- Email: [email protected]
Investigations on the effect of high surface albedo on erythemally effective UV irradiance: Results of a campaign at the Salar de Uyuni, Bolivia
Journal of Photochemistry and Photobiology B: Biology 87 (2007) 1–8 www.elsevier.com/locate/jphotobiol
Investigations on the effect of high surface albedo on erythemally effective UV irradiance: Results of a campaign at the Salar de Uyuni, Bolivia Joachim Reuder
b
a,*
, Flavio Ghezzi b, Eduardo Palenque b, Rene Torrez b, Marco Andrade b, Francesco Zaratti b
a University of Bergen, Geophysical Institute, Alle´gaten 70, N-5007 Bergen, Norway Universidad Mayor de San Andre´s, Laboratorio de Fı´sica de la Atmo´sfera, Campus Cota Cota, La Paz, Bolivia
Received 23 October 2006; received in revised form 1 December 2006; accepted 1 December 2006 Available online 16 January 2007
Abstract Measurements and model calculations have been performed to study the effect of high surface albedo on erythemally effective UV irradiance. A central part of the investigation has been a one week measurement campaign at Salar de Uyuni in the Southern part of the Bolivian Altiplano. The Salar de Uyuni, the largest salt lake of the world, is characterized by largely homogeneous surface conditions during most of the year. Albedo measurements performed by an UV radiometer result in a reflectivity for erythemally effective radiation of 0.69 ± 0.02. The measurements show hardly any dependency on solar elevation, indicating the homogeneity of the surface and nearly isotropic reflection properties of the Salar. The effects of the high albedo surface on the erythemally effective irradiance, i.e. the UV index (UVI), has been experimentally determined by simultaneous measurements of several UV radiometers located at different sites around and on the Salar. In this context a method for the minimization of systematic deviations between the individual detectors used for the investigation is presented. It ensures the intercomparability of the performed UV measurements within ±2% which is a distinct improvement compared to the typical absolute accuracy of UV irradiance measurements in the order of ±5%. For solar elevations around 50 the UVI measured close to the center of the Salar is typically enhanced by 20% compared to the values determined outside. Towards lower solar elevations this increase becomes slightly weaker. The measurements agree well with both, own corresponding 1D and previously published 3D radiative transfer calculations from literature. 2006 Elsevier B.V. All rights reserved. Keywords: UV irradiance; UV index; Albedo; Model validation; Salar de Uyuni
1. Introduction There is wide evidence about the effects of excessive exposure to UV radiation on the human health [1,2]. Most important in this context are acute (e.g. sun burn, snow blindness) and chronic (e.g. skin cancer, eye cataracts) irritations or damages of human skin and eyes [3–5] as well as a potential suppression of the immune system [6,7]. People
*
Corresponding author. Tel.: +47 55588433. E-mail address: joachim.reuder@gfi.uib.no (J. Reuder).
1011-1344/$ - see front matter 2006 Elsevier B.V. All rights reserved. doi:10.1016/j.jphotobiol.2006.12.002
living in the tropics are especially at risk to these effects due to the naturally low levels of total ozone and the high solar elevations reached at those latitudes. Moreover the average UV exposure increases markedly with altitude (e.g. [8,9]). In addition, high surface albedo may reinforce the irradiance for both horizontal and inclined surfaces due to multiple scattering processes of reflected radiation. Bolivia, where this present study has been carried out, has approximately three million people living above 3000 m a.s.l., making it one of the most densely populated regions in the world at such high altitudes [10]. Hence, it is worthwhile to investigate the radiative properties of tropical
2
J. Reuder et al. / Journal of Photochemistry and Photobiology B: Biology 87 (2007) 1–8
regions located at high altitude that have extreme surface albedo values. Due to its high elevation, its high albedo over a large area and its homogeneous surface reflection properties, the Salar de Uyuni is an ideal natural laboratory to carry out experiments on albedo effects on UV radiation. Model simulations on the basis of three-dimensional radiative transfer calculations for the determination of the albedo effect on UV irradiance [11] indicate that the regional reflection properties of the surface within several tenth of kilometers around the measurement site should affect the UV irradiance level significantly. These investigations based on two different surface albedo values of 0.03 for vegetation and 0.85 for snow have been performed for various horizontal distribution patterns of high and low surface albedo areas in a model test region of 200 km · 200 km. The results for the UV wavelength indicate that the maximum increase of UV irradiance due to the albedo change from 0.03 to 0.85 will be close to 50% for 330 nm and around 35% for 300 nm. The magnitude and spectral course of the effect is in excellent agreement with published theoretical investigations [12]. The diameter of the high albedo area around a detector to produce 80% of this maximum effect ranges from 45 km at a wavelength of 330 nm to 12 km for 300 nm. Until now the effects of surface albedo inhomogeneities on UV radiation have been studied experimentally by measurements at the ice edge of Antarctica [13,14] and in glaciated areas in the Alps [15]. However these investigations encountered difficulties with respect to inhomogeneities in the terrain and access to the boundary between low and high surface albedo due to the polar environment. Changes in UV irradiance in comparable magnitude and of similar distance dependence can be expected in the surroundings of the Salar de Uyuni. In contrast to the model simulations mentioned above that have been performed at sea level, the Salar de Uyuni is located around 3700 m a.s.l. To compare measurements and 3D model simulations, additional sensitivity experiments have been conducted by using a 1D radiation transfer model. The basic data for the present study on albedo effects on erythemally effective UV irradiance have been obtained during a measurement campaign in May 2005, using several UV broadband instruments. The results from the present study might contribute to enrich our current knowledge in albedo and albedo effects and expected to be for the public benefit with respect to the welfare of the local population and the relevant tourist industry of the region. The measured UV irradiances and albedos will provide important data for the validation of radiative transfer model simulations and for the interpretation and evaluation of satellite-borne data. 2. Campaign and instrumentation The Salar de Uyuni is located in the Southern region of the Bolivian Altiplano between 19.7 and 20.7S and 66.9
and 68.3W at a mean altitude of 3700 m a.s.l. It covers an area of around 12,000 km2, with a maximum meridional and zonal extension of 120 km by 150 km. The actual size of the Salar may vary slightly due to seasonal changes at its outer borderline which is characterized by a variable mixture of mud, salt and water. Except for the rainy season (December–March), when a thin layer of water spreads across its surface, the Salar is covered homogeneously by salt. Although only few measurements of atmospheric parameters have been performed at the Salar [16,17], the atmosphere may be characterized as very unpolluted. The area is sparsely inhabited, with a few small settlements at the shoreline of the Salar. Uyuni, the nearest town with about 10,000 inhabitants, is located about 10 km outside the Salar to the southwest. Typical surface ozone values are around 20 ppb, a low value that can be attributed to the high concentrations of BrO radicals released by the salt surface of the Salar [17]. The total ozone content for the region can either be retrieved from the Total Ozone Mapping Spectrometer (TOMS) data or be estimated from Brewer measurements at the Bolivian Altiplano in La Paz (16.5S, 67.1W). From both sources the average values for May are around 250 DU. Consequently, a low aerosol optical depth, low levels of total and surface ozone, and a rather high effective albedo can be used for modelling purposes described in Section 3. From May 10 to May 16, 2005 a one week field campaign on the albedo effect has been performed in the area. The erythemally effective UV irradiance, represented by the UV index (UVI), has been measured by three broadband UV radiometers. Two instruments of the same type (UVS-E-T, one manufactured by Kipp & Zonen, serial number 001, one manufactured by SCINTEC, serial number 399) have been provided by the Meteorological Institute of the University of Munich (MIM). The third instrument (UVB-1, Yankee Environmental Systems, serial number 137) is owned by the Atmospheric Physics Laboratory of the University in La Paz (LFA). Another UVB-1 (serial number 138) has been running operationally at LFA. Additional measurements of the total ozone content by a MICROTOPS II ozonometer [18], and of the aerosol optical depth at four wavelength by a NOLL sunphotometer [19] have been routinely conducted during the campaign in the region of the Salar. Simultaneous UV measurements have been performed at various sites on and outside the Salar in different distances to its edge. At two half days the local albedo on the Salar has been determined by one UV instrument alternately directed upward and downward. In addition one UV radiometer has been mounted on the roof top of a four-wheel drive during two days to measure variations in UV irradiance while driving from the edge towards the center of the Salar. The distribution of the various UV instruments during the campaign is summarized in Table 1. The locations of UV measurements are shown in Fig. 1.
