UV Doses Worldwide


  • Posted on the website on 7 April 2005

  • This paper is dedicated to the memory of Frederick Urbach, M.D., a renowned dermatologist and pioneering photobiologist; he won the prestigious Finsen Medal of the Association of Internationale de Photobiologie and the Lifetime Achievement Award of the American Society for Photobiology (deceased July 2004 at the age of 82).

*To whom correspondence should be addressed: Center for Devices and Radiological Health, Food and Drug Administration, 9200 Corporate Boulevard (HFZ-120), Rockville, MD 20850, USA. Fax: 301-796-9826; e-mail: DEG@CDRH.FDA.GOV


UV radiation affects human health. Human exposure to UV radiation causes a few beneficial health effects like vitamin D3 formation but it causes many detrimental health effects: sunburn, ocular damage, photoaging, immune suppression, DNA damage and skin cancer. In countries with fair-skinned populations, skin cancer is the most diagnosed of all cancers. In the United States in 2002, there were over one million new skin cancer cases. That means one out of every 285 people got skin cancer. Skin cancer of fair-skinned individuals is increasing at an alarming rate (4–6% per year) around the world and has now reached so-called “pandemic” proportions. Thus, it is important to know what UV doses people around the world get throughout their lives. This review covers how the outdoor UV doses are weighted for different biological effects, the most commonly used measuring devices for terrestrial and personal UV doses, the natural and other effects on terrestrial and personal UV doses, the time people spend outside, their ambient exposures and the terrestrial and personal UV doses of adult outdoor and indoor workers as well as children and adolescents around the world. Overall, outdoor-working adults get about 10%, while indoor-working adults and children get about 3% (2–4%) of the total available annual UV (on a horizontal plane). People's UV doses increase with increasing altitude and decreasing latitude; most indoor-working adult Europeans get 10 000–20 000 J/m2 per year, Americans get 20 000–30 000 J/m2 per year and Australians are estimated to get 20 000–50 000 J/m2 per year (excluding vacation, which can increase the dose by 30% or more).


basal cell carcinoma


Commission Internationale de l'Eclairage


cutaneous malignant melanoma


Environmental Protection Agency


minimal erythemal dose


National Human Activity Pattern Survey


nonmelanoma skin cancer




squamous cell carcinoma


standard erythemal dose


sun protection factor


ultraviolet-A radiation (320–400 nm)


ultraviolet-A1 radiation (340–400 nm)


ultraviolet-B radiation (290–320 nm)


ultraviolet radiation


Ultraviolet radiation (UVR) affects human health (1). It primarily causes detrimental health effects: sunburn (2); photoaging (3,4); eye damage, especially cataracts (5); immune suppression (6,7); DNA damage and mutations (8) and skin cancers (9–11). However, UV exposure can also affect human health in beneficial ways. It is used to treat skin and other diseases, is necessary for vitamin D3 formation (12), possibly lowers hypertension (13,14) and reduces the occurrence of some internal cancers such as prostate cancer (15). The public considers a tan to be the major benefit of UV exposure but some scientists view a tan, along with sunburn, as a warning that too much UV exposure and subsequent damage has occurred to the skin.

UVR exposure can cause skin cancer (16). Most personal exposures to UVR occur from outdoor activities in the sun; however, some UV exposures can come from the sun through windows and from indoor devices that are either for the medical treatment of skin disorders (e.g. psoriasis) or for the cosmetic purpose of tanning (17). The UV waveband regions reaching the earth's surface are only ultraviolet-A (UV-A, 320–400 nm) and ultraviolet-B (UV-B, 290–320 nm). The extreme UV (10–120 nm), far UV (120–200 nm), vacuum UV (< 240 nm) and middle UV or ultraviolet-C radiation (200–290 nm) waveband regions are screened out by the oxygen in the atmosphere; most of the UV-B is screened out by the stratospheric ozone layer (18–20). The UV-B wavelengths have the most carcinogenic potential according to the photocarcinogenesis action spectrum (21,22) accepted by the Commission Internationale de l'Eclairage (CIE). Nonmelanoma skin cancers (NMSC), basal cell carcinomas (BCC) and squamous cell carcinomas (SCC) are almost exclusively caused by cumulative UV exposure, while cutaneous malignant melanoma (CMM) has one or more additional contributing factors: sunburns, number of nevi, genetic background, chemical exposure or other factors.

Skin cancers are a worldwide problem for fair-skinned individuals. The incidence of skin cancers among fair-skinned individuals has been increasing by 4–6% every year (23,24). The incidence of NMSC correlates with the annual erythemally weighted UV doses fair-skinned people get in different parts of the world: about 1100/million per year (25) with 12.5 kJ/m2 per year in the Netherlands (26), about 2300/million per year (27) with 25 kJ/m2 per year in the United States (28) and about 8200/million per year (29) with about 37 kJ/m2 per year in Australia (30–32). In 2002 there were over one million new cases of nonmelanoma and melanoma skin cancers in the United States (33). That means one out of every 285 people got a new skin cancer in the United States in one year. CMM has been increasing in fair-skinned individuals at a logarithmic rate over the past seven decades (F. Urbach, personal communication). It has the highest fatality rate among the skin cancers because it is very invasive and metastasizes to other organs. The risk for an American getting CMM during their lifetime was one out of 250 in 1980, increased to one out of 87 in 1996, further increased to one out of 74 in 2000, and reached one out of 67 in the year 2003 (34,35).

Because skin cancer in fair-skinned populations has now reached so-called “pandemic” proportions, it is important to know what UV doses people around the world get throughout their lives. This review is by no means an exhaustive presentation of all the UV exposure data that is available, but rather covers most outdoor, erythemally weighted UV doses (for indoor tanning see reference 17). However, to understand how these UV doses are ultimately obtained, this review also includes the most common action spectra used to weight the different biological effects, the most common measuring devices of terrestrial and personal UV doses, the natural and other effects on the available terrestrial and personal UV doses, the time people spend outside, their personal ambient exposures and finally the terrestrial and personal erythemally weighted UV doses of adult indoor and outdoor workers as well as of children and adolescents around the world.


