Enhanced UV exposure on a ski-field compared with exposures at sea level

Abstract

Personal erythemal UV monitoring badges, which were developed to monitor the UV exposure of school children, were used to measure UV exposures received by one of the authors (MA) at the Mt Hutt ski-field, in New Zealand. These were then compared with measurements taken at the same times from a nearby sea level site in Christchurch city. The badges were designed to give instantaneous readings of erythemally-weighted (i.e., “sun burning”) UV radiation and were cross-calibrated against meteorological grade UV instruments maintained by the National Institute of Water & Atmospheric Research (NIWA). All skiing and calibration days were clear and almost exclusively cloud free. It was found that the UV maxima for horizontal surfaces at the ski-field (altitude ≈2 km) were 20–30% greater than at the low altitude site. Larger differences between the sites were observed when the sensor was oriented perpendicular to the sun. The personal doses of UV received by a sensor on the skier's lapel during two days of skiing activity were less than those received by a stationary detector on a horizontal surface near sea level. The exposures depended strongly on the time of year, and in mid-October the maximum UV intensity on the ski-field was 60% greater than in mid-September. The UV exposure levels experienced during skiing were smaller than the summer maxima at low altitudes.

---

1 Introduction

Ambient intensities of erythemally-weighted (i.e., “sun burning”) UV radiation¹ are relatively high in New Zealand.^2–4 Peak values occur in summer when ozone values are relatively low and when solar elevations are highest. During spring, UV intensities are significantly lower: firstly, because solar elevations are lower and secondly, because ozone levels are relatively high. The large scale ozone losses in Antarctica occur well to the south of the country, and their influence is seen in New Zealand only later in the spring, or in the summer months when ozone-poor air can move over the region after the break-up of the ozone hole.⁵ UV intensities are highest in the north of the country during summer, and are lowest in the south of the country during winter.⁶

Studies of UV radiation in mountainous areas have been carried out in Europe since the 1960s.^7,8 More recently, several groups in different continents have investigated the detailed dependence of UV radiation on altitude. These studies have shown the altitudinal gradients depend on the wavelength of the radiation, the solar elevation, tropospheric pollution (from ozone and aerosols), surface albedo (from snow, and clouds below the observation point), and surface topography.^9,10 Over snow-covered surfaces in particular, the intensity of erythemally-weighted UV can increase appreciably with altitude.

Activities such as skiing can therefore significantly increase our exposure to UV radiation. One mitigating factor is that most of the body is covered. Against that though, the exposure period can be relatively long on skin that may have been unexposed to harmful UV for several months over the winter period. Furthermore, unlike the usual situation where the UV dose is heavily dominated by downwelling radiation, for exposures on ski-fields the upward component may be comparable with the downwelling component.¹¹ Consequently, body-sites that are normally well protected (e.g., under the chin or nose) may be exposed to much larger levels of UV than are usually experienced in summer. At higher altitudes, enhancements in downwelling UV arising from increased reflectivity are smaller than they would be at lower altitudes because Rayleigh scattering is less important at the lower atmospheric pressures. However, for irradiances incident on inclined surfaces, reflections from the snow can lead to large increases.

Here we use data from personal erythemal UV monitoring badges that have recently been developed, to compare the UV exposure of a skier with that near sea level at the same times. We then compare the exposures on the ski-field with those at the peak of summer throughout New Zealand. To our knowledge this is the first study of its kind to investigate such differences in real time-resolved personal exposures rather than just irradiances on a horizontal plane. Further, most previous studies of UV increases with altitude have been carried out at locations where sea level UV intensities are lower, and where the concentrations of tropospheric aerosols and ozone are larger.¹² The data obtained provides a useful baseline for more comprehensive measurements, which are currently being undertaken with sets of similar detectors, in behavioural studies of UV exposure involving school students in New Zealand (Wright et al., manuscript in preparation, 2005).

2 Measurements

This study makes use of UV data from two sources. The first is from the UV monitoring badges, which were designed to measure personal erythemal UV exposure. The badges are miniaturized battery-powered UV detectors that contain on-board data logging capability. Their design and construction details will be described elsewhere (Allen, manuscript in preparation 2005).

These electronic detectors have several advantages over polysulfone film which has been used previously in UV exposure studies. They are re-useable; their response is linear with exposure (rather than logarithmic); and they do not suffer from short term saturation issues. Furthermore, they have a higher data sampling rate, which is necessary in this case to quantify the increased UV exposure whilst engaged in a realistic physical activity, such as skiing. Because these UV monitoring badges are new, we provide some information here about their construction and performance. UV radiation is detected using Schottky photodiodes, fabricated by SVT Associates Inc, USA, from aluminium gallium nitride (AlGaN) ternary alloys. The percentage of Al can be controlled, allowing the responsivity of the photodiodes to be varied across the whole UV spectrum. AlGaN photodiodes fabricated with an Al content of 26% were used in this study. These, according to data published by the manufacturers,¹³ have a spectral response that closely matches the erythemal action spectrum, combined with greater than four orders of magnitude rejection of visible and infra-red radiation. There is no need for additional filtering, and as a result we expect that their temperature dependence will be less than for other sensors, which use combinations of broader band detectors together with interference filters. The actual temperature response is discussed later. Signals are sampled to 10-bit accuracy, and stored in EEPROM memory. The photodiode and electronics are encapsulated in a weather proof case made from shaped PTFE to provide a good cosine response. The diameter of the package is 35 mm, and the thickness is 13 mm. Because of its small size and weight it can be conveniently pinned to apparel for use as a personal UV exposure monitor.