J. Reuder et al. / Journal of Photochemistry and Photobiology B: Biology 87 (2007) 1–8 Table 1 Overview of the location of UV measurements during the campaign in May 2005 at the Salar de Uyuni (m: morning, a: afternoon) Date
UV-S-E-T 001
UV-S-E-T 399
YES UVB-1
Total ozone (DU)
11.05 12.05 13.05 14.05 15.05 16.05
Uyuni Colchani Colchani (m) Jeep (a) Isla Incahuasi Jeep (m) Salar albedo (a) Salar albedo
Uyuni Colchani Colchani Colchani Uyuni Uyuni
Uyuni Uyuni Uyuni Uyuni Uyuni Uyuni
251 253 254 246 240 238
The given ozone values are daily averages of measurements taken by our MICROTOPS II ozonometer.
latitude [˚S]
-19.5
13.05.2005 15.05.2005
-20.0
Colchani
3
3.1. Radiometer intercomparison For the detection of the albedo effect on UV radiation mainly the two detectors of identical type have been used (Kipp & Zonen UV-S-E-T, serial number 001 and SCINTEC UV-S-E-T, serial number 399). Both detectors have been running parallel during several days, and at different locations before, during and after the campaign. These intercomparisons are summarized in Table 2. All one minute values of R, the ratio between the UVI measured by detector UV-S-E-T 399 (UVI399) and the UVI measured by detector UV-S-E-T 001 (UVI001) during these days have been plotted in Fig. 2. Both instruments coincide very well within the estimated absolute accuracy of broadband UV radiometers of ±5% [25]. However the ratio shows a marked dependence on solar elevation with values around 0.95 for low solar elevations and around 1.05 for 55, the maximum solar elevation reached at noon during the campaign. This systematic deviation between both instruments, mainly due to individual differences in spectral sensitivity and cosine response of the detectors, indicates an additional uncertainty factor in the determination of albedo effects by simultaneous measurements with two instruments at different sites. To minimize these
Isla Incahuasi Salt Hotel
-20.5
Uyuni
50 km
-68.0
-67.5
-67.0
longitude [˚W] Fig. 1. Map of the Salar de Uyuni with UV measurement sites (triangles) and tracks on the UV cross sections during the campaign in May 2005.
Table 2 Overview of the simultaneous and parallel measurements of broadband UV radiometers used in this study Station
Latitude Longitude Altitude n
Date
Munich (MIM) La Paz (LFA) Uyuni Colchani La Paz (LFA)
48.13N 16.54S 20.46S 20.30S 16.54S
23.04.–02.05.2005 07.05.–09.05.2005 11.05.2005 12.05.–13.05.2005 18.05.–20.05.2005
11.57O 68.07W 66.82W 66.93W 68.07W
530 m 3450 m 3700 m 3690 m 3450 m
10 3 1 2 3
n denotes the number of UV measurement days.
3. UV measurements 1.10 1.05 1.00
R
UV irradiances have been measured using broadband radiometers described in Section 2. These instruments are designed to coincide in their spectral response with the erythema action spectrum defined by CIE [20] to measure the erythemally effective UV irradiance, i.e. the UV index (UVI) [21,26]. Due to inevitable deviations of the detectors spectral response from the CIE definition, elaborate calibration processes and data processing are required to provide reliable UVI measurements. All radiometers have been calibrated accordingly, including detailed laboratory characterizations of the spectral response and the cosine dependency of the instruments, providing a calibration matrix depending on total ozone content and solar elevation [22,23]. In the following the measured intensity of the UV radiation will be described by the UVI. To facilitate the interpretation of the data, the UVI will be used here as a physical unit and therefore given with decimal places, although the authors are aware that this deviates from the recommendations of the WMO [24].
0.95 0.90 0.85 0
10
20
30
40
50
60
solar elevation [˚] Fig. 2. Ratio R between the UVI measured by the detector UV-E-S-T 399 and the UVI measured by the detector UV-E-S-T 001. The data are based on 1 min values during 9 days of parallel measurements with both instruments. The bold line represents a quadratic fit on the one minute data, the thin lines denote a ±2% interval around this fit.
4
J. Reuder et al. / Journal of Photochemistry and Photobiology B: Biology 87 (2007) 1–8
uncertainties as far as possible, the measured data have been corrected for the observed average deviation. A quadratic fit has been applied to all one minute values of R and is shown as bold line in Fig. 2 together with a ±2% interval around the fit. After application of this correction function, the UVI values derived by both instruments coincide better than ±2% for solar elevations above 15. An example of the improvement due to the method is presented in Fig. 3 for one day of instrument intercomparison at Uyuni. For May 11 the lower panel shows the time series of measured and corrected UVI values. In the upper panel the corresponding ratios of UVI399 to UVI001 can be found. Without correction the marked dependence from solar elevation described above can be seen clearly in the ratio. By application of the correction function derived from Fig. 2 this dependency largely disappears. All measurements collected by the UV-S-E-T 001 instrument during the campaign and used later for the determination of albedo effects on UVI have been corrected accordingly. 3.2. Albedo measurements Measurements of the local albedo at Salar de Uyuni have been performed close to the Salt Hotel at 20.333S and 67.046W, about 15 km inside the Salar, and 20 km to the east of the village of Colchani. One UV radiometer (UVS-E-T 001) and one pyranometer have been mounted on a stand about 2 m above the surface (see Fig. 4). Roughly every 15 min the orientation of both instruments has been switched between upward and downward direction by manually turning the extension arms of the mount. The levelling of the instrument has been controlled by the instruments water level measuring the downwelling radiation and by applying an external water level to the mount in case of
Fig. 4. Experimental setup for albedo determination on the Salar de Uyuni.
upwelling radiation measurement. The measurements have been carried out during afternoon of May 15, starting around 17:30 UTC, about 1 h after local noon, and in the morning of May 16, from around 12:30 until 18:00 UTC. The two measurement days have been merged to provide the diurnal course presented in Fig. 5. It shows the dark current corrected output signals of the detector UV-S-E-T 001 in V taken at a sampling rate of 10 s. The data points measured during the times of switching the instrument have been removed from the data set. During the afternoon of
3.0
2.5
1.05 ratio
1.00 0.95 14
0.90 399 001 uncorrected 001 corrected
12
UVI
10 8
signal UV-S-E-T 001 [V]
1.10 uncorrected corrected
downward 16.05. upward 16.05. downward 15.05. upward 15.05.
2.0
1.5
1.0
0.5
6 4
0.0
2
10
0 10
12
14
16
18
20
22
time [UTC] Fig. 3. Illustration of the effect of the applied correction function on comparability of the UVI determined by the radiometers UV-S-E-T 001 and 399 for May 11 at Uyuni.
12
14
16 18 time [UTC]
20
22
Fig. 5. Compilation of the detector signals of the UV-S-E-T instrument 001 recorded during May 15 and 16, 2005 on Salar de Uyuni. The direction of the detector has been changed manually at intervals of 10– 20 min. The grey lines denote the fit functions applied to the measured data for a continuous determination of albedo values given in Fig. 6.
J. Reuder et al. / Journal of Photochemistry and Photobiology B: Biology 87 (2007) 1–8
May 15 the sky was covered by 2–4 octas thin cirrus clouds which did not noticeably effect the UV radiation. The morning of May 16 started with a cloud cover of around 4 octas cirrus that increased to 6 octas towards noontime. In addition the cloud characteristics changed towards more optical dense cirrostratus fields. Their effects can clearly be associated with the increased variability in measured UV radiation between 16:00 and 18:00 UTC. Fit functions have been applied to the data sets of reflected upwelling and incoming downwelling UV radiation. The fitting procedure has been performed using the statistical software package TableCurve (SPSS Scientific Software). The beta peak function (TableCurve number 8055), resulting in the highest correlation coefficient, has been applied to the upward and downward directed radiation fluxes. Both fit curves are shown as grey lines in Fig. 5 together with the measurements. These fit functions have been used to calculate a continuous diurnal course of a broadband UV albedo value, valid for the erythemally effective UV irradiance, i.e. the UVI. The resulting time series of the local surface UVI albedo is presented in Fig. 6. From around 12:30 to 20:30 UTC, corresponding to solar elevations above 15, the albedo is quite stable at a value of 0.69 ± 0.02. It should be kept in mind that the albedo determination is based on the raw signal of the UV detector. Usually the cosine correction for broadband UV radiometers is done by applying a total ozone and solar elevation dependent calibration function on the raw data to transfer the output signal in volts into erythemally weighted UV irradiance. Due to missing information with respect to a calibration function for the barely diffuse upwelling radiation we decided to derive the albedo from the uncorrected raw signal of the detector. Therefore the slight diurnal variation with a weak maximum around local noon and especially the increase towards lower solar elevations should not be over-interpreted. Such small effects in the order of a few percent could be an artefact caused by deviations of the instrument from ideal cosine response dis-
0.85 0.80
albedo
0.75 0.70
5
cussed above and from a general decrease in absolute accuracy of UV radiation measurement towards low intensities during morning and evening. 3.3. UV radiation on inclined surfaces During 3 h of the albedo measurements (from around 15:00 to 18:00 UTC on May 16) the UV irradiance has also been measured for two additional sensor orientations. At every shift between upward and downward orientation of the UV sensor, it has been manually oriented first for around 1 min in a vertical position, pointing in azimuth towards the sun, and afterwards also for around one minute normally directed towards the sun. The latter position can be expected to give the highest possible UV irradiance value for the actual solar elevation and the actual atmospheric conditions. The maximum 10 s average value during the 1 min period was taken as representative for maximum UV irradiance on a surface normal to the solar position. For the vertically oriented detector the average measured value during the one minute of orientation has been selected. The 3 h measurement period covers a rather constant solar elevation range between 45 and 50. The UV irradiances measured on a vertical plane, oriented normal to the solar azimuth angle, reach values that are on average 12% (single values vary between 9% and 15%) higher than the UV irradiances with respect to the horizontal reference plane. This result indicates that the UV irradiance reflected by the salt surface distinctly exceeds the amount of diffuse UV irradiance of the sky for the altitude of Salar de Uyuni. The maximum UV irradiances measured with the detector oriented normally towards the sun show an average increase of 18% (15–21%). Due to the short measurement period and potential inaccuracies due to the manual positioning procedure, the data should of course not be over-interpreted. Nevertheless the data show interesting features, which encourage the future installation of ASCARATIS, an automatic measurement system for the determination of UV irradiances on inclined surfaces [27,28]. Corresponding long-term measurements, at least for one year, at one site on the Salar de Uyuni, will provide an unique data set on reflection properties of the Salar de Uyuni and resulting UV irradiances on arbitrarily oriented surfaces as a function of solar elevation and season. 3.4. Simultaneous measurements inside and outside the Salar
0.65 0.60 0.55 10
12
14
16
18
20
22
time [UTC] Fig. 6. Diurnal course of albedo for erythemally effective UV irradiance derived from measurements performed on Salar de Uyuni during May 15 and 16 (see also Fig. 5).
