Cold season dose rate contributions from gamma, radon, thoron or progeny in legacy mines with high natural background radiation

Abstract In areas with high natural background radiation, underground cavities tend to have high levels of airborne radionuclides. Within mines, occupancy may involve significant exposure to airborne radionuclides like radon (222Rn), thoron (220Rn) and progeny. The Fen carbonatite complex in Norway has legacy mines going through bedrock with significantly elevated levels of uranium (238U) and especially thorium (232Th), and significant levels of their progeny 222Rn and 220Rn. There are also significantly elevated levels of gamma radiation in these mines. These mines are naturally chimney ventilated and release large volumes of air to the outdoors giving a large local outdoor impact. We placed alpha track detectors at several localities within these mines to measure airborne radionuclides and measured gamma radiation of bedrock at each locality. The bedrock within the mines shows levels up to 1900 Bq kg−1 for 238U, 12 000 Bq kg−1 for 232Th and gamma dose rates up to 11 μSv h−1. Maximum levels of airborne radionuclides were 45 000 Bq m−3 for 220Rn and 6900 Bq m−3 for 222Rn. In addition, we measured levels of thoron progeny (TnP). In order to estimate radiation dose contribution, TnP should be assessed rather than 220Rn, but deposition-based detectors may be biased by the airflow of mine-draft. We present dose rate contributions using UNSCEAR dose conversion factors, and correcting for airflow bias, finding a combined cold season dose rate within these mines of 17–24 μSv h−1. Interestingly, fractional dose rate contributions vary from 0.02 to 0.6 for gamma, 0.33 to 0.95 for radon and 0.1 to 0.25 for TnP.


Introduction
Radon ( 222 Rn) and thoron ( 220 Rn) originate from radium isotopes ( 226 Ra and 224 Ra), respectively, in the uranium ( 238 U) and thorium ( 232 Th) decay series. Mining activity in areas with elevated naturally occurring background radiation is therefore often associated with radiological issues with these radioactive gasses (1)(2)(3) . Dose is given by the progeny of 222 Rn (RnP) and 220 Rn (TnP) that deposit in the respiratory tract (4)(5)(6) . Within uranium mines, levels of 222 Rn and RnP can be very high (7,8) , and the health risk to miners has been well documented (2,9,10) . Because of 220 Rn halflife (55 s), measurements and equilibrium with TnP are uncertain (11)(12)(13) and dose rates should be based on TnP measurements (14)(15)(16)(17) . For legacy mines within bedrock rich in 232 Th, it has been shown that the level of TnP can be high (2.4-2.9 WL) (3) and high levels of 220 Rn (40 KBq m −3 ) naturally chimney ventilate (1) . Knowledge about variation of 222 Rn, 220 Rn and TnP within such mines and natural ventilation and exhalation to the surface is thus needed.
In the Fen igneous complex in Norway (Figure 1), legacy mines run through bedrock where background radiation is elevated for radionuclides in the decay series of 238 U and especially 232 Th (3,18) . In the eastern part of the complex and running through redrock, iron legacy mines older than 100 years are found in Mining Hill ( Figure 2). In the western part, the Søve legacy niobium mine, also called Tuftestollen, from the 1960s runs through mainly søvite but also Fe-Dolomite carbonatite (FDC), formerly known as rauhaugite (19) . Redrock activity concentrations range up to 12 000 Bq kg −1 of 232 Th and søvite range up to 1400 Bq kg −1 of 226 Ra (3) . The natural chimney ventilation of the iron legacy mines involves a seasonal direction of mine draft with sinking airflow in summer and cold air flowing out of low lying mine-openings. In winter, the mine draft is warmer than outside air and Figure 1. Map of Fen igneous complex with radioactive bedrock and the two legacy mining sites (Søve mine and Mining Hill) but see also Dahlgren (19) for an updated and more detailed distribution of FDC. Axes give UTM coordinates (32N). Holocene deposits in gray hachure, anthropogenic deposits in orange hachure and densely populated areas in blue.
ventilates up through higher lying mine openings (1) . FDC has intermediate levels of radioactivity (3) and is because of its contents of rare earth elements (REE) attractive for new mining activities. Regarding any occupancy in existing mines or future ones because of REE mining in FDC and søvite, dose rate pathways should be assessed for gamma radiation, radon and thoron progeny in such mines. Since mine draft has opposite directions between seasons, the air at different places within mines may differ between seasons, and exposure pathways must be assessed for each season.
