The ultraviolet (UV) solar radiation in the Earth's lower atmosphere and at the surface has been widely studied in the last decades due to its involvement in chemical and biological processes. Wavelengths in the range 100–400 nm constitute the UV spectral region that is attenuated by scattering and absorption processes occurring in the atmosphere. UV radiation (200–400 nm) at the top of the atmosphere represents less than 8% of the total irradiance.
The shortest wavelengths (UV-C, 100–280 nm) are absorbed by stratospheric ozone; wavelengths in the UV-B range (280–315 nm) are absorbed only weakly by ozone and therefore, a small part of UV-B and most of the UV-A (315–400 nm) radiation reach the Earth's surface.
Solar UV reaching the ground is governed by astronomical and geographical parameters such as sun height and station altitude that change the path length of sunlight through the atmosphere; the atmospheric constituents such as ozone, clouds and aerosols can lead to UV variations at the Earth's surface; as a consequence, at ground level 94% of the UV energy corresponds to UV-A and 6% to UV-B.
In relation to the effect of UV, it is known that a marked increase in the incidence of diseases is associated with the exposure to UV-B solar radiation (WMO, 2002); it can produce sunburn, ocular damage and allergies, skin cancer and other effects on living beings. Many biological responses to the exposure to the UV range turn to be greater at the shortest wavelengths; thus even small increments of UV-B radiation can lead to substantial biological effects.
Therefore the UV effects on skin can be quantified by some different indices; the erythemal UV-Index, simply called the UV-Index (UVI), is an estimation of the effectiveness of solar radiation to produce harmful effects on human skin, where one unit is equal to 25 mWm−2. The UVI is usually given at local solar noon, when the Sun is high in the sky (Marin et al., 2005).
Nevertheless, the erythemal UVI is an artificial quantity derived from the erythemal irradiance, which can be obtained as an integration of the UV irradiance at ground level, weighted by the International Commission on Illumination (CIE) action spectrum. This spectrum is a model for the susceptibility of the Caucasian skin to sunburn (erythemal). It was proposed by McKinlay and Diffey (1987) and adopted as a standard by the Commission Internationale de l'Éclairage (International Commission on Illumination). The CIE spectrum of human erythemal describes the energy efficiency at different wavelengths to produce the particular biological effect and shows an absolute maximum at 297 nm; the erytemal UV irradiance consists of 17% UV-A and 83% UV-B, at the Earth's surface. Therefore, small changes in UV-B may produce strong biological effects.
UVER can be modelled by radiative transfer models and validated against UVER measurements. Experimental UVER data are also useful for developing UV radiation climatology and establishing geographical and seasonal distribution of UVER exposure that, as it has been explained, are essential in many fields including human health.
In the last years, therefore, the interest of the scientific community has been focussed on the study of UVER radiation, principally due to its harmful effects, on biological systems; in the same way (Diffey, 1991) studied the biological effect of solar UV on humans. Fioletov et al. (2009) studied the relationship of UVER and vitamin-D. The following authors compared the biological active UV radiation in two different high altitude stations: Ambach et al. (1991) and Blumthaler et al. (1997). Di Sarra et al. (2002) recorded data in the Mediterranean area to study the effects of different factors. Several authors in Spain analysed the UVER in different regions of the country: Foyo-Moreno et al. (1998), Serrano et al. (2006), Cañada et al. (2008), among others. In Slovakia, Pribullová and Chmelik (2008) proposed the isoline-maps (curves where the function has the same particular value) for solar erythemal studies. Badarinath et al. (2008) examined the influence of natural and anthropogenic activities on UV over a tropical region in India.
National and international organisations recommend the set-up of UV-B networks and regular measurements are also advised by the World Meteorological Organisation (WMO, 2007) but few radiometric stations measure systematically and, as a consequence, continuous data series are not numerous enough to provide a global climatology of UV-B solar radiation. In relation with measurements, specifically, their quality and sensors uncertainties, Hülsen and Gröbner (2007) suggested the characterisation and calibration of UV sensor procedures for a required quality of measurements; Webb et al. (2007) devised a practical guide to derive erythemal weighted solar UV irradiance from measurements with broadband sensors.