On May 14 simultaneous UVI measurements have been performed with the detector UV-S-E-T 399 at Colchani (about 5 km outside of the Salar) and with detector UVS-E-T 001 at Isla Incahuasi close to its center (see Fig. 1). Isla Incahuasi has an extension of around 0.75 km by 1 km and its highest elevation above the salt surface of the Salar is about 50 m. The UV radiometer has been installed on the northeasterly tip on a small ridge at an altitude of about 20 m above the salt surface. The
6
J. Reuder et al. / Journal of Photochemistry and Photobiology B: Biology 87 (2007) 1–8
0.9
1.2 ratio
1.0
0.8
0.8
2
4
6 8 10 12 14 16
18
20
22
0.7
14
surface albedo
12
UVI
10 8 6
0.6 0.5
2
4
6
20
8 10 12 14 16 18
0.4 0.3
4 Colchani Isla Incahuasi
2
0.2 2
0 12
14
16 18 time [UTC]
20
22
Fig. 7. Intercomparison of UVI measured at Colchani (UV-S-E-T 399, outside the Salar) and Isla Incahuasi (UV-S-E-T 001) close to the center of the Salar de Uyuni on May 14, 2005.
diurnal course of the UVI measured at both sites is presented in Fig. 7 together with the ratio of the measurements taken at Isla Incahuasi and Colchani. The data of the detector UV-S-E-T 001 have been corrected for the solar elevation dependent difference given in Fig. 2. Consequently the deviation of both data sets can be mainly attributed to the albedo effect. Most of the time the UV radiation level on the Salar at Isla Incahuasi is about 20% higher than at the station located outside, but close to its border line. However, the ratio shows a slight decrease towards lower solar elevation in the morning. Potential reasons for this behavior could be an azimuthal dependency of the UV instrument or a non-ideal diurnal course in UV irradiance measured at the site Colchani located only about 5 km outside to the East of the Salar, leading to differences in UV irradiance at comparable solar elevation between morning and evening. Unfortunately this assumption could not be verified by evening measurements because the data sampling had to be stopped around 19:00 UTC for logistic reasons. 4. Model simulations The model simulations of albedo effects on UVI have been performed by the radiation transfer model STAR [29,30]. Fig. 8 presents a sensitivity study on the UVI as function of solar elevation and surface albedo for average conditions of the Bolivian Altiplano. Calculations have been performed for an altitude of 3700 m a.s.l. and a corresponding pressure of 650 hPa. A total ozone content of 250 DU, representing average Altiplano conditions, has been used. Furthermore the aerosol type ‘clean continental’ [31] has been selected as model input. Corresponding to measurements with the NOLL sunphotometer an aerosol
0
6
8 10 12 14 16
18
20
40
60
80
solar elevation [˚] Fig. 8. Modelled UVI as a function of solar elevation and surface albedo. Total ozone content: 250 DU; pressure 650 hPa; altitude 3700 m; aerosol optical depth at 550 nm: 0.05.
optical depth of 0.05 at 550 nm has been chosen for the simulations. In general the UVI strongly increases with increasing solar elevation due to reduced optical pathlength through the atmosphere and therefore reduced atmospheric absorption. There is also a marked increase of the UVI with enhanced surface reflectivity. Multiple scattering processes, mainly at air molecules, redirect a portion of the reflected upward directed radiation towards the ground, thus also increasing the downward directed UV irradiance. Fig. 9 shows the percentage effect of an increase of ground albedo on erythemally effective UV irradiance represented by the UVI. Two basic scenarios have been inves50 40 increase in UVI [%]
10
4
0.1
30 20 albedo increase from 0.03 to 0.85, 3700 m a.s.l. albedo increase from 0.2 to 0.7, 3700 m a.s.l. albedo increase from 0.03 to 0.85, sea level albedo increase from 0.2 to 0.7, sea level
10 0 10
20
30
40
50
60
70
80
90
solar elevation [˚] Fig. 9. Percentage increase in UVI due to enhancement of surface albedo as a function of solar elevation derived from model calculations. Total ozone content: 250 DU; pressure 650 hPa; altitude 3700 m; aerosol optical depth at 550 nm: 0.05.
J. Reuder et al. / Journal of Photochemistry and Photobiology B: Biology 87 (2007) 1–8
tigated. The first is associated with an increase in UV surface albedo from 0.03 that is typical for vegetated surfaces [32,33] to 0.85 a value representative for fresh snow away from polar and high mountain regions [34,35]. This scenario has been chosen for comparison with the results obtained by [11]. The second describes an albedo increase from 0.2 to 0.7. These values are expected to fit the albedo conditions around and on the Salar de Uyuni best. While own albedo measurements have been performed on the Salar (see Fig. 6) the lower value of 0.2 has been estimated from literature values (see e.g. compilation by [36]) for the dry, partly sandy and rocky surroundings of the salar nearly without any vegetation. Both scenarios have been modelled for sea level and the typical altitude of the Bolivian Altiplano in the region of the Salar de Uyuni, at 3700 m a.s.l., resulting in the four curves presented in Fig. 9. The increase of UV irradiance due to albedo effects shows only a weak variation with solar elevation. A slight maximum occurs around 20. The albedo effect at sea level (open symbols) is markedly higher than at 3700 m a.s.l. (filled symbols), representing the strong dependence of multiple scattering processes on air density. An increase in albedo from 0.03 to 0.85 results in an increase of the UVI of around 45% at sea level and of around 25% at the altitude of the Bolivian Altiplano. The results for the UVI at sea level conditions agree with the estimated effect derived by three-dimensional radiative transfer calculations [11] that predict a wavelength dependent range of 35% (at 300 nm) to 50% (at 330 nm). For comparison it should be taken into account that the central wavelength for the spectrally integrated UVI lies in the region between 305 nm and 315 nm, depending on the solar elevation and the extinction properties of the atmosphere. 5. Summary and outlook The effect of a high albedo surface on erythemally effective UV irradiance, i.e. the UVI, has been investigated by a combination of measurements and model calculations. A central part of the investigation has been a one week measurement campaign at the Salar de Uyuni. The erythemally effective UV irradiance has been measured by three broadband UV radiometers simultaneously at various sites on and outside the Salar at different distances to its edge. For an improved intercomparability of the individual UV radiometers, a new correction method has been developed and applied. It increases the relative accuracy of two UV radiometers to ±2% which is a distinct improvement compared to the typical absolute accuracy of UV irradiance measurements in the order of ±5%. Due to multiple scattering of UV radiation reflected by the high albedo surface, the UVI measurements taken on the Salar show a distinct enhancement compared to areas outside its limits. At Isla Incahuasi in the central part of the Salar, the radiation increase reaches values of the order of 20% for an solar elevation of 50. The measurements also indicate a moderate reduction of the enhancement
7
with decreasing solar height. There is an excellent agreement between the measured increase in UVI at solar elevations between 25 and 50, the highest solar elevation occurring during the campaign, and 1D radiation transfer calculations for an increase in surface albedo from 0.2 to 0.7. However, in contrast to the measurements, the model simulations predict a weak maximum around 20 and no distinct decrease towards lower solar elevation. One reason for this lack of agreement could be a general decrease in accuracy of UV measurements towards lower solar elevation and a corresponding increase in individual systematic differences between the used radiometers [22,25]. In addition there could be a solar elevation dependent variation in the contribution of local albedo, i.e. in the direct vicinity of the radiometer (rocks, bare soil and sparse vegetation), and regional albedo, determined by the surface properties in an area of a few tenth of kilometers around the measuring site (homogeneous salt surface). A detailed study of this problem can only be addressed by extended measurements in the future. In this context a detailed interpretation of the UV data measured from roof top of a four-wheel drive during two transects across the Salar will be of particular interest. These results will be presented in a follow-on paper. To our knowledge we have presented here the first direct measurements of UV surface albedo of the Salar de Uyuni. The reflectivity for erythemally effective radiation has been determined as 0.69 ± 0.02. Hardly any dependency on solar elevation has been found, indicating the homogeneity of the salt surface and therefore the presence of nearly isotropic reflection properties on the Salar. Together with its size, the rather constant atmospheric conditions with respect to total ozone content and low aerosol load, and the low average cloud cover, these nearly ideal reflection properties make the Salar an unique site for calibration purposes of space-borne radiometers in satellite remote sensing [16]. During the albedo measurements UV irradiances on tilted surfaces have also been determined for solar elevations between 45 and 50. The UV irradiances measured on a vertical plane, oriented normal to the solar azimuth angle, have shown values that are on average 12% higher than the UV irradiances with respect to the horizontal reference plane. The maximum UV irradiances measured with the detector oriented normally towards the sun have given an average increase of 18%. Due to the rather short measurement period and potential inaccuracies due to the manual positioning procedure, the data should of course not be over-interpreted. Nevertheless the data show interesting features, which encourage the future installation of ASCARATIS [27,28], an automatic measurement system for the determination of UV irradiances on inclined surfaces. Acknowledgements The authors are very grateful to Mr. Gonzalo Gutie´rrez for experimental help. We acknowledge the Volkswagen Foundation grant support which enabled this collaborative project.