Here we have assessed the winter season with a mainly rising mine draft and the exposure pathways of 222 Rn, 220 Rn and TnP within the legacy iron mines (including one adit in FDC) and in the niobium legacy mine.

Study area and measurement points
The Fen complex (59.275 • N, 9.304 • E) contains bedrocks with different background radiation. Redrock in the eastern part (Figures 1 and 2) has especially elevated levels of 232 Th and was mined because of associated iron ore (hematite). Activity concentrations range from 670 to 12 000 Bq kg −1 of 232 Th and from 44 to 550 Bq kg −1 of 238 U, involving exhalation rates of 2-25 Bq s −1 m −2 for 220 Rn and 0.008-0.08 Bq s −1 m −2 for 222 Rn (3) . By comparison, søvite in the western part ( Figure 1) was mined because of niobium ore and has levels from 10 to 780 Bq kg −1 of 232 Th and from 10 to 1400 Bq kg −1 of 226 Ra (3,20) . However, the niobium mine also run through sections of FDC and it should be noted that the søvite bedrock is very heterogenous with regard to distribution of elements and radionuclides (21) . FDC is otherwise found in the middle of Fen complex ( Figure 1) with a transition zone toward redrock ( Figure 2) (19) . FDC has levels from 160 to 7000 Bq kg −1 of 232 Th and from 20 to 290 Bq kg −1 of 226 Ra (3,20) . FDC is also enriched in REE, which is attractive for new mining activities (see Reference (19) for more details).
The Mining Hill iron legacy mines form a complex system with open pits and mine openings running through mainly red rock with iron ore (Figures 2  and 3). The natural chimney ventilation of the iron legacy mines involves large volumes of air ventilating with high activity concentrations. Seasonal directions of the mine draft are mainly upwards in winter and mainly downwards in summer but variable in spring and autumn because of diurnal variations of outdoors temperatures (1) . Approximate air temperatures within the iron mines are 4-6 • C, which explains the main chimney ventilation with a rising draft in winter months with outdoor sub-zero temperatures (1) . The one-adit Søve niobium mine has two openings, one main and one higher lying opening, and field observations suggest natural chimney ventilation where the draft direction changes between seasons, being inwards from the entrance opening in winter.
The legacy iron mines at Mining Hill are dangerous with rock fall and are difficult to access, with climbing equipment being required at several places because of vertical shafts. This puts restrain on fieldwork within these mines, both regarding the number of possible entries and on how long one can stay. Therefore, measurements are limited and alpha track detectors, which are fast to deploy and collect, were used. The detectors were hung from the roof and the distance to the wall was recorded. The Søve niobium mine is one-adit and was used both to mine ore and for transport. Much of the mine system within the legacy iron mines was used for transport and would normally be termed haulage adits and haulage shafts, respectively, as for the main adit and the Bolla shaft. We chose localities spread out through much of this haulage system and these will be termed adits and shafts within this paper. Localities 1-19 are within Mining Hill. Of these, only the Hov mine has a true mine gallery, but ore was also taken out from Karup shaft. Only locality nr 1 is within the FDC zone, and it was by far the driest and most dusty locality, probably because of a combination of the near absence of cracks in the bedrock for ground water to enter and because this adit is far away from any shaft where surface water enters. At all other localities the air was damp, and water was dripping from the roof and walls at different magnitudes. The main transport (haulage) adit of Mining Hill runs from, and at level 3. Lichtloch (L) shaft base at crossing with L adit. 4-6. Lichtloch cross-cut adit at 54, 106 and 129 m E of the crossing with AC adit; transport (haulage) between western and central (Bolla) parts of Mining Hill. Nr 6 is close to the Bolla shaft and rich in iron ore.
7. Bolla shaft. Iron ore was taken out in Bolla adit which crosses this shaft, which was also used for transport. 8. AC adit at shaft base. 9. AC adit at crossing with the western adit (below). 10 Western adit, which is the northwestern most part of the adit system, and which belongs to the adit system between the Karup and AC shafts.