UVR can be measured directly in W/m2 for every wavelength, which can then be weighted for different biological effects if an action spectrum exists for that endpoint. To obtain the “effective” W/m2 one simply multiplies the emission spectrum of a given source, wavelength by wavelength, by the appropriate biological action spectrum. Once weighted, or convoluted, the area under the resulting curve is integrated to get the total contribution toward that biological effect in W/m2. To get the “effective” UV dose one simply multiplies the effective W/m2 by the number of seconds of exposure to that particular source and this is called the effective dose for only that biological endpoint. The most common action spectra used for weighting biological effects are the CIE erythemal action spectrum (36) and the CIE photocarcinogenesis action spectrum (19,20) shown in Fig. 1a. However, unweighted UV readings can be convoluted by other action spectra: tanning or melanogenesis (37), SCC (19,20,38,39), DNA damage (40,41), immune suppression (42,43), photoaging (44,45), ocular damage (46) and previtamin D3 production (47). All of these action spectra show that UV-B is primarily responsible for most biological effects. UV-A also contributes toward these effects, but usually to a lesser degree. UV-B appears to be three to four orders of magnitude more effective than UV-A but is really only one order of magnitude more effective after weighting by the sun's emission spectrum because there is much more UV-A than UV-B reaching the earth's surface. For example, although UV-A contributes toward erythema, tanning and SCC (21,22,36,37), it is about three to 10 times less effective than UV-B after convoluting by the solar spectrum. But this is not always the case, because another action spectrum suggests that solar UV-A may contribute almost as much as solar UV-B toward the induction of melanoma in a fish model (48). However, this action spectrum can only be used to calculate the “worst-case” scenario for the influence of UV-A toward human melanoma.

Figure 1a.

Figure 1a.

Action spectra for erythema and NMSC (SCC) for terrestrial wavelengths.

Figure 1b. Standard CIE air mass 1 (AM1) solar spectral irradiance measured at sea level at noon on a clear day in July.

Figure 1c. Erythema or photocarcinogenesis (SCC) action spectra (Fig. 1a) multiplied by the solar spectral irradiance (Fig. 1b).

UV-A is also known to affect different biological endpoints than those affected by UV-B. For example, UV-A contributes towards photoaging but mainly causes skin sagging rather than wrinkling, which is largely caused by UV-B (4). In addition, different cell death mechanisms are induced by UV-A1 (340–400 nm) and UV-B (49–55). Another example is UV-A1 phototherapy, which decreases systemic lupus erythematosus symptoms (56,57), where-as UV-B increases them. Furthermore, because UV-A induces collagenases (58) and UV-B does not, UV-A is used to treat scleroderma (59). The different wavelength regions can have similar or opposite effects, which are not only important in phototherapeutic approaches, but are also important when assessing the risks associated with UV exposure by weighting the different biological effects using action spectra.

When weighting data using action spectra it is important to recognize that the interaction between wavelength regions, i.e. UV-A and UV-B, is not incorporated into the calculations because the data were originally collected at discrete wavelength intervals. The interaction between UV-A and UV-B can change the outcome of the data because UV-A could negate or enhance the effects of UV-B and vice versa. Another caveat concerns the model used to obtain the action spectrum. For example, a fish or a mouse is not a man, and they not only have different skin thicknesses, but also have different amounts and types of skin chromophores. Although skin thickness and chromophore amounts and differences between the species can be mathematically corrected using transmission data (22), the possible different biological outcomes from their presence during irradiation cannot be reconciled. Only experiments on humans can truly answer this question, and they are unethical.

Figure 1a shows the CIE erythemal and photocarcinogenesis action spectra from 288 to 400 nm. If the area under both curves is integrated, a ratio close to unity is obtained (1.03). But once the solar irradiance (Fig. 1b) is multiplied by the erythemal or SCC action spectrum, a weighting difference is clearly seen (Fig. 1c). If one now integrates the area under both curves and forms a ratio, one finds the ratio of SCC to erythema is actually 2.2 at noon. So, the effective dose toward SCC is about 2.08 times the effective dose toward one minimum erythemal dose (MED) from 10:00 A.M. to 4:00 P.M. (D. E. Godar, unpublished calculations), but note that this ratio will change throughout the day and year as the ratio of UV-B to UV-A changes. This is an important point to remember because most of the terrestrial and personal UV doses are weighted by the erythemal action spectrum. Therefore, unless specified otherwise, all the UV doses in this review are erythemally weighted UV doses.


Although many devices and modifications of devices exist today, the two most commonly used for terrestrial measurements are radiometers and spectroradiometers. The former makes broadband measurements, while the latter makes spectrally resolved measurements. In addition, portable data-loggers and handheld monitors are available. An example of a handheld monitor is the Model 500-C (National Biological Corporation, Twinsburg, OH), which was used by Rigel et al. (60) to measure changes in UVR with altitude and latitude.


An example of a radiometer commonly used to monitor changes in the UVR reaching the earth's surface is the Robertson-Berger meter (R-B meter) (61). R-B meters measure UVR at discrete wavelength intervals and are time-integrating radiometers that compound the readings taken during the day into one representative daily graph (62). The total amount of UVR reaching the earth's surface is not measured by these meters, but rather only the “burning” UV-B rays, 290–330 nm (63). Thus, these meters have a spectral sensitivity similar to the human erythemal response. However, they deviate from the human erythemal action spectrum for wavelengths greater than 300 nm and solar zenith angles below 40° (64). R-B meters are also sensitive to temperature changes for wavelengths below 320 nm, deviating up to 6% for every 10°C change in sensor temperature (65). They are fairly reliable for longterm field studies, showing little drift over long periods of time, although some drift was detected over a 10-year period (65,66). These meters report the data in “R-B counts” that have to be calibrated and normalized in order to be converted into erythemally effective J/m2 doses. An R-B count was originally determined to be equivalent to 0.2 J/m2, extrapolated from zenith angles greater than 34° and relative to 297 nm (64). However, since then, different values for R-B counts have been published, ranging from 0.16 J/m2 per count (67) to 0.357 J/m2 per count (68). One study in New Zealand calibrated the R-B meters using spectroradiometers, normalized the data to 300 nm, and found that one R-B count was equivalent to 0.25 ± 0.02 J/m2 (69).

The U.S. Department of Agriculture, in cooperation with the National Park Service, uses R-B meters to monitor terrestrial UVR throughout the United States (70).