8
J. Reuder et al. / Journal of Photochemistry and Photobiology B: Biology 87 (2007) 1–8
References [1] WHO, Global solar UV Index – A practical guide, A joint recommendation of WHO, WMO, UNEP, and ICNIRP, 2002. [2] J. Longstreth, F.R.d. Gruil, M.L. Kripke, S. Abseck, F. Arnold, H.I. Slaper, G. Velders, Y. Takizawa, J.C.v.d. Leun, Health risks, J. Photochem. Photobiol. B 46 (1–3) (1998) 20–39. [3] B.K. Armstrong, A. Kricker, The epidemiology of UV induced skin cancer, J. Photochem. Photobiol. B 63 (2001) 8–18. [4] H.S. Black, F.R.d. Gruil, P.D. Forbes, J.E. Cleaver, H.N. Ananthaswamy, E.C. deFabo, S.E. Ullrich, R.M. Tyrell, Photocarcinogenesis: an overview, J. Photochem. Photobiol. B 40 (1997) 29–47. [5] J.E. Roberts, Ocular phototoxicity, J. Photochem. Photobiol. B 64 (2001) 136–143. [6] M. Norval, Effects of solar radiation on the human immune system, J. Photochem. Photobiol. B 63 (2001) 28–40. [7] D.X. Nghiem, N. Kazimi, G. Clydesdale, H.N. Ananthaswamy, M.L. Kripke, S.E. Ullrich, Ultraviolet A radiation suppresses an established immune response: implications for sunscreen design, J. Invest. Dermatol. 117 (2001) 1193–1199. [8] M. Pfeifer, P. Koepke, J. Reuder, Effects of altitude and aerosols on UV radiation, J. Geophys. Res. 111 (2006) D01203, doi:10.1029/ 2005JD006444. [9] F. Zaratti, R. Forno, J. Garcia-Fuentes, M. Andrade, Erythemally weighted UV variations at two altitude locations, J. Geophys. Res. (2003), doi:10.1029/2001JD000918. [10] INE, National Institute of Statistics, Bolivian census, 2001. [11] M. Deguenther, R. Meerkoetter, A. Albold, G. Seckmeyer, Case study on the influence of inhomogeneous surface albedo on UV irradiance, Geophys. Res. Lett. 25 (19) (1998) 3587–3590. [12] J. Lenoble, Modeling of the influence of snow reflectance on ultraviolet irradiance for cloudless sky, Appl. Opt. 37 (12) (1998) 2441–2447. [13] I. Smolskaia, M. Nunez, K. Michael, Measurements of erythemal irradiance near Davies Station, Antarctica: effect of inhomogeneous surface albedo, Geophys. Res. Lett. 26 (10) (1999) 1381–1384. [14] S. Wuttke, G. Seckmeyer, G. Ko¨nig-Langlo, Measurements of spectral snow albedo at Neumayer, Antarctica, Ann. Geophys. 24 (2006) 7–21. [15] M. Huber, M. Blumthaler, J. Schreder, B. Schallhart, J. Lenoble, Effect of inhomogeneous surface albedo on diffuse UV sky radiance at a high-altitude site, J. Geophys. Res. 109 (2004) D08107, doi:10.1029/ 2003JD004013. [16] R.A.C. Lamparelli, F.J. Ponzoni, J. Zullo Jr., G.Q. Pellegrino, Y. Arnaud, Characterization of the Salar de Uyuni for in-orbit satellite calibration, IEEE Trans. Geosci. Remote Sens. 41 (6) (2003) 1461– 1468. [17] G. Honninger, N. Pobrowski, E.R. Palenque, R. Torrez, U. Platt, Reactive bromine and sulfur emissions at Salar de Uyuni, Bolivia, Geophys. Res. Lett. 31 (2004) L04101. [18] M. Morys, F.M. Mims III, S. Hagerup, S. Anderson, A. Baker, J. Kia, T. Walkup, Design, calibration and performance of MICROTOPS II handheld ozone monitor and Sun photometer, J. Geophys. Res. 106 (D13) (2001) 14573–14582.
[19] G.A. d’Almeida, R. Jaenicke, P. Roggendorf, D. Richter, New sunphotometer for network operation, Appl. Opt. 22 (23) (1983) 3796–3801. [20] A. McKinley, B. Diffey, A reference spectrum for ultraviolet induced erythema in human skin, Commission Int. Eclairage (CIE) 6 (1987) 17–22. [21] WMO, Report of the WMO meeting of experts on UV-B measurements, data quality and standardization of UV indices, WMO Technical Report No. 95, 1994. [22] A. Oppenrieder, P. Hoeppe, P. Koepke, J. Reuder, J. Schween, J. Schreder, Simplified calibration for broadband solar ultraviolet radiation measurements, Photochem. Photobiol. 78 (6) (2003) 603– 606. [23] B.A. Bodhaine, E.G. Dutton, R.L. McKenzie, P.V. Johnston, Calibrating broadband UV instruments: ozone and solar zenith angle dependence, J. Atmos. Ocean. Technol. 15 (1998) 916–926. [24] WMO, Report of the WMO–WHO meeting of experts on standardization of UV indices and their dissimination to the public, WMO Technical Report No. 143, 1997. [25] WMO, Report of the LAP/COST/WMO intercomparison of erythemal radiometers (Thessaloniki, Greece, 13–23 September 1999), WMO Technical Report No. 141, 2001. [26] WMO, Instruments to measure solar ultraviolet radiation, part II: broadband instruments measuring erythemally weighted solar irradiance, WMO-GAW Report No. 164, 2005. [27] A. Oppenrieder, P. Hoeppe, P. Koepke, Routine measurement of erythemally effective UV irradiance on inclined surfaces, J. Photochem. Photobiol. B 78 (6) (2004) 85–94. [28] A. Oppenrieder, P. Hoeppe, P. Koepke, J. Reuder, Long term measurements of the UV irradiance on inclined surfaces and visualization of UV exposure of the human body, Meteorol. Z. 14 (2) (2005) 285–290. [29] A. Ruggaber, R. Dlugi, T. Nakajima, Modelling radiation quantities and photolysis frequencies in the troposphere, J. Atmos. Chem. 18 (1994) 171–210. [30] H. Schwander, P. Koepke, A. Kaifel, G. Seckmeyer, Modification of spectral UV irradiance by clouds, J. Geophys. Res. 107 (D16) 4296. doi:10.1029/2001JD001297. [31] M. Hess, P. Koepke, I. Schult, Optical properties of aerosols and clouds: the software package OPAC, Bull. Am. Met. Soc. 79 (1998) 831–844. [32] M. Blumthaler, W. Ambach, Solar UVB-albedo of various surfaces, Photochem. Photobiol. 48 (1) (1988) 85–88. [33] R.L. McKenzie, M. Kotkamp, Upwelling UV spectral irradiances and surface albedo measurements at Lauder, New Zealand, Geophys. Res. Lett. 23 (14) (1996) 1757–1760. [34] T.C. Grenfell, S.G. Warren, P.C. Mullen, Reflection of solar radiation by the Antarctic snow surface at ultraviolet, visible, and near infrared wavelength, J. Geophys. Res. 99 (1994) 18669–18684. [35] S.G. Warren, W.J. Wiscombe, A model for the spectral albedo of snow. II: snow containing atmospheric aerosols, J. Atmos. Sci. 37 (1980) 2734–2745. [36] P. Koepke, J. Reuder, H. Schwander, Solar UV radiation and its variability due to the atmospheric components, Recent Res. Devel. Photochem. Photobiol. 6 (2002) 11–34.