We demonstrate below an excellent calibration against a meteorological grade instrument at the start of the field work and immediately after the field work, and we investigate possible temperature dependences. In this study the UV monitoring badges were configured to record instantaneous UV readings every 8 s. At this sampling interval, approximately 14 days of data can be stored. At the end of each day's observations, data were downloaded to a computer for further analysis. Because most fluctuations in UV (e.g., due to changes in cloud cover) have periodicities longer than the 8 s sampling interval, the data can be integrated to provide an accurate measure of the daily dose of UV.

The second data source is from commercially-available broadband instruments that are designed to continuously measure erythemally weighted UV radiation. The primary comparison is with an instrument located in Christchurch, which logs data at 10 min intervals. This instrument is a second-generation “Roberston–Berger” (RB) meter,¹⁴ model UVB-1 manufactured by Yankee Environmental Systems (YES), and it is owned, operated and calibrated by NIWA.

The results are expressed in terms of the UV index (UVI), which is an internationally-agreed measure of the erythemally-weighted UV radiation incident on a horizontal surface.¹⁵ One unit of UVI corresponds to 25 mW m⁻² of erythemally-weighted UV radiation. In the case of the UV monitoring badge readings, there will be deviations from the usual definition of UV index because the badge orientation is generally not horizontal when worn by a person.

In this initial study, the NIWA instrument was used as a reference calibration standard and the UV monitoring badges were used to investigate: (1) differences between the doses on a horizontal surface and on a badge pinned to the lapel of the subject, and (2) differences in UV exposure on a ski-field compared with a near sea level location.

The two sites for the comparison are Christchurch city and the Mt Hutt ski-field, which is the same latitude as Christchurch, but is 90 km west, and 2 km higher.

The UV intensities are also compared with those measured near sea level at other sites in New Zealand. Details of all the measurement sites are shown in Table 1.

Table 1 Details of measurement sites

Location	Sensor	Latitude/°S	Longitude/°E	Altitude/m
Auckland (Leigh)	Biometer 501	36.5	175.0	30
Wellington (Paraparaumu)	UVB-1	40.9	175.0	5
Mt Hutt ski-field	AlGaN	43.5	171.5	2080
Christchurch	UVB-1	43.5	172.6	30
Lauder Central Otago	UVB-1	45.0	170.0	370
Invercargill	RB meter	46.4	168.3	5

---

3.0 Calibration of UV monitoring badges

3.1 Regression analysis

The personal UV monitoring badge used in this study was calibrated by comparison with measurements from the RB meter at Christchurch, which is operated by NIWA. In turn that instrument is cross calibrated annually against a spectroradiometer at Lauder whose calibration is maintained by reference to lamps whose outputs are traceable to the US National Institute of Standards and Technology (NIST).

Cross-calibrations between the badge and the RB meter in Christchurch were carried out on one clear day near the start of the measurement campaign (15 Sept, 2003), and two clear days immediately after the campaign (16 and 17 Oct, 2003), which were cloud free. The range of solar zenith angles (SZA) included in these calibrations was similar to that during the ski-field campaign. Because the RB meter logs data at 10 min intervals, the badge data were averaged similarly for these calibrations. During the latter two days, the badge was pointed directly towards the sun at regular intervals to assess the importance of orientation. Data from these periods were omitted from the regression analysis. The data show the excellent cross-calibration accuracy for a range of SZA, and the seasonal change in UV over the period of the study. The results of the regression analysis are shown in Fig. 1.

Fig. 1 Regression plot for the calibration data at Christchurch. Badge calibration data with quadratic fits through the origin (10 min means, with normal incident data removed).

There are slight differences in the regression curves between the three calibration days, but over the range of conditions encountered in this study the simple regression relation in eqn. (1) was derived:

UVI = 0.015 × count + 4.1 × 10⁻⁶ × (count)² (1)

The near-linearity of the regression curves, and the repeatability from day to day indicate that the spectral response and cosine response of the detector are adequate for the purposes of the present study, which is limited to SZA in the range 35–65°.