In spite of the important role of UVER, few radiometric stations are installed in our Autonomous Community (Castilla and Leon) and continuous measurements are not numerous to provide a global climatology of UV-B and erythemal UV irradiance.
The aim of this paper is to analyse erythemal UV solar irradiance measured in the wavelength interval 280–400 nm, in order to know its characteristics. The paper also examines the magnitude and nature of the annual variability using the isoline maps, which are related to the optical properties of the atmosphere. Erythemal UV may be absorbed and scattered in its passage through the atmosphere. The attenuation experienced by UVER, including hydrometeors and impurities is investigated under cloud-free conditions. The final objective is to understand the role of the atmospheric constituents in modulating the annual cycle of UVER irradiance at ground level.
The interest of this paper is that similar data series have not been recorded before and no similar work has previously been done in Central Spain. Furthermore, previous studies have observed a decrease in stratospheric ozone levels in the region Bilbao et al. (2008), and as a consequence, higher levels of UVER radiation could reach the troposphere and the surface.
This work constitutes the first erythemal UV analysis in the region, which is essential due to the fact that the atmospheric conditions in the area (dry summers and cold winters, high air temperatures and low-vapour pressure values at midday from June to August) lead us to think that the solar radiation received is higher than in other regions in the country (Bilbao and Miguel, 2010). The study is recommended from the point of health protection as the region is an area with a number of natural parks and courses of recreational activities.
In the following sections a description of the place, data base and the methodology is given; the statistical analysis of hourly, daily and monthly UVER values is described and, in the same way, the interannual variability and the solar UV-B attenuation in the atmosphere are explained.
2. Data and methodology
2.1. Location, instrumentation and data collection
The measurements used in this study have been taken in a rural area near Valladolid, Spain, 38 km NW of the city (41°40′N; − 4°50′W and 849 m above sea level). The climate in the area is usually warm with maximum and minimum temperatures of 38 °C and 10 °C in summer, respectively; whereas in winter months, the weather is influenced by Atlantic frontal systems.
The UV-B irradiance was recorded by a pyranometer YES UVB-1 (Yankee Environmental Systems, Inc.) and the global solar irradiance (G) was measured by a pyranometer CM-6B (Kipp & Zonen), both on horizontal surface (Bilbao and Miguel, 2010).
According to the manufacturer, the UVB-1 radiometer has a spectral response between 280 and 320 nm, which is very similar to the erythemal (sunburn) and DNA damage action spectra; its cosine response is better than ± 5% for 0–60° of solar zenith angle (SZA) and its sensitivity is 2.04 Wm−2V−1. The instrument was calibrated every two years, and the sensitivity variation between 2002 and 2008 was around 0.5%. The last sensor calibration was developed at the National Institute for Aerospace Technology (INTA) in Spain. This calibration consisted of the measurement of the spectral response of the sensor indoors and a comparison with a Brewer MKIII spectroradiometer outdoors. After this process, a double input matrix with the calibration factors depending on the SZA and the total ozone column (TOC) was obtained using a radiative transfer model.
In order to convert the output-voltage signal of the sensor into units of erythemal irradiance (Wm−2), the signal voltage was multiplied by the calibration factor which is obtained from the matrix taking into account the average SZA during the 10-min measure and the daily TOC. The experimental uncertainty of the sensor according to the results of Hülsen and Gröbner (2007) is in the range 4.6–7%.
Global solar data were recorded by a CM-6B (Kipp & Zonen) pyranometer that provides measurements in a spectral range 305–2800 nm. Another CM-6B sensor was calibrated at the Institute of Renewable Energy (CIEMAT) in Madrid, Spain, by an outdoor inter-comparison with the reference sensor CM22 Kipp & Zonen calibration traceable by the World Radiometric Reference through the World Radiometric Centre in Davos, Switzerland. The calibration constant has uncertainties lower than 2%. The last CM-6B instrument is only used to calibrate every year the CM-6B instrument that takes ground measurements. The variation in the calibration factor along the analysed period was around 3%.
