Ultraviolet radiation in the Atacama Desert
The world’s highest levels of surface ultraviolet (UV) irradiance have been measured in the Atacama Desert. This area is characterized by its high altitude, prevalent cloudless conditions, and a relatively low total ozone column. In this paper, we provide estimates of the surface UV (monthly UV index at noon and annual doses of UV-B and UV-A) for all sky conditions in the Atacama Desert. We found that the UV index at noon during the austral summer is expected to be greater than 11 in the whole desert. The annual UV-B (UV-A) doses were found to range from about 3.5 kWh/m2 (130 kWh/m2) in coastal areas to 5 kWh/m2 (160 kWh/m2) on the Andean plateau. Our results confirm significant interhemispherical differences. Typical annual UV-B doses in the Atacama Desert are about 40% greater than typical annual UV-B doses in northern Africa. Mostly due to seasonal changes in the ozone, the differences between the Atacama Desert and northern Africa are expected to be about 60% in the case of peak UV-B levels (i.e. the UV-B irradiances at noon close to the summer solstice in each hemisphere). Interhemispherical differences in the UV-A are significantly lower since the effect of the ozone in this part of the spectrum is minor.
KeywordsUV Spectroradiometry UV radiance Atacama
The ultraviolet (UV) solar spectrum is divided into three regions: the UV-C (< 280 nm wavelength), which is strongly absorbed by the atmosphere and is therefore undetectable by ground-based measurements, UV-A (315–400 nm), and UV-B (280–315 nm) (Cordero et al. 2016).
The UV radiation is known to have adverse effects on the biosphere including terrestrial and aquatic ecosystems as well as on public health; several plants react to increased UV radiation with reduced growth or diminished photosynthetic activity (Tevini 1993). Although UV radiation is necessary for the synthesis of vitamin D in the human skin (Schrempf et al. 2016), exposure to UV radiation is associated with skin cancer, accelerated ageing of the skin, cataract and other eye diseases. It may also affect people’s ability to resist infectious diseases and compromise the effectiveness of vaccination programmes (Slaper et al. 1996).
The UV index (UVI) is used as an international standard measure of the surface UV irradiance that can lead to an erythemal or sunburning response in humans (Cordero et al. 2008). The UVI is computed by calculating the integral of the spectral UV irradiance in the range 250–400 nm weighted by using the so-called McKinlay-Diffey Erythema action spectrum (McKinlay and Diffey 1987). Since the McKinlay-Diffey Erythema action spectrum assigns a relatively low weight to the UV-A radiation in the UVI calculation, most of our attention in this paper is paid to the surface UV-B. The World Health Organization considers that values of the UVI greater than 11 stand for extreme risk of harm from unprotected sun exposure (Cordero et al. 2008).
The surface UV depends on the solar zenith angle (SZA), the total ozone column (TOC), cloudiness, aerosols, ground reflectivity (albedo), the Sun–Earth distance, and the altitude. SZA is the most important factor determining natural variations in the surface UV since it varies through the day more than any atmospheric constituent (including the TOC). The SZA determines the optical path length (OPL) of the solar radiation through the ozone and other absorbers and scatterers, changing in turn the surface irradiance; a shorter OPL leads to higher irradiance. The minimum daily SZA (and in turn the maximum daily UV irradiance) depends on the geographical location or latitude; in tropical areas close to the equator, the SZA can be zero at noon.
The world’s highest levels of surface UV irradiance have been measured in the Atacama Desert where ground-based measurements have shown the UV index to exceed 20 (Cordero et al. 2014a). This area is characterized by its location (relatively close to equator), its high altitude (a high altitude is associated with a shorter OPL through the atmosphere), relatively low TOC values, prevalent cloudless conditions, and low aerosol loading (McKenzie et al. 2015).
The TOC over the Atacama Desert is low when compared to locations in the northern hemisphere. Satellite daily estimates of the TOC value typically range from 240 Dobson Units (DU) to 300 DU in the Atacama Desert (Cordero et al. 2016).