There is some evidence for a higher regression slope on the first day. The reasons for this are not yet understood fully. It is possible that local gradients in UV could contribute to these differences because the badge was located on the roof of an eight storey building approximately 2 km from the RB meter. In the winter period, strong inversion layers can build up in the city, but this would tend to have a greater effect on the RB meter. We verified that the difference is not due to a temperature coefficient in the data (see below). We also tried regressing the badge data against model calculations, and RB data against model calculations. In each case there were similar differences in regression slopes between days, and this indicates that the badge performance is at least comparable with the research-grade RB instrument. One possibility, which has not yet been fully explored, is the effect of a slight mismatch between the instrument response¹³ and the erythemal action spectrum.¹ Such errors would result in differences that depend on SZA and ozone. For the purposes of the present study, this is not important since the ozone amounts and SZAs at both sites during the campaign were similar to those on the calibration days. The estimated 2σ uncertainty, relative to the reference instrument, resulting from this approximation in eqn. (1) is ±6%.

3.2 Temperature dependence

Possible temperature dependences of the UV monitoring badge were investigated in two ways.

Firstly, during the calibration periods, we recorded the ambient temperatures at 10 min intervals at the calibration site. The ranges of temperature during all of the badges measurements discussed are shown in Table 2, which also includes information about the ozone amounts and minimum SZA on the observation days. In the regression analysis we included a term to allow for temperature-dependence in the badge output. Because the temperature range was relatively small, we were able to confirm only that the temperature coefficient must be less than 0.5% per °C.

Table 2 Minimum SZA, ozone amounts, and temperatures during all measurement periods with the UV monitoring badge. For Mt Hutt, the first value is for the base altitude, and the second value (in parentheses) is for the top of the ski-field. Ozone values were from the NASA E-P TOMS satellite instrument

Date (and day), 2003	Site	SZAmin	Ozone/DU	Temperature at selected times (NZST)/°C
				09:00	12:00	15:00
15 Sept (258) calibration 1	Christchurch	47.1	364	6.5	12.9	14.1
12 Sept (255) comparison 1	Christchurch	48.3	361
	Mt Hutt	48.3	345	−0.7 (−4.1)	6.7 (−4.3)	4.2 (−3.7)
15 Oct (288) comparison 2	Christchurch	35.5	356
	Mt Hutt	35.5	353	2.6 (1.8)	9.9 (4.4)	9.4 (5.8)
16 Oct (289) calibration 2	Christchurch	35.2	379	10.9	14.1	15.1
17 Oct (290) calibration 3	Christchurch	34.8	385	16.6	14.6	13.6

---

Because the temperature difference between calibration conditions and mountain conditions is larger than the range of temperatures available for calibrations, we carried out a further check of the temperature stability. Under clear sky conditions at Christchurch on 9 June 2004 from 12.20 to 12.40 NZST, the UV monitoring badge, which had been stored in a fridge at approximately 8 °C for two hours, was placed on a hot plate with a large thermal mass, pre-heated to approximately 80 °C, which was then placed in the sun and allowed to return to ambient temperature. By heating the whole badge in this manner we include any temperature coefficient of the electronics as well as the sensor. Model calculations showed that any change in UV intensity over the period of measurement was less than 1%. The results were as follows.

Ambient badge reading (full scale 1024 counts): 73 ± 2 counts

Badge reading after 10 min on the hot plate: 78 ± 2 counts.

Thus, even with this extreme change in temperature, the change in output was relatively small: 7 ± 3%. The ambient temperature at the observation time varied from 9.5 °C to 10.2 °C. If the badge reached 80 °C, the temperature coefficient would have been ≈0.1% per °C. In reality, the real equilibrium temperature would be intermediate between the ambient air temperature and the plate temperature. Based on these results we estimate that the temperature coefficient should be less than 0.3% per °C.

The temperatures at Mt Hutt are at most 15 °C lower than during the calibration period. Based on the above sensitivity tests, we estimate that the observations at Mt Hutt may be underestimated by up to 5%.

3.3 Cosine response

The angular response of the UV monitoring badge was measured at Lauder using a quartz-halogen lamp at a distance of 1.3 m. We found that it showed a reasonable approximation to the cosine response. For incident angles up to 30°, it underestimates the true cosine response by up to 5%. The error is largest at 60°, where it underestimates the true response by 12%. For larger angles the error reduces and at 80° the error is less than 5%. Although these response figures do not represent the state of the art, they are comparable with those of several research grade UV instruments,¹⁶ and they are significantly better than some commercially-available UV detectors. In the present application, where the instrument orientation is non-horizontal in a snow-covered, mountainous area, the effects of any errors in cosine response become more important.

3.4 Directional effects

Unlike previous studies of the effects of altitude and snow cover, the present study investigates the effects on personal exposures for a sensor located on the skier's lapel. Individual UV badge readings, which have arbitrary orientation, will be designated as UVI_b. In this case, the peak UVI readings correspond to times when the badge orientation is normal to the sun. These will be designated UVI_n.

During calibrations at Christchurch, the dosimeter badge was generally positioned to measure the irradiance falling on a horizontal surface. To quantify the difference between UVI on a horizontal surface with that normal to the sun, the badge was occasionally re-positioned to point directly towards the sun on the 16 October, resulting in higher data points, as shown in Fig. 2.

