for higher values of solar zenith angles) and the enhanced scattering because of cloudiness result in strongly reduced VDED rates, relatively to the erythemal values. In contrast, the solar irradiance at lower UV-B wavelengths can reach the ground during summer, affecting more the vitamin D than the erythemal dose rates.

Fig. 2 Vitamin D effective dose rate (VDED) as a function of erythemal dose rate for Jokioinen (upper panel, left), Bilthoven (upper panel, right) and Thessaloniki (lower panel, left). Thick solid lines represent a quadrature polynomial fit on the data. The percentage differences in vitamin D dose rates between the polynomial and a linear fit are also presented for each site (lower panel, right).

A polynomial fit of second degree is proposed for the estimation of vitamin D dose rates at the three sites. The uncertainties in dose rates, imposed by using the fitting, are within 25% and 10% for erythemal values less than 0.05 W m⁻² and more than 0.1 W m⁻² respectively. The percentage difference between the proposed polynomial fits and simple linear fits is presented in Fig. 2 (lower panel, right). Differences from −50% to +30% are revealed at the three sites, as the erythemal dose rates increases. The above mentioned differences are comparable with the uncertainty of polynomial fit only for very low erythemal dose rates and enhance the statement that the use of a linear fit between the two radiometric quantities could introduce significant uncertainty in results.

The proposed polynomial fits could be used at sites with similar geographical and atmospheric conditions, where there are no available spectral UV and ozone measurements. In recent years, a great effort has been implemented for reconstruction of UV erythemal irradiance at previous decades at many sites around the world.^26–28 A possible future step would be the application of the proposed empirical relationships to the reconstructed erythemal data, in order to estimate the variability and the long-term changes of VDED at these sites.

Vitamin D dose rates during the day

The average VDED rates at the three sites around local noon are presented in Fig. 3. The dose rates are below the detection threshold value from mid-November to end of January at Jokioinen, while they lie just above the threshold value at Bilthoven during winter. At the same time, the average VDED rates are almost 10 times higher at Thessaloniki These results are in accordance with similar studies¹⁵, proposing that, when taking into account the proposed threshold limit, the cutaneous vitamin D production cannot be sustained for latitudes above 50 degrees during winter, the so-called “vitamin D winter”.

Fig. 3 The average vitamin D effective dose (VDED) rates at local noon throughout the year for the selected sites. The proposed detection threshold value (from Engelsen et al.¹⁵), for photoconversion to previtamin D is also presented (dash line).

The exposure of humans to the Sun around the local noon is not always feasible, especially for people who work indoors. For this reason, the average VDED rates at the three sites are examined also around 9 and 16 local time (LT). These hours are selected as indicatives, since they are related to the beginning and the end of the working time period for the majority of the population. For all years of measurements, the wintertime was assumed for the November-March time period and the summer time for the rest of the year. At 9 LT the average dose rate at Thessaloniki is below the threshold limit from the beginning of December to the mid of February (Fig. 4). The cutaneous production of previtamin D in skin is possible from early March till the end of September at Bilthoven and Jokioinen. The VDED rates at 9 LT are almost the same at these two sites, because they are on different time zones, while the solar irradiance at Jokioinen at local noon is lower than Bilthoven. Data are not presented for some days during wintertime at Jokioinen, since the Sun at 9 LT is too close or below the horizon. The VDED rates at 16 LT are higher than the threshold value at Thessaloniki from mid of January to mid of November. In contrast, the recommended dose rates at Jokioinen cannot be attained from mid October to early March. This time period is suppressed by almost one month at Bilthoven.

Fig. 4 The average vitamin D effective dose (VDED) rates during the year at Thessaloniki, Bilthoven and Jokioinen around 9 (upper panel) and 16 (lower panel) local time.

The above-mentioned results about the vitamin D dose rates are only guidelines for these or other similar sites. Environmental factors like air pollution, surface albedo, altitude and even the surroundings at the place of living (the UV irradiance at an open area is considerably higher than at an urban neighbourhood), would strongly modify our findings. In addition, the skin orientation relative to the Sun and the geometry of the human body could modify our results, which are based on UV irradiance measurements on a horizontal surface.

McKenzie et al.²⁰ concluded that even for more limited exposures, the vitamin D produced would be non-zero. He also concluded, that any received threshold for vitamin D production is actually caused by the inability of the method to detect smaller amounts produced. It is evident that the relationship between UV irradiance and vitamin D cutaneous production should be investigated more extensively. In this case, the efficiency of the measured VDED rates, presented in this study, should be also re-examined.

Minimum recommended exposure for vitamin D production

A standard vitamin D dose (1 SDD) for different skin types, based on exposure of ¼ body surface area and corresponds to a UV equivalent of an oral dose of about 1000 IU vitamin D, has been calculated by Webb and Engelsen.¹⁷ A summary of those doses and skin types is presented in Table 2. In our study, we calculate the VDED values for three time intervals around local noon (15, 30 and 60 minutes) and the recommended exposure time to achieve 1 SDD at the three sites is estimated.