As with other desert areas in subtropical continental regions, the Atacama Desert is characterized by clear conditions (Fournier et al. 2006). Nevertheless, at mid-latitude locations, UV irradiances recorded under broken cloud conditions can be enhanced compared to clear sky conditions, as for example when both direct and cloud-scattered sunlight (e.g. in presence of bright broken clouds) reach the surface (Piacentini et al. 2011; Schwander et al. 2002). Indeed, record high levels of surface UV have been measured under broken cloud conditions in Mauna Loa, Hawaii (3400 m altitude, 19°30′N) (Bodhaine et al. 1997), in Tibet (5000 m altitude, 29°N) (Dahlback et al. 2007), and on the Chajnantor Plateau (Atacama Desert, 5100 m altitude, 23°00′S) (Cordero et al. 2014a).
Due to the dry and arid conditions, surface UV in the Atacama Desert is expected to be significantly affected by wind-blown dust. However, satellite-estimates of the aerosol optical depth (AOD) consistently show climatological values relatively low (ranging from 0.25 in coastal area in the northern part of the Atacama desert to 0.05 over the Andean plateau). Retrievals from the moderate resolution imaging spectroradiometer (MODIS) (see Bellouin et al. 2013; Kato et al. 2011) allow estimating an annual regional climatological value of about 0.1 for the AOD in the UV part of the spectrum in the Atacama Desert. This relatively low AOD roughly agrees with ground-based measurements carried out at the Paranal Observatory (2635 m altitude, 24°37′S, 70°24′W) (Patat et al. 2011). For comparison, the AOD is typically higher than 0.15 at 500 nm at sites in North Africa (Toledano et al. 2009), while the AOD in the visible range measured at desert sites in northern China ranges from 0.24 to 0.36 (Xin et al. 2007).
The albedo of bright deserts affects the surface irradiance (especially in the UV part of the spectrum). Ground-based measurement of the spectral albedo of desert surfaces in western and central China have shown that the albedo in the UV part of the spectrum range from 0.05 to 0.11 (Aoki et al. 2005). These values roughly agree with the data of Lambertian equivalent reflectivity (LER) recorded in the UV range over the Atacama Desert by the ozone monitoring instrument (OMI) aboard the Aura satellite (Damiani et al. 2014).
The earth is closer to the sun in the southern hemisphere (SH) summer compared with the corresponding season in the northern hemisphere (NH). Only the effect of the elliptical orbit of the earth around the sun, leads to about 7% difference in the interhemispherical peak irradiances (McKenzie et al. 2006). As shown below, the different geographical distribution of the TOC, cloudiness, and aerosols, further strengthen these inter-hemispherical differences in the surface irradiance.
In this paper, we discuss the influences of aerosols, ozone, cloudiness, and albedo on the surface UV in the Atacama Desert and how these influences explain the significant interhemispherical differences shown in Fig. 1. Our analysis is based on an integrated approach that includes ground-based measurements, satellite-derived data, as well as data computed by using radiative transfer models. These data allowed us to produce climatological maps of the surface UV (monthly UV index at noon and annual doses of UV-B and UV-A) for all sky conditions in the Atacama Desert. Methodological details are provided below.
Materials and methods
Some of the results shown below are based on ground-based spectral measurements of solar irradiance. These measurements were carried out by using double monochromator-based spectroradiometers. These instruments are recommended for quality-controlled measurements of the spectral irradiance (particularly in the UV part of the spectrum). We used double monochromator-based instruments (fitted with a photomultiplier detector) developed according to the specifications issued by the World Meteorological Organization (WMO) (Seckmeyer et al. 2001) and the network for the detection of atmospheric composition change (NDACC) (Wuttke et al. 2006).