Fig. 2 Time series of UVI observations in Christchurch on 16 Oct, 2003. The upper smooth curve shows the normal-incidence UV (UVI_n) derived as discussed in the text.

The ratio (UVI_n/UVI) of erythemal UV radiation on a surface normal to the sun compared with that on a horizontal surface has been measured previously at Lauder,¹⁷ and the results are shown in Fig. 3, along with the corresponding measurements from the calibration sequences at Christchurch.

Fig. 3 Relationship between erythemal UV incident on a horizontal surface compared with that incident on a surface perpendicular to the sun.

For the measurements at Lauder, which is a pristine site at altitude 370 m, the function could be adequately represented by a 4th order polynomial. Slight differences would be expected due to differences in atmospheric conditions (e.g., Rayleigh scattering and aerosol scattering), and departures from the ideal cosine-weighted angular response of the detector.

The range of SZA for which data are available in the present study is rather restricted, so we simply assumed the curve had a similar form, and adjusted the offset parameter in the polynomial from 0.86 to 0.94 to match the badge measurements. The smooth dashed upper curve in Fig. 2 is then derived from the 10 min means measured from the RB meter, and shows the resulting agreement between the derived function and measured normal incidence radiation (UVI_n) measured by the UV monitoring badge. With this function the normal-incidence radiation may be over estimated by up to 10% at SZA greater than 50° (see Fig. 2). However, for the more critical midday periods, any conversion errors are less than ≈5%.

4.0 Measurements at Mt Hutt ski-field compared with Christchurch

For measurements at Mt Hutt ski-field, increases in UV arise mainly from increased reflectivity from snow and increased altitude. The altitude range of the skiing area at Mt Hutt is from 1585 m to 2075 m.

UV measurements were made on two days at Mt Hutt using the personal UV monitoring badge and compared with those at Christchurch, which is close to sea level. The badge at Mt Hutt was worn on the lapel of the skier's ski-suit. The UV badge measurements (UVI_b) for both days are shown in Fig. 4. In this figure, the smooth curves correspond to the measured UVI and derived UVI_n (using the function described in section 3.4) on the same day in Christchurch.

Fig. 4 Time series of UVI_b observations at Mt Hutt (12 Sept and 15 Oct 2003). The smooth curves are the corresponding clear sky values and normal incidence values (UVI and UVI_n) measured with the YES RB meter at Christchurch. The weather was clear at both sites throughout both days. The Mt Hutt data are not generally for normal incidence since the sensor was mounted on the lapel of the skier. However, near local solar noon, the sensor was oriented horizontally at the top of the ski run (altitude = 2 km).

Because skiing involves movement up and down complex terrain, combined with frequent changes of direction, the UVI_b recorded by the UV monitoring badge contains data for arbitrary orientations relative to the sun. Therefore it is reasonable to assume, given the high sampling rate of the badge, that the maximum data points in Fig. 4 correspond to times that the orientation of the badge was towards the sun (i.e., giving values of UVI_n at those times). As with most ski-fields in New Zealand, the ski runs at Mt Hutt have a southerly aspect, pointing away from the midday sun. The peak values therefore tend to occur on the north facing ski lifts where skiers lean back as they are towed up the mountain. The grouping of readings below the UVI values for Christchurch corresponds to periods when the skier was on downhill runs, or waiting for the lift. The skier executed 15 to 20 runs on each of these days, as indicated by a closer inspection of the longer period fluctuations. Examples of various skiing-related activities are identified for the period around midday of 15 October in Fig. 5. The badge was deliberately oriented horizontally during the period from 12:33 to 12:35 to allow the peak UVI at the summit of Mt Hutt to be measured.

Fig. 5 UVI_b observations at Mt Hutt compared with UVI and UVI_n at Christchurch over the midday period of 15 Oct 2003.

The results are summarized in Table 3, in which peak results and daily doses for the two days of UV monitoring badge measurements at Mt Hutt ski-field are compared with corresponding measurements in Christchurch. The period of skiing activity was from 09:45 to 14:40 NZST (NZST = UTC +12 hour) on 12 September; and from 09:00 to 15:00 on 15 October. The doses quoted are for the integrals at each site that correspond to those skiing periods. The table also shows the percentage increases at each site between the two observation days, and the percentage increases between the two sites on each observation day.