Investigations on the effect of high surface albedo on erythemally effective UV irradiance: Results of a campaign at the Salar de Uyuni, Bolivia Joachim Reuder
b
a,*
, Flavio Ghezzi b, Eduardo Palenque b, Rene Torrez b, Marco Andrade b, Francesco Zaratti b
a University of Bergen, Geophysical Institute, Alle´gaten 70, N-5007 Bergen, Norway Universidad Mayor de San Andre´s, Laboratorio de Fı´sica de la Atmo´sfera, Campus Cota Cota, La Paz, Bolivia
Received 23 October 2006; received in revised form 1 December 2006; accepted 1 December 2006 Available online 16 January 2007
Abstract Measurements and model calculations have been performed to study the effect of high surface albedo on erythemally effective UV irradiance. A central part of the investigation has been a one week measurement campaign at Salar de Uyuni in the Southern part of the Bolivian Altiplano. The Salar de Uyuni, the largest salt lake of the world, is characterized by largely homogeneous surface conditions during most of the year. Albedo measurements performed by an UV radiometer result in a reflectivity for erythemally effective radiation of 0.69 ± 0.02. The measurements show hardly any dependency on solar elevation, indicating the homogeneity of the surface and nearly isotropic reflection properties of the Salar. The effects of the high albedo surface on the erythemally effective irradiance, i.e. the UV index (UVI), has been experimentally determined by simultaneous measurements of several UV radiometers located at different sites around and on the Salar. In this context a method for the minimization of systematic deviations between the individual detectors used for the investigation is presented. It ensures the intercomparability of the performed UV measurements within ±2% which is a distinct improvement compared to the typical absolute accuracy of UV irradiance measurements in the order of ±5%. For solar elevations around 50 the UVI measured close to the center of the Salar is typically enhanced by 20% compared to the values determined outside. Towards lower solar elevations this increase becomes slightly weaker. The measurements agree well with both, own corresponding 1D and previously published 3D radiative transfer calculations from literature. 2006 Elsevier B.V. All rights reserved. Keywords: UV irradiance; UV index; Albedo; Model validation; Salar de Uyuni
1. Introduction There is wide evidence about the effects of excessive exposure to UV radiation on the human health [1,2]. Most important in this context are acute (e.g. sun burn, snow blindness) and chronic (e.g. skin cancer, eye cataracts) irritations or damages of human skin and eyes [3–5] as well as a potential suppression of the immune system [6,7]. People
*
Corresponding author. Tel.: +47 55588433. E-mail address: joachim.reuder@gfi.uib.no (J. Reuder).
1011-1344/$ - see front matter 2006 Elsevier B.V. All rights reserved. doi:10.1016/j.jphotobiol.2006.12.002
living in the tropics are especially at risk to these effects due to the naturally low levels of total ozone and the high solar elevations reached at those latitudes. Moreover the average UV exposure increases markedly with altitude (e.g. [8,9]). In addition, high surface albedo may reinforce the irradiance for both horizontal and inclined surfaces due to multiple scattering processes of reflected radiation. Bolivia, where this present study has been carried out, has approximately three million people living above 3000 m a.s.l., making it one of the most densely populated regions in the world at such high altitudes [10]. Hence, it is worthwhile to investigate the radiative properties of tropical
2
J. Reuder et al. / Journal of Photochemistry and Photobiology B: Biology 87 (2007) 1–8
regions located at high altitude that have extreme surface albedo values. Due to its high elevation, its high albedo over a large area and its homogeneous surface reflection properties, the Salar de Uyuni is an ideal natural laboratory to carry out experiments on albedo effects on UV radiation. Model simulations on the basis of three-dimensional radiative transfer calculations for the determination of the albedo effect on UV irradiance [11] indicate that the regional reflection properties of the surface within several tenth of kilometers around the measurement site should affect the UV irradiance level significantly. These investigations based on two different surface albedo values of 0.03 for vegetation and 0.85 for snow have been performed for various horizontal distribution patterns of high and low surface albedo areas in a model test region of 200 km · 200 km. The results for the UV wavelength indicate that the maximum increase of UV irradiance due to the albedo change from 0.03 to 0.85 will be close to 50% for 330 nm and around 35% for 300 nm. The magnitude and spectral course of the effect is in excellent agreement with published theoretical investigations [12]. The diameter of the high albedo area around a detector to produce 80% of this maximum effect ranges from 45 km at a wavelength of 330 nm to 12 km for 300 nm. Until now the effects of surface albedo inhomogeneities on UV radiation have been studied experimentally by measurements at the ice edge of Antarctica [13,14] and in glaciated areas in the Alps [15]. However these investigations encountered difficulties with respect to inhomogeneities in the terrain and access to the boundary between low and high surface albedo due to the polar environment. Changes in UV irradiance in comparable magnitude and of similar distance dependence can be expected in the surroundings of the Salar de Uyuni. In contrast to the model simulations mentioned above that have been performed at sea level, the Salar de Uyuni is located around 3700 m a.s.l. To compare measurements and 3D model simulations, additional sensitivity experiments have been conducted by using a 1D radiation transfer model. The basic data for the present study on albedo effects on erythemally effective UV irradiance have been obtained during a measurement campaign in May 2005, using several UV broadband instruments. The results from the present study might contribute to enrich our current knowledge in albedo and albedo effects and expected to be for the public benefit with respect to the welfare of the local population and the relevant tourist industry of the region. The measured UV irradiances and albedos will provide important data for the validation of radiative transfer model simulations and for the interpretation and evaluation of satellite-borne data. 2. Campaign and instrumentation The Salar de Uyuni is located in the Southern region of the Bolivian Altiplano between 19.7 and 20.7S and 66.9
and 68.3W at a mean altitude of 3700 m a.s.l. It covers an area of around 12,000 km2, with a maximum meridional and zonal extension of 120 km by 150 km. The actual size of the Salar may vary slightly due to seasonal changes at its outer borderline which is characterized by a variable mixture of mud, salt and water. Except for the rainy season (December–March), when a thin layer of water spreads across its surface, the Salar is covered homogeneously by salt. Although only few measurements of atmospheric parameters have been performed at the Salar [16,17], the atmosphere may be characterized as very unpolluted. The area is sparsely inhabited, with a few small settlements at the shoreline of the Salar. Uyuni, the nearest town with about 10,000 inhabitants, is located about 10 km outside the Salar to the southwest. Typical surface ozone values are around 20 ppb, a low value that can be attributed to the high concentrations of BrO radicals released by the salt surface of the Salar [17]. The total ozone content for the region can either be retrieved from the Total Ozone Mapping Spectrometer (TOMS) data or be estimated from Brewer measurements at the Bolivian Altiplano in La Paz (16.5S, 67.1W). From both sources the average values for May are around 250 DU. Consequently, a low aerosol optical depth, low levels of total and surface ozone, and a rather high effective albedo can be used for modelling purposes described in Section 3. From May 10 to May 16, 2005 a one week field campaign on the albedo effect has been performed in the area. The erythemally effective UV irradiance, represented by the UV index (UVI), has been measured by three broadband UV radiometers. Two instruments of the same type (UVS-E-T, one manufactured by Kipp & Zonen, serial number 001, one manufactured by SCINTEC, serial number 399) have been provided by the Meteorological Institute of the University of Munich (MIM). The third instrument (UVB-1, Yankee Environmental Systems, serial number 137) is owned by the Atmospheric Physics Laboratory of the University in La Paz (LFA). Another UVB-1 (serial number 138) has been running operationally at LFA. Additional measurements of the total ozone content by a MICROTOPS II ozonometer [18], and of the aerosol optical depth at four wavelength by a NOLL sunphotometer [19] have been routinely conducted during the campaign in the region of the Salar. Simultaneous UV measurements have been performed at various sites on and outside the Salar in different distances to its edge. At two half days the local albedo on the Salar has been determined by one UV instrument alternately directed upward and downward. In addition one UV radiometer has been mounted on the roof top of a four-wheel drive during two days to measure variations in UV irradiance while driving from the edge towards the center of the Salar. The distribution of the various UV instruments during the campaign is summarized in Table 1. The locations of UV measurements are shown in Fig. 1.