TOC estimates shown below were retrieved from OMTO3 products from the ozone monitoring instrument (OMI) aboard theAura satellite. OMI-derived estimates of the UV were retrieved from OMUVB products (version 1.3). We selected overpass OMI data minimizing the distance between the point of interest and the center of the satellite pixel. OMI estimates have a 0.5 × 0.5 degree resolution.
Broadband albedo data over the period 2000–2005 were produced by the surface meteorology and solar energy (SSE-release 6.0); a renewable energy resource web site sponsored by NASA’s Applied Sciences (see LARC 2017), while the Lambertian equivalent reflectivity (LER) at 340 nm wavelength was computed by using daily OMI estimates over the period 2004–2007 (see Kleipool et al. 2008).
Estimates of the AOD were retrieved from different satellite instruments including the moderate resolution imaging spectroradiometer (MODIS) (aboard the Terra satellite), from MODIS and multi-angle imaging spectroradiometer (MISR) (both aboard the Terra satellite), and from the Sea-viewing Wide Field-of-view Sensor (SeaWIFS), processed with the Deep Blue algorithm (Hsu et al. 2012). In the case of MODIS, we used the algorithm “Combined Dark Target and Deep Blue AOD at 0.55 micron for land and ocean (MOD08_M3 v6)” (Sayer et al. 2014), while in the case of MISR, we used the MIL3DAE daily level 3 product (Diner et al. 2001).
Ground-based measurements of the AOD were obtained from the aerosol robotic network (AERONET) (Holben et al. 1998) in Arica (the northernmost Chilean city) and from the Euroskyrad network (Campanelli et al. 2007) on the Chajnantor plateau (5100 m altitude, 23°00′S, 67°45′W).
As radiative transfer model, we used the UVSPEC model, which is the main tool of the libRadtran package for radiative transfer calculations (Mayer and Kylling 2005). In the UV spectral range, this model has been validated by systematic comparisons with ground-based measurements under cloudless conditions (Badosa et al. 2007). We used the discrete ordinates radiative transfer (DISORT) solver (Dahlback and Stamnes 1991) and the extraterrestrial spectrum by Gueymard (2004).
We also analyzed climatological estimations computed from outcomes of MIROC-ESM-CHEM i.e. a climate-coupled model with modules that represent the whole Earth system (Watanabe et al. 2011).
Ozone has been depleted gradually since 1980 and is about 3% lower over the globe (WMO 2014). The depletion, which exceeds the natural variations of the TOC, is due to the action of the so-called halogen source gases (also known as ozone-depleting gases) (see Molina and Rowland 1974). The production and consumption of the main halogen source gases by human activities are regulated worldwide under the montreal protocol (WMO 2014).
The ozone depletion is very small near the equator and increases with latitude toward the poles. As a result of the Montreal Protocol, the total abundance of ozone-depleting gases in the atmosphere has begun to decrease in recent decades (Hu et al. 2017; Strahan and Douglass 2017). It is expected that the decrease will continue throughout the 21st century if compliance with the Montreal Protocol continues (Solomon et al. 2016; McKenzie et al. 2011). In Antarctica, ozone depletion has been the dominant factor for surface UV anomalies. However, in the rest of the world, especially at locations close to equator (such as the Atacama Desert), the natural variations of the other factors also affecting the surface UV normally obscure the effect of the ozone depletion (McKenzie et al. 2011).
The amount of ozone is well known to be lower in the southern hemisphere than in the northern hemisphere (McKenzie et al. 2006). For example, Fig. 2b shows the monthly average of the OMI-derived TOC values over the period 2005–2015 on the Chajnantor Plateau (see red line) and at the Izaña Observatory (2367 m altitude, Tenerife, Canary Islands, 28°18′N, 16°30′W; see blue line). Seasonal variations in the ozone values are apparent. Monthly averages peak in spring. Moreover, peak ozone values are usually higher at the Izaña Observatory than on the Chajnantor Plateau. The TOC usually ranges from 240 to 270 Dobson units (DU) on the Chajnantor Plateau, while it ranges from 270 to 330 DU at the Izaña Observatory. The average of the OMI-derived total ozone column at the Izaña Observatory in June (300 DU) is typically about 23% higher than on the Chajnantor Plateau in January (245 DU); the differences are significantly lower (about 16%) when comparing annual averages.