In Europe, the European Light Dosimeter Network monitors UVR and photosynthetically active radiation (400–700 nm) using R-B meters (71). The network has about 33 stations with 40 broadband instruments, which covers all of Europe and some locations around the world, including Argentina, Brazil, Egypt, India, Israel, Japan and New Zealand.

In New Zealand, the Meteorological Service has used R-B meters to monitor solar UVR since 1981 but the full potential of those readings cannot be exploited because calibration uncertainties exist (69).

In Australia, the Australian Radiation Laboratory set up a network of modified R-B meters in the early 1980s to monitor solar UVR at 16 sites in Australia and four in Antarctica (72,73).


An example of a spectroradiometer commonly used to monitor the changes in UVR reaching the earth's surface is the Brewer spectroradiometer (Kipp and Zonen Instruments, Inc., Saskatoon, Saskatchewan, Canada). Brewer spectroradiometers are automated instruments that not only measure the amount of UVR that reaches the earth's surface but also infers the total column of ozone in the atmosphere (in Dobson units). They measure full-sky, spectrally resolved solar radiation in the UV-B and UV-A waveband regions. Brewers spectrally resolve UVR between 286.5 and 363 nm wavelengths, usually measuring 0.5 nm wavelength increments. Each reading is stored separately and not integrated over the entire day, so that changes throughout the day can be detected. The consolidated daily readings are weighted by the erythemal action spectrum.

The U.S. Environmental Protection Agency (EPA) had a program operating a network of Brewer MKIV spectroradiometers monitoring UVR throughout the United States, with 14 sites in national parks and seven in urban areas (74). The readings were taken 20–30 times during the daylight hours for every day of the year and the spectroradiometers were in operation from the early 1990s to 2004.


Surveys, questionnaires and time diary reports can give the time people spend outside during the day, which can be converted into a personal ambient exposure (the fraction of the total available UVR that a person gets relative to a horizontal plane) and then into a personal UV dose, if the terrestrial UV doses are known for that time frame. On the other hand, people's ambient exposures can be measured with badges, using polysulphone or spore-containing biofilms or portable data-loggers. An example of a portable datalogger is the “Sunsaver™” wristband, which gives time-stamped readings every 10 minutes (with 75 compounded readings). The Sunsaver™ wristband houses a digital watch, a data-logger and a silicon carbide photodiode (model JECF1-IDE; Laser Components, Olching, Germany) as the sensor, which has a spectral response similar to the CIE erythemal action spectrum (75).

Surveys, questionnaires, and time diary reports

There have been large-scale (>200 people) and small-scale (<200 people) surveys on time spent outdoors and exposure to the sun. Usually telephone surveys are used, but sometimes written surveys are given. Surveys are based on either long-term, e.g. years to decades, or short-term, e.g. weeks to months (76,77), memory recall. Long-term memory recall is not as reliable as short-term memory recall. In addition, people's perceptions of their tendencies to burn and their abilites to tan does not agree with their Fitzpatrick skin types (78). An example of a large survey (about 10 000 people) based on short-term memory recall is the EPA's National Human Activity Pattern Survey (NHAPS). NHAPS used telephone surveys based on previous-day recall to obtain minute-by-minute activity and location data for 24 h, every day of the year for 2 years. From that survey, only the outdoor daylight data was extracted and used to get the UV doses of Americans (28).

Questionnaires are usually based on long-term memory recall, but some have been based on shorter terms: weeks, months or seasons. To check their accuracy, time diary reports can be kept and compared to questionnaires (79). Time diary reports are kept throughout the day by the person and/or by a child's parent or guardian who keeps track of when the child is actually outside and the length of time they spend outdoors. Time diary reports correlate well with actual UVR measurements of people (75,80). However, time diary reports and telephone surveys, aided by badge measurements of people's ambient UV doses, showed a poor correlation between actual UV exposures and people's short-term memory recall of their previous week's outdoor activities (81).

Besides actual measurements, time diary reports and previous-day recall surveys are the most reliable approaches for ultimately determining people's UV doses. However, to assure accuracy, children 10 years or younger must be supported by parental or caretaker observations either to keep their time diary reports for them or to confirm as many of the child's outdoor activities as possible (82).


The two common types of badges used to measure people's ambient UV exposures are polysulphone film badges (83) and spore-containing biofilms (84). Polysulphone film badges and biofilms can be calibrated using a variety of UV sources, but the best calibration sources are broad-spectrum sources that emit radiation similar to the sun, i.e. solar simulators. To obtain personal ambient exposures, a film badge or biofilm is left out on a rooftop with a full-sky view all day to get the total available outdoor UV irradiance. The UV exposure a person gets is expressed as the fraction of that total available UVR, and this value is called their personal ambient exposure. Their personal ambient exposure can be converted into a personal dose by multiplying by the available terrestrial UV dose. The doses can be expressed as J/m2, as MED (200–250 J/m2 for skin type II) or as standard erythemal dose (SED; 100 J/m2).

Most studies have used polysulphone film badges to measure people's outdoor UV exposures. Polysulphone film (P1700, Union Carbide Corporation, Somerset, NJ) is mounted as a 40-μm-thick sheet in standard 35-mm transparency mounts and worn as a badge. These badges are usually placed on the lapel site (83) but they can also be placed on various parts of the body to get doses to those sites. Polysulphone film badges are most responsive to wavelengths below 315 nm and have sensitivity close to the human erythemal response. The change in optical absorbance is measured using a spectrophotometer at 330 nm, where the change in absorbance of the film is proportional to the erythemal UV exposure.

One type of biofilm used to measure personal UV exposures is made using spores from Bacillus subtilis (84). The spores are isolated and immobilized in agarose on a biofilm. These spores have an action spectrum for inactivation of germination similar to the absorption spectrum of DNA and can be modified to reflect the erythemal response of human skin by using appropriate filters. Each badge has its own “dark control” and both the dark control and the exposed parts of the badge are incubated in a nutrient broth for about 4 h to allow spore germination and synthesis of proteins. After staining with Coomassie blue for the presence of proteins, the absorbance at 590 nm is measured using a spectrophotometer. Some studies used biofilms to measure UV exposures in extreme conditions: sports (85,86), high elevations (87), extremely cold regions like Antarctica (88) and in a spacecraft orbiting around the earth (89). The UV-B passing through car windows was also tested but, as expected, only UV-A passed through the glass (90). This is also true for windows in buildings (Godar et al., unpublished results).


