Introduction

Atmospheric 7Be has been widely studied at different places around the world, like Europe [eg. 1, 2], America [3], or Asia [4, 5]. This radionuclide is a potentially useful tracer of aerosols due to its short half-life (T 1/2 = 53 days), ease of measurement and well-defined source term. Most of 7Be is produced in the lower stratosphere and slowly transported to the surface. 7Be plays a role of atmospheric tracer and its measurements provide an important clue on atmospheric air mass motions. Thus, several works have used the radionuclide to analyze vertical transport of air masses of stratospheric origin, atmospheric long-range transport, as well as determine the air-sea exchange process [6–8]. Others have focused on estimating soil erosion rates using 7Be [9, 10].

7Be is a natural cosmogenic radionuclide produced in the upper atmosphere when cosmic-ray-produced neutrons and protons interact with the nuclei of nitrogen and oxygen and disintegrate them into lighter fragments in a process known as spallation [11–14]. Recent research points out the photo-nuclear reaction as another possible mechanism of 7Be formation in the upper atmosphere [15]. Once the radionuclide is formed, it is attached quickly to aerosols with aerodynamic diameter between 0.4 and 2 μm [16–20].

Previous studies [3, 21–24] have reported 7Be production rate has altitude dependence because of the atmospheric attenuation of the radiation (around 70 and 30 % of 7Be is produced in the stratosphere and in the troposphere respectively), and latitude dependence due to the deflect effect of the magnetic field of the Earth (increasing at high latitude). Moreover, variations of the 7Be concentration in the air show a seasonal pattern with a spring–summer maximum and an autumn–winter minimum [e.g. 2, 3].

Thus, galactic cosmic rays and solar energetic particles entering the heliosphere are affected by interplanetary magnetic field and solar activity [13, 14, 22, 25]. In addition, meteorological parameters such as precipitation scavenging, air exchange between the stratosphere and the troposphere and vertical transport influence on atmospheric 7Be concentrations [1, 3, 5, 24, 26–29].

The aim of the present study was to investigate the effect of the solar activity, cosmic rays and the presence of airborne particulate matter on the variation of airborne 7Be radioactivity surface levels during a complete solar cycle.

Materials and methods

Study area

The sampling site is located in the north-west of the city of Málaga, approximately 5 km from the coastline, close to the airport and surrounded by roads with traffic exhaust (Fig. 1). Málaga is the major coastal city of the Andalusia region, South Spain, with a population of approximately 570,000 inhabitants. The city lies on the Mediterranean coast and it is bordered to the north by a high mountain range. Thus, Málaga is affected by a Mediterranean climate with dry and hot summers, temperate winters, scarce rainfall with mean annual precipitation of around 550 mm year−1, and a well developed land-sea breeze regime [30].

Fig. 1
figure 1

Location of the study area

Full size image

Sampling procedure

In the period January 1997–December 2007, 132 aerosol samples were collected weekly on the roof of a three-story building on the Faculty of Sciences of the University of (4.49°W; 36.73°N) at a height of 10 m above the ground to minimize the resuspension from soil. Aerosol samples were collected in cellulose membrane filters with a pore size of 0.8 μm and a diameter of 47 mm diameter as part of a national monitoring programme coordinated by the Consejo de Seguridad Nuclear. The collection efficiency amounted to 99.99 % with an air sampler (Radeco, mod AVS-28A) at a flow rate of 40 l/min. A monthly composite sample containing 4–5 filters was formed (average volume of 1,600 m3) for the 7Be determination. Measurements by high-resolution gamma-ray spectrometry were performed to determine the 7Be activities of the samples using a coaxial-type germanium detector (Canberra Industries Inc., USA). The peak analysis of 7Be (I = 10.52 %, 477.7 keV) was done using SPECTRAN at peak analysis software. The spectra acquisition time was 172,800 s. Details of the low-background gamma-ray detection system used and the calibration of the resulting samples for gamma spectrometry have been previously described by Dueñas et al. [27, 31].

Data sets

In order to describe the 11-year cycle of the solar activity, the monthly sunspot number and the Solar energetic proton (SEP) obtained from NOAA’s Space Weather Prediction Center (SWPC) were used. Shielding effect was shown using magnetic field data measured by various spacecraft near the Earth’s orbit. These data are provided by the Space Physics Data Facility (SPDF) at the Omni2 directory. Cosmic ray data were obtained from the neutron monitor belonging to Sodankyla Geophysical Observatory of the University of Oulu in Finland. Finally, atmospheric particulate matter was indirectly derived from the Aerosol Optical Depth measured by the Sea-viewing Wide Field-of-view Sensor (SeaWiFS) Deep Blue Level 3 operated by NASA.