Due to multiple reflections between the ground and the scattering atmosphere, high surface albedo significantly enhances surface UV (Cordero et al. 2014b). The albedo is normally taken as being equal to the ratio between the up- and downwelling global irradiance.
Figure 4b depicts the annual average of the lambertian equivalent reflectivity (LER) in the UV range, computed from OMI estimates over the period 2004–2007. LER is the required reflectance of an isotropic surface needed to match the observed top of the atmosphere (TOA) reflectance in a pure Rayleigh scattering atmosphere (Damiani et al. 2014). Therefore, it can be used to roughly estimate the actual surface albedo in the UV. Indeed, the LER in the UV range is taken as a proxy of the albedo by the OMI algorithm to compute the surface UV irradiance. Higher values of UV albedo are shown in the northern part of the Atacama Desert while the reflectivities diminish southward.
In the case of the Atacama Desert, estimates of the AOD retrieved from different satellite instruments show significant differences (Chin et al. 2014). Satellite-retrieved data of AOD differ between each other likely due to the complex morphology of the Andes. Indeed, our comparisons of AOD estimates from different satellites over the Atacama Desert also show significant differences.
Despite the strong need for ground-based measurements of AOD aimed at validating satellite estimates, few instruments have been deployed in the Atacama desert. The AOD at 380 nm measured by a sunphotometer affiliated with AERONET and set up in Arica (the northernmost Chilean city) ranges from 0.2 to 0.3 (with values close to the upper limit of this range in winter). Our team also deployed in September 2016 a sunphotometer affiliated with the Euroskyrad network on the Chajnantor plateau (5100 m altitude, 23°00′S, 67°45′W). AOD retrievals at 400 nm from this instrument during 2017 ranged from 0.02 (in winter) to 0.08 (in summer).
The ratio between the UV irradiance measured under all-sky conditions and the UV irradiance computed by using the UVSPEC radiative transfer model assuming cloudless conditions, can be used to compute the so-called cloud modification factor (CMF). The CMF accounts for the attenuation of UV radiation due to clouds. Although the CMF can be estimated at any wavelength, it is often computed by using satellite estimates of UVI both under clear-sky conditions and under all-sky conditions (see Cordero et al. 2013). In that case, the CMF is referred to as UVI-based CMF.
UV in Atacama
As shown in Fig. 7, the seasonal change in the SZA leads to relatively lower UV values from April to August; the UVI at noon ranges in June from 4 (in the southern part of the Atacama desert) to 9 at locations closer to equator. The UVI boosts in the austral summer such that, from November to February, it is typically greater than 11 in the whole desert. At high altitude locations (like the Chajnantor plateau), the UVI peaks at about 20 at noon close to the summer solstice under cloudless conditions. As pointed out above, these records are due to the high altitude, which is associated with a shorter OPL through the atmosphere and in turn with a lower attenuation. The frequent broken cloud conditions on the Chajnantor plateau in January led to the highest values of the surface UV ever measured (when both direct and cloud-scattered sunlight reaches the surface) (Cordero et al. 2014a).
As shown in Fig. 8, the annual UV-B (UV-A) doses were found to range from about 3.5 kWh/m2 (130 kWh/m2) in coastal areas to about 5 kWh/m2 (160 kWh/m2) on the Andean plateau. The peak UV doses in the Atacama Desert occur at high altitude locations (such as the Chajnantor plateau). Although northern coastal locations (such as Arica or Iquique) are located at lower latitudes, the average daily UV doses tend to be lower there than on the Chajnantor plateau due to the cloudiness and the significant effect of the altitude. Our estimates of the annual UV-B doses in the Atacama Desert (see Fig. 8a), roughly agree with the estimates computed by using outcomes of the MIROC-ESM-CHEM model (see Fig. 1).