11-14 Karup mine shaft, which is the shaft down to the northern end of the adit leading to the AC shaft. 16-17. Main adit transport (haulage) system between Bolla mine shaft and Mundloch (Nr 16, weakly inclined). 18. Hov adit; crossing near the Hov mine gallery, which is too dangerous to enter. The Hov mine gallery is connected by a branch of the adit, which joins the main adit near Mundloch (the main mine opening near Lake Norsjø.
19. Bolla-Bolladalen cross-cut; early transport (haulage) adit, ∼50 meters above main adit (nr [16][17] and ca 24 m beneath surface. Short and with two entrances, and in such a way, standing out from the other localities above.

Gamma measurements
Measurements of gamma radiation within the legacy mines were done through gamma spectrometry using a Georadis GT-30 held at around 1 meter height at the position of the alpha track detector, directed toward the roof ( Figure 4). Six were at localities within a transition zone of mixed bedrock between the redrock and the FDC zone in Mining Hill (nr 2-4 and 8-10, Figure 2), Figure 3. Cross section of the iron legacy mines rewritten from historical records (43) , illustrating the large mine openings on the top of in Mining Hill and natural chimney ventilation through these and lower openings (1) . one within the FDC zone (nr 1), 11 measurements at localities within the redrock zone (nr 5-7, 11-14 and [16][17][18][19] and one within the søvite zone (nr 20). The instrument measures 214 Bi and 228 Ac and, assuming secular equilibrium, gives the concentrations of total uranium (where the main fraction is 238 U but 0.72% is 235 U) and 232 Th in bedrock, respectively, as well as an estimate of the ambient gamma dose rate (μSv h −1 ). For the typical activity concentrations of Fen bedrock, uncertainties (1 sigma) in gamma measurements are for 238 U up to 10% at the lowest activity concentrations but below 1% at mean bedrock levels, and for 232 Th up to 17% at the lowest measurements but below 2% at mean bedrock levels. Gamma measurements were made at all localities where alpha track detector measurements were done.

Alpha track detector measurements
To assess variation in levels of 222 Rn, 220 Rn and TnP within these legacy mines, short-period time-integrated measurements were done with alpha track detectors at different mine localities at known distances from the walls. These were exposed from 27 to 28 November in the western part of Mining Hill (  in November and 21 h in December. Detectors hung at distances ranging from 40 to 150 cm (median: 60, mean: 69, SD: 31) from the nearest wall (source). The airflow at each locality was estimated qualitatively from observation of a ca 1 meter long, thin plastic strip hanging from the roof, fluttering more or less, indicating the direction and speed of the draught.
Levels of 222 Rn and 220 Rn were measured at all assessed mine localities with a pair of diffusion chamber solid-state nuclear track detectors (SSNTD) (Rad-track2 CR-39). One of these only lets in and measures 222 Rn, whereas the second detector of the pair has holes for measurement of both 222 Rn and 220 Rn. The difference in track density between the two detectors is used to estimate 220 Rn. These were provided and analysed by Radonova Laboratories (www.radonova. no). Because of the very short exposure time compared with the 3-month measurement time for which these detector measurements are calibrated, and because very high levels of 220 Rn only affect the detector with holes but not the detector without holes during the short exposure time, detection limits (MDA) were higher than the calculated activity concentration in 15 of 20 detector pairs. For 220 Rn, we therefore used the calculated values (based on alpha track numbers). Uncertainties (1 SD) ranged for 222 Rn from 13 to 100% (median: 50%) and for 220 Rn from 9 to 49% (median: 35%).