Results and discussion

7Be concentration levels in surface air

7Be in ground level air over a 11-year period, was measured in Málaga (Spain). Table 1 gives the summary of the descriptive statistics such as arithmetic mean (AM), geometric mean (GM), standard deviation (SD), skewness (SK), maximum and minimum values and the coefficient of variation (CV). The activity concentrations of the 7Be ranged between 2.47 and 8.12 mBq m−3, with an average of 4.71 ± 1.13 mBq m−3. These values are coincident with those measured in the regions of similar latitude reported by Chae et al. [5] at Daejeon (36°N), Piñero-García et al. [32] at Granada (37°N), or by Narazaki and Fujitaka [33] at Dazifu (33°N).

Table 1 Descriptive statistical parameters of monthly 7Be activity concentrations (mBq m−3)
Full size table

The frequency distribution of the concentrations of 7Be is illustrated in Fig. 2. As seen in this figure, the plot of 7Be specific activity measured in aerosols at surface during this period showed skewed histogram which fits well with a log-normal distribution. The goodness of the fit was tested using the Kolmogorov–Smirnov test with a confidence level of 90 %.

Fig. 2
figure 2

Frequency histogram for monthly 7Be atmospheric concentrations for the data collected from 1997–2007. Black line represents lognormal fitting

Full size image

Annual and monthly variation of 7Be

Inter and intra-annual variability was expected and observed. Atmospheric 7Be was below average in 2000 and 2004, close to average in 1999, 2001, 2002 and 2003, and above average in 1997, 1998, 2005, 2006 and 2007 (Fig. 3a). These above average values correspond with the initial and final phases of the 23rd solar cycle. The significance of the differences in 7Be concentrations among the different years was assessed by the corrected Mann–Whitney test. The lowest concentrations of 7Be were found during the year 2000 (corresponding to the solar maximum activity), showing significant differences with the start or the end period of the solar cycle (1997, 1998, 2006 and 2007), when the highest concentrations are registered. Piñero-García and Ferro-García [34] for example, found that during the ascending phase of the solar cycle, solar activity played an important role in the production rate of 7Be, whereas during the descending phase, the role of the solar activity was secondary. Moreover, concentrations of 7Be show seasonal pattern with maximum values being observed in the spring–summer months and minimum in the autumn–winter months (Fig. 3b), though this variation is weaker than the one found in other studies (e.g. [2, 35]). This behaviour is representative of middle latitudes, such Málaga, due to their cosmogenic origin and the characteristics of the sampling point. Several authors have attributed the observed seasonality to a combination of the precipitation scavenging which removes from the atmosphere the aerosols where 7Be is attached, or the stratosphere-to-troposphere exchange and vertical mixing within the troposphere [1–3, 29, 36, 37]. During warm months, the troposphere is less stable due to the solar heating of the earth’s surface, which leads to upward transport of tropospheric species and hence a downward transport of stratospheric air parcels because of the convective movements.

Fig. 3
figure 3

a Boxplot of the variation in 7Be activity on a yearly basis. b Boxplot of the mean monthly atmospheric concentrations of 7Be averaged over the study period

Full size image

Influence of solar activity in 7Be variability

The solar activity modulates the heliospheric condition, which modulates the cosmic ray flux, which in turn modulates the cosmogenic isotopic production. Therefore, as 7Be is a cosmogenic isotope primarily produced by nuclear reactions of cosmic ray and solar energetic particles with atmospheric gases such nitrogen and oxygen, 7Be concentrations in surface air are affected by solar activity [38].

The sunspot number is commonly used as a solar parameter related with the solar activity [39]. Figure 4 shows the monthly average values of 7Be specific activity and sunspot number. Solar activity reaches the maximum during the 2000, while 7Be atmospheric concentration has the lowest values. Many studies have found that 7Be concentrations are strongly inverse correlated with sunspot activity (e.g. [1, 2, 4, 35]). In the present study monthly 7Be activity is not well correlated with the sunspot number and a very weak Pearson correlation coefficients (r = −0.14, p value = 0.110) was found. The weak inverse correlation observed was also found by Chae et al. [5] and Piñero-García et al. [32], who stated that meteorological factors could reduce this correlation.