As pointed out above, the typical annual UV-B doses in the Atacama Desert is about 40% greater than typical annual UV-B doses in northern Africa. The UVSPEC radiative transfer model allowed us to disentangle the effects of the parameters leading to these difference. We found that interhemispherical differences in the TOC (about 15%) and the AOD (about 5%) account for about half of the difference (since both the TOC and the AOD are lower in the Atacama Desert than in northern Africa). The rest of the difference arises from the influence of the latitude (the Atacama Desert is closer to the equator than northern Africa) and the altitude (the Atacama Desert is on average significantly higher than northern Africa).
The differences between the Atacama Desert and northern Africa are expected to be even greater in the case of peak UV-B levels (i.e. UV-B irradiances at noon close to the summer solstice in each hemisphere). We estimated that under cloudless conditions, the typical peak UV-B irradiance in the Atacama Desert is about 60% greater than typical peak UV-B irradiance in northern Africa. Nearly half of this difference (about 28%) comes from the TOC; although annual averages of the TOC are only 10% apart, the difference between typical values of the TOC close to the summer solstice in each hemisphere is significantly greater: 20% (250 DU vs. 300 DU). The elliptical orbit of the earth around the sun (which makes the earth closer to the sun in the SH summer compared with the corresponding season of the NH) adds 7%. As in the case of the annual doses, the rest of the difference arises from the influence of the latitude, the altitude and the AOD.
Summary and conclusions
The world’s highest levels of surface UV have been measured in the Atacama Desert. This area is characterized by its high altitude, prevalent cloudless conditions, and a relatively low total ozone column.
Satellite retrieved data of the AOD show significant differences between different instruments, especially in the coastal northern part of the Atacama Desert and the eastern slope of the Andes. These differences underline the strong need for ground-based measurements of the AOD aimed at validating satellite estimates. Despite the differences in satellite retrievals, we estimate that the AOD in the UV part of the spectrum typically ranges from 0.25 (in the coastal northern part of the Atacama Desert) to 0.05 (on the Andean plateau). Comparing regional averages, we estimate that the typical AOD in the Atacama Desert is about 50% lower than typical AOD in northern Africa.
The Atacama Desert is characterized by clear conditions. Estimates of the UVI-based CMF typically are greater than 0.9 in the Atacama Desert although they can slightly lower in January and February. The lower CMF values in the austral summer are due to an increase in the broken clouds conditions (overcast conditions are infrequent). Broken cloud conditions can significantly enhance the surface UV (compared to clear sky) and produce record peak UV levels. Ground-based measurements on the Andean plateau (above 5000 m altitude) have shown that these conditions often lead to the UVI values greater than 20 in summer.
Satellite retrieved data of the LER allowed us to estimate that the climatological values of the albedo in the UV part of the spectrum, ranging from 0.11 in the northern part of the Atacama Desert to 0.08 in the southern part of the desert. Although the surface UV can be enhanced by occasional snowfalls over The Andes, these events are not expected to significantly affect the climatological values of the UV albedo.
OMI-derived estimates show that, despite significant seasonal variations, the TOC over the Atacama Desert is low compared to locations in the northern hemisphere. We estimated that the climatological values of the TOC range in the Atacama Desert from 250 to 280 DU The TOC values also change with the altitude such that it can be up to 15 DU higher over coastal locations compared to the Andean plateau. This difference (along with the shorter OPL through the atmosphere associated with high altitude locations) contributes to the increment observed in the UV with the altitude. We estimate that in the range 0–2500 m altitude, surface UV increases with the altitude by about 4% per km in the UV-A range, and 9% per km in the UV-B range. In the range 2500–5000 m altitude, surface UV increases with the altitude by about 2% per km in the UV-A range, and 4% by km in the UV-B range.