J. Reuder et al. / Journal of Photochemistry and Photobiology B: Biology 87 (2007) 1–8 Table 1 Overview of the location of UV measurements during the campaign in May 2005 at the Salar de Uyuni (m: morning, a: afternoon) Date
UV-S-E-T 001
UV-S-E-T 399
YES UVB-1
Total ozone (DU)
11.05 12.05 13.05 14.05 15.05 16.05
Uyuni Colchani Colchani (m) Jeep (a) Isla Incahuasi Jeep (m) Salar albedo (a) Salar albedo
Uyuni Colchani Colchani Colchani Uyuni Uyuni
Uyuni Uyuni Uyuni Uyuni Uyuni Uyuni
251 253 254 246 240 238
The given ozone values are daily averages of measurements taken by our MICROTOPS II ozonometer.
latitude [˚S]
-19.5
13.05.2005 15.05.2005
-20.0
Colchani
3
3.1. Radiometer intercomparison For the detection of the albedo effect on UV radiation mainly the two detectors of identical type have been used (Kipp & Zonen UV-S-E-T, serial number 001 and SCINTEC UV-S-E-T, serial number 399). Both detectors have been running parallel during several days, and at different locations before, during and after the campaign. These intercomparisons are summarized in Table 2. All one minute values of R, the ratio between the UVI measured by detector UV-S-E-T 399 (UVI399) and the UVI measured by detector UV-S-E-T 001 (UVI001) during these days have been plotted in Fig. 2. Both instruments coincide very well within the estimated absolute accuracy of broadband UV radiometers of ±5% [25]. However the ratio shows a marked dependence on solar elevation with values around 0.95 for low solar elevations and around 1.05 for 55, the maximum solar elevation reached at noon during the campaign. This systematic deviation between both instruments, mainly due to individual differences in spectral sensitivity and cosine response of the detectors, indicates an additional uncertainty factor in the determination of albedo effects by simultaneous measurements with two instruments at different sites. To minimize these
Isla Incahuasi Salt Hotel
-20.5
Uyuni
50 km
-68.0
-67.5
-67.0
longitude [˚W] Fig. 1. Map of the Salar de Uyuni with UV measurement sites (triangles) and tracks on the UV cross sections during the campaign in May 2005.
Table 2 Overview of the simultaneous and parallel measurements of broadband UV radiometers used in this study Station
Latitude Longitude Altitude n
Date
Munich (MIM) La Paz (LFA) Uyuni Colchani La Paz (LFA)
48.13N 16.54S 20.46S 20.30S 16.54S
23.04.–02.05.2005 07.05.–09.05.2005 11.05.2005 12.05.–13.05.2005 18.05.–20.05.2005
11.57O 68.07W 66.82W 66.93W 68.07W
530 m 3450 m 3700 m 3690 m 3450 m
10 3 1 2 3
n denotes the number of UV measurement days.
3. UV measurements 1.10 1.05 1.00
R
UV irradiances have been measured using broadband radiometers described in Section 2. These instruments are designed to coincide in their spectral response with the erythema action spectrum defined by CIE [20] to measure the erythemally effective UV irradiance, i.e. the UV index (UVI) [21,26]. Due to inevitable deviations of the detectors spectral response from the CIE definition, elaborate calibration processes and data processing are required to provide reliable UVI measurements. All radiometers have been calibrated accordingly, including detailed laboratory characterizations of the spectral response and the cosine dependency of the instruments, providing a calibration matrix depending on total ozone content and solar elevation [22,23]. In the following the measured intensity of the UV radiation will be described by the UVI. To facilitate the interpretation of the data, the UVI will be used here as a physical unit and therefore given with decimal places, although the authors are aware that this deviates from the recommendations of the WMO [24].
0.95 0.90 0.85 0
10
20
30
40
50
60
solar elevation [˚] Fig. 2. Ratio R between the UVI measured by the detector UV-E-S-T 399 and the UVI measured by the detector UV-E-S-T 001. The data are based on 1 min values during 9 days of parallel measurements with both instruments. The bold line represents a quadratic fit on the one minute data, the thin lines denote a ±2% interval around this fit.
4
J. Reuder et al. / Journal of Photochemistry and Photobiology B: Biology 87 (2007) 1–8
uncertainties as far as possible, the measured data have been corrected for the observed average deviation. A quadratic fit has been applied to all one minute values of R and is shown as bold line in Fig. 2 together with a ±2% interval around the fit. After application of this correction function, the UVI values derived by both instruments coincide better than ±2% for solar elevations above 15. An example of the improvement due to the method is presented in Fig. 3 for one day of instrument intercomparison at Uyuni. For May 11 the lower panel shows the time series of measured and corrected UVI values. In the upper panel the corresponding ratios of UVI399 to UVI001 can be found. Without correction the marked dependence from solar elevation described above can be seen clearly in the ratio. By application of the correction function derived from Fig. 2 this dependency largely disappears. All measurements collected by the UV-S-E-T 001 instrument during the campaign and used later for the determination of albedo effects on UVI have been corrected accordingly. 3.2. Albedo measurements Measurements of the local albedo at Salar de Uyuni have been performed close to the Salt Hotel at 20.333S and 67.046W, about 15 km inside the Salar, and 20 km to the east of the village of Colchani. One UV radiometer (UVS-E-T 001) and one pyranometer have been mounted on a stand about 2 m above the surface (see Fig. 4). Roughly every 15 min the orientation of both instruments has been switched between upward and downward direction by manually turning the extension arms of the mount. The levelling of the instrument has been controlled by the instruments water level measuring the downwelling radiation and by applying an external water level to the mount in case of
Fig. 4. Experimental setup for albedo determination on the Salar de Uyuni.
upwelling radiation measurement. The measurements have been carried out during afternoon of May 15, starting around 17:30 UTC, about 1 h after local noon, and in the morning of May 16, from around 12:30 until 18:00 UTC. The two measurement days have been merged to provide the diurnal course presented in Fig. 5. It shows the dark current corrected output signals of the detector UV-S-E-T 001 in V taken at a sampling rate of 10 s. The data points measured during the times of switching the instrument have been removed from the data set. During the afternoon of
3.0
2.5
1.05 ratio
1.00 0.95 14
0.90 399 001 uncorrected 001 corrected
12
UVI
10 8
signal UV-S-E-T 001 [V]
1.10 uncorrected corrected
downward 16.05. upward 16.05. downward 15.05. upward 15.05.
2.0
1.5
1.0
0.5
6 4
0.0
2
10
0 10
12
14
16
18
20
22
time [UTC] Fig. 3. Illustration of the effect of the applied correction function on comparability of the UVI determined by the radiometers UV-S-E-T 001 and 399 for May 11 at Uyuni.
12
14
16 18 time [UTC]
20
22
Fig. 5. Compilation of the detector signals of the UV-S-E-T instrument 001 recorded during May 15 and 16, 2005 on Salar de Uyuni. The direction of the detector has been changed manually at intervals of 10– 20 min. The grey lines denote the fit functions applied to the measured data for a continuous determination of albedo values given in Fig. 6.
J. Reuder et al. / Journal of Photochemistry and Photobiology B: Biology 87 (2007) 1–8
May 15 the sky was covered by 2–4 octas thin cirrus clouds which did not noticeably effect the UV radiation. The morning of May 16 started with a cloud cover of around 4 octas cirrus that increased to 6 octas towards noontime. In addition the cloud characteristics changed towards more optical dense cirrostratus fields. Their effects can clearly be associated with the increased variability in measured UV radiation between 16:00 and 18:00 UTC. Fit functions have been applied to the data sets of reflected upwelling and incoming downwelling UV radiation. The fitting procedure has been performed using the statistical software package TableCurve (SPSS Scientific Software). The beta peak function (TableCurve number 8055), resulting in the highest correlation coefficient, has been applied to the upward and downward directed radiation fluxes. Both fit curves are shown as grey lines in Fig. 5 together with the measurements. These fit functions have been used to calculate a continuous diurnal course of a broadband UV albedo value, valid for the erythemally effective UV irradiance, i.e. the UVI. The resulting time series of the local surface UVI albedo is presented in Fig. 6. From around 12:30 to 20:30 UTC, corresponding to solar elevations above 15, the albedo is quite stable at a value of 0.69 ± 0.02. It should be kept in mind that the albedo determination is based on the raw signal of the UV detector. Usually the cosine correction for broadband UV radiometers is done by applying a total ozone and solar elevation dependent calibration function on the raw data to transfer the output signal in volts into erythemally weighted UV irradiance. Due to missing information with respect to a calibration function for the barely diffuse upwelling radiation we decided to derive the albedo from the uncorrected raw signal of the detector. Therefore the slight diurnal variation with a weak maximum around local noon and especially the increase towards lower solar elevations should not be over-interpreted. Such small effects in the order of a few percent could be an artefact caused by deviations of the instrument from ideal cosine response dis-
0.85 0.80
albedo
0.75 0.70
5
cussed above and from a general decrease in absolute accuracy of UV radiation measurement towards low intensities during morning and evening. 3.3. UV radiation on inclined surfaces During 3 h of the albedo measurements (from around 15:00 to 18:00 UTC on May 16) the UV irradiance has also been measured for two additional sensor orientations. At every shift between upward and downward orientation of the UV sensor, it has been manually oriented first for around 1 min in a vertical position, pointing in azimuth towards the sun, and afterwards also for around one minute normally directed towards the sun. The latter position can be expected to give the highest possible UV irradiance value for the actual solar elevation and the actual atmospheric conditions. The maximum 10 s average value during the 1 min period was taken as representative for maximum UV irradiance on a surface normal to the solar position. For the vertically oriented detector the average measured value during the one minute of orientation has been selected. The 3 h measurement period covers a rather constant solar elevation range between 45 and 50. The UV irradiances measured on a vertical plane, oriented normal to the solar azimuth angle, reach values that are on average 12% (single values vary between 9% and 15%) higher than the UV irradiances with respect to the horizontal reference plane. This result indicates that the UV irradiance reflected by the salt surface distinctly exceeds the amount of diffuse UV irradiance of the sky for the altitude of Salar de Uyuni. The maximum UV irradiances measured with the detector oriented normally towards the sun show an average increase of 18% (15–21%). Due to the short measurement period and potential inaccuracies due to the manual positioning procedure, the data should of course not be over-interpreted. Nevertheless the data show interesting features, which encourage the future installation of ASCARATIS, an automatic measurement system for the determination of UV irradiances on inclined surfaces [27,28]. Corresponding long-term measurements, at least for one year, at one site on the Salar de Uyuni, will provide an unique data set on reflection properties of the Salar de Uyuni and resulting UV irradiances on arbitrarily oriented surfaces as a function of solar elevation and season. 3.4. Simultaneous measurements inside and outside the Salar
0.65 0.60 0.55 10
12
14
16
18
20
22
time [UTC] Fig. 6. Diurnal course of albedo for erythemally effective UV irradiance derived from measurements performed on Salar de Uyuni during May 15 and 16 (see also Fig. 5).
