Indications for solar influence on radon and thoron in the atmosphere, Arad, Romania

Abstract

Radon (Rn-222; Rn) and Thoron (Rn-220; Tn) from the lowermost atmosphere in Arad, Romania, were investigated from 1993 to 2015. Daytime and night-time measurements exhibit (i) systematic differences; (ii) large annual variations; (iii) semi-annual and possibly ternary annual periodicities; (iv) amplitudes of the annual and semi-annual periodicities differ between daytime and night-time measurements. The day of year of the summer peak time (SPT) of the annual variation of composite Rn and Tn data depict smoothly covarying SPT patterns separated around 50 days. SPT patterns of daytime and night-time of Rn and Tn depict additional separations of gradually varying multi-year patterns. For both radionuclides, the night-time SPT precedes the daytime SPT—by around 105 days for Tn and around 60 days for Rn. In the case of temperature, the SPT of daytime and night-time measurements overlap at around day 195. Thus, for Rn and Tn, daytime and night-time differences depend on the time of measurement within the 24 h cycle. This implies that Rn and Tn measurements are influenced by the rotation of the Earth around its axis. It is suggested that these relations, observed in coexisting Rn and Tn, are a further indication for a solar (extraterrestrial) influence on nuclear radiation, previously suggested for radon.

1. Introduction

Radon (Rn-222; Rn) and thoron (Rn-220; Tn) are radioactive noble gases derived, respectively, in the nuclear decay chains of U-238 and Th-232 with half-lives of 3.82 days (Rn) and 55 s (Tn). Comprehensive compilation of the decay properties of Rn and Th are given by Baskaran [1]. Both occur in nature, in the crust, air and hydrosphere. Being radioactive noble gases promotes the suggestion of their application for tracing processes in the natural environment, relating mainly to transfer and transport phenomena.

The main conventional frames of interpretation ([1,2], and references therein) try to associate radon variability to factors and processing in media near the surface of the Earth affecting the release, transfer and transport of radon and thoron. Temporal patterns of radon and thoron are composed of periodic signals having wavelengths spanning from annual to subdaily and non-periodic variation of multiday to subdaily scale. The periodic phenomena in the annual and daily frequency bands often demonstrate similarities to variation patterns of atmospheric phenomena, mainly ambient temperature and also barometric pressure. The fact that observation of these radioactive noble gases is performed in air led, naturally, to trying to attribute the variations primarily to the influence of atmospheric variations.

Numerous observations of Rn and Tn indicate unique and complicated variation patterns which are difficult to account for in terms of simple physico-chemical terms. An overview of works in the last decades shows that interpretation of radon and thoron time series is challenging when trying to present coherent interpretation schemes that can account for the physical parameters of radon, the observed variability and the boundary conditions imposed by the environmental situations. These difficulties limit their use.

Concerning radon and thoron in air the phenomena we are dealing with are temporal variations which have in common fluctuations of a mode similar to those occurring in atmospheric variations, primarily in terms of wavelength and periodicity. Cleary the fundamental periodicities underlying these variations are of diurnal and annual scale which are generated by the well-known rotational relations in the Earth and Sun system. In recent years, a different approach has been proposed for radon variation where the periodic components of variation are attributed to an influence of a component of solar radiation which is modulated by the rotational relations in the Earth–Sun system [3]. In this approach, the rotational relations of the Earth–Sun system independently drive the annual and daily periodicities in both atmospheric variation patterns and in radon. The similarities in the evolving patterns, mainly in the time domain, lead unsurprisingly to frequent suggestions that attribute the apparent patterns in the time domain as indicative of causality.

This contribution presents a further case where eventual extraterrestrial influence is observed in a multi-year experiment using simultaneous measurement of the radionuclides of radon and thoron.

2. Methods

Measurements were performed at the Arad Radioactivity Laboratory (ARL), in the city of Arad (Western Romania, elev. 125 m). The laboratory belongs to the National Surveillance Network of Environmental Radioactivity. Full details are given by Florea & Duliu [4].

Air uptakes, collected 2 m above ground, were performed for 5 h twice in the 24 h cycle:

• (1) between 2.00 and 7.00 (night uptake)

• (2) between 8.00 and 1.00 (day uptake).