December measurements at some Mining Hill localities and in the niobium mine included TnP deposition detectors with an SSNTD covered by a Mylar film so only 212 Po among TnP has sufficient energy to penetrate. The National Institute of Radiological Sciences (Japan) developed these detectors, laboratory calibrated total deposition rate of TnP (22) and analysed the exposed detectors. Within mines there is often a draft and there may be an upward bias on deposition velocities from air movement (23,24) . Even so, one study from a mine with drafts up to more than 0.4 m s −1 that used deposition-based detectors got similar EECRn values from these as measured with in situ instruments (25) . Normally, indoor calibration is done at 0.02 m s −1 and it has been shown that increasing ventilation rates (modest airflow) involve a bias with an approximate linear increase in deposition velocity (26,27) . An outdoors study at even higher airflow velocities confirms the bias but suggest it may be too stringent to standardise by dividing the detector-estimated EEC thoron value with the ratio between airflow at laboratory calibration and measured wind velocity, since airflow velocity can be lower and must be measured along the detector surface (28) . This latter study used a statistical model to demonstrate the deposition detector bias from airflow but also identified the simultaneous importance of variation in 220 Rn and 222 Rn. The study showed that a mean airflow velocity along detectors of 0.2 m s −1 involved a mean bias from airflow that comprised a 0.87 fraction of the measurement (SD = 0.05).
The data used in this study can be made available at request to the corresponding author.

Dose rate contributions from 222 Rn, 220 Rn and TnP
We apply the UNSCEAR dose conversion factors (DCF) for 222 Rn and 220 Rn progeny of 9 and 40 nSv (Bq m −3 h) −1 , respectively (12,13) . For 222 Rn, we assume an equilibrium factor, F, like the average found in a manganese mine of 0.36 (29) , and for comparison, we calculate the 220 Rn cold season dose rate contribution using an F-value of 0.04 for a centered position in the wolframite mine (25) . 220 Rn measurements involve large uncertainties and dose rates should rather be calculated from TnP (14)(15)(16)(17) . We also calculate the dose rate contribution from TnP, but regarding mine draft, consider both dose rates potentially biased by the mine draft as well as dose rates corrected for mine draft assuming a 0.87 fraction of airflow bias from the mine draft.

Calculations, statistics, weather and maps
Calculations and statistics were performed with R (30) . TnP values are expressed as equilibrium equivalent concentrations for 220 Rn (i.e. EECTn), and the equilibrium factor, F, was calculated as the ratio between these. Bedrock was grouped according to main bedrock, but with differentiation between redrock localities with or without iron ore. Hourly weather data were downloaded (www.eklima.no) for the deployment periods from a nearby weather station (St.no = 32060, 59.380644 • N, 9.201890 • E). Map layers of bedrock and loose masses (Holocene and anthropogenic deposits) were downloaded from the Geological Survey of Norway and map layers of populated areas et cetera were downloaded from Kartverket (all via www.kartka talog.geonorge.no). The map in Figure 1 was made in R (30) , using packages sp (31) and rgdal (32) .

Results
Gamma spectrometry showed gamma radiation dose rates ranging from 0.1 to 11 μSv h −1 (median: 4.3, mean: 5.1, SD: 3.6) and bedrock activity concentrations ranging from 34 to 1900 Bq kg −1 (median: 930, mean: 1000, SD: 560) of 238 U and from 28 to 12 000 Bq kg −1 (median: 4900, mean: 5700, SD: 4200) of 232 Th. Variation among localities were up to 4-fold for 238 U, 20-fold for gamma and 40-fold for 232 Th, with the niobium mine standing out with lowest concentrations (Table 1). We notice increasing levels of gamma radiation and 232 Th toward redrock in the eastern part of Mining Hill (Figures 1 and 2 and Table 1). In redrock with iron ore, the levels of gamma and 232 Th were highest with means of 8.1 (SD: 2.0) μSv h −1 and 9300 (SD: 2400) Bq kg −1 , and a 238 U mean of 1300 (SD: 300) Bq kg −1 . In redrock without ore, these means were similar for gamma radiation and 232 Th (8.0 (SD: 2.2) μSv h −1 and 9000 (SD: 2500) Bq kg −1 ), whereas 238 U was somewhat higher (1500 (SD: 480) Bq kg −1 ). In the FDC zone even lower gamma radiation, 232 Th and 238 U were found (1.3 μSv h −1 , 1300 and 360 Bq kg −1 , respectively). In the transition zone, intermediate means were found (2.6 (SD: 0.5) μSv h −1 , 2700 (SD: 