On May 14 simultaneous UVI measurements have been performed with the detector UV-S-E-T 399 at Colchani (about 5 km outside of the Salar) and with detector UVS-E-T 001 at Isla Incahuasi close to its center (see Fig. 1). Isla Incahuasi has an extension of around 0.75 km by 1 km and its highest elevation above the salt surface of the Salar is about 50 m. The UV radiometer has been installed on the northeasterly tip on a small ridge at an altitude of about 20 m above the salt surface. The
6
J. Reuder et al. / Journal of Photochemistry and Photobiology B: Biology 87 (2007) 1–8
0.9
1.2 ratio
1.0
0.8
0.8
2
4
6 8 10 12 14 16
18
20
22
0.7
14
surface albedo
12
UVI
10 8 6
0.6 0.5
2
4
6
20
8 10 12 14 16 18
0.4 0.3
4 Colchani Isla Incahuasi
2
0.2 2
0 12
14
16 18 time [UTC]
20
22
Fig. 7. Intercomparison of UVI measured at Colchani (UV-S-E-T 399, outside the Salar) and Isla Incahuasi (UV-S-E-T 001) close to the center of the Salar de Uyuni on May 14, 2005.
diurnal course of the UVI measured at both sites is presented in Fig. 7 together with the ratio of the measurements taken at Isla Incahuasi and Colchani. The data of the detector UV-S-E-T 001 have been corrected for the solar elevation dependent difference given in Fig. 2. Consequently the deviation of both data sets can be mainly attributed to the albedo effect. Most of the time the UV radiation level on the Salar at Isla Incahuasi is about 20% higher than at the station located outside, but close to its border line. However, the ratio shows a slight decrease towards lower solar elevation in the morning. Potential reasons for this behavior could be an azimuthal dependency of the UV instrument or a non-ideal diurnal course in UV irradiance measured at the site Colchani located only about 5 km outside to the East of the Salar, leading to differences in UV irradiance at comparable solar elevation between morning and evening. Unfortunately this assumption could not be verified by evening measurements because the data sampling had to be stopped around 19:00 UTC for logistic reasons. 4. Model simulations The model simulations of albedo effects on UVI have been performed by the radiation transfer model STAR [29,30]. Fig. 8 presents a sensitivity study on the UVI as function of solar elevation and surface albedo for average conditions of the Bolivian Altiplano. Calculations have been performed for an altitude of 3700 m a.s.l. and a corresponding pressure of 650 hPa. A total ozone content of 250 DU, representing average Altiplano conditions, has been used. Furthermore the aerosol type ‘clean continental’ [31] has been selected as model input. Corresponding to measurements with the NOLL sunphotometer an aerosol
0
6
8 10 12 14 16
18
20
40
60
80
solar elevation [˚] Fig. 8. Modelled UVI as a function of solar elevation and surface albedo. Total ozone content: 250 DU; pressure 650 hPa; altitude 3700 m; aerosol optical depth at 550 nm: 0.05.
optical depth of 0.05 at 550 nm has been chosen for the simulations. In general the UVI strongly increases with increasing solar elevation due to reduced optical pathlength through the atmosphere and therefore reduced atmospheric absorption. There is also a marked increase of the UVI with enhanced surface reflectivity. Multiple scattering processes, mainly at air molecules, redirect a portion of the reflected upward directed radiation towards the ground, thus also increasing the downward directed UV irradiance. Fig. 9 shows the percentage effect of an increase of ground albedo on erythemally effective UV irradiance represented by the UVI. Two basic scenarios have been inves50 40 increase in UVI [%]
10
4
0.1
30 20 albedo increase from 0.03 to 0.85, 3700 m a.s.l. albedo increase from 0.2 to 0.7, 3700 m a.s.l. albedo increase from 0.03 to 0.85, sea level albedo increase from 0.2 to 0.7, sea level
10 0 10
20
30
40
50
60
70
80
90
solar elevation [˚] Fig. 9. Percentage increase in UVI due to enhancement of surface albedo as a function of solar elevation derived from model calculations. Total ozone content: 250 DU; pressure 650 hPa; altitude 3700 m; aerosol optical depth at 550 nm: 0.05.
J. Reuder et al. / Journal of Photochemistry and Photobiology B: Biology 87 (2007) 1–8
tigated. The first is associated with an increase in UV surface albedo from 0.03 that is typical for vegetated surfaces [32,33] to 0.85 a value representative for fresh snow away from polar and high mountain regions [34,35]. This scenario has been chosen for comparison with the results obtained by [11]. The second describes an albedo increase from 0.2 to 0.7. These values are expected to fit the albedo conditions around and on the Salar de Uyuni best. While own albedo measurements have been performed on the Salar (see Fig. 6) the lower value of 0.2 has been estimated from literature values (see e.g. compilation by [36]) for the dry, partly sandy and rocky surroundings of the salar nearly without any vegetation. Both scenarios have been modelled for sea level and the typical altitude of the Bolivian Altiplano in the region of the Salar de Uyuni, at 3700 m a.s.l., resulting in the four curves presented in Fig. 9. The increase of UV irradiance due to albedo effects shows only a weak variation with solar elevation. A slight maximum occurs around 20. The albedo effect at sea level (open symbols) is markedly higher than at 3700 m a.s.l. (filled symbols), representing the strong dependence of multiple scattering processes on air density. An increase in albedo from 0.03 to 0.85 results in an increase of the UVI of around 45% at sea level and of around 25% at the altitude of the Bolivian Altiplano. The results for the UVI at sea level conditions agree with the estimated effect derived by three-dimensional radiative transfer calculations [11] that predict a wavelength dependent range of 35% (at 300 nm) to 50% (at 330 nm). For comparison it should be taken into account that the central wavelength for the spectrally integrated UVI lies in the region between 305 nm and 315 nm, depending on the solar elevation and the extinction properties of the atmosphere. 5. Summary and outlook The effect of a high albedo surface on erythemally effective UV irradiance, i.e. the UVI, has been investigated by a combination of measurements and model calculations. A central part of the investigation has been a one week measurement campaign at the Salar de Uyuni. The erythemally effective UV irradiance has been measured by three broadband UV radiometers simultaneously at various sites on and outside the Salar at different distances to its edge. For an improved intercomparability of the individual UV radiometers, a new correction method has been developed and applied. It increases the relative accuracy of two UV radiometers to ±2% which is a distinct improvement compared to the typical absolute accuracy of UV irradiance measurements in the order of ±5%. Due to multiple scattering of UV radiation reflected by the high albedo surface, the UVI measurements taken on the Salar show a distinct enhancement compared to areas outside its limits. At Isla Incahuasi in the central part of the Salar, the radiation increase reaches values of the order of 20% for an solar elevation of 50. The measurements also indicate a moderate reduction of the enhancement
7
with decreasing solar height. There is an excellent agreement between the measured increase in UVI at solar elevations between 25 and 50, the highest solar elevation occurring during the campaign, and 1D radiation transfer calculations for an increase in surface albedo from 0.2 to 0.7. However, in contrast to the measurements, the model simulations predict a weak maximum around 20 and no distinct decrease towards lower solar elevation. One reason for this lack of agreement could be a general decrease in accuracy of UV measurements towards lower solar elevation and a corresponding increase in individual systematic differences between the used radiometers [22,25]. In addition there could be a solar elevation dependent variation in the contribution of local albedo, i.e. in the direct vicinity of the radiometer (rocks, bare soil and sparse vegetation), and regional albedo, determined by the surface properties in an area of a few tenth of kilometers around the measuring site (homogeneous salt surface). A detailed study of this problem can only be addressed by extended measurements in the future. In this context a detailed interpretation of the UV data measured from roof top of a four-wheel drive during two transects across the Salar will be of particular interest. These results will be presented in a follow-on paper. To our knowledge we have presented here the first direct measurements of UV surface albedo of the Salar de Uyuni. The reflectivity for erythemally effective radiation has been determined as 0.69 ± 0.02. Hardly any dependency on solar elevation has been found, indicating the homogeneity of the salt surface and therefore the presence of nearly isotropic reflection properties on the Salar. Together with its size, the rather constant atmospheric conditions with respect to total ozone content and low aerosol load, and the low average cloud cover, these nearly ideal reflection properties make the Salar an unique site for calibration purposes of space-borne radiometers in satellite remote sensing [16]. During the albedo measurements UV irradiances on tilted surfaces have also been determined for solar elevations between 45 and 50. The UV irradiances measured on a vertical plane, oriented normal to the solar azimuth angle, have shown values that are on average 12% higher than the UV irradiances with respect to the horizontal reference plane. The maximum UV irradiances measured with the detector oriented normally towards the sun have given an average increase of 18%. Due to the rather short measurement period and potential inaccuracies due to the manual positioning procedure, the data should of course not be over-interpreted. Nevertheless the data show interesting features, which encourage the future installation of ASCARATIS [27,28], an automatic measurement system for the determination of UV irradiances on inclined surfaces. Acknowledgements The authors are very grateful to Mr. Gonzalo Gutie´rrez for experimental help. We acknowledge the Volkswagen Foundation grant support which enabled this collaborative project.