Solar zenith angle, time of day, season, and hemisphere

At any given latitude and altitude, the solar zenith angle has the most profound effect on the terrestrial UV-B readings (91). The solar zenith angle changes with time of day and season; the smaller the zenith angle (higher the sun appears in the sky) the more UV-B reaches the earth's surface because there is less air and ozone for it to pass through.

Because the sun's zenith angle changes throughout the day, UV intensity also changes; UV-B changes much more dramatically than does UV-A because UV-B is readily absorbed and reflected by the clouds and atmosphere and is attenuated at dawn and dusk when a thicker layer of ozone and air has to be penetrated. UV-B is most intense from 11:00 A.M. to 1:00 P.M. all year long (92).

Other than time of day, seasons have the most profound effect on the terrestrial UV readings: summer>spring>fall>winter. For example, the erythemal dose during the winter in the contiguous United States can be anywhere from 44 000 J/m2 (Boston, Massachusetts) to 102 000 J/m2 (Riverside, California), and in the summer from 272 000 J/m2 (Boston) to 393 000 J/m2 (Riverside) (28). The summer seasonal doses can exceed the winter seasonal doses by four to six times or more.

The earth's distance from the sun, especially during the summer, also affects the intensity of UV-B reaching the surface. The Northern Hemisphere is 1.7% farther away from the sun in the summer than the Southern Hemisphere is and, as a result, represents a 3.5% variation in distance that results in about a 7% decrease in UV-B intensity (73).

Clouds, particulate matter/air pollution and ozone

Cloud cover can reduce UV-B values 25–30% (93) but reflections off the edges of cumulus clouds near the solar disk can increase UV-B values 25–30% (94). However, the decrease caused by clouds blocking the sun is rarely compensated for by the increase in UV-B reflected off cumulus clouds near the solar disk. The mean reduction in UV-B radiation by clouds is usually 15–30% (91).

Particulate matter or air pollution can reduce the UV-B radiation reaching the earth's surface because particles absorb, scatter and reflect the shorter wavelengths much more than they affect longer wavelengths. A famous example of apparent decreases in terrestrial UV-B doses was reported to occur from 1974 to 1985 in the United States; these readings were actually caused by an increase in air pollution from airports near the R-B meters (68). In reality, the stratospheric ozone had slightly decreased during that time. At the same latitude and altitude, air pollution can cause urban sites to have 20–50% lower UV-B readings than nearby rural sites, but they are usually below 20% (95).

Stratospheric ozone column differences at different locations or on different days can contribute to significant variations in UV-B readings (96). In New Zealand, differences in stratospheric ozone can change the terrestrial UV-B radiation by as much as 10% (93). From 1979 to 2000, the summertime UV-B radiation in New Zealand apparently increased by about 20% (91). From 1976 to 1997, an increase in annual UV-B radiation of about 6.1%± 2.9% per decade was noted from all-weather R-B meter readings in Belsk, Poland (67). Using the all-weather averaged UV readings from R-B meters in the United States from 1974–1979 (68) and from Brewer spectroradiometers in the United States from 1996–1998 (28), after correcting for altitude and latitude differences between the four quadrant sites, one can calculate an average increase in UV-B radiation of about 6% (±5%) per decade. A decrease of 1% in the total column of ozone can result in a 2% (1.25–3%) increase in UV-B radiation (97,98) and a 2.5–2.7% increase in the incidence of NMSC (for the U.S. population: 3.5% SCC, 2% BCC and 2.5% NMSC) (19,99). Cloud cover, climate, aerosols and particulate matter make it difficult to know exactly how much stratospheric ozone depletion has occurred over the decades (18,100,101).

Latitude effects

At midlatitudes (28–46°) around the world the increase in erythemally effective UVR for every degree of latitude toward the equator is between 3% and 3.6% (Table 1a). At higher latitudes (60–68°), the change is even bigger: 4.2%. The change is greater toward the poles because the ozone layer gets thinner. The readings vary because they were taken with different meters in different countries that had varying cloud cover and stratospheric ozone levels.

Table 1a.  Latitude effects on terrestrial UV radiation
Northern HemisphereSitesLatitude (°N)Altitude (km)J/m2 per yearChange in J/m2 per year per °N% increase UV/°N decrease
Godar et al. (28) East CoastAtlanta, GA33.650.315743 030*174903
 Boston, MA42.37<0.1590514  
Godar et al. (28) West CoastRiverside, CA33.93<0.1967926364746.8
 Bozeman, MT45.781.36535711*  
Rigel et al. (60) East CoastOrlando, FL28.54<0.13
 New York40.71<0.1   
 Ono (102,103)Tsukuba363.6
 Jokela et al. (104)Helsinki60.2587200183174.2
 Saariselkä68.4 437000  
Southern HemisphereSitesLatitude (°S)Altitude (km)J/m2 per yearChange in J/m2 per year per °S% increase UV/°S decrease
  1. *Altitude corrected values: 3.8%, 4% and 2.1% were used to adjust the data of Godar et al., Roy et al. and McKenzie et al., respectively.

Roy et al. (73) East CoastSydney33.9<0.1950200252143.5
New Zealand (clear-sky)      
 McKenzie et al. (93)Leigh36.3<0.11281150344563.5
New Zealand (all weather)      
 McKenzie et al. (93)Leigh36.3<0.1923450223593.1

At midlatitudes in the Northern Hemisphere, one can calculate a 3% increase in UVR/°N decrease (28,60) from the data of Godar et al. (28). Rigel et al. (60) measured UVR using a handheld monitor on the east coast of the United States, in New York City and in Orlando, Florida, and also got a 3% change in UVR/°N. In the western United States, there is an increase of 6.8% in UVR/°N decrease in latitude (28). However, note that it is not valid to compare these western sites because the cloud cover differences between the southwest and northwest in the United States are quite dramatic (0.48 versus >0.6) unlike the northwest and northeast where they are very similar (>0.6 cloud cover) (68). In Japan, when similar latitudes as those in the United States are compared, there is about a 3.6% increase in UVR/°N decrease in latitude (102,103). At higher latitudes in northern Europe, two Finland sites, one at 60.2°N and the other at 68.4°N, gave a 4.2% increase in UVR/°N decrease (104).