Table 2 The standard vitamin D dose (SDD) for different skin types¹⁷

Skin type	Color	1 SDD/kJ m^−2
I	Caucasian; blonde or red hair, freckles, fair skin, blue eyes	0.0372
II	Caucasian; blonde or red hair, freckles, fair skin, blue eyes or green eyes	0.0465
III	Darker Caucasian, light Asian	0.0558
IV	Mediterranean, Asian, Hispanic	0.0836
V	Middle Eastern, Latin, light-skinned black, Indian	0.1114
VI	Dark-skinned black	0.1851

---

The calculated average doses and the standard deviation, based on spectral measurements at the three sites, are presented in Fig. 5. For average atmospheric conditions, the recommended SDD cannot be attained at Jokioinen for skin types V and VI, when the exposure time is limited to 15 minutes. This exposure time could be considered as appropriate during summertime at Bilthoven for people of skin type I to V and occasionally for skin type VI. For skin types I-III, IV, V and VI, 1 SDD can be received at Thessaloniki during almost 9.5, 8, 6.5 and 5 months respectively.

Fig. 5 Average (thick lines) and range of variation (±1σ, shaded areas) of vitamin D effective dose (VDED) received on a horizontal surface exposed for 15 min (upper panel), 30 min (middle panel) and 60 min (lower panel) around local noon at Thessaloniki, Bilthoven and Jokioinen. Horizontal lines mark the dose thresholds for different skin types (I-VI).

If the exposure time is extended to 30 minutes around local noon, 1 SDD can be received under average atmospheric conditions at Jokioinen from mid-March to mid-October for people with skin types I to III. The time period is reduced by 1 and 2 months for skin types IV and V respectively. For people with skin type VI, it is possible to receive 1 SDD only during summertime. At Bilthoven, the above-mentioned time periods are extended by almost one month for skin types I-V and two months for skin type VI. Finally, people with skin types I and VI can receive 1 SDD during 12 and 7 months respectively at Thessaloniki with an exposure time of 30 minutes around local noon.

At the same site and for skin types I-IV, 1 SDD can be received throughout the year with a double exposure time (60 minutes). The time period is reduced to 11 and 9 months for skin types V and VI respectively. Even for skin types I-III, 1 SDD is still not feasible from mid-November to mid-February with an exposure time of 60 minutes under physiological atmospheric conditions at Bilthoven. The above-mentioned time period is extended by almost one month at Jokioinen.

The exposure times shown are not intended to be used as a guideline of the recommended solar exposure. The risk of sunburn, as indicated by the actual UV Index, must be simultaneously taken into account. The production of vitamin D and erythema of the skin are two independent effects on the human body, and they are neither additive nor preclusive in any way.

The recommended exposure time to achieve a standard vitamin D dose for different skin types has been calculated at Australia.²⁹ and New Zealand,²⁰ based on measurements under real or optimum atmospheric conditions. Although a detailed comparison is not feasible, it is evident that for the same skin type the recommended exposure time in the mid-latitude summer of the northern hemisphere is considerably higher than in the southern hemisphere summer. This result is in accordance to previous studies, that confirm the relatively lower UV-B irradiance in Europe, when compared to similar sites of the southern hemisphere^30–32

Differences between the old and the new action spectrum, as derived from measurements

For many years, the vitamin D action spectrum, proposed by McLaughlin et al.,²³ had been used in relevant studies. An updated action spectrum, based on the old one and extended to the UVA spectral region, has been proposed by CIE.²¹ The average difference (%) in daily values of VDED at the three sites for the two action spectra is presented in Fig. 6. At all sites, the difference is below 2% from May to September. When using the CIE action spectrum, the daily value of VDED is higher by 5 and 8% at Bilthoven and Jokioinen respectively. As mentioned before, the solar irradiance at the lower UVB wavelengths is diminished during winter at these sites. As a consequence, the action spectrum extension into UVA spectral region affects mainly the calculated wintertime daily doses, although the differences lie within the uncertainty of spectral measurements.

Fig. 6 The average percentage difference on daily values of VDED calculated with the CIE action spectrum²¹ and the one proposed by McLaughlin et al.,²³ using spectral solar irradiance measurements at Thessaloniki, Bilthoven and Jokioinen.

Conclusions

In this study, spectral UV irradiances measured by three UV spectroradiometers at different sites in Europe (Jokioinen, Bilthoven and Thessaloniki) are used to calculate the vitamin D effective dose (VDED) rates and integrals.