8
J. Reuder et al. / Journal of Photochemistry and Photobiology B: Biology 87 (2007) 1–8
References [1] WHO, Global solar UV Index – A practical guide, A joint recommendation of WHO, WMO, UNEP, and ICNIRP, 2002. [2] J. Longstreth, F.R.d. Gruil, M.L. Kripke, S. Abseck, F. Arnold, H.I. Slaper, G. Velders, Y. Takizawa, J.C.v.d. Leun, Health risks, J. Photochem. Photobiol. B 46 (1–3) (1998) 20–39. [3] B.K. Armstrong, A. Kricker, The epidemiology of UV induced skin cancer, J. Photochem. Photobiol. B 63 (2001) 8–18. [4] H.S. Black, F.R.d. Gruil, P.D. Forbes, J.E. Cleaver, H.N. Ananthaswamy, E.C. deFabo, S.E. Ullrich, R.M. Tyrell, Photocarcinogenesis: an overview, J. Photochem. Photobiol. B 40 (1997) 29–47. [5] J.E. Roberts, Ocular phototoxicity, J. Photochem. Photobiol. B 64 (2001) 136–143. [6] M. Norval, Effects of solar radiation on the human immune system, J. Photochem. Photobiol. B 63 (2001) 28–40. [7] D.X. Nghiem, N. Kazimi, G. Clydesdale, H.N. Ananthaswamy, M.L. Kripke, S.E. Ullrich, Ultraviolet A radiation suppresses an established immune response: implications for sunscreen design, J. Invest. Dermatol. 117 (2001) 1193–1199. [8] M. Pfeifer, P. Koepke, J. Reuder, Effects of altitude and aerosols on UV radiation, J. Geophys. Res. 111 (2006) D01203, doi:10.1029/ 2005JD006444. [9] F. Zaratti, R. Forno, J. Garcia-Fuentes, M. Andrade, Erythemally weighted UV variations at two altitude locations, J. Geophys. Res. (2003), doi:10.1029/2001JD000918. [10] INE, National Institute of Statistics, Bolivian census, 2001. [11] M. Deguenther, R. Meerkoetter, A. Albold, G. Seckmeyer, Case study on the influence of inhomogeneous surface albedo on UV irradiance, Geophys. Res. Lett. 25 (19) (1998) 3587–3590. [12] J. Lenoble, Modeling of the influence of snow reflectance on ultraviolet irradiance for cloudless sky, Appl. Opt. 37 (12) (1998) 2441–2447. [13] I. Smolskaia, M. Nunez, K. Michael, Measurements of erythemal irradiance near Davies Station, Antarctica: effect of inhomogeneous surface albedo, Geophys. Res. Lett. 26 (10) (1999) 1381–1384. [14] S. Wuttke, G. Seckmeyer, G. Ko¨nig-Langlo, Measurements of spectral snow albedo at Neumayer, Antarctica, Ann. Geophys. 24 (2006) 7–21. [15] M. Huber, M. Blumthaler, J. Schreder, B. Schallhart, J. Lenoble, Effect of inhomogeneous surface albedo on diffuse UV sky radiance at a high-altitude site, J. Geophys. Res. 109 (2004) D08107, doi:10.1029/ 2003JD004013. [16] R.A.C. Lamparelli, F.J. Ponzoni, J. Zullo Jr., G.Q. Pellegrino, Y. Arnaud, Characterization of the Salar de Uyuni for in-orbit satellite calibration, IEEE Trans. Geosci. Remote Sens. 41 (6) (2003) 1461– 1468. [17] G. Honninger, N. Pobrowski, E.R. Palenque, R. Torrez, U. Platt, Reactive bromine and sulfur emissions at Salar de Uyuni, Bolivia, Geophys. Res. Lett. 31 (2004) L04101. [18] M. Morys, F.M. Mims III, S. Hagerup, S. Anderson, A. Baker, J. Kia, T. Walkup, Design, calibration and performance of MICROTOPS II handheld ozone monitor and Sun photometer, J. Geophys. Res. 106 (D13) (2001) 14573–14582.
[19] G.A. d’Almeida, R. Jaenicke, P. Roggendorf, D. Richter, New sunphotometer for network operation, Appl. Opt. 22 (23) (1983) 3796–3801. [20] A. McKinley, B. Diffey, A reference spectrum for ultraviolet induced erythema in human skin, Commission Int. Eclairage (CIE) 6 (1987) 17–22. [21] WMO, Report of the WMO meeting of experts on UV-B measurements, data quality and standardization of UV indices, WMO Technical Report No. 95, 1994. [22] A. Oppenrieder, P. Hoeppe, P. Koepke, J. Reuder, J. Schween, J. Schreder, Simplified calibration for broadband solar ultraviolet radiation measurements, Photochem. Photobiol. 78 (6) (2003) 603– 606. [23] B.A. Bodhaine, E.G. Dutton, R.L. McKenzie, P.V. Johnston, Calibrating broadband UV instruments: ozone and solar zenith angle dependence, J. Atmos. Ocean. Technol. 15 (1998) 916–926. [24] WMO, Report of the WMO–WHO meeting of experts on standardization of UV indices and their dissimination to the public, WMO Technical Report No. 143, 1997. [25] WMO, Report of the LAP/COST/WMO intercomparison of erythemal radiometers (Thessaloniki, Greece, 13–23 September 1999), WMO Technical Report No. 141, 2001. [26] WMO, Instruments to measure solar ultraviolet radiation, part II: broadband instruments measuring erythemally weighted solar irradiance, WMO-GAW Report No. 164, 2005. [27] A. Oppenrieder, P. Hoeppe, P. Koepke, Routine measurement of erythemally effective UV irradiance on inclined surfaces, J. Photochem. Photobiol. B 78 (6) (2004) 85–94. [28] A. Oppenrieder, P. Hoeppe, P. Koepke, J. Reuder, Long term measurements of the UV irradiance on inclined surfaces and visualization of UV exposure of the human body, Meteorol. Z. 14 (2) (2005) 285–290. [29] A. Ruggaber, R. Dlugi, T. Nakajima, Modelling radiation quantities and photolysis frequencies in the troposphere, J. Atmos. Chem. 18 (1994) 171–210. [30] H. Schwander, P. Koepke, A. Kaifel, G. Seckmeyer, Modification of spectral UV irradiance by clouds, J. Geophys. Res. 107 (D16) 4296. doi:10.1029/2001JD001297. [31] M. Hess, P. Koepke, I. Schult, Optical properties of aerosols and clouds: the software package OPAC, Bull. Am. Met. Soc. 79 (1998) 831–844. [32] M. Blumthaler, W. Ambach, Solar UVB-albedo of various surfaces, Photochem. Photobiol. 48 (1) (1988) 85–88. [33] R.L. McKenzie, M. Kotkamp, Upwelling UV spectral irradiances and surface albedo measurements at Lauder, New Zealand, Geophys. Res. Lett. 23 (14) (1996) 1757–1760. [34] T.C. Grenfell, S.G. Warren, P.C. Mullen, Reflection of solar radiation by the Antarctic snow surface at ultraviolet, visible, and near infrared wavelength, J. Geophys. Res. 99 (1994) 18669–18684. [35] S.G. Warren, W.J. Wiscombe, A model for the spectral albedo of snow. II: snow containing atmospheric aerosols, J. Atmos. Sci. 37 (1980) 2734–2745. [36] P. Koepke, J. Reuder, H. Schwander, Solar UV radiation and its variability due to the atmospheric components, Recent Res. Devel. Photochem. Photobiol. 6 (2002) 11–34.