At midlatitudes in the Southern Hemisphere, the all-weather annual averages between Sydney (33.9°S) and Hobart (42.8°S), which are about the same distances away from the equator as the U.S. sites, give a slightly higher value of 3.5% increase in UVR/°S decrease (73). In New Zealand, clear-sky data give a value of 3.5% increase in UVR/°S decrease (93), but the all-weather data give a value similar to the United States, or a 3.1% increase in UVR/°S decrease. The difference between the clear-sky data and the all-weather data exemplifies the importance of weather, or cloud cover, on the terrestrial UVR readings.

These values are needed for adjusting latitude differences between sites.

Altitude effects

The northwestern and northeastern U.S. sites can be compared for altitude effects because the cloud cover is very similar (>0.6) (68). From the data of Godar et al. (28) using a 3% increase in UVR for every degree decrease in °N, one can calculate a 3.8% decrease in UVR/300 m (or 1000 feet) of descent in elevation from averaged readings between 1996 and 1998 (Table 1b). From the data of Scotto et al. (68), one can calculate a 3.6% decrease in UVR/300 m of descent in elevation in the southwestern United States from averaged readings between 1974 and 1985 (Albuquerque, NM compared to Oakland, CA; cloud cover 0.44 and 0.48, respectively). Comparing Hawaii and New Zealand, McKenzie et al. (105) only got a 2.1% change in UVR/300 m for the same overhead ozone column. The greatest increases occurred at solar zenith angles of 60–70° on a clear day. They accounted for surface reflections (albedo) and other artifacts that affect the readings.

Table 1b.  Altitude effects on terrestrial UV radiation
United StatesSitesAltitude (km)Latitude (°N)J/m2 per yearChange in J/m2 per year per 300 m% decrease UV/300 m descent
Godar et al. (28)Bozeman, MT1.3645.78714 30827 3083.8
 Boston, MA<0.142.37590 514  
Scotto et al. (68)Albuquerque, NM1.5135.0703 21425 1943.6
 Oakland, CA<0.137.7576 404  
McKenzie et al. (105)Mauna Loa, HI3.419.52.1 (±0.6%)
 Lauder, New Zealand0.445.0   
     Change in% decrease
AustraliaSitesAltitude (km)Latitude (°N)J/m2 per yearChange in J/m2 per year per 300 m% decrease UV/300 m descent
  1. Latitude corrected values: 3% was used to adjust the data of Godar et al. and Scotto et al.

Robertson (63)4

In Australia, Robertson (63) got a 4% change in UVR/300 m, similar to the U.S. readings (68).

These values are useful for adjusting sites to the same altitude, which is usually sea level.

Longitudinal changes or climate effects

The northwestern and northeastern U.S. sites have very similar cloud cover (68), making these sites best for finding altitude UVR changes because very little longitudinal difference is observed. The northeast and southeast are best for obtaining latitudinal differences in the United States.

In the United States (28), after correcting for latitude and altitude differences between sites, a longitudinal change of 0.126% occurs in the northern United States, whereas a 31% increase occurs in the southern United States (Table 1c), where the cloud covers are very different from east to west. However, from the study of Scotto et al. (68), from 1974 to 1985 one can calculate a difference of only 7.37% between the southeast and the southwest.

Table 1c.  Longitude (cloud cover) effects on terrestrial UV radiation
United StatesSitesLongitude (°W)J/m2 per yearChange in J/m2 per year per °long% change/year per °long% increase from east to west
Godar et al. (28)Boston, MA71.0359051418.50.003130.126
 Bozeman, MT111.15591257   
Godar et al. (28)Atlanta, GA84.4374303068600.92331.36
 Riverside, CA118.4976057   
Scotto et al. (68)Tallahassee, FL84.460535711820.1957.37
 Oakland, CA122.2649988   
AustraliaSitesLongitude (°W)J/m2 per yearChange in J/m2 per year per °long% change/year per °long% increase from east to west
  1. Latitude- and altitude-corrected values using 3% and 3.8% to correct for latitude and altitude, respectively, for the data of Godar et al., 3% to correct for latitude for the data of Scotto et al. and 3.5% to correct for latitude for the data of Roy et al.

Roy et al. (73)Sydney & Perth151.2101338862260.61421.73

In Australia, after correcting for latitude differences between Sydney (southeast) and Perth (southwest), the longitudinal change from southeast to southwest is about 22% (73). The summer cloud cover in the west is low compared to that in the east, similar to the pattern observed in the United States.

Although the total column of ozone is usually thinner in the Southern Hemisphere, Sydney, Australia gets about the same annual terrestrial UV dose (950 200 J/m2 per year) as does Riverside, California (967 926 J/m2 per year), and both sites are about the same distance away from the equator. In addition, Berger and Urbach (106) found values within 1% for Melbourne, Australia and Oakland, California during an earlier decade. These observations, once again, exemplify the importance of cloud cover on terrestrial UV readings.

Surface reflections (albedo)

Surface reflections can increase the terrestrial readings of UVR (see Table 1d). For example, snow reflects 60–80% of the UVR as measured from the ground (104), though airborne measurements show only a 30% reflection (107). Ice can reflect 7–75% (108). Gypsum sand reflects up to 25%, whereas desert sand only reflects about 5% of the incident UVR, similar to water. About 40% of the incident UVR can penetrate some water to a depth of 50 cm, or 20 inches. Of course, ripples, waves and particles in the water will alter the amount of UV that penetrates to that depth. A pine forest, green grass and farmland reflect 2–4% and black asphalt reflects 4–11% of the incident UVR.

Table 1d.  Surface UV reflections§
Surface% reflected UV (ground)% reflected UV (air)(107)
  1. §Surface UV reflection values depend on solar zenith angle and surface roughness.

