Utilising shade to optimize UV exposure for vitamin D

Numerous studies have stated that humans need to utilise full sun radiation, at certain times of the day, to assist the body in synthesising the required levels of vitamin D3. The time needed to be spent in the full sun depends on a number of factors, for example, age, skin type, latitude, solar zenith angle. Current Australian guidelines suggest exposure to approximately 1/6 to 1/3 of a minimum erythemal dose (MED), depending on age, would be appropriate to provide adequate vitamin D3 levels. The aim of the study was to determine the exposure times to diffuse solar UV to receive exposures of 1/6 and 1/3 MED for a changing solar zenith angle in order to assess the possible role that diffuse UV (scattered radiation) may play in vitamin D3 effective UV exposures (UVD3). Diffuse and global erythemal UV measurements were conducted at five minute intervals over a twelve month period for a solar zenith angle range of 4 to 80 at a latitude of 27.6S. For a diffuse UV exposure of 1/3 MED, solar zenith angles smaller than approximately 50 can be utilised for exposure times of less than 10 minutes. Spectral measurements showed that, for a solar zenith angle of 40, the UVA (315-400 nm) in the diffuse component of the solar UV is reduced by approximately 62% compared to the UVA in the global UV, whereas UVD3 wavelengths are only reduced by approximately 43%. At certain latitudes, diffuse UV under shade may play an important role in providing the human body with adequate levels of UVD3 (290-315 nm) radiation without experiencing the high levels of UVA observed in full sun.

exposure to approximately 1/6 to 1/3 of a minimum erythemal dose (MED), depending on age, would be appropriate to provide adequate vitamin D 3 levels. The aim of the study was to determine the exposure times to diffuse solar UV to receive exposures of 1/6 and 1/3 MED for a changing solar zenith angle in order to assess the possible role that diffuse UV (scattered radiation) may play in vitamin D 3 effective UV expo- 10 sures (UV D3 ). Diffuse and global erythemal UV measurements were conducted at five minute intervals over a twelve month period for a solar zenith angle range of 4 • to 80 • at a latitude of 27.6 • S. For diffuse UV exposures of 1/6 and 1/3 MED, solar zenith angles smaller than 60 • and 50 • respectively can be utilised for exposure times of less than 10 min. Spectral measurements showed that, for a solar zenith angle of 40 • , the UVA 15 (315-400 nm) in the diffuse component of the solar UV is reduced by approximately 62% compared to the UVA in the global UV, whereas UV D3 wavelengths are only reduced by approximately 43%. At certain latitudes, diffuse UV under shade may play an important role in providing the human body with adequate levels of UV D3 (290-330 nm) radiation without experiencing the high levels of damaging UVA observed in full sun.