Fig. 4
figure 4

Monthly concentrations of atmospheric 7Be and sunspot number. Dash-dotted lines indicate experimental values and solid lines show trend component

Full size image

The time series have four components: trend, cyclical, seasonal and irregular [40, and references therein]. The first one corresponds to the long term pattern, and the last one represents the unpredictable component, i.e. the meteorological factors. We have decomposed our data in these components. When the correlation between the trends of 7Be and sunspot number is made, a statistically significant anti-correlation is found (r = −0.59, and p value = 0). Therefore, that means that when removing the seasonal and irregular components, the relation between both variables increases and hence, the weak inverse correlation of the observed data may be elucidated by meteorological factors.

When performing that correlation on a seasonal basis, it is found that summertime has the highest correlation coefficient (r = −0.63), while winter, spring and autumn periods do not show a significant correlation. It is noteworthy that the correlation between trend components of 7Be and sunspots during the summer does not show an appreciable increase, whereas for the rest of seasons this correlation increases reaching a significant correlation. That could mean that the most apparent factor contributing to 7Be variations in summertime is the solar activity while meteorological factors have greater influence on the radionuclide variations during the rest of the year.

Solar activity is associated also with the onset of solar energetic particles, both by coronal mass ejection (the weakest events), and by prompt events originating on the disk linked to flares. The geosynchronous operational environmental satellite (GOES) provides energetic particle population at geostationary orbit. Figure 5a shows monthly variations of SEP and sunspot number during the study period. The largest pick events correspond to the maximum solar activity and the descending phase, i.e. the active periods of the solar cycle. When removing the outliers detected with Grubbs’ test, the SEP time series (Fig. 5b) shows a similar behaviour than cosmic rays (Fig. 7a).

Fig. 5
figure 5

a Time series of >100 MeV proton fluence and sunspot number. b Monthly variation of solar energetic protons without outliers

Full size image

The larger solar flares can cause massive ejection of solar matter which are related with these outliers, and correspond mainly with optical flares of importance three, that is, optical flares which corrected area in heliospheric square degreesFootnote 1 at maximum brightness range from 12.5 to 24.7 square degrees (almost 60 % of this flares are related to these outliers). Correlation between >100 MeV proton fluence without outliers and 7Be concentrations shows a weak significant positive correlation with a correlation coefficient r = 0.30. This correlation is weaker for lower energetic SEP, and it could be explained by the cut-off effects of the geomagnetic field, that only high-energy SEPs can reach the mid- and low-latitudes [33, 41].

Influence of cosmic rays in 7Be variability

It is well known that interactions of cosmic rays with the terrestrial atmosphere produce radionuclides such as 7Be, therefore there should be a positive correlation between 7Be activity concentration and the flow of cosmic rays. Cosmic rays can be deflected by both the interplanetary magnetic field and solar wind which have an effect of shielding on the Earth. Solar activity and magnetic field magnitude have similar behaviour (Fig. 6) showing a strong positive correlation (r = 0.69). In turn, magnetic field magnitude and cosmic rays are anti-correlated with a correlation coefficient of r = −0.77. Thus, a lineal model to describe the relationship between both variables shows that the 59.97 % of the cosmic rays could be explained by fluctuations of the magnetic field. Consequently, when solar activity presents the maximum, cosmic ray amount reaching the Earth is the lowest (Fig. 7a), resulting in a reduced 7Be production rates with increasing solar activity. This is in agreement with results of previous works [42–44].

Fig. 6
figure 6

Monthly variation of sunspots number and magnetic field magnitude

Full size image
Fig. 7
figure 7

a Comparison of monthly number of sunspots and flow of cosmic rays arriving at Earth for the period 1997–2007. b Monthly 7Be activity concentration plotted against cosmic ray amount. The solid line represents the linear least squares’ regression line

Full size image

The amount of 7Be that reaches the Earth’s surface is dependent on its atmospheric production rate, stratosphere–troposphere air exchange, and tropospheric circulation [45–47], and as a result, the flux of 7Be from the atmosphere to Earth and cosmic rays show large scatter and rather low significant correlation (r = 0.24, p value = 0.007) (Fig. 7b). The correlation increases considerably (r = 0.52) when performing the correlation between the trends of 7Be and cosmic rays. Then we can state that again meteorological factors may affect the relationship between both variables.