Daily TOC values were remote sensing recorded by Total Ozone Mapping Spectrometer (TOMS) and Ozone Monitoring Instrument (OMI) sensors. They were provided by National Aeronautics and Space Administration (NASA) through the web (http://macuv.gsfc.nasa.gov/). Daily TOC data from June 2002 to December 2008 have been used in this study. The used data version from TOMS (before 2005) has been the TOMS-V8-corrected. This version shows a deviation lower than 1% for the latitudes similar to Valladolid when it was compared with ground measurements (Antón et al., 2010).
Solar sensors were connected to a Campbell CR10X Data Logger, which was programmed to take measurements every 10 s and computed the average values every 10 min; then, hourly and daily irradiances were evaluated from the 10-min average values, with respect to the daily values, the sunshine and sunset (hours with no 6 measures) were taken into account to better consider the length of the day. The measures are collected in Greenwich Mean Time (GMT).
UVER irradiance data series was composed of 88 278 10-min diurnal data which are used to evaluate 13 490 and 1488 UVER hourly and daily values, respectively.
The global solar irradiance set consisted of 167 378 values as 10-min average diurnal data. Global irradiance was checked taking into account the detection limits of the sensor and the following quality control tests: G≤1.2 G0, where G is the horizontal global solar irradiance and G0 is the horizontal extraterrestrial solar irradiance. This criterion was used to guaranty real G values because, in principle, G values higher than 1.2 G0 cannot be possible. For further details about the followed quality control see Mateos et al. (2010). After this process, from 164 813 10-min global solar data, a total of 29 464 hourly and 2732 daily values were obtained. Some month by data were missing in 2006, 2007 and 2008, due to calibration activities and technical problems.
Few studies about the performance of solar erythemal radiation have been carried out in Spain and even fewer in Castile and Leon region, Spain, although the territory is one of the country that receives really high levels of solar radiation (Miguel et al., 2001).
As different sky conditions have to be taking into account for comparing models and measurements, the hours and the days of the experimental period have been classified according to the sky conditions. Hourly, kh, and daily, kd, clearness indices were evaluated for each hour and day, respectively, during the measurement period, by using the following expressions:
where Gh is the hourly global solar irradiation; Gd is the daily global solar irradiation; G0h and G0d are the hourly and the daily extraterrestrial solar global irradiation, respectively, which are given by the following expressions:
where δ is the sun declination; ϕ is the site latitude; ωi is the sun hourly angle; ωs is the sunset hourly angle; Isc is the solar constant (1367 Wm−2); E0 is the eccentricity of the Earth's orbit around the Sun (Iqbal, 1983).
Table I shows the criteria set that was established by the authors for the classification of different sky conditions according to the clearness index intervals. The sky condition intervals have been selected according to Mateos et al. (2010), considering that the limit value for clear conditions is lower in the daily values than in the hourly ones because the hourly values are calculated with 6 data, and it is easier to find a clearness index higher than 0.75; however the daily values are the average of much more data and find values higher than 0.75 is more difficult.
Table I. Criteria set by the authors for the classification of different sky conditions according to clearness index intervals
kd ≥ 0.70
0.34≤kd < 0.70
0 < kd < 0.34
kh ≥ 0.75;
0.34≤kh < 0.75
0 < kh < 0.34
With the aim of investigating the characteristics of the UVER dataset, a statistical analysis was carried out. The hourly irradiance data for each month of the measurement period were arranged in ascending order and the following statistical parameters were evaluated: arithmetic mean (M), median (Md) maximum (Mx), standard deviation (SD), the first and third quartiles, Q1 and Q3, coefficient of quartile variation V(%) = 100(Q3− Q1)/(Q3 + Q1), kurtosis (Ku), skewness (Sk) and percentiles: P0, P5, P25, P50, P75 and P95. A similar calculation has been made with the daily data. The statistical parameters of M, SD, Sk and Ku give information about the data distribution characteristics, such that M is the central tendency; SD shows the dispersion; Sk represents the degree of asymmetry with relation to the central value and Ku is the degree of the inclination of the peak of the distribution in relation to its centre. These parameters are necessary for characterising the location and variability of the data set and evaluating a suitable probability distribution that could be used to generate synthetic data or for assessing model results.