Using a high-volume air sampler the outdoor air was pumped through a glass fibre filter at an uptake rate of 5 m³ h⁻¹. ²²²Rn and ²²⁰Rn concentrations are based on the activities of their progeny, which were determined by measuring the gross beta activity of particulate matter retained on the filters. Filter activity was determined by using an NaI(Tl) scintillator system for 20 min total counting interval. The first 3 min (first measurement) are used to calculate the activity of ²²⁰Rn and a second 3 min measurement at 20 min after collection to determine the activity of ²²²Rn. The determination of both ²²⁰Rn and ²²²Rn concentration took into account the half-lives of the short-time progeny. The estimated errors for the activities are 3–11% for ²²²Rn and 2–25% for ²²⁰Rn [4].

For processing, the date of the measured data was converted to decimal days where day 0 = 1.1.1992, as presented in electronic supplementary material-1. This decimal time scale (as used by the GSI) is convenient for the numerical processing of geophysical time series. The dataset of each time series was normalized.

3. Results

Florea & Duliu [4] presented initial results on an 18 year (1993–2010) long and continuous time series of coeval measurement of radon and thoron in the lowermost atmosphere. Extended time series are used and re-examined for indications of an influence of a component in solar radiation. The possibility of such influence in radon was first raised based on observations from the geological environment [5,6] and later substantiated in laboratory simulation experiments [3,7].

For an overview, the separate daytime and night-time series were recombined to a single composite time series consisting of two measurements per day. In the upper plots, figure 1 presents the overall variation pattern of Rn and Tn over 23 years. The observed highest peaks occurred on 4 December 2000 and 28 September 2011. The normalized values, shown in the lower plots, indicate an overall similar relative variability (similar scale of y-axis) in spite of the large difference in count rate (upper plots).

Figure 1.

(a–d) Daytime and night-time series (composite data) of radon (Rn) and thoron (Tn) from Arad, Romania. Measured values are shown in the upper plots and normalized values (i.e. linear detrending and division by the standard deviation) in the lower plots.

Figure 2 presents the detrended and normalized time series of the same data separated to the daytime and night-time data. It shows similar variation patterns in both time series.

Figure 2.

Time series of measured values of Rn (a,c) and Tn (b,d) at daytime (a,b) and night-time (c,d).

Using data collected in the years 1993–2010 Florea & Duliu [4] showed, for Rn and Tn at the annual scale, that systematically different radiation patterns occur at night-time and daytime. This feature is further elaborated below. Figure 3 presents the time series of daytime and night-time measurements in the years 1993–2015. These plots correspond to the plots in figure 1 of Florea & Duliu [4]. For both radioactive isotopes, the overall night-time values are high relative to the daytime values.

Figure 3.

Normalized time series of Rn (a,c) and Tn (b,d) at daytime (a,b) and night-time (c,d). The large differences observed in the measured values (figure 2) are diminished when using normalized values.

Spectra (FFT) of the de-trended and normalized values are shown in figure 4 for the composite time series and in figure 5 for the separated daytime and night-time measurements. The annual variation (1 yr⁻¹) is the dominant frequency in all sets. Further periodicities are evident: (i) a semi-annual (2 yr⁻¹) in Rn-daytime and in Tn; (ii) a ternary annual (3 yr⁻¹) periodicity in Tn, both at daytime and at night-time. Comparison of daytime and night-time power spectra of Rn indicates a significantly stronger (×2) annual periodicity at night-time, and a relatively stronger semi-annual periodicity during daytime. The annual periodicity of Tn is also significantly stronger (×5) at night-time, and a clear semi-annual periodicity is indicated (and possibly also ternary annual periodicity).

Figure 4.

(a,b)Power spectra of Rn and Tn composite time series shown in figure 1 (normalized values). Horizontal line indicates power of 1%.

Figure 5.

(a–d) Power spectra of day time and night-time variation of Rn and Tn, shown in figure 3. Horizontal line indicates power of 1%.

The temporal variation patterns of Rn and Tn show a maximum in summer. The specific timing, within the annual scale (year), of the primary annual variation is examined below. The summer peak time (SPT) is determined within each annual cycle and presented as the day of year (DOY). To this end, the annual variation was derived from the time series of the normalized values. Using the continuous wavelet transform (CWT; Morlet 8 wavelet) allows a good location in time of the summer maxima at a frequency of 1 yr⁻¹. This variation was reconstructed in the range of 0.950–1.028 yr⁻¹, which overlaps the periodicity of 1 yr⁻¹. The CWT reconstruction of the annual signal for the composite measurements for Rn and Tn is shown in figure 6 (1992–2011). A clear time offset is observable over the whole time range, with the Tn signal preceding the Rn signal. Using these curves, the timing of the SPT is extracted from the annual maximum.