Influence of airborne particulate matter in 7Be variability

As 7Be becomes an aerosol associated species, a correlation between its specific activity and aerosol concentration is expected to be found. Aerosol optical depth (AOD) provides an estimation of sun light extinction due to atmospheric aerosols, and it is therefore an indirect measurement of aerosols distributed within a column of air. We have used the AOD derived from Sea-viewing Wide Field-of-view Sensor (SeaWiFS) Deep Blue Level 3 satellite data with the retrieval algorithm over land and ocean [48, 49] at the reference wavelength of 550 nm and 0.5° × 0.5° resolution to determine whether there is a significant correlation between aerosol concentration and 7Be atmospheric concentration.

Figure 8a depicts the comparison between AOD and 7Be activity concentrations. Both variables show a seasonal cycle with higher values in the spring and summer where differences do not reach statistical significance for 7Be values according to the corrected Mann–Whitney test, and lower values in winter and autumn with non statistical differences in both variables. This pattern was also found by Dueñas et al. [7] and Cesnulyte et al. [50]. The higher AOD in spring–summer may be explained by several contributions such as stagnant synoptic meteorological patterns or secondary aerosol photochemical formation [51], as well as the high frequency of African dust outbreaks which occurs during these months in this area [52]. 7Be spring–summer maximum could be attributed to the air exchange between the stratosphere and troposphere due to the thinning of the tropopause [6, 7] and the high occurrence of African dust outbreaks [7]. Figure 8b presents the performance of 7Be specific activity versus AOD at 550 nm. There is a significant positive correlation with a correlation coefficient of r = 0.48, ratifying the attachment of 7Be on airborne particles. When performing that correlation seasonally, it is found that only autumn does not show a statistically significant relationship. As AOD is an indirect measure of atmospheric particulate matter, and it depends on several factors like vertical structure, composition, size distribution and water content of atmospheric aerosol, further research is needed to assess the relationship between 7Be and atmospheric particulate matter.

Fig. 8
figure 8

a Monthly mean AOD550 and 7Be specific activity for the period 1997–2007. b Scatter plot of monthly 7Be atmospheric concentration versus monthly average SeaWiFS AOD550. The linear regression line is shown as the black line

Full size image

Multiple correlation analysis

A multiple linear regression model was developed to explain the variability of 7Be activity concentrations at Málaga using all the cosmogenic factors studied above (sunspot number, cosmic rays and SEPs >100 MeV), as well as the presence of atmospheric particulate matter. When using the observed values, only the SEP removing the outliers and AOD are statistical significant predictors of the concentration of 7Be measured at ground level in Málaga, explaining about 34 % of the variance. However, when the trend component of the 7Be variability is studied, cosmic rays is the unique variable which does not become a significant predictor, explaining around 56 % of the variability of 7Be trend component. Therefore, these results may indicate that cosmic rays have not strong influence on controlling the variability of 7Be activity concentration at this coastal city. Furthermore, as the explained variance is greater when trend components are used, it could suggest that meteorological parameters may mask the influence of the cosmogenic factor on the 7Be variations.

Conclusions

7Be atmospheric concentrations were measured during an 11-year period in Málaga, Spain (4.49°W; 36.73°N) covering the 23rd solar cycle. Monthly 7Be values range between 2.47 and 8.12 mBq m−3 with an average of 4.71 mBq m−3. Concentrations of the cosmogenic radionuclide display seasonal trend with the highest values being observed in the spring–summer and the lowest in autumn–winter period.

The present study contributes to the understanding the relationship between the 7Be and the variables solar activity (displayed by sunspot number and SEPs), cosmic ray and atmospheric particulate matter. During the active phase of the solar cycle, the sunspot number increases as well as the number of solar flare injections. A weak anti-correlation (r = −0.14) between 7Be and sunspot number is found in good agreement with previous reported results. When removing the seasonal and the irregular component of the time series, the correlation increases (r = −0.59) suggesting that meteorological factors may mask the solar influence.

The intense solar flares can cause massive ejection of SEPs being outliers in the time series. If these outliers are eliminated, the SEP >100 MeV time series display a similar pattern than the one of the cosmic rays. Both variables show minimum values when maximum values of sunspot number (i.e. solar active phase) are found. There is a significant positive correlation between these variables and 7Be (r = 0.30 and r = 0.24 for SEP and cosmic rays respectively), and hence an anti-correlation with sunspot number. This is due to the fact that the Sun’s magnetic field is stronger during the solar maximum, and shields the Earth from cosmic rays and solar particles. However, the larger solar flares related mainly with optical flares of importance three, are not deflected by the magnetic field and can reach the Earth’s atmosphere. Finally, a positive significant correlation is found between the AOD and 7Be atmospheric concentrations (r = 0.48).