Table 3 Peak UV and daily doses over the period of skiing for two days at Mt Hutt ski-field compared with corresponding values for Christchurch. Doses are given in terms of the standard erythemal dose (SED), where 1 SED = 100 J m⁻² of erythemally weighted UV.¹⁸

	Christchurch			Mt Hutt			Increase (%)
	Max UVI	Max UVIn	Dose/SED	Max UVI	Max UVIn	Dose/SEDb	Max UVI	Max UVIn
12 Sept 2003	3.3	4.3	12.5	4.3	6.1	10.6	30	42
15 Oct 2003	5.6	6.4	24.4	6.9	8.0	19.7	23	25
Increase (%)	70	48	87	60	31	86

---

4.1 Comparing conditions

At Mt Hutt the snow coverage and weather conditions were similar for both days. The 15th October 2003 was completely cloud free while the 12th September 2003 was completely cloud free until 14:00 NZST. The snow base was in excess of 120 cm on both days, with complete snow coverage across the whole ski-field. The snow conditions on the 12th September were particularly fresh and dry (“powdery”) being one day after a very large snowfall. This was in marked contrast with the “spring snow” conditions of the 15th October, when there were frozen icy patches in the morning which reverted to more “slushy” wet snow after the thaw. On the first day the snow-line was about two thirds of the way up the access road, at an altitude of ≈1000 m. As a result, although the immediate area of the ski-field is snow-covered, the coverage of all terrain within a 20 km radius of the ski field would be less than 50%. The snow line was higher on the second day. During all ski periods, the ski-field was in direct sunlight, despite its southerly aspect.

At Christchurch, the weather was clear both days. As expected, the total ozone column is generally greater over the low altitude site because of the additional tropospheric component. On 12 September, the ozone column was 20 DU greater in Christchurch than at Mt Hutt, but on 15 October the ozone amount at Christchurch was only 6 DU greater than at Mt Hutt (see Table 2).

Since the latitudes of the sites are the same, the noon SZA is the same at each site, but it occurs approximately 4 min later at Mt Hutt.

4.2 Comparing UVI

Because the badge was deliberately oriented horizontally at noon at the top of the ski-field, direct comparisons of peak UVI values are possible. The increase between Christchurch and Mt Hutt was 30% and 23%, respectively, on the two ski-days. The larger increase on 12 September is attributable to the lower ozone amount at Mt Hutt that day. The 5% lower ozone is expected to lead to an increase in UVI of 6%, since for small changes in ozone, sun burning UV increases by approximately 1.2% for each 1% reduction in ozone.¹⁹ The increase in peak UVI over the 5 weeks period between the two ski-days was 70% at Christchurch and 60% at Mt Hutt.

These observed differences in UVI were compared with calculated values using the TUV radiative transfer code.²⁰ In these calculations, we adjusted the aerosol extinction parameters so that there was a match between the measured and calculated UVI at the low altitude site on the first day. We then separately calculated the increase in UVI expected at Mt Hutt for each factor that contributes to these differences (i.e., ozone, altitude, albedo, and aerosol optical depth). The results are shown in Table 4.

Table 4 Calculated peak UVI, with percentage increases from base case in parentheses. Model input parameters were as in Tables 1 and 2. The baseline aerosol model parameters to match measurements at Christchurch were as follows: Ångström aerosol extinction coefficient (β) 0.1, and wavelength dependence (α) 1.3; single scattering albedo 0.95; aerosol asymmetry parameter 0.9. The surface albedo was 0.05. The last line of the table shows corresponding values deduced from the measurements

Conditions	12 Sept	15 Oct
Christchurch: base calculation (matches observed)	3.3	5.7
Mt Hutt:
Base parameters but altitude increased to 2000 m	3.7 (+12%)	6.3 (+10%)
Base parameters but new altitude, ozone decreased	3.9 (+18%)	6.4 (+12%)
Base parameters but new altitude, ozone: albedo increased to 0.70	4.8 (+45%)	7.8 (+37%)
Base parameters but new altitude, ozone, albedo: β reduced to 0.05	5.0 (+51%)	8.1 (+42%)
Mt Hutt: observed (from Table 3)	4.3 (+30%)	6.9 (+23%)

---

The calculated increase in UVI at Christchurch between 12 Sept 2003 and 15 Oct 2003 matches the measured increase at that site, with an increase of ≈70%. However, the calculated increase from Christchurch to Mt Hutt was larger than observed on both days The discrepancy cannot be explained by the temperature-dependence of the UV monitoring badge, since this is at most a 5% effect, and furthermore, that effect would be larger on the first day, which was much colder than the second (Table 2). The most likely explanation is that the effective surface albedo at the mountain site was significantly less than the value of 0.7 assumed in the calculation, especially on the second day. On the second day the snow was less fresh and less extensive. Previous studies have demonstrated that the effective surface albedo for UV measurements at any given measurement site is significantly influenced by surfaces up to 40 km away.²¹ At least half of that region was not snow-covered, especially on the second ski-day. Terrain roughness and consequent shading of some surfaces could also reduce the reflected component. In order to match measured and modelled changes in UVI with altitude, the effective albedo would have to be lower than 70% on the first day, and less than 50% on the 2nd day. An important implication is that if the snow coverage had been more extensive, the increases in UVI would have been significantly larger than observed here.