Figure 6.

Continuous Wavelet Transform (CWT) reconstruction of annual variation (0.950–1.023 yr⁻¹) of Rn and Tn (composite) time series (shown for the years 1993–2010). See text.

Figure 7 shows for the composite time series of Rn and Tn the variation of SPT in an interval of 18 years (1993–2010). A distinct separation of around 80 days within the annual cycle occurs between SPT of Rn and Tn. Furthermore, multi-annual variation of SPT occurs, partly concordant among the SPT curves.

Figure 7.

The variation of summer peak time (SPT) as a function of the day of year (DOY; bottom) and as a function of annual phase (top) for Rn and Tn. SPT are derived from the normalized composite time series (figure 1). See text.

The CWT reconstructed time series of the annual variation is used to determine the maximum in summer presented as the DOY of the SPT in each year (table 1). Figures 7 and 8 show the variation of DOY of Rn and Tn over the span of 23 years. Rn and Tn depict non-random smoothly covarying DOY patterns separated by around 50 days. Moreover, multi-year and longer-term variation are also observed. As a remark at this early stage—if the dissimilar DOY patterns are confirmed with further observations—this may be related to different properties of the nucleus of Rn and Tn and their response to the assumed remote influence.

Table 1.

Summer peak time (SPT) of radon and thoron (composite values).

	radon (Rn)		thoron (Tn)
year	phase (yr−1)	DIY	phase (yr−1)	DIY
1993	0.8255	302	0.6188	226
1994	0.8227	301	0.6434	235
1995	0.8118	297	0.6297	230
1996	0.7995	292	0.5859	214
1997	0.7967	291	0.5339	195
1998	0.8063	295	0.5421	198
1999	0.8131	297	0.6023	220
2000	0.8131	297	0.6133	224
2001	0.7940	290	0.5749	210
2002	0.7598	278	0.5311	194
2003	0.7379	270	0.5229	191
2004	0.7461	273	0.5366	196
2005	0.7625	279	0.5558	203
2006	0.7639	279	0.5667	207
2007	0.7488	274	0.5667	207
2008	0.7351	269	0.5722	209
2009	0.7337	268	0.5832	213
2010	0.7529	275	0.6105	223
2011	0.7734	283	0.6407	234
2012	0.7762	284	0.6461	236
2013	0.7693	281	0.6078	222
2014	0.7830	286	0.5613	205
2015	0.8118	297	0.5722	209
mean	0.7799	285	0.5834	213
s.d.	0.0298	11	0.0379	14

Figure 8.

(a–d) Compilation of SPT multi-year patterns in terms of DOY and phase (x-axis). See text.

A similar analysis procedure was applied to the daytime and night-time series of Rn and Tn (figure 3) as well as ambient temperature (table 2). Using a joint DOY scale, figure 8 presents a compilation and comparison of DOY patterns of the different components. The overall pattern for Rn and Tn (shown in figure 7) is reproduced in the upper-left plot. The DOY patterns of daytime and night-time of Rn are shown in the upper right, and for Tn in the lower right. The DOY patterns of daytime and night-time ambient temperature is shown in the lower left.

Table 2.

Summer-peak-time (SPT) of daytime and night-time time series.