We have produced climatological maps of the surface UV (monthly UV index at noon and annual doses of UV-B and UV-A) for all sky conditions in the Atacama Desert. The annual UV-B (UV-A) doses were found to range from about 3.5 kWh/m2 (130 kWh/m2) in coastal areas to 5 kWh/m2 (160 kWh/m2) on the Andean plateau. Due to the seasonal change in the SZA, the UVI at noon ranges in June from 4 (in the southern part of the Atacama desert) to 9 at locations closer to the equator. The UVI boosts in the austral summer such that, from December to February, it is typically greater than 11 in the whole desert.
We confirmed that the typical annual UV-B doses in the Atacama Desert is about 40% greater than typical annual UV-B doses in northern Africa. Interhemispherical differences in the TOC (about 15%) and the AOD (about 5%) account for about half of the difference (since both the TOC and the AOD are lower in the Atacama Desert than in northern Africa). The rest of the difference arises from influence of the latitude (the Atacama Desert is closer to the equator than northern Africa) and the altitude (the Atacama Desert is on average significantly higher than northern Africa). Interhemispherical differences in the UV-A are significantly lower since the effect of the ozone in this part of the spectrum is minor.
We also found that the differences between the Atacama Desert and northern Africa are expected to be about 60% in the case of peak UV-B levels (i.e. the UV-B irradiances at noon close to the summer solstice in each hemisphere). The increment with respect to the interhemispherical differences in annual UV-B doses arises from the elliptical orbit of the earth around the sun (which makes the earth closer to the sun in the SH summer compared with the corresponding season of the NH), and the seasonal variations in the TOC (which increases the difference between the Atacama Desert and northern Africa when comparing TOC values corresponding to the summer solstice in each hemisphere).
The data are available under request.
The support of the Consejo Nacional de Ciencia y Tecnología (CONICYT, Preis ACT1410, 1171690, 1161460 and 1151034) and the Corporación de Fomento de la Producción (CORFO, Preis 17BPE-73748 and 16BPE2-66227), Centro de Nanotecnología (CEDENNA), and the Universidad de Santiago de Chile (USACH, Preis USA1555), is gratefully acknowledged.
Conceived and designed the experiments: RRC, AD and JC Performed the experiments: JJ., and ES. Analyzed the data: RRC, AD, SF, MC, SF, FL, DL and PL Wrote the paper: RRC, AD, JC and SF.
Although the support of the several agencies is gratefully acknowledged (see details below), the funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.
Compliance with ethical standards
Conflict of interest
The authors declare neither competing interests nor conflict of interests.
Informed consent was obtained from all individual participants included in the study.
Not applicable since this research did not involved humans, animals, plants or any form of life.
- Bodhaine BA, Dutton EG, Hofmann DJ, McKenzie RL, Johnston PV (1997) UV measurements at Mauna Loa: July 1995 to July 1996. J Geophys Res 102(19265):19273Google Scholar
- LARC NASA (2017) Surface meteorology and Solar Energy (SSE-release 6.0): A renewable energy resource web site sponsored by NASA’s Applied Sciences. https://power.larc.nasa.gov/common/php/SSE_ExSummary.php. Accessed 20 July 2017
- McKinlay AF, Diffey BL (1987) A reference action spectrum for ultraviolet induced erythema in human skin. CIE J 6:17–22Google Scholar
- Seckmeyer G, Bais A, Bernhard G, Blumthaler M, Booth CR (2001) Part 1: Spectral instruments. Instruments to Measure Solar Ultraviolet Radiation. WMO-GAW 125 World Meteorological Organization- Global Atmosphere Watch. Secretariat of the World Meteorological OrganizationGoogle Scholar
- Tevini M (1993) UV-B radiation and ozone depletion: effect on humans, animals, plants, microorganisms and materials. Lewis, New YorkGoogle Scholar
- WMO/UNEP Scientific Assessment of Ozone Depletion (2014) Global ozone research and monitoring project. Switzerland, GenevaGoogle Scholar