As expected, the maximum average daily VDEDs (from 3.5 to 8 kJ m⁻²) are observed during the second half of June and reveal the dominant effects of solar zenith angle and cloud variability. The winter average values are 20 and 250 times lower than the summer ones at the southern (Thessaloniki) and the northern (Jokioinen) site respectively.

A linear fit between the vitamin D and the erythemal dose rates at each site could introduce uncertainties from −50% to +30% in the calculation of VDED rates. In this study, a polynomial fit is proposed at each site.

The calculation of the average values of VDED rates around local noon reveals that the cutaneous production of vitamin D Thessaloniki and Bilthoven can take place throughout the year. The dose rates are below the detection threshold value (as proposed by Engelsen et al.)¹⁵ from mid-November to end of January at Jokioinen. The average VDED rates at the three sites are examined also around 9 and 16 local time (LT). According to results, the beneficial for vit-D production time period lies between 6.5 (at Jokioinen) and 9.5 months (at Thessaloniki). However, the risk of sunburn must be simultaneously taken into account.

The VDEDs at three different exposure time intervals (15, 30 and 60 minutes) around local noon is calculated and compared with the standard vitamin dose (SDD) for different skin types (as proposed by Webb and Engelsen).¹⁷ Even for skin types I-III, 1 SDD is still not feasible from mid-November to mid-February with an exposure time of 60 minutes under physiological atmospheric conditions at Bilthoven. The above-mentioned time period is extended by almost one month at Jokioinen.

The daily values of VDED are decreased from 2.5 to 8% at all sites during wintertime, when using the action spectrum of MacLaughin et al.²³ These differences are comparable with spectral measurement uncertainties. At all sites, the above mentioned difference is less than 2% during summertime.