Introduction
The health effects of solar UV radiation vary significantly, from initiating the synthesis of vitamin D 3 to the severe degradation of the human body. Excessive solar UV radiation exposure led to the premature deaths of approximately 60 000 people in the year 2000 (Lucas et al., 2006). It is well known that exposure to small amounts of UV radiation are Introduction EGU excessive exposure to sunlight is known to cause skin cancer, immune suppression, DNA damage, erythema and sun-related eye disorders (Baron et al., 2003;Glerup et al., 2000;Terenetskaya, 2000;Sliney, 2004;Setlow et al., 1993). Approximately 90-95% of our vitamin D 3 requirement comes from exposure to the sun (Holick, 2004;Holick, 1998). The synthesis of pre-vitamin D 3 is initiated through exposure of human 5 skin to terrestrial UV radiation from 290 to 330 nm. The longer wavelength UVA radiation (greater than 330 nm) plays no part in the synthesis of pre-vitamin D 3 in humans. Although, over exposure to UV wavelengths is known to cause damage, vitamin D 3 synthesis occurs at doses far below those needed for erythema to arise (Webb, 1993). Current Australian guidelines suggest exposure to approximately 1/6 to 1/3 of a MED 10 (minimum erythemal dose), depending on age, would be appropriate to provide adequate vitamin D 3 levels (Samanek et al., 2006;CDHAA, 2004). Older people require more frequent exposures as vitamin D 3 production rates are about 1/5 as high for the elderly as for the young (Holick et al., 1989). Also, people with highly pigmented skin would require exposures 3-4 times greater than those necessary for fair skinned peo- 15 ple (Pathak et al., 1999). Recent research suggests that the optimal serum vitamin D 3 level for bone health and protection against other diseases is between 800 and 2000 IU per day (Aloia and Li-Ng, 2007;Lappe et al., 2007;Vieth et al., 2007;Dawson-Hughes et al., 2005); with 1000 IU equivalent to exposing 15% of the human body to approximately 1/3 MED (PS, 2006). A position statement was recently approved by 20 the Working Group of the Australian and New Zealand Bone and Mineral Society, Endocrine Society, Osteoporosis Australia, Australian College of Dermatologists and the Cancer Council Australia recommending five minutes solar UV exposure either side of the peak UV periods on most days of the week in summer and approximately 2-3 h solar UV exposure over a week in winter (PS, 2006). This position statement is still behind 25 emerging scientific findings (e.g. Giovannucci et al., 2006;Grant et al., 2007;Gorham et al., 2007;Garland et al., 2007). Nevertheless, it is still the most up to date statement that has been distributed to the public. The position statement is also a generalized report as sun exposure varies significantly with latitude and atmospheric conditions. EGU Solar UV radiation is incident on the earth's surface in two distinct components, direct and diffuse. The direct component is straight forward to minimise by simply blocking its path, whereas the diffuse component is incident from all directions and is difficult to minimise. Atmospheric scattering which is influenced by factors such as solar zenith angle (SZA), clouds and atmospheric constituents plays a major role 5 in determining the proportions of UVB (280-315 nm) and UVA (315-400 nm) in diffuse UV, with greater scattering occurring at the shorter UVB wavelengths than at the longer wavelength UVA. Turnbull et al. (2005) found that UVA was reduced by as much as 65% when utilising a shade umbrella, whereas UVB exposures were only reduced by approximately 48%. Therefore, at certain SZA, the diffuse UV may provide adequate levels of the effective wavelengths needed for vitamin D 3 production (UV D3 ), without producing erythema. In comparison, mutagenic and carcinogenic effects on human skin have been shown to be caused by exposure to UVA radiation (Agar et al., 2004;Moan et al., 1999;Garland et al., 2003). Therefore, reducing personal exposures to UVA radiation while still receiving sufficient amounts of the UV D3 wavelengths required 15 to produce recommended vitamin D 3 levels is essential.
The importance of this research is that it provides a basis for public health campaigns aimed at reducing the incidence of hypovitaminosis D and its attendant disorders, as well as reducing over exposure to harmful solar UV radiation. In the light of the considerable body of evidence regarding the harm done by excessive UV exposure and 20 the public awareness of skin cancer, any new public health program would need to avoid confusing the public with seemingly opposing messages. Utilizing diffuse solar UV radiation to obtain beneficial amounts of UV, while at the same time minimizing personal overexposure to UVA radiation may prove to be absolutely necessary for an improvement in public health. Studies have shown that shade structures are important 25 as a UV minimisation strategy; however, shade settings may also play an important role in providing exposures to the human body with adequate levels of UV radiation for vitamin D 3 production without experiencing the higher levels of UVA experienced in full sun. The lengths of exposure to solar UV at noon in major Australian cities to produce EGU exposures of 1/3 and 1/6 MED has previously been reported (Samanek et al., 2006). This current paper extends this previous research to report on the lengths of exposure periods required to produce these exposures for diffuse erythemal UV compared to global erythemal UV for a range of SZA and cloud conditions encountered over a year.
2 Materials and methods 5 2.1 Integrated automated cloud and spectral UV measurement system An integrated automated cloud and spectral UV measurement system was employed at a Southern Hemisphere site at Toowoomba, Australia (27.6 • S, 151.9 • E, 693 m above sea level) for automatically measuring every five minutes both the cloud fraction and the UV spectrum. The system was located on the roof of a building with relatively unobstructed views and the surrounding topography does not provide any significant obstruction to the field of view of the instruments. The cloud detection system is a Total Sky Imager (TSI) (model TSI-440, Yankee Environmental Systems, MA, USA) with approximately 160 • field of view of the sky. The TSI consists of a CCD camera suspended over a hemispherical reflective dome 15 that points upwards on a horizontal plane. For each image, digital analysis provides the fraction of cloud cover as a number between 0 and 1 where 1 represents total cloud cover. The uncertainty in the determination of the amount of cloud has been reported as ±10% at least 95% of the time (Sabburg and Long, 2004).
The spectral UV measurement system is located within three metres of the cloud 20 detection system and is housed in an environmentally sealed box with a fibre optic connecting the diffuser to the input slit of the monochromator through a hole with a seal around the fibre to protect the equipment inside the box from the environment.
The system is based on a double monochromator system model (DTM300, Bentham Instruments, Reading, UK) with the gratings having 2400 grooves/mm. The input and EGU bandwidth of 0.5 nm. The spectral UV data was recorded from 280 to 400 nm in 0.5 nm increments. The input optics are provided by a diffuser (model D6) with the manufacturer determined error associated with the cosine response as less than ±0.8% for a SZA up to 70 • and ±3.3% for an SZA of 80 • . The system is scheduled to scan every five minutes from 05:00 EST (Eastern Standard Time) to 17:00 EST and the start of 5 the initialisation of a spectral scan is one minute before the TSI obtains an image of the sky. This allows the spectral scan to start at approximately the same time that the TSI takes the sky image. Wavelength and irradiance calibration of the spectral UV system was undertaken by employing the 365 nm mercury spectral line and a 150 Watt quartz halogen lamp with calibration traceable to the National Physical Laboratory, UK 10 standard.