Contour lines, also called isoline-maps, have been used to plot the percentile, mean, skewness and kurtosis variables. The isoline is a function of the time (hourly) and the month that connects points of same value. The gradient of a variable is always perpendicular to the isolines. These configurations allow inferring relative gradients of the variable and estimating that variable value at specific atmospheric conditions.
3. Results and discussion
3.1. Data variation
The UVER and global solar irradiation data series are studied in this work. Figure 1 shows the variation of the measured UVER values during the measurement campaign. Data reveal a common evolution shape with maxima in summer and minima in winter, mainly due to the daily minimum SZA and day-length (astronomical factors) variation during the year. Large fluctuations in the months of March and November are mainly due to unstable meteorological conditions during the transition from cold to warm weather and vice versa; the TOC, cloud and aerosol changes along the year can also explain the UVER variations (Miguel et al., 2010).
The absolute daily maximum UVER irradiation was 6307 Jm−2 (25th June 2008) and the minimum 164 Jm−2 (1st November 2008). The maximum daily UVER values were the following for the other years: 5706 Jm−2 (11th July 2002), 5936 Jm−2 (19th June 2003), 5810 Jm−2 (14th July 2004) and 5699 Jm−2 (4th June 2005).
The fluctuations may be due to clouds, aerosols and TOC; however, the annual variation is also influenced by astronomic factors, such as the SZA, which modifies the optical path through the atmosphere. It can be seen that the annual maximum is higher when it occurs near the summer solstice date, because in this period, the extraterrestrial solar radiation and the day length reach their upper values.
The cloud effects are significantly appreciated in Figure 1 when high differences occur between two consecutive days. Days of the measurement campaign have been classified according to the clearness index values; July is the month with more clear days and winter and autumn are, principally, partly cloudy seasons. The cloudiest months are January, December and November. January has approximately as many cloudy days as partly cloudy.
3.2. Statistical estimators
3.2.1. Analysis of hourly average UVER
A statistical study of the most representative erythemal indices for each month of the year has been carried out and the UVER accumulated values have been evaluated because they are very useful in the studies of effects on human beings.
The statistical parameters of the hourly average irradiance values for each month (all the years together) of the experimental period were assessed. Table II shows the hourly statistical estimators for June. It can be observed that the median values are higher than the average ones, which suggests that the UVER distribution is not normal.
Table II. Statistical estimators of the mean hourly UVER (mWm−2) in June, Valladolid, Spain, under all-sky conditions, for the period 2002–2008 (GMT time)
The differences between the superior quartiles and the maximum values are low, with the highest ones not exceeding 33 mWm−2. This result means that the maximum values are representative of the UVER irradiance at midday. The SD shows low values that confirm high stability in the same period.
The UVER variability has been studied by means of the index V. From this, it can be seen that June represents high stability with low values of V that fluctuate between 13% at midday and 15% in the afternoon. The coefficient of quartile variation V reaches its lowest value before solar noon in summer time; this indicates a high stability along these hours in summer months. It can be appreciated that the inferior quartile is close to the mean value and the SD is higher along midday hours, and symmetrically distributed around solar noon during the summer months. This could be explained by a minor presence of clouds in summer that leads to a high stability.
There is a great difference between the minimum values and the percentile P5, more remarkable in summer time around solar noon. It can be concluded that the absolute minimum values cannot be considered as representative of UVER at the measurement station and correspond to atypical values, particularly in summer. Similar results have been obtained by Martinez-Lozano et al. (1996), Foyo-Moreno et al. (1998) and Bilbao et al. (2008).
The differences between Mx values and P95 are quite small; for instance in June they are smaller than 7%. So, the maximum values may be considered as representative of UVER irradiation characteristics in Valladolid along the year.
3.2.2. Analysis of monthly average hourly UVER
The monthly average hourly UVER values have been calculated. The results show that they fluctuate between 29.33 mWm−2 in December and 208.41 mWm−2 in July at solar noon. Figure 2 shows the mean monthly hourly annual evolution and a high symmetry can be observed around the month of June when the irradiance reaches its maximum, while it decreases in spring and autumn and reaches its minimum in winter months. The results can be explained taking into account the relation between the summer and winter solstices. In Figure 2 the TOC effects can be deduced because the UVER values are lower in spring than in the autumn months, (April and September).
