3.1.1. Concentration of 222Rn in Homes

Regarding the concentration of 222Rn in homes, the 11,500 data points on the concentration of 222Rn in air were analyzed using as a starting point the central trend measures and

dispersion, and according to the shape of the distribution sample. Table 2 offers the updated data regarding previous publications [52,53] showing the main parameters obtained:

**Table 2.** Statistics on radon concentration data for Spain.


The central trend measures show that the data do not follow a normal distribution, since the arithmetic mean (101 Bq/m3) is far from the median (56 Bq/m3). Furthermore, the standard deviation of the arithmetic mean is 260.6 Bq/m3, which shows a high dispersion of the data.

Analysis of the shape of the sample distribution shows a high coefficient of kutorsis (K = 1497), indicating a leptokurtic distribution, whereas the asymmetry coefficient (CS = 31.5) indicates a positive asymmetry: The distribution of measurements is log-normal. The distribution of measurements is log-normal, as shown in the histogram (Figure 2). This distribution is usual for radon concentration measurements since most of the measurements obtained are in low concentrations, whereas only a few measurements appear in the high concentration range.

**Figure 2.** Histogram of the radon concentration data.

Analyzing these 11,500 data points, it is seen that 27% of the samples exceed the level of 100 Bq/m3 to initiate action plans (5% of the measurements exceed 300 Bq/m<sup>3</sup> and 22% are in the range of between 100 and 300 Bq/m3).

On transposing these data to the cell system, it is seen (Table 3) that the majority (76%) are in the range of low concentrations (<100 Bq/m3), 22% between 100–300 Bq/m3, and that the percentage of cells in high concentrations (>300 Bq/m3) is reduced to 2%.


**Table 3.** Mean concentration of 222Rn and number of measurements per 10 <sup>×</sup> 10 cell.

It is also clear that the sampling in Spain is not heterogeneous, since there are areas where the measurement density is much higher than others; this is because the sampling in Spain was defined based on a series of criteria that concentrated the number of measurements in areas with potentially high radon concentrations. The decision of how many measurements to be carried out in each 10 km × 10 km cell was made by the CSN, taking into account the general objectives established by the EC-JRC in the creation of the European Radon Map, considering superficial, population and lithostratigraphic criteria, and according to the rate of exposure to terrestrial gamma radiation [51,52].

Despite efforts to try to cover the entire country with at least one measurement per 10 km × 10 km cell, it can be seen that a large part of its surface does not have any measurements (40%). Of the cells for which measurements are available, it is representative that a large percentage of Spain is covered with only one measurement (47%) or with two measurements (19%), whereas cells with more than 6 measurements represent 15% of the total.

Analyzing cell percentage according to concentration category and measurements, it is seen that 68% of cells with concentrations greater than 400 Bq/m3 have more than two measurements (in 52% of cases from 2 to 6 measurements, and in 16% more than 6 measurements), and that in the concentration range between 301 and 400 Bq/m3 this percentage of cells is also high (66%). The percentage is slightly lower for the intermediate concentrations (101–300 Bq/m3) where 63% of cells have more than 2 measurements. Low concentrations (<100 Bq/m3), despite being the most numerous category, is the one with the fewest measurements. Half of its cells have a single measurement, reducing the number of cells with more than 6 measurements to 9%.

It is clear that as the number of measurements per cell increases, the concentration ranges are better defined.

#### 3.1.2. Exposure Rate to Terrestrial Gamma Radiation

From the analysis of the rates of exposure to terrestrial gamma radiation (Figure 3a), 95% of the peninsula is found to have medium and low rates: 2% of the country is below 44 nGy/h, and 93% between 45 and 122 nGy/h. The areas with rates higher than 122 nGy/h are few (5%), and are mainly in the northwest area of the peninsula and in the Central System. In these areas, there is a clear correspondence between the presence of high radon concentrations and high rates of gamma exposure [43,54].

Analyzing the data of this variable with respect to radon concentrations (Figure 3b), a positive linear relationship is observed between the two parameters starting at 78 nGy/h, the relationship becoming clearer in the case of the identified areas of high rates. These areas correspond once again to those previously mentioned, along with areas of the Catalan Coastal Cordillera and the Pyrenees, corroborating the correspondence of high radon concentrations and high rates of gamma exposure.

**Figure 3.** (**a**) Figure Terrestrial gamma radiation rates (nGy/h) and (**b**) Relationship between terrestrial gamma radiation rates and radon concentration.

#### 3.1.3. Lithostratigraphies

Regarding the lithostratigraphy variable, it is known that the main indicator in determining a higher or lower probability of high concentrations of radon in an area is the presence of uranium in soils and rocks, for which reason the lithological formations with a high proportion of uranium will generate a high proportion of radium and therefore a higher proportion of radon. In general, the highest uranium values (>2.88 ppm) [55], are associated with acidic intrusive plutonic rocks such as granites.

The analysis of lithostratigraphies (Figure 4a) in Spain suggests that the geologies most commonly found in Spain are acidic plutonic rocks such as granites, granodiorites, and quartz diorites (8% of the territory). Due to the large number of lithostratigraphies present in Spain, the legend of this figure only shows the most numerous (more than 2% of the territory), the complete legend is available in the IGME [56]. It is also noteworthy that 4% of the territory is made up of slates and greywacke. Both shales (metamorphic rocks produced by silt-clay sedimentary rocks) and greywacke (detrital sedimentary rocks derived from the dismantling of acidic plutonic rocks) generally also have high uranium content [43].

In the areas where these two formations are present, there may be a high probability of finding high radon concentrations, which was confirmed after performing the correlation analyses of the two variables.