	thoron (Tn)				radon (Rn)				temperature
	night-time		daytime		night-time		daytime		night-time		daytime
year	phase (yr−1)	DIY	phase (yr−1)	DIY	phase (yr−1)	DIY	phase (yr−1)	DIY	phase (yr−1)	DIY	phase (yr−1)	DIY
1993	0.6188	226	0.8994	328.5	0.7639	279	0.9158	334.5	0.5366	196	0.5394	197
1994	0.6434	235	0.8775	320.5	0.7639	279	0.9158	334.5	0.5394	197	0.5448	199
1995	0.6297	230	0.8638	315.5	0.7474	273	0.9158	334.5	0.5394	197	0.5476	200
1996	0.5859	214	0.8939	326.5	0.7255	265	0.9185	335.5	0.5421	198	0.5503	201
1997	0.5339	195	0.9487	346.5	0.7201	263	0.9131	333.5	0.5366	196	0.5503	201
1998	0.5421	198	0.9432	344.5	0.7392	270	0.9103	332.5	0.5366	196	0.5476	200
1999	0.6023	220	0.9240	337.5	0.7529	275	0.9103	332.5	0.5366	196	0.5448	199
2000	0.6133	224	0.9185	335.5	0.7502	274	0.9158	334.5	0.5366	196	0.5448	199
2001	0.5749	210	0.9076	331.5	0.7201	263	0.9185	335.5	0.5339	195	0.5421	198
2002	0.5311	194	0.8747	319.5	0.6762	247	0.9240	337.5	0.5311	194	0.5448	199
2003	0.5229	191	0.8282	302.5	0.6571	240	0.9213	336.5	0.5339	195	0.5503	201
2004	0.5366	196	0.8392	306.5	0.6735	246	0.9076	331.5	0.5394	197	0.5558	203
2005	0.5558	203	0.8583	313.5	0.6982	255	0.8912	325.5	0.5394	197	0.5530	202
2006	0.5667	207	0.8720	318.5	0.7091	259	0.8775	320.5	0.5394	197	0.5476	200
2007	0.5667	207	0.8665	316.5	0.6982	255	0.8693	317.5	0.5366	196	0.5421	198
2008	0.5722	209	0.8227	300.5	0.6899	252	0.8665	316.5	0.5394	197	0.5394	197
2009	0.5832	213	0.7899	288.5	0.6899	252	0.8720	318.5	0.5394	197	0.5366	196
2010	0.6105	223	0.8090	295.5	0.7036	257	0.8912	325.5	0.5394	197	0.5366	196
2011	0.6407	234	0.8255	301.5	0.7173	262	0.8966	327.5	—	—	—	—
2012	0.6461	236	0.8337	304.5	0.7146	261	0.8939	326.5	—	—	—	—
2013	0.6078	222	0.8665	316.5	0.6954	254	0.8912	325.5	—	—	—	—
2014	0.5613	205	0.9213	336.5	0.6982	255	0.8994	328.5	—	—	—	—
2015	0.5722	209	0.9240	337.5	0.7365	269	0.9103	332.5	—	—	—	—
mean	0.5834	213	0.8743	319	0.7148	261	0.9020	329	0.5375	196	0.5454	199
s.d.	0.0379	14	0.0444	16	0.0293	11	0.0176	6	0.0027	1	0.0054	2

The variation of ambient temperature is a well-known environmental parameter. Its annual variation contains a distinct maximum in summer. As expected, the DOY patterns of daytime and night-time temperature in Arad overlap, and present a stable pattern around day 195. On the other hand, the SPT patterns of Rn and Tn show markedly different patterns. In all cases of Rn and Tn gradually varying multi-year patterns occur. The separation among variation patterns of the composite measurements of Rn and Tn (figure 8a) is accompanied by a further separation among the daytime and night-time measurements (figure 8b,d). In the case of both these components, the night-time DOY precedes the daytime DOY, around 105 days for Tn and around 60 days for Rn (table 2). This result clearly demonstrates that the determination of the DOY depends on the time of measurement within the 24 h cycle. A 24 h periodicity is a fundamental feature of the Earth–Sun system. Thus the dependence on the time of measurement (at a 24 h scale) implies that Rn and Tn measurements are influenced by the rotation of the Earth around its axis.

4. Discussion

For several decades radon (and thoron) in air have studied in natural environs and in dwellings. Large temporal variations are recorded consisting of periodic signals of annual and daily scale as well as non-periodic signals of mainly multi-day scale. These signal types have similarity with variations encountered in atmospheric parameters, primarily temperature and pressure. This similarity and the fact that radon is hosted in air led investigators to assume a connection between variability of radon and variation patterns of atmospheric parameters. The general approach and assumption of atmospheric variation as drivers of radon variability turns out to be highly unclear when assessing the large accumulated sets of observations from very different scenarios. Cases interpreted to be indicate of positive, negative and non-correlation for the influence of P and T are abundant. Analysis of measurements relied mainly on visual inspection in the time domain, rarely in the frequency domain. Correlation diagrams (when applied) led to inconclusive patterns which certainly cannot be interpreted to prove causality. In these works, such conclusions are assumed as there is no other frame of interpretation. Substantial evidence especially in terms of experimental simulation is not presented. The disconformity in the raised interpretation is recognized in some of works, leading even to the statement that radon variability is ‘unexplainable’ and therefore also useless. Extensive investigations at the GSI indicated the need for a different approach. A discussion of the issue is detailed in electronic supplementary material-2.