Isoline configurations have been used in order to show a more detailed evolution of UVER (Miguel et al., 1994). Two symmetries are observed with respect to the solar noon and the solstices. Maximum values at solar noon depend on the SZA that modifies the optical path of the radiation through the atmosphere and on the day length. The plots show a similar pattern for all atmospheric conditions except for cloudy skies. Similar results were obtained by other authors in Spain: Foyo-Moreno et al. (1998) in Granada, Martinez-Lozano et al. (1996) in Valencia, and Serrano et al. (2006) in Badajoz.
Figure 3 presents the isoline diagram of the hourly erythemal irradiance percentiles in Wm−2 for four different types of sky: clear, partly cloudy, overcast and all conditions. Lines of equal intervals of 0.02 Wm−2 have been drawn (Bilbao et al., 2003). The diagram provides information about solar UVER percentile during each hour of the day and for each month. The diurnal variation of the percentile for a month can be obtained from the figure by drawing a vertical line for that particular month. A horizontal line drawn for a specific hour can give the annual variation of the percentile.
For all situations, the mean radiation flux at midday is the greatest in the summer months when it exceeds 0.24 Wm−2 for clear skies, while the lowest midday flux occurs in the month of January about 0.02 Wm−2. The irradiance tends to be less in the afternoon than in the morning for the autumn months of September and October, probably due to the cloud development after midday.
It can be seen in Figure 3 that the slope of the isolines along months for fixed hour is smaller for low percentiles than for the highest ones; this result can be explained by the prevailing atmospheric conditions in the region with air masses that lead to cloudy sky in winter, spring and autumn, and clear days in summer with cloud diurnal evolution and persistent high pressures; for clear-sky conditions, the slope in the summer months shows a different pattern due to lack of clouds.
For all-sky conditions and low percentile values (P0 and P25), a big gradient can be observed for fixed hour; this pattern is not relevant under clear-sky conditions, which could confirm a relation with cloud cover. With respect to cloudy partial hours, the effect of the SZA is the most significant one.
The average hourly UVER irradiances for all, clear, partial cloudy and overcast conditions are shown in Figure 4. It can be observed that the average values are highly statistically significant; maximum values appear in spring and autumn when ozone column is maximum or minimum, respectively. It is also appreciated the gradient increases in summer months and that the average erythemal values are influenced by the optical path through the atmosphere and by the clouds attenuation.
Figure 5 shows the SD isolines for the monthly average hourly UVER values. For all-sky conditions, the gradient increases in summer and spring months in the central hours of the day. The distribution is near bimodal with two maxima (one of them is a relative maximum) under all sky conditions. The maximum SD for clear sky is obtained in May; this trend can be caused by several attenuation sources. In fact, they can be influenced by cloudiness evolution and by the atmospheric turbidity-aerosol loadings that can produce scattering process in the lower atmosphere, Jacovides et al. (2009).
The isolines of the monthly average hourly erythemal solar UV skewness and kurtosis explain the asymmetry of the distribution and are shown in Figure 5. It can be observed that the skewness for all-sky conditions are negative for summertime and positive in winter, which is associated with highly asymmetrical distribution with a left tail during summer; the tail could be due to the low values attenuated by clouds and by the TOC, in spring and summer months.
Kurtosis isolines of the monthly average hourly erythemal solar UV are also shown in Figure 5; it is the fourth moment of the distribution and it is a sign of whether the data are peaking or are flat related to a normal distribution. Datasets with positive kurtosis tend to have a distinct peak near the mean that declines rather quickly. It can be seen in Figure 5, for all sky conditions, that the data have positive kurtosis in May, June and July; these values correspond with a leptokurtic distribution, left tail. This fact is due to the clear conditions which are responsible for the narrow peak distribution and owing to some cloudy days which are responsible for the left tail. For the rest of the months the kurtosis values are near zero or negative (October, November and December) and, therefore, the distributions are mesokurtics. Similar results were obtained in Southwester Spain by Serrano et al. (2006).