Diffuse and global erythemal UV
Two erythemal UV meters (UV-Biometer Model 501, Solar Light Co., Philadelphia, PA, USA) were employed to monitor the diffuse and global erythemal UV (UV ery ). The diffuse UV ery meter is a global UV ery meter that has been specifically set up to measure 15 the diffuse erythemal radiation. It utilizes a shadow band, which is aligned east-west, to block the sun during the day. The shadow band is 0.076 m wide and the distance from the shadow band to the top of the diffuse UV ery meter varies from 0.25 m to 0.27 m as it is moved with the seasons. The shadow band blocks out part of the sky view and the amount blocked has been measured at approximately 10%. A correction 20 has been applied to all of the data to account for this. The diffuse and global UV ery meters were set up to record data every five minutes.

Diffuse and global erythemal UV
An example of diffuse and global UV spectra is shown in Fig. 1a  UV plotted on the right hand axis of Fig. 1b gives an indication of the change in ratio when comparing global and diffuse UV wavelengths. Figure 1b provides the global and diffuse solar UV spectra (from Fig. 1a) weighted with the pre-vitamin D (CIE, 2006) action spectrum. Pre-vitamin D 3 exposure was reduced by 43% whereas UVA exposure was reduced by 64%. These spectral mea- 15 surements were obtained using the spectral measurement system described earlier.
3.2 Exposure times for 1/6 and 1/3 MED The times required for an exposure of 1/3 MED on a horizontal plane for SZA greater than 4 • for the atmospheric and cloud conditions during 2003 are shown in Fig. 3a for diffuse and Fig. 3b for global UV ery . Over 25 000 data points are provided in each figure   20 for diffuse and global UV ery for all SZA, sky and atmospheric conditions encountered over this period. The minimum time for an exposure of 1/3 MED due to diffuse UV ery was 4.2 min for a SZA of 4 • . As shown in Fig. 3a and b, this data has been broken down into separate cloud levels of 0 to 80% and 85 to 100%, respectively. This was undertaken to illustrate that 0 to 80% of cloud cover has a lesser effect on the diffuse

EGU
The times required for an exposure of 1/3 MED due to the diffuse and global UV ery on a horizontal plane expressed as a percentage of occurrence for 2003 are shown in Fig. 4. For diffuse UV ery , the maximum incidence for 1/3 MED was in the range of 5 to 10 min with an occurrence of 34.9%, followed by 29.5% for 10 to 15 min, and 17.8% for 15 to 20 min. For global UV ery , the maximum incidence of the times for 1/3 MED was in 5 the range of 5 to 10 min with an occurrence of 32.9%, followed by 26.7% for less than or equal to 5 min, and 17.9% for 10 to 15 min.

Discussion
Studies on the levels of UV observed in the shade of different shade environments have shown that the relative proportions of UVA and UVB in the shade are significantly 10 different to those in full sun (Turnbull et al., 2005;Parisi et al., 2001). The relative proportions of the UVA in shade to those in full sun are decreased compared to the relative proportions of the UVB in shade to those in full sun. At certain latitudes and SZA, unprotected exposure for short periods to the UV under shade structures may be the best course of action as it will contribute 15 more toward vitamin D 3 production than toward erythema compared to exposing the body to full sun UV due to the reduced relative component of the UVA in the shade.
The minimum time for an exposure of 1/3 MED due to diffuse UV ery was 4.2 min for a SZA of 4 • . These times are longer for larger SZA and for overcast conditions. For exposures of 1/6 and 1/3 MED, SZA smaller than 60 • and 50 • respectively can be utilised 20 for exposure times of less than 10 min. For SZA between 4 • to 50 • , 94% of the data shows that an exposure of 1/6 MED requires less than 10 min of exposure to diffuse UV. This extends to 20 min for exposure to 1/3 MED. Although the measurements are at one latitude, the data was collected for the SZA range of 4 • to 80 • . This will allow the results to be applicable at latitudes that experience these SZA. 25 The values provided in this paper are for pale skin and for a sky view that is not blocked, apart from the blocking of the solar disc. For larger structures the amount

EGU
of sky view will be reduced due to the structural components of the shade structure, causing the exposure times required for the respective exposures due to the diffuse UV ery to be slightly longer. Nevertheless, the findings in this paper may have significant ramifications for future public health policy regarding sun exposure, which is currently being debated in many countries (Diffey, 2006;Gillie, 2006). Also, the findings of this 5 research may prove useful to those that work from 09:00 a.m. to 05:00 p.m., where solar UV exposure at noon or during a lunch break may be the only option for the necessary UV exposures to initiate the synthesis of vitamin D 3 . Therefore, advice to stay out of the sun in the middle of the day, which is best for vitamin D 3 synthesis in the skin, can be adjusted to advocate the possible use of appropriate shade environments.   8,[781][782][783][784][785][786][787][788][789][790][791][792][793][794][795][796]2008 Utilising shade to optimize UV exposure for vitamin D  Fig. 4. Histogram of the percentage of the occurrences when the time required for an exposure of 1/3 MED due to diffuse (grey bar) and global (black bar) UV ery falls within a certain time range provided in units of minutes.