720) and 910 (SD: 290) Bq kg −1 , respectively). Notice the low Table 1. Gamma radiation (γ ) dose rate (DR, μSv h −1 ) and bedrock content (Bq kg −1 ) of 238 U and 232 Th in Fen legacy iron mine localities and Søve Niobium mine. Type of bedrock is given per measurement site (MS) but also the main one per adit or shaft is given (zone). Fe-Dolomite carbonatite is abbreviated FDC. In Karup shaft, the locality name is marked with number of meters below surface of the measurement point. Control situated in School office in area is indoors and above Holocene clay deposits (HCD). gamma dose rate contributions in the niobium mine (Søve) because of the low 232 Th level here (Table 1). Outside temperatures ranged during the November deployment from −9 to −3 • C (mean: −6 • C) and during the December deployment from −6 to −1 • C (mean: −4 • C). During both periods the observed mine draft within the mines was rising. Outdoor wind velocities ranged from 0.3 to 2.8 m s −1 (mean: 1.2 m s −1 ) mainly from northeast but also from northwest. The activity concentrations in mine air for 222 Rn ranged from 500 to 6900 Bq m −3 (median: 1300, mean: 2000, SD: 2100) and for 220 Rn from 700 to 45 000 Bq m −3 (median: 3100, mean: 5500, SD: 9300). There was some variation among localities for 222 Rn but especially for 220 Rn ( Table 2). Especially in the niobium mine (Søve) and in the Hov and western minet in Mining Hill, 222 Rn was high. Levels of 220 Rn were markedly higher at localities within redrock with iron ore but with large variation (mean: 11 000, SD: 17 000 Bq m −3 ). Interestingly, levels of 220 Rn were much lower in localities with redrock without ore (mean: 4800, SD: 1600 Bq m −3 ). The most extreme 220 Rn value was in the Bolla shaft, which is a part of the old transport system leading into the other mines, whereas Bolla adit that lies above the shaft had relatively lower values of 220 Rn and 222 Rn since it is relatively short (50-60 m) and has openings in both ends leading to the draft pulling trough and dilution of airborne radionuclides. In mines running through redrock without ore, 222 Rn levels were markedly higher but with large variation (mean: 2700, SD: 2300 Bq m −3 ) compared with in mines running through redrock with iron ore (mean: 900, SD: 300 Bq m −3 ). Levels were somewhat lower in FDC (Table 2) and the transition zone (means: 2000 (SD: 2400) and 2800 (SD: 2100) Bq m −3 for 222 Rn and 220 Rn). For all the 220 Rn measurements, if the outlier (Bolla shaft) is disregarded, there is no correlation between these measurements and distance to the mine wall (r = −0.12, p > 06, Figure 5). There was, however, a near significant correlation between 220 Rn measurements and bedrock 232 Th levels (r = 0.40, p < 0.09, Figure 6), which was not affected by removal of the outlier in 220 Rn. Levels of 222 Rn were not correlated to the bedrock 238 U measurements (r = −0.31, p > 0.2). All in all, the cold season dose rates calculated for inhalation during occupancy are high because of the 222 Rn contribution in the Søve niobium mine, whereas in Mining Hill, it is the very high levels of 220 Rn that are the main contributors (Tables 2). Localities had different mine draft, ranging from hardly notable (estimated to be <0.2 m s −1 from the movement of plastic strings), to weak (estimated 0.2-0.3 m s −1 ) at most localities, but intermediate (estimated 0.5-1 m s −1 ) in adits of Søve (niobium mine),   (median: 320, mean: 380, SD: 290) when not controlled for any mine draft airflow bias. After our correction of the bias, which probably reduced uncertainty, TnP ranged from 11 to 98 Bq m −3 (median: 42, mean: 49, SD: 37), still high levels. By comparison, when a correction for the wind bias of mine draft was done by correcting with the assumed airflow velocity divided by 0.02 m s −1 (lab conditions), corrected TnP values were reduced by another 67% compared with the correction of airflow bias done in this paper. In other words, the Table 3. Airborne TnP uncorrected with bias (BiasEECTn) and corrected for mine draft airflow bias (CorrEECTn) measured with deposition detectors at some Fen complex mine localities, and compared with measured 220 Rn, the equilibrium factor, F, with the ones most similar to another mine study in bold correction used here involves three times higher values than the most stringent way to correct for airflow because of mine draft (results not presented as this is only a factor difference). Among the three assessed mines, the highest TnP level was found in the main adit (Table 3). In the main adit, the TnP dose rate contribution, when not