3.2.3. Analysis of the accumulated daily UVER
In the study of the biologically active erythemal solar UV radiation, it is very interesting to know the necessary time to have a fixed amount of irradiation, this magnitude is the accumulated erythemal solar irradiation through a period of time (kJm−2); this information is relevant in studies on the biological effects of UV radiation and vitamin-D synthesis production.
From the monthly mean hourly erythemal irradiance along the period 2002–2008 the accumulated hourly UVER has been evaluated. Figure 6 shows the evolution of accumulated hourly UV values for each month. Adding up these for all the days and months, the accumulated UVER along a year presents a value of 29 kJm−2.
3.2.4. Analysis of the daily UVER
Daily and monthly average erythemal UV irradiation values have been calculated. Figure 7 shows the results with the greatest fluctuations in spring season. It can be seen that the variation of the monthly values (continuous smooth curve) is quite regular, with the maximum taking place in July and the minimum in December. Daily UVER increases in spring to summer being its slope smaller than the decrease in autumn. The different slops can be explained by the total ozone effects. During the summer months, when SZA leads to high UVER irradiance, the TOC is declining; this effect drives the maximum levels from June to July. Similar results were obtained by Frederick et al. (2000) at different stations in the United States of America.
Table III shows the statistical characteristics of daily UVER dataset for all-sky conditions. The median values are higher than the average ones and the SD increases in May and in summer months; the coefficient of variation shows the lowest values in July and August, which means that the highest stability is observed in the summer months and the attenuation by aerosols could be compensated by the diminution of TOC.
Table III. Statistical estimators of the daily UVER (kJm−2) in Valladolid, Spain, for the period 2002–2008
The differences between the minimum and the percentile P5 are quite high, exceeding 40% of the absolute minimum values in September; this result means that minimum values cannot be considered as representative of the erythemal radiation characteristics received in Valladolid, (Spain), and correspond to atypical values.
3.3. Attenuation of the UVER
In order to know the attenuation that UVER experiences in its passage through the atmosphere, the irradiation corresponding to clear days has been evaluated. For this reason, first the number of clear days for each month in the experimental period was calculated using the clearness index criteria, and then, the number of measured UVER data under clear conditions are shown in Table IV, being a total of 377 clear days with UVER measurements along the period.
Table IV. Number of clear days with UVER measurements in Valladolid, Spain, for the period 2002–2008
From the UVER daily data, the decadal (average of 10-day consecutive values) clear-sky UVER, called UVERclear, were evaluated. Figure 8 shows the evolution of erythemal extraterrestrial, called UVER0, UVERclear and measured UVER monthly average decadal values of erythemal radiation, respectively. UVER0 was calculated using the erythemal solar constant of 10.1 Wm−2 obtained using the erythemal action spectrum (McKinlay and Diffey, 1987) and the solar spectrum proposed by Gueymard (2004). The variation of UVER in Figure 8 shows slight relative minimum, the first occurred in the last ten days of February; the second one took place in the third decade of March and the third one in the second decade of April. This evolution is according to the TOC influence (McKenzie et al., 1991).
The maximum corresponded to the third decade of June with a value of 5 kJ m−2. The difference between the UVERclear and UVER measured values is greater in the spring and summer time than in autumn and winter.
Comparing the variation of UVER extraterrestrial and UVERclear, it can be seen in Figure 8 that both curves are symmetrical, although with different values. At the end of spring the difference diminishes because of UVERclear values are not affected by cloudiness but the influence of TOC which reaches its maximum in spring can be observed.
The ratio UVER measured to UVER extraterrestrial, UVER/UVER0, has been evaluated from the decadal values shown before. This ratio represents the percentage of energy which, on the average, is transmitted by the atmosphere and may be considered as the atmospheric transparency under average conditions (Bilbao et al., 2008).