Snow60–80% (104)30%
Gypsum sand25%
Desert sand5%
Water5–10% (40% penetrates to 50 cm) (108)6%
  (open ocean)
Pine forest2%
Green grass3%
Green or brown farmland4%
Blacktop asphalt4–11% (108)


The primary reason people do not get all the available terrestrial UVR when they are outside is that trees provide a lot of shade. Trees screen out most of the UVR while a person is outside; however, a person can still get some UVR from reflection even under the complete shade of trees (109). Trees are important in rural areas; both trees and buildings play important roles in urban areas. In addition, awnings, umbrellas, porches and other structures also decrease a person's UV exposure while they are outside. Overall, an individual's UV doses are highly dependent on their personal choices of outdoor activities (110).

In the United States, on average, a person gets about 30% of the available terrestrial UVR while they are outside (28). The range is 27.6–33.2% for children in the northeastern United States (60). About the same percentages were found in Australia, where children get about 32% (20.9–34.6%) of the available UVR while they are outside (111), and outdoor-working adults are exposed to about the same extent, 27–36% (112). However, in England, the children and adolescents only get about 25% of the available UVR while they are outside (113).


In the United States, an indoor-working adult goes outside about 10% of the time during the daylight hours. Ten percent is an average annual value for all Americans: males and females of all ages in the north and south (28). Americans spend the most time outside during the summer (north) and spring (south) and the least time outside during the winter. The percent time an American indoor-working adult spends outside ranges from an average of 3.33% in the winter to 13.3% in the summer. Northern indoor-working men over 40 spend the most time outside during the summer; they go outside about 17% of the time during the daylight hours. American children have similar outdoor times as adults, with the youngest children (0–5 years) going out a bit more during the summer or spring (114). The males go outside more often than the females, and the adolescents go outside the least during the day (28,114).

In England, no difference between boys and girls in either primary (9–10 years) or secondary (14–15 years) schools is observed during the school week, but boys go outside more than girls on the weekends, and adolescents go outside less than children (113).

In Australia, 12-year-old boys go outside more often during the day than do girls and this was not only the case for weekends but for school days as well (111).


A personal UV ambient is that fraction of UVR a person gets out of the total available UVR, on a horizontal plane (see Table 2).

Table 2.  Personal ambient percent exposure of males and females around the world
United States (28) (39°N)Men (>21 years)Women (>21 years)Boys (0–5 years)Girls (0–5 years)Boys (6–12 years)Girls (6–12 years)Adolescent males (13–19 years)Adolescent females (13–19 years)
South (34°N, annual)3.6%2.5%3.5%2.8%3.6%2.6%3.3%1.9%
North (44°N, annual)3.6%2.5%3.4%2.8%3.2%3.1%2.8%2.3%
South (34°N, summer)3.6%3.2%3.5%4.2%3.6%2.6%3.3%2.9%
North (44°N, summer)3.9%3.5%4.4%3.1%3.7%4.3%3.2%2.8%
 Indoor workersOutdoor workersChildrenAdolescents
United States (28,114) (34°N, annual)3.1% (>21 years) (0–12 years) 3.1%(13–19 years) 2.6%
United States (28,114) (44°N, annual)3.1% (>21 years) 3.1%2.6%
Netherlands (26,116) (52.5°N, annual)2.5%7%  
England (50–55°N)(annual) (117,118)(annual) (118) 10%(summer, 9–10 years) (113)(summer, 14–15 years) (113)
 3% (2–4%) females 6.1%, males 6.9%females 4%, males 4.2%
Denmark (75)3.1% & 4.2%6.6% (gardeners)(4–15 years) 4.1%(16–19 years) 4.7%
 (55.41°N, annual)(with vacation) (with vacation)(with vacation)
Sweden (58–60°N, summer)6% (119)10% (119)6.4% (0–5 years) (126) 
Japan (120) (40°N, annual)  3.1% (9–12 years) 
 Adult (120) indoor home workers (summer, 27.5°S)Adult (120) outdoor yard workers (summer, 27.5°S)Adult (127) farmers (summer, 27.5°S)Children (121) (summer)Adolescents (121) (summer)Adolescent boys (120) (summer, 27.5°S)Boys (111) (3/4 years & vacation included, 21–28°S)Girls (111) (3/4 years & vacation included, 21–28°S)
Australia (21–28°S)4.2%9.8%14%4.1%4.5%4.7%8%4.9%
    (7–12 years)(13–19 years)(15–16 years)(12 years)(12 years)
    whole bodywhole body   

In the United States, the average annual personal ambient for all people in the north and south is 3.1% (28). In fact, the annual personal ambients are 3.1% in the north (44°N) and 3.1% in the south (34°N), which shows that latitude is not an important factor. Men over 40 in the north and south have the highest annual personal ambients (4%), which vary from 4% to 5% during all the seasons, except winter. Adolescents (13–19 years) have lower personal ambients (2.6%) than do adults or children (3.1%) (114) though some teenagers make up for this decrease in outdoor UV exposure by using indoor tanning devices (17,115). Unlike males, females have a strong seasonal preference for outdoor activities: summer>spring>fall>winter.

In Europe, Dutch indoor-working adults have annual ambients of about 2.5% (26) and outdoor-working adults have annual ambients of about 7% (116). English indoor-working adults have an annual personal ambient of 3% with a 2–4% range (117) and outdoor-working adults have an annual personal ambient of 10%, while English indoor laboratory workers only get about 2% of the available ambient UVR (118). English children 9–10 years have summer personal ambients of 6.1% and 6.9% for girls and boys, respectively, and adolescents 14–15 years have summer personal ambients of 4% and 4.2% for girls and boys, respectively (113). Danish indoor-working adults have annual personal ambients of 3.1% (like those of Americans and Australians) not including vacation and a 4.2% ambient including vacation (75). The vacation doses can increase the annual indoor-working adult doses by about 30% as found by Thieden et al. (75) and estimated by Godar et al. (114). Danish outdoor-working adult gardeners have a 6.6% personal ambient (75), similar to the Dutch (7%). In Sweden, indoor workers and young children (0–5 years) have ambients of 6% and 6.4%, respectively, and outdoor workers have ambients of about 10% (119).

In Asia, Japanese children have an average annual ambient of 3.1% (120), exactly like children in the United States.

In Australia, indoor-working adults also have personal ambients of 2–4% (121,122) and outdoor-working adults have ambients of 10–14% (122). Australian primary school children (7–12 years) have about a 4% summer ambient and adolescents (13–19 years) have about a 4.5% summer ambient (121).