4.3 Comparing UVI_n

On the two observation days, the maximum UVI_n at Mt Hutt was higher than that in Christchurch by 42% and 25%, respectively. On the first day especially, this increase was larger than for the increase in UVI. We speculate that this is due to a larger contribution from reflected radiation. On this day the surface albedo was higher, and the SZA was larger so that for these peak values, a larger fraction of the snow-covered surface would be within the sensor's field of view. Within the period of skiing activity, the largest relative enhancements of UVI_n, compared with those at Christchurch, occurred at larger SZA. This is partly because the geometric factor UVI_n/UVI increases as a function of SZA for SZA < 65° (Fig. 3), even over surfaces with a low albedo.

4.4. Comparing doses

The UVI_b measurements on the ski-field are highly variable. To investigate how the daily total dose received on the ski-field compares with that at sea level, we integrated the readings at each site over the period of skiing activity, and we express the results in terms of the standard erythemal dose (SED).¹⁸

On the two observation days, the total UV dose (SED_b) measured by the UV monitoring badge at Mt Hutt was 15–20% less than the SED incident on a horizontal surface over the same period in Christchurch (see Table 3), and was ≈40% less than the total daily doses at Christchurch. However, we emphasise that the dose measured at the ski-field (SED_b) is not for horizontal incidence, and the lapel siting of the sensor would not lead to the maximum dose possible because of the shielding effect of the body from some surfaces. The lapel site was selected so that the data are comparable with data from follow-up behavioural studies involving school students wearing these sensors on their lapels. The dose would have been greater, for example, if the sensor had been placed on the forehead,²² but such a site would be impractical for the wider studies planned. Nevertheless, our comparison gives at least a sense of the differences in UV dose received by a skier compared with the maximum dose that could be realistically expected (i.e., on a horizontal surface) near sea level. To fully assess differences in UV exposure, a much more extensive campaign would be necessary, using simultaneous measurements from a large number of monitoring badges on different individuals and body-sites. These should then be compared with a corresponding set of measurements from individuals engaged in outdoor activities that are typical for the sea level site. This more extensive investigation was not possible at the time of this study, but may be possible in the future, since a set of ≈70 such sensors has now been developed for use by school children in behavioural studies.

The UV dose measured by the UV monitoring badge on 15th October was 86% greater than on 12th September, though the extra hour of skiing on the second day contributes approximately 10% of that change. Note that 1 minimum erythemal dose (MED) is approximately equal to 2 SED. At both sites the daily MED is exceeded so, there is risk of sunburn, especially during the second day (15 Oct) when the daily dose was ≈20 SED.

It is perhaps surprising, in view of the sunburn and eye damage that sometimes follows skiing, that the dose received by the badge on the ski-field was less than the dose received by a horizontal surface at sea level. However, such a continuous personal UV exposure at sea level would be unlikely. Also, as noted above, the UV dose received on the skier's nose–top of the head would be significantly higher than the dose recorded here. The more continuous and prolonged UV exposure from spring skiing still presents a relatively harsh UV environment and a sudden increase in UV exposure for skin that may have received little UV during the winter months. Eye damage during skiing is more likely to be the result of the different angles of incidence at the ski-field, with a much larger reflected component, rather than the peak values of the UV radiation.^8,23

5 Results from broad band meters throughout New Zealand

We now put the results from the ski-field study into a wider context by comparing them with UV irradiances measured throughout the year at several sites in New Zealand. In each case the data are from instruments similar to the RB meter at Christchurch. These instruments are located at Leigh (North Auckland), Paraparaumu (Wellington), Lauder (Central Otago) and Invercargill (Southland). Details of the sites were shown in Table 1.

For all of the ski-field measurements, the maximum UVI was ≈7 (with UVI_n ≈8). Although this is significantly greater than at the sea level site at this time of year, it is significantly less than the peak values that occur in summer, as has been seen previously for sites in Europe.²⁴

Fig. 6 displays UV data from three near-sea level sites in New Zealand that span a wide range of latitudes. The data shown are for 30 min integrations. The plots on the left show peak irradiances, and the plots on the right show corresponding daily doses. In these plots, the y-axes on the left use standard SI physical units, and the y-axes on the right are in terms of the UV Index (UVI) and the standard erythemal dose (SED), which are more commonly used in behavioural sciences.^1,18 For example, in North Auckland the peak UVI frequently exceeds 12, and even in the south of the country, the peak UVI can exceed 10. Results, which are similar to previously published values for earlier years,⁶ are summarized in Table 5. Data at Lauder are available at 1 min resolution, and for this site, results are also shown for 1 min integrations. Note that for the shorter integration period, cloud enhancement effects can increase the peak UVI values significantly, in this case by ≈14%, but the effect on clear days, and for daily integrals of UV, is small. With the exception of Lauder (altitude 370 m) all of these sites are close to sea level.