The ratio of UVERclear to UVER extraterrestrial, UVERclear/UVER0, gives information about the atmospheric transparency without clouds but with aerosols, although in low proportion. The ratio of UVER measured to UVERclear (UVER/UVERclear) is also interesting and represents the observed UVER irradiation fraction, which corresponds to a cloud-free sky and mean normal transmission; due to the fact that UVER is smaller than UVERclear, this attenuation is attributed to cloudiness and aerosols in quantity superior to the value used in the definition of UVERclear.
The mean decadal values of the UVER irradiation attenuation at the measurement station site for each month of the year are shown in Figure 9. It can be seen that fluctuations of the greatest amplitude are shown in the UVER/UVERclear decadal values; it is normal as UVERclear is smaller than UVER0 and the quotient is more sensitive to UVER variations. In addition, UVER/UVER0 is in phase with UVER/UVERclear although the latter shows higher values.
Figure 9 also shows that both UVER/UVER0 and UVERclear/UVER0 follow the same pattern. As a result of the higher values of UVERclear (mentioned above), the values of the cloud-free atmospheric transparency are higher. The transparency UVER/UVER0 increases from spring to summer and decreases from summer to winter showing relative maxima in spring and September.
With regard to the curve UVERclear/UVER0, the mean decadal values show slight fluctuations in summer with a maximum of 2%; it diminishes in winter and shows a remarkable maximum in March. This study could be concluded saying that the atmosphere, considering all days or only days with clear solar UVER radiation, would have similar transparency, 2%, and it could absorb at least 98% of it. It can be said that the total attenuation would be 98%, due to all atmospheric constituents.
The UVER/UVERclear ratio gives information about the contribution of cloudiness to UVER attenuation. The transparency in summer is about 90% (Figure 9), which means that clouds cause 10% of attenuation in this season. In spring, March and April, the transparency of the atmosphere due to clouds is 70% or the attenuation by clouds is 30%.
4. Summary and conclusions
The biologically active UV solar radiation, UVER, has been analysed to characterise the statistical distribution parameters and the attenuation by atmospheric constituents.
The study has permitted to know that the measurement period (1488 days) has 25% clear, 57% partly cloudy and 18% overcast days, respectively, being the total of clear days in the same period 377.
With respect to the hourly average irradiance values, the analysis shows that the maxima may be considered as representative of the UVER characteristics in Valladolid, whereas the minima are quite atypical, mainly during the summertime.
The variability index, V, shows that the highest stability corresponds to the summer time at midday hours, mainly during June. The study shows that the optical properties of the atmosphere are more variable in spring than in summer with May the month with the largest variations.
The accumulated values of erythemal UV solar radiation show an annual irradiation of 29 kJm−2. The percentile distributions were analysed by isoline plots for different sky conditions and we can conclude that cloudiness is an attenuation factor affecting the low percentiles on the symmetry and the gradient. Monthly hourly average erythemal isolines present a high symmetry for clear sky and a 95% confidence interval with high statistical significance.
Data series have positive kurtosis values in May, June, July and autumn months, that corresponds with a leptokurtic distribution left tail. This is because of clear conditions, responsible for narrow peak distribution and also owing to some cloudy days responsible for the left tail; for the rest of the months, kurtosis is near or just below zero (October, November and December) and, therefore, the distributions are mesokurtic.
In relation to the attenuation of erythemal solar radiation in its passage through the atmosphere, it has been found that UVER transparency along its path to the ground is about 90%. This means that clouds cause 10% of attenuation in summer; in March and April, the transparency of the atmosphere due to clouds effect is 70% and the attenuation is obviously 30%.
The results have been compared with the ones obtained by Serrano et al. (2006), in Badajoz, Spain, and they showed that the performance is similar though the UVER levels in Valladolid are a little higher than theirs. The dissimilarities could be attributed to the different altitude and atmospheric constituents in both places. On the other hand, solar station at Valladolid has fewer clear days than Badajoz. Therefore, this study can help understand the evolution and characteristics of UVER radiation in Spain.
The authors gratefully acknowledge the financial support extended by the Education and Research Spanish Ministry through the Project CGL2009-08097 (CLI) and the Autonomous Government through the project G220 (2008–2011). The anonymous reviewers for their useful comments and suggestions in improving the paper are highly acknowledged. The authors also thank OMI International Science Team for the remote sensing data used in this study.