Thus, around the world, excluding vacation, indoor-working adults and children get about 3% (2–4%) and outdoor workers get about 10% of the annual terrestrial UVR.


For some erythemally weighted terrestrial UV doses in the Northern and Southern hemispheres see Tables 3a and 3b, respectively, and Fig. 2. Note that the effect of altitude was corrected in Fig. 2 (the measured numbers and the altitude corrected numbers are given in Table 3).

Table 3a.  Terrestrial UV doses in the Northern Hemisphere from 1974–1998
Country/source/yearsPlace (longitude)LatitudeAltitudeReading RB or MEDJ/m2 per yearAltitude corrected*
  1. *Altitudes were corrected by bringing the site to sea level using a 3.6% decrease for every 300 m of descent for the data of Berger and Urbach (106) and Scotto el al. (68) and a 3.8% decrease in UV for every 300 m of descent for the data of Godar et al. (28).

United States/Godar et al. (28)/for 1996–1998Atlanta, GA (84.43°W)33.65°N315 m or 1000 feet3871 MED774 103743 030
 Riverside, CA (118.4°W)33.93°N<0.1 km4840967 926967 926
 Boston, MA (71.03°W)42.37°N<0.1 km2953590 514590 514
 Bozeman, MT (111.15°W)45.78°N1.36 km3240648 016535 711
United States/Scotto et al. (68)/for 1974–1985Tallahassee, FL (84.4°W)30.4°N<0.1 km169 500 RB605 357605 357
 El Paso, TX (106.4°W)31.8°N1.194 km211 600755 714647 956
 Fort Worth, TX (97.0°W)32.8°N0.164 km163 500583 214571 792
 Albuquerque, NM (106.6°W)35.0°N1.51 km196 900703 214576 404
 Oakland, CA (122.2°W)37.7°N<0.1 km149 300533 214533 214
 Philadelphia, PA (75.2°W)39.9°N<0.1 km108 400387 143387 143
 Minneapolis, MN (93.2°W)44.9°N0.255 km104 500373 214361 849
 Bismarck, ND (100.7°W)46.8°N0.502 km111 900398 929375 013
United States/Berger andMauna Loa, HI19.5°N3.38 km7078 MED1 415 600844 194
 Urbach (106)/for 1974–1979Tallahassee, FL30.4°N<0.1 km3825765 000765 000
 El Paso, TX31.8°N1.14 km4889977 800844 680
 Fort Worth, TX32.8°N0.25 km3583716 600695 205
 Albuquerque, NM35.0°N1.51 km4511902 200739 507
 Oakland, CA37.7°N<0.1 km3426685 200685 200
 Philadelphia, PA40.0°N<0.1 km2441488 200488 200
 Honey Brook, PA (75°W)40.1°N0.21 km2566513 200500 330
 Des Moines, IA (93°W)41.5°N0.29 km2759551 800532 690
 Minneapolis, MN44.9°N0.25 km2403480 600466 251
 Bismarck, ND46.8°N0.51 km2609521 800490 019
Australia/Berger and Urbach (106)/for 1974–1979Melbourne, Australia (144.58°E)38.0°S<0.1 km3388677 600677 600
Europe/Berger andDavos, Switzerland (9.49°E)46.8°N1.58 km2436487 200395 271
 Urbach (106)/for 1974–1979Belsk-Duzy, Poland (20.49°E)51.8°N<0.1 km1521304 200304 200
Netherlands/Schothorst et al. (116)/for 1983Leiden (4.5°E)52.2°N<0.1 km2240 MED448 000448 000
Table 3b.  Terrestrial UV doses in the Southern Hemisphere for 1991 and 1994
Country/source/yearsPlace (longitude)LatitudeAltitudeReading (MED)J/m2 per yearAltitude corrected
Australia/Roy et al. (73)/for 1991Darwin (130.83°E)12.4°S083751 675 0001 675 000
 Alice Springs (133.88°E)23.7°S0.608 km74951 499 0001 377 481
 Brisbane (153.02°E)27.5°S<0.1 km58491 169 8001 169 800
 Perth (115.83°E)32.0°S<0.1 km61181 233 6001 233 600
 Sydney (151.2°E)33.9°S04751950 200950 200
 Melbourne (145°E)37.8°S0.114 km4364872 800859 533
 Hobart (147.33°E)42.8°S0.110 km3683736 600725 797
New Zealand/McKenzie et al. (93)/for 1994Leigh (174.82°E)36.5°S0923 450923 450
 Lauder (169.67°E)45.0°S0.369 km748 250728 923
Figure 2.

Figure 2.

Erythemally weighted and altitude-corrected terrestrial UV doses in the Northern and Southern hemispheres.

The UVR data in the United States from 1974–1979 (106) were higher than the UVR data from 1974–1985 (68) so at first it appeared that the stratospheric ozone levels might be getting higher over time rather than lower. (Compare the slope of the thin line with the triangles to the dashed line with the squares in Fig. 2.) However, after careful examination of the R-B meters, it was concluded that their positions near airports were responsible for the apparent decrease in UVR due to increases in air pollution over the years.

Notice that dry, arid places with little cloud cover like Riverside, California (U.S. west coast) and, at the same latitude away from the equator, Sydney, Australia (east coast), have almost identical terrestrial UV doses. In addition, although New Zealand is closer to the ozone hole and has a thinner total ozone column on average compared to Australia, it falls on the line along with Australia, probably because it has higher annual rainfall. On average, the Southern Hemisphere gets more UVR than does the Northern Hemisphere, primarily because it is closer to the sun during the summer months, but also because the total ozone column is usually thinner.


See Tables 4a and 4b for the following discussion and note that vacation UV doses are not included unless specified otherwise.

Table 4a.  UV doses of males and females in the Northern Hemisphere in J/m2 per year (excludes vacation doses, unless specified)
United States (39°N) (28)Men 22–40 yearsMen 41–59 yearsMen 60+ yearsWomen 22–40 yearsWomen 41–59 yearsWomen 60+ years
South (34°N)260003500034000210002600025000
North (44°N)180002800024000160001900021000
United States (39°N) (114)0–5 years6–12 years13–19 years0–5 years6–12 years13–19 years
South (34°N)330003200029000280002400020000
North (44°N)240002100018000180002200015000
  1. *Estimates based on 10% personal ambients for outdoor workers and averaged UV terrestrial doses for the United States (1996–1998) from Table 3a.