Table 5 Daily UV maxima for year 2003 compared for several sites. Values are based on 30 min averages. Corresponding values based on 1 min averages are shown in parentheses for Lauder

Location	UVI	UV dose/SED
Auckland (Leigh)	13.1	72
Wellington (Paraparaumu)	12.9	67
Christchurch	11.7	66
Lauder Central Otago	11.3 (12.9)	69 (68)
Invercargill	9.8	57

---

Fig. 6 Time series showing seasonal variations in UV measured at Leigh, North Auckland (36.5 °S), Christchurch (43.5 °S), and Invercargill (46.4 °S). The peak value is the maximum mean value over a 30 min period. Periods of missing data during March at Invercargill and Leigh correspond to times when the instrument was in Lauder for calibration.

Finally, we remark that in more polluted regions, such as in Europe, extinctions by aerosols and ozone in the lower atmosphere cause larger reductions in the UV radiation received at the surface near sea level. Consequently, the UV intensities there tend to be significantly less than in New Zealand, and the gradients with altitude are correspondingly larger.

6 Conclusions

The utility of recently developed personal badges, which are low-cost, light weight, and have a high data sampling rate has been demonstrated by measuring UV exposures during skiing activities on two clear sky days in the southern hemisphere spring.

The peak UVI measured by the personal UV monitoring badge on the ski-field at altitude 2.1 km was found to be 20–30% greater than that measured on the same day at a nearby sea level location. Larger percentage increases were seen for UVI_n, when the sensor is pointed directly towards the sun. The relative enhancements were largest for larger SZA.

The increase is due to the combined effects of: (1) less air to scatter the radiation, which causes an increase of ≈10% here, (2) enhanced reflection from fresh snow, and (3) reduced tropospheric ozone. The increases would have been larger if the snow had extended over a wider area. Also, in more polluted locations, the contrast between low latitude and high altitude would be larger.

The integrated dose of erythemally weighted UV received on the skier's lapel over a day's skiing activity was less than that received by stationary horizontal sensor at the sea level site, integrated over the same period. The latter case however is extreme as human body sites are very unlikely to be continuously exposed to the sun in exactly the same horizontal position for 5 or 6 hours. To fully assess differences in UV dose, multiple sensors would be needed. Such measurements are now possible, since a set of identical detectors has been manufactured for behavioural studies involving UV exposure monitoring of school children wearing similar lapel badges. The present data provide a useful baseline for those behavioural studies and for further studies into monitoring the UV exposure during different outdoor activities and occupations.

Over the period of the ski season there are strong increases in peak UV intensities, which are primarily related to seasonal changes in the peak solar elevation, rather than changes in ozone. The ozone column tends to increase during the ski season (in contrast to the situation in Antarctica).

At both locations, but especially at the high altitude site, the exposure to UV for surfaces oriented directly towards the sun can be significantly greater than on horizontal surfaces for which the UV Index is defined.

Although the increases with altitude in UV are significant, the exposure levels experienced on ski-fields are not as intense as those experienced at sea level during the summer months when solar elevations are higher and ozone amounts are lower. However, the sudden increases in UV may be physiologically important.

Some individuals in New Zealand may not receive enough UV for vitamin D production in the winter months, it is not clear whether the increased UV from skiing would be beneficial or detrimental to health. More work is necessary to clarify this issue.

Acknowledgements

Temperatures at Christchurch were obtained from New Zealand's National Climate data base (http://cliflo-niwa.niwa.co.nz/). Temperatures at the ski-field were provided by Ross Lawrence, Mountain Manager, Mt Hutt Ski Area. Ozone data were from the NASA total ozone mapping spectrometer (TOMS) on the Earth-probe satellite (http://jwocky.gsfc.nasa.gov/). We thank Paul Johnston and Greg Bodeker, from NIWA Lauder for helpful comments on the revised manuscript. One of the authors (MA) received funding from the Royal Society of New Zealand Science, Mathematics and Technology Teacher Fellowship Scheme whilst developing the badges at the University of Canterbury, Department of Electrical and Computer Engineering.