Notes and references

1. M. J. Newchurch, E.-S. Yang, D. M. Cunnold, G. C. Reinsel, J. M. Zawodny and J. M. Russell, III, Evidence for slowdown in stratospheric ozone loss: First stage of ozone recovery, J. Geophys. Res., 2003, 108, 4031.
2. A. F. Bais, D. Lubin, A. Arola, G. Bernhard, M. Blumthaler, N. Chubarova, C. Erlick, H. P. Gies, N. Krotkov, K. Lantz, B. Mayer, R. L. McKenzie, R. D. Piacentini, G. Seckmeyer, J. R. Slusser and C. S. Zerefos, Surface ultraviolet radiation: Past, present, and future, Scientific Assessment of Ozone Depletion: 2006. Global Ozone Research and Monitoring Project. Report no. 50, World Meteorological Organization, Geneva, 2007, ch. 7, 7.1–7.54.
3. H. Slaper, G. J. M. Velders, J. S. Daniel, F. R. Degruijl and J. C. Vanderleun, Estimates of ozone depletion and skin cancer incidence to examine the Vienna convention achievements, Nature, 1996, 384(6606), 256–258.
4. R. M. Lucas, A. J. McMichael, B. K. Armstrong and W. T. Smith, Estimating the global disease burden due to ultraviolet radiation exposure, Int. J. Epidemiol., 2008, 37(3), 654–67.
5. J. Elwood, Melanoma and sun exposure, Semin. Oncol., 1996, 23(6), 650–666.
6. M. Berwick, B. K. Armstrong, L. Ben-Porat, J. Fine, A. Kricker, C. Eberle and R. Barnhill, Sun exposure and mortality from melanoma, J. Natl. Cancer Inst., 2005, 97, 195–9.
7. M. F. Holick and D. Vitamin, Importance in the prevention of cancers, type 1 diabetes, heart disease and osteoporosis, Am. J. Clin. Nutr., 2004, 79, 362–371.
8. C. F. Garland, F. C. Garland, E. D. Gorham, M. Lipkin, H. Newmark, S. B. Mohr and M. F. Holick, The role of vitamin D in cancer prevention, Am. J. Public Health, 2006, 96, 252–61.
9. W. B. Grant, Epidemiology of disease risks in relation to vitamin D insufficiency, Prog. Biophys. Mol. Biol., 2006, 92, 65–79.
10. M. F. Holick, Vitamin D deficiency, New Engl. J. Med., 2007, 357(3), 266–81.
11. E. Hyppönen and C. Power, Hypovitaminosis D in British adults at age 45 y: nationwide cohort study of dietary and lifestyle predictors, Am. J. Clin. Nutr., 2007, 85(3), 860–8.
12. J. MacLaughlin and M. F. Holick, Aging decreases the capacity of human skin to produce vitamin D3, J. Clin. Invest., 1985, 76(4), 1536–8.
13. T. C. Chen, F. Chimeh, Z. Lu, J. Mathieu, K. S. Person, A. Zhang, N. Kohn, S. Martinello, R. Berkowitz and M. F. Holick, Factors that influence the cutaneous synthesis and dietary sources of vitamin D, Arch. Biochem. Biophys., 2007, 460(2), 213–7.
14. A. R. Webb, L. Kline and M. F. Holick, Influence of season and latitude on the cutaneous synthesis of vitamin D3: Exposure to winter sunlight in Boston and Edmonton will not promote vitamin D3 synthesis in human skin, J. Clin. Endocrinol. Metab., 1988, 67, 373–378.
15. O. Engelsen, M. Brustad, L. Aksnes and E. Lund, Daily duration of vitamin D synthesis in human skin with relation to latitude, total ozone, altitude, ground cover, aerosols and cloud thickness, Photochem. Photobiol., 2005, 81(6), 1287–1290.
16. O. Engelsen and A. Kylling, Fast simulation tool for ultraviolet radiation at the Earth's surface, Opt. Eng., 2008, 44(4), 041012.
17. A. R. Webb and O. Engelsen, Calculated Ultraviolet Exposure Levels for a Healthy Vitamin D Status, Photochem. Photobiol., 2006, 82(6), 1697–1703.
18. M. G. Kimlin, William J. Olds and Michael R. Moore, Location and Vitamin D synthesis: is the hypothesis validated by geophysical data?, J. Photochem. Photobiol., B, 2007, 86(3), 234–239.
19. R. L. McKenzie, UV radiation Climatology and Trends, Proceedings of the UV Conference “One Century of UV radiation Research”, 2007, Davos, Switzerland, 18–20 September 2007, pp. 13–14.
20. R. L. McKenzie, J. B. Liley and L. O. Björn, UV Radiation: Balancing Risks and Benefits, Photochem. Photobiol., 2008, DOI: 10.1111/j.1751-1097.2008.00400.x.
21. CIE (International Commission on Illumination), Action spectrum for the production of previtamin D3 in human skin, CIE, 174, 2006.
22. A reference action spectrum for ultraviolet induced erythema in human skin, ed. A. F. MacKinley and B. L. Diffey, CIE J., 1987, 6(1), 17–22.
23. J. A. MacLaughlin, R. R. Anderson and M. F. Holick, Spectral character of sunlight modulates photosynthesis of previtamin D3 and its photoisomers in human skin, Science, 1982, 216, 1001–1003.
24. J. Gröbner, et al., Traveling reference spectroradiometer for routine quality assurance of spectral solar ultraviolet irradiance measurements, Appl. Opt., 2005, 44(25), 5321–5331.
25. H. Slaper, H. A. J. M. Reinen, M. Blumthaler, M. Huber and F. Kuik, Comparing ground-level spectrally resolved solar UV measurements using various instruments: A technique resolving effects of wavelength shift and slit width, Geophys. Res. Lett., 1995, 22(20), 2721–2724.
26. G. E. Bodeker and R. L. McKenzie, An algorithm for inferring surface UV irradiance including cloud effects, J. Appl. Meteorol., 1996, 35(10), 1860–1877.
27. P. N. den Outer, H. Slaper and R. B. Tax, UV radiation in the Netherlands: Assessing long-term variability and trends in relation to ozone and clouds, J. Geophys. Res., 2005, 110 , D02203, doi:10.1029/2004JD004824.
28. A. Lindfors, J. Kaurola, A. Arola, T. Koskela, K. Lakkala, W. Josefsson, J. A. Olseth and B. Johnsen, A method for reconstruction of past UV radiation based on radiative transfer modeling: Applied to four stations in northern Europe, J. Geophys. Res., 2007, 112, D23201,  DOI:10.1029/2007JD008454.
29. A. J. Samanek, E. J. Croager, P. Gies, E. Milne, R. Prince, A. J. McMichael, R. M. Lucas and T. Slevin, Estimates of beneficial and harmful sun exposure times during the year for major Australian population centres, Med. J. Aust., 2006, 184(7), 338–41.
30. G. Seckmeyer, B. Mayer, G. Bernhard, R. L. McKenzie, P. V. Johnston, M. Kotkamp, C. R. Booth, T. Lucas, T. Mestechkina, C. R. Roy, H. P. Gies and D. Tomlinson, Geographical differences in the UV measured by intercompared spectroradiometers, Geophys. Res. Lett., 1995, 22, 1889–1892.
31. G. Seckmeyer and R. L. McKenzie, Increased ultraviolet radiation in New Zealand (45° S) relative to Germany (48° N), Nature, 1992, 359, 135–137.
32. G. Seckmeyer, M. Glandorf, C. Wichers, R. L. McKenzie, D. Henriques, F. Carvalho, A. Webb, A. M. Siani, A. Bais, B. Kjeldstad, C. Brogniez, P. Werle, T. Koskela, K. Lakkala, J. Groebner, H. Slaper, P. N. denOuter and U. Feister, Europe's darker atmosphere in the UV-B, Photochem. Photobiol. Sci., 2008, 7, 925–930.

---