The derived analysis patterns in time and frequency domains portray systematic multi-year variations of two radioactive noble gas components in the lowermost atmosphere, which are derived in two independent nuclear decay chains. They indicate that the measurements/results cannot be attributed to reflect an instrumental (measurement) artefact. Furthermore, the patterns cannot be attributed to influence environmental atmospheric parameters, such as temperature or pressure. The clear manifestation of a semi-annual (and possibly a ternary-annual) periodicity alongside the annual periodicity supports this view.

Nuclear radiation from Rn and Tn contains a large annual signal which is manifested relatively strongly in daytime measurements. The phase of the annual signal, mirrored in the SPT, varies gradually at a multi-year scale for each radionuclide. The composite daily variation of Rn and Tn demonstrates a significantly different phase. The phase (SPT) of the composite measurements is composed of distinctly separate variation patterns when examined for daytime and night-time measurements. The occurrence of a day–night effect in Rn and Tn clearly indicates a relation to the rotation of the Earth around its axis. These considerations suggest that the observed variation of Rn and Tn does not reflect a simple variation of their concentration but probably rather a variation of nuclear radiation from Rn and Tn in the atmosphere.

The overall time-varying patterns of SPT and their relations suggest the influence of a component in solar radiation. The specific considerations leading to this assumption are:

• (1) non-relation to atmospheric variation

• (2) different SPT for Rn and Tn

• (3) indication for a daily modulation of the annual variation, due to the rotation of the Earth around its axis.

Timing and spacing of the highest peaks in the composite time series (figure 1) correspond to the timing and spacing of the peaks in solar activity as measured by sunspot number (see https://www.swpc.noaa.gov/news/solar-cycle-24-status-and-solar-cycle-25-upcoming-forecast). Our data show that the highest peaks occurred on 4 December 2000 and 28 September 2011, corresponding to peaks in sunspot activity. This observation supports our inference that the periodicities we have detected in our Rn-222 and Rn-220 data have an extraterrestrial influence.

The outcome of this study should be evaluated in the frame of the overall characteristics of the radon signals in geogas. Extensive investigations performed in Israel led to raising the unconventional possibility that the periodic components (daily, annual) of the variation are linked directly to the rotational relations of the Earth–Sun system [5,6,8]. This conclusion led to the suggestion that a component (unidentified) in solar radiation influences these signals—i.e. an extraterrestrial influence. Results, performed along similar lines, from field sites outside Israel [9,10] support the latter idea of an extraterrestrial influence. This serves as an indication that such influence operates at different locations on the Earth, at very different geologic scenarios and to a depth of hundreds of meters and even 1 km in the crust. Results from laboratory simulation experiments by GSI of radon signals support this proposition [3,11–13]. A key simulation is a long-term multi-sensor reference experiment (EXP #1), acquiring data at a resolution less than 1 h. Using time series spanning 3.5 years a linkage of radon signals in the simulation experiment and in geogas is established, based on similar geophysical statistical characteristics of the signals in the time, frequency and frequency-time domains [3,14,15]. The results of the simulation experiments indicate again, and probably also confirm, that a component in solar radiation influences the periodic signals in radon time series.

The outcome of the present investigation conforms with the above-mentioned recent observations on the Rn system in the geological environment and in laboratory simulations. This investigation indicates in addition:

• (1) Observation of eventual solar influence is extended also to Thoron, in addition to radon. Thus, such processes are not unique to radon.

• (2) Decay of Rn and Thoron occurs in separate and independent radioactive chains. This further opens the possibility that other isotopic species are also influenced by solar radiation.

• (3) The specific response of Rn and Tn to the solar influence indicates that this may be related to different intrinsic properties of the nucleus of Rn and Tn and their response to the assumed remote influence.

Data accessibility

The datasets supporting this article are available via file: electronic supplementary material-1 ARAD data file.

Authors' contributions

N.F. collected the data, organized it and preprocessing. G.S. and O.P. performed the data analysis and laid out the manuscript. All the authors gave their final approval for publication.

Competing interests

We have no competing interests.

Funding

The Geological Survey of Israel supported the data analysis.

Dedication

This manuscript is dedicated to prof. Gideon Steinitz, of blessed memory, the main author of this work who passed away shortly before publication, a special open minded individual who was an unlimited source of new ideas.