controlled for airflow, contributes between half to twothirds of the combined cold season dose rate. Consequently, when controlling for airflow bias on deposition, the TnP dose rate contributions are around one quarter of the combined cold season dose rate (Table 4). Compared with the equilibrium factor of F = 0.04 given by others for mines (25) , the inclined main adit (nr 16) and Hov adit (nr 18) have the most similar F-values when using TnP values controlled for airflow bias (Table 3). This suggests a significant bias from the airflow of mine draft at these localities, which was ∼4fold compared with our other localities with deposition detector deployment. By comparison, the main adit (nr 17), Bolla adit (nr 19) and the niobium mine (nr 20) have similar F-values as in the literature when their TnP values are not controlled for airflow bias. Also, the corresponding cold season dose rate contributions calculated from 220 Rn levels in the inclined main adit (nr 16) and adit (nr 18) are similar to the corrected TnP cold season dose rate contributions (bold in Table 4). By comparison, the cold season dose rate contributions calculated from 220 Rn levels in the main adit (nr 17), Bolla adit (nr 19) and in the niobium mine (nr 20) are similar to the non-corrected TnP cold season dose rate contributions (Table 4), suggesting there may not be much bias from mine draft at these mine localities. In total, between mines, among cold season dose rate contributions, gamma radiation contributes a fraction of 0.02-0.6, radon contributes a fraction of 0.33-0.95 and TnP a fraction of 0.1-0.25.

Discussion
In this study, gamma radiation measurements were made perpendicularly to the roof at some distance.
These results are therefore not entirely representative in relation to the gamma dose rates received by humans during occupancy in the mine. The dose rate from gamma radiation measurements should therefore be verified with a gamma dosemeter suitable for gamma from all directions. Moreover, even though the assumed secular equilibrium may hold for rock not exposed to weathering (33) , deviations may occur where there is leaching of uranium and radium isotopes with weathering through fissures (34,35) . Since gamma measurements target progeny, any large deviations from secular equilibrium would involve even higher levels of progenitor radionuclides than measured. The large difference in gamma from the Mining Hill localities within redrock (with or without iron ore) compared with the transition zone and the FDC zone is clearly because of the amount of redrock and its level of 232 Th. The lower value of the Søve niobium mine is because of the Fenite mix bedrock measured here. The niobium mine elsewhere runs through søvite. Søvite is usually not so enriched in 232 Th but rather 238 U, which emits much less gamma but explains the high levels of 222 Rn in Søve.
The levels of 222 Rn and 220 Rn within the assessed mines during the cold season with rising mine draft are relatively high and are clearly relevant regarding dose rate contribution ( Table 2). The levels seem to be related to the type of bedrock each mine runs through and the bedrock level of primordial radionuclide. Clearly, levels of 220 Rn are higher in mines with redrock and iron ore, lower when redrock has less or no ore, and even less in the transition and FDC zones. This is supported by the near significant correlation between 220 Rn and bedrock levels of 232 Th. The ventilation pattern is probably also important and may in places with significant draft explain why these correlations are not stronger for 220 Rn. However, with a mine draft of 0.1 m s −1 and no exhalation, a parcel of air would in five half-lives only move 27 meters, meaning that any high 220 Rn levels must originate from nearby bedrock with high levels of progenitor radionuclides. By comparison, 222 Rn has a much longer Table 4. For some Fen complex mine localities; uncorrected potentially air-flow-biased cold season dose rate contribution of TnP (BiascDR (μSv h −1 )), the corresponding cold season dose contribution of TnP corrected for airflow of mine draft (CorrcDR (μSv h −1 )) and combining also with gamma and radon dose rate, the combined cold season dose rate of these mine localities at minimum (Min ccDR) and maximum (Max ccDR), depending on whether airflow correction has been performed. Most probable values in bold.  (21) . These heterogeneities are not apparent in the single gamma spectrometric measurement from the niobium mine, but with mine draft, which was observed as intermediate, it may explain the high 222 Rn level measured.