 Indoor workersOutdoor workersChildrenAdolescents
United States (34°N) (28)2800087102*(0–12 years) 29000(13–19 years) 25000
United States (44°N) (28)2100061927*2100017000
Netherlands (52.2°N) (26)13800 (69 MED)30000 (150 MED)  
Denmark (55.41°N) (75)13200 (with vacation)22400 gardeners(4–15 years)(16–19 years)
   14700 (with vacation)18900 (with vacation)
Sweden (60°N) (119)1000020000–60,000  
Table 4b.  UV doses of males and females in the Southern Hemisphere in J/m2 per year (excluding vacation doses)
AustraliaIndoor workersOutdoor workers
 Infants (30,31) 1 yearChildren (30,31) 2.5 yearsTeenagers (32) 13–14 years
  1. Estimates based on 3% personal ambient for indoor workers and 10% personal ambients for outdoor workers and UV terrestrial doses in Table 3b or extrapolated from Fig. 2.

19°S Townsville146003285036680

On average, people living in the contiguous United States get about 25000 J/m2 of erythemally weighted UVR per year not including vacation or about 33000 J/m2 per year including a conservative continental U.S. vacation (7800 J/m2 per year). Men get more UVR than women do in both the north and south, and men over 40 get the highest UV doses compared to any other age group. Unlike men, women go outside to about the same extent throughout their lives (see Table 4a). Based on an average personal ambient of 10% for outdoor workers in other countries and the terrestrial U.S. doses (see Fig. 2), the average American outdoor worker (39°N) is estimated to get about 75 kJ/m2 per year, not including a vacation.

In Europe, Dutch indoor workers at 52.5°N get 13 800 J/m2 per year (26), Danish indoor workers at 55°N get 13 000 J/m2 per year (75) and Swedish indoor workers at 60°N only get about 10 000 J/m2 per year (119). Depending on their jobs, outdoor workers can get two to 10 times the UV dose that indoor workers get and they also display a similar latitudinal pattern (see Tables 4a and 4b).

Because indoor workers around the world all have about a 3% personal ambient, they have increasing UV doses with decreasing latitude as shown in Fig. 3. Thus, estimates for Australian indoor and outdoor workers can be made by knowing their personal ambients (3% and 10%, respectively), and by knowing the annual terrestrial UV doses in Australia (Table 3b). From the data in Table 3b, we can estimate the UV doses an indoor-working adult should get in Townsville (19°S; 45 000 J/m2 per year), Sydney (33.9°S; 28 500 J/m2 per year) and Melbourne, Australia (44°S; 26 000 J/m2 per year), not including vacation. Using the same approach, we can estimate the doses of an Australian outdoor-working adult (see Table 3b). If a conservative vacation is included, doses are usually at least 30% higher.

Figure 3.

Figure 3.

Indoor-working adults’ annual erythemally weighted UV doses in the Northern Hemisphere and one estimate in the Southern Hemisphere (Townsville, Australia at 19°S).


Few studies actually measured the UV doses of very young children under the age of five using any sort of personal monitoring devices. However, a couple of studies in Australia measured the doses of infants and very young children (30,31). Moise et al. (30) found 1-year-old infants get 8400 J/m2 per year and 2.5-year-old children get about 39 400 J/m2 per year (31). In another study by Moise et al. (32) about the same doses were observed for 13–14-year-old teenagers: they get 36 680 J/m2 per year. From the data in Fig. 3 and assuming that Australian children have about the same ambients as Japanese and American children (3.1%), one can estimate Australian children's doses to be about 37 000 J/m2 per year, almost exactly that measured by Moise et al. (30–32).

People's outdoor behavioral patterns are somewhat subject to cultural changes, so that children and adolescents are evidently now going out less today than in previous generations. Now young adults get about the same UV doses as adults (Table 4). Children in the United States and Denmark are now getting about the same doses as adults (28,75,114,123,124). This situation is apparently different than in previous generations, when children and adolescents were thought to go out about three times more than adults (125). Children now go outside less and have personal ambients and doses similar to adults, so that they now get less than 25% of their lifetime UV dose by the age of 18 (123,124). In fact, a person gets about 25% of their lifetime UV dose for every two decades of life, assuming they live to be about 80 (Fig. 4).

Figure 4.

Figure 4.

Percent lifetime UV dose.

Thus, the new public message should be, “Start young, but continue to practice good UV protection throughout your life.” In the United States, the Public Health Service recommends that people avoid the sun from 10:00 A.M.-4:00 P.M., seek shade whenever possible and use UV protection like sunscreens (SPF 15 or higher), hats (especially broad-brimmed), sunglasses and protective clothing.


For the public, one established benefit of UV-B exposure is the formation of vitamin D3 in the skin, but there are many established risks: sunburn, photoaging, DNA damage, mutations, immune suppression, eye damage and skin cancers. Many factors influence the UVR reaching the earth's surface: solar zenith angle, season, time of day, hemisphere, latitude, altitude, clouds, air pollution, surface reflections and stratospheric ozone. The major factors influencing a person's UV dose, besides avoidance (staying indoors), is everything mentioned above, but especially shade from trees and shade or protection from buildings, awnings, umbrellas and other sources (e.g. clothing, hats and sunscreens).

Around the world with few exceptions males usually go out more than females, have higher personal ambients, get higher annual UV doses and consequently have higher rates of skin cancer than do females. Excluding vacation, indoor-working adults and children around the world get about 3% (±1%) of the available annual terrestrial UVR, while outdoor-working adults get an average of 10% (6.6–17.7%), ranging from two to nine times what indoor workers get. Children now get similar outdoor UV doses as indoor-working adults: teenagers can get the lowest outdoor UV doses, while men over 40 can get the highest UV doses. Most people get about 25% of their lifetime UV dose for every two decades of life, assuming they live to be about 80.


The author would like to acknowledge Dr. Frederick Urbach, M.D. and Professor Jan van der Leun, Ph.D. for critically reviewing this manuscript prior to submission. In addition, I would like to acknowledge Thomas A. Fuchs for literature and other online searches, and Steven Wengraitis and Dr. David Sliney for providing helpful information.