References

1. A. F. McKinlay and B. L. Diffey, A Reference Action Spectrum for Ultra-violet Induced Erythema in Human Skin, in Human Exposure to Ultraviolet Radiation: Risks and Regulations, ed. W. F. Passchier and B. F. M. Bosnajakovic, Elsevier, Amsterdam, 1987, pp. 83-87.
2. G. Seckmeyer and R. L. McKenzie, Increased ultraviolet radiation in New Zealand (45 °S) relative to Germany (48 °N), Nature, 1992, 359, 135–137.
3. 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.
4. R. L. McKenzie, B. J. Connor and G. E. Bodeker, Increased summertime UV observed in New Zealand in response to ozone loss, Science, 1999, 285, 1709–1711.
5. J. Ajtic, B. J. Connor, C. E. Randall, B. N. Lawrence, G. E. Bodeker and D. N. Heuff, Antarctic Air over New Zealand Following Vortex Breakdown in 1998, Ann. Geophys., 2003, 21, 2175–2183.
6. R. L. McKenzie, G. E. Bodeker, D. J. Keep, M. Kotkamp and J. H. Evans, UV radiation in New Zealand: measured North to South differences, and relationship to other latitudes, Weather Climate, 1996, 16, 17–26.
7. P. Bener, Approximate values of intensity of natural ultraviolet radiation for different amounts of atmospheric ozone, Report No. AD-752 115, US Dept. of Commerce, National Technical Information Service, Springfield, VA, 1972.
8. M. Blumthaler and W. Ambach, Human Solar Ultraviolet Radiant Exposure in High Mountains, Atmos. Environ., 1988, 22, 749–753.
9. J. R. Herman, R. L. McKenzie, S. B. Diaz, J. B. Kerr, S. Madronich and G. Seckmeyer, Ultraviolet radiation at the Earth's surface, in ch. 5 of WMO Scientific Assessment of the Ozone Layer: 1998, Report No. 44, ed. D. L. Albritton, P. J. Aucamp, G. Megie and R. T. Watson, World Meteorological Organization, Global Ozone Research and Monitoring Project, Geneva, 1999.
10. J. B. Kerr, G. Seckmeyer, A. F. Bais, G. Bernhard, M. Blumthaler, S. B. Diaz, N. A. Krotkov, D. Lubin, R. L. McKenzie, A. A. Sabziparvar and J. Verdebout, Surface ultraviolet radiation: past and future, in in ch. 5 of WMO Scientific Assessment of Ozone Depletion: 2002, Report No. 47, ed. A.-L. Ajavon, D. L. Albritton, G. Megie and R. T. Watson, World Meteorological Organization, Global Ozone Research and Monitoring Project, Geneva, 2003.
11. R. L. McKenzie, K. J. Paulin and S. Madronich, Effects of snow cover on UV radiation and surface albedo: a case study, J. Geophys. Res., 1998, 103, 28785–28792.
12. R. McKenzie, D. Smale, G. Bodeker and H. Claude, Ozone profile differences between Europe and New Zealand: Effects on surface UV irradiance and its estimation from satellite sensors, J. Geophys. Res., 2003, 108, 4179 , DOI: 4110.1029/2002JD002770.
13. P. P. Chow, J. J. Klaassen, J. M. VanHove, A. Wowchak, C. Polley and D. King, Group III - nitride materials for ultraviolet detection applications, Proc. SPIE, 2000, 3948, 295–303.
14. D. F. Robertson, Solar Ultraviolet Radiation in Relation to Sunburn and Skin Cancer, Med. J. Aust., 1968, 2, 1123.
15. WHO, Global solar UV Index: A practical guide, World Health Organisation (WHO), World Meteorological Organisation (WMO), United Nations Environment Program (UNEP), and International Commission on Non-Ionising Radiation Protection (ICNRP), Geneva, 2002.
16. K. Leszczynski, K. Jokela, L. Ylianttila, R. Visuri and M. Blumthaler, Erythemally weighted radiometers in solar UV monitoring: results from the WMO/STUK intercomparison, Photochem. Photobiol., 1998, 67, 212–221.
17. R. L. McKenzie, K. Paulin and M. Kotkamp, Erythemal UV irradiances at Lauder New Zealand: relationship between horizontal and normal incidence, Photochem. Photobiol., 1997, 66, 683–689.
18. B. L. Diffey, C. T. Jansén, F. Urbach and H. C. Wulf, The Standard Erythema Dose: a new photobiological concept, Photodermatol. Photoimmunol. Photomed., 1997, 13, 64–66.
19. S. Madronich, R. L. McKenzie, L. O. Björn and M. M. Caldwell, Changes in biologically-active ultraviolet radiation reaching the Earth's surface, J. Photochem. Photobiol., B, 1998, 46, 5–19.
20. S. Madronich and S. Flocke, Theoretical estimation of biologically effective UV radiation at the Earth's surface, in Solar Ultraviolet Radiation, ed. C. S. Zerefos and A. F. Bais, NATO, ASI Series I: Global Environmental Change, Vol. 52, Springer-Verlag, Berlin–Heidelberg, 1995, pp. 23–48.
21. M. Degünther, R. Meerkötter, A. Albold and G. Seckmeyer, Case study of the influence of inhomogeneous surface albedo on UV irradiance, Geophys. Res. Lett., 1998, 25, 3587–3590.
22. P. Hoeppe, A. Oppenrieder, C. Erianto, P. Koepke, J. Reuder, M. Seefeldner and D. Nowak, Visualization of UV exposure of the human body based on data from a scanning UV-measuring system, Int. J. Biometeorol., 2004, 49(1), 18–25.
23. D. H. Sliney, Ultraviolet radiation effects upon the eye: problems of dosimetry, Rad. Protect. Dosim., 1997, 72, 197–206.
24. G. Seckmeyer, B. Mayer, G. Bernhard, R. Erb, A. Albold, H. Jäger and W. R. Stockwell, New maximum UV irradiance levels observed in Central Europe, Atmos. Environ., 1997, 31, 2971–2976.

---