Nr
In the Fen legacy iron mines, mine draft must also be significant in places, as both 220 Rn and 222 Rn have been recorded at high levels in air ventilating from the mines in Mining Hill with important local outdoor impact (1,28,36) . The observed differences in the levels of these gases between iron legacy mine localities thus probably reflects the differences in both bedrock radionuclide levels and mine draft. However, the levels of these gasses at the different localities and the whole pattern may be different with the opposite direction of mine draft in summer, which is characterised by a sinking draft. By comparison, another legacy metalliferous (wolframite) mine had much lower levels of both airborne radionuclides, with 222 Rn ranging from 120 to 150 Bq m −3 and 220 Rn ranging from 270 to 960 Bq m −3 , and bedrock levels of primordial radionuclides of 140 Bq kg −1 of 226 Ra and 120 Bq kg −1 of 232 Th (25) . That study involved dose rates of 0.3, 1.2 and 0.5 μSv h −1 from RnP, TnP and gamma, respectively. As mentioned previously, there was significant mine draft, but the study did not assess bias on the measured TnP levels, which ranged from 8 to 10 Bq m −3 (25) , even though uncertainties in these measurements were in the order of 10 Bq kg −1 and a bias cannot be excluded. The equilibrium factor F in the Kleinschmidt et al.'s (25) study is similar to the F-values of three localities of the study at hand with TnP levels not corrected for wind bias (one main adit location, Bolla and Søve mines) but similar to the F-values of two localities with TnP values corrected for wind bias (inclined main adit and Hov adit). The difference corresponds well with the approximate 4-fold difference in felt mine draft between these mine localities (uncorrected versus corrected). This suggests a bias from airflow, at least at higher airflow velocities, as suggested by previous studies (28) . The characterisation of the natural ventilation of the Mining Hill mines strongly suggests a mainly seasonal chimney ventilation (1) , which gives the seasonal mine draft directions but no prediction of the magnitude of draft. Such natural ventilation can be large (37)(38)(39)(40)(41) . There is therefore much potential for different residence times for 222 Rn, 220 Rn and potentially different disequilibrium factor values.
Regardless the size of the potential bias on depositionbased TnP detectors, in the Mining Hill mines there is significant gamma radiation, which dose rate contribution must be considered for any occupancy. Notice, however, that because of the lower levels of 232 Th, gamma radiation dose rate was lower in the FDC zone and especially much lower in the søvite zone. For Mining Hill localities and especially the niobium Søve mine, the cold season dose rate contribution of 222 Rn is significant and must also be considered. Notice that we have applied the UNSCEAR DCFs, whereas ICRP DCFs (42) for the 222 Rn dose rate contribution would near double in significance and the 220 Rn dose rate contribution would triple. From the most probable equilibrium factors calculated, either for potentially biased or for corrected TnP values (Tables 3 and 4), we assume a combined cold season dose rate of ∼17-24 μSv h −1 . The most probable TnP cold season dose rate contribution being from a tenth to a fourth of the combined cold season dose rate would with ICRP DCFs change to constituting from a third to over half of the combined cold season dose rate. Since new mining activities for REE and 232 Th are being planned for the Fen complex, exposure pathways for both the cold and warm season dose rate contributions are needed to consider occupancy. Exposure pathways and dose rate contributions should therefore also be assessed during the warm season and during the transient seasons, which could have diurnal changes in draft direction.
Unfortunately, measurements of TnP were not done in the School adit. We must stress that only this one of our localities was situated truly within the FDC bedrock unit, and because it was situated in a deadend adit, which is anomalously dry and not ventilated, it is not comparable to the other localities neither in bedrock nor ventilation. Hence, other localities within the FDC unit must be investigated to validate this result. Given the levels of 222 Rn and 220 Rn at the FDC location, the cold season dose rate contributions of these gasses are probably less significant than within the redrock zone. However, it was noted within the FDC zone adit a much higher level of fine airborne dust, and since the DCF's of inhalation of the 238 U and 232 Th associated with such dust is several orders of magnitude larger than for progeny of 222 Rn and 220 Rn, the real combined cold season (and possibly warm season) dose rates in mine within FDC may be larger and should be assessed in the future.