The results of the relation between lithostratigraphies and radon concentrations (Figure 4b) show that 100 lithostratigraphies show a positive relationship. The clearest relation (>+0.75) appears in a single case, in the geologies corresponding to acidic, Hercynian plutonic rocks (granites, granodiorites, quartz diorites). It has been confirmed that this geology is associated with a high presence of radon.

In addition, six other lithostratigraphies show a marked relationship with the presence of high radon concentrations (between +0.51 and +0.75), representing another 8% of the peninsular surface. They mainly correspond to metamorphic rocks such as shales, gneiss, schists, or quartzites (these are rocks with high concentrations of uranium) [43] and detrital sedimentary rocks such as greywacke derived from acidic plutonic rocks.

On analyzing the geographical distribution of these areas, it is seen that they correspond to the northwest area of the peninsula and the area of the Central System. A close relationship is also observed with the geological formations in the west of the peninsula and their extension towards Sierra Morena, specific areas of the Pyrenean Range, and in the area of the Catalan Coastal Cordilleras.

**Figure 4.** (**a**) Figure Lithostratigraphies 1:200,000 and (**b**) Figure Relationship lithostratigraphies and radon concentration.

#### *3.2. Comparison of Radon Potential Maps Generated*

As shown in Figure 5, where the P90 Radon Potential Map generated by the CSN (hereafter P90 Potential Map) and the Radon Potential Map Calculated in the present study are compared, both maps show a similar percentage of cells in the range of radon concentrations greater than 400 Bq/m3 (17% in the case of the Calculated Map and 16% in the P90 Potential Map). In both maps, the areas defined in this range correspond once again to the northwest of the peninsula, the Central System area, the west of the peninsula extending towards Sierra Morena, south of the Pyrenees, and in the area of the Catalan Coastal Cordilleras.

The increase in the weight of the cells with concentrations between 301 Bq/m3 and 400 Bq/m3 is significant: it goes from 2% in the P90 Potential Map to 19% in the case of the Calculated Map. As will be seen later, the calculation of this new zoning fits possible radon concentrations more reliably. It is mainly seen in the west of the peninsula surrounding the highest concentrations. The area of Sierra Morena up to the border with Portugal is also clearly defined in this range, and in some areas to the north of the peninsula, areas of the Penibaetic System or areas to the west of the Ebro valley.

Both maps have a similar percentage of cells in the concentration range between 201–300 Bq/m3 (20% in the case of the Calculated Map and 21% in the P90 Potential Map). However, analyzing Figure 4 shows changes in zoning: in the CSN P90 Potential Map these areas were defined as mainly bordering the areas of greater concentrations in the west of the peninsula and certain areas in the south of the Ebro valley, while with the Calculated Map these areas are mainly found in the south of the Iberian System and the south of the Ebro valley.

**Figure 5.** Potential Radon Map. (**a**) Calculated Map and (**b**) Map based on Spanish Nuclear Safety Council (CSN) data.

The most significant change occurs in the cells between 101 and 200 Bq/m3, where it drops from 59% in the case of the P90 Potential Map to 35% in the Calculated Map. The reduction is this percentage is accompanied by a large group of the cells located in this category switching, on the Calculated Map, to the range between 301–400 Bq/m3 and cells of less than 100 Bq/m3.

Regarding the range of cells with concentrations below 100 Bq/m3, an increase in the number of cells is observed, going from 2% on the P90 Potential Map to 7% on the Calculated Map. This increase in cells is due to a large number of the cells identified in the CSN map as in the range between 101–200 Bq/m3 having moved to this range of lower concentrations. The area in this category lies mainly in the south of Spain in the Guadalquivir Valley and the Levante.

#### *3.3. Assessment of the Degree of Identification of the Maps*

To quantify the degree of identification of the Radon Potential Calculated Map with respect to radon concentrations, the cells are analyzed, identifying for each variable the percentage of failures or successes in each of the ranges. Table 4 reflects the degree of identification of each of the study variables.


**Table 4.** Degree of identification between the variables of the maps Potential Radon Map Calculated and P90 CSN.

#### 3.3.1. Degree of Identification Regarding Radon Concentrations

Regarding radon concentrations, both maps show a high capacity to identify cells with concentrations greater than 400 Bq/m3, but the Calculated Potential Map improves the data obtained with respect to the P90 Potential Map: it returns 68% correct identification of these areas compared to 64%.

The increase in the identification capacity of the Calculated Potential Map in areas with concentrations between 301–400 Bq/m<sup>3</sup> is of special interest, from 20% to 3% reported by potential radon map CSN, Directive 2013/59/Euratom sets the first value as a reference level to be considered when devising National Action Plans against radon gas in order to define Radon Priority Areas that, with the proposed method, becomes easier to define.

In the intermediate concentration ranges, the identification capacity of the P90 Potential Map is superior to that of the Calculated Potential Map: In the range 201–300 Bq/m3 and 101–200 Bq/m3 it identifies appropriately 30% and 37% of the cells. The Calculated Potential Map correctly identifies 15% of cases.

The greater capacity of identification of the P90 Potential Map in these ranges of mean concentrations is mainly due to the scarcity of measurements made in these areas. As previously shown, a higher sampling density per cell more precisely defines the concentrations, and many of the cells identified as having average concentrations would move to another range of concentrations as the number of measurements in them increases. This greater identification capacity is also due to the fact that the CSN map assigned a high weight to the gamma radiation variable, while when creating the Calculated Potential Map, the weight of the variables was homogenized.

In the lower concentration ranges, the identification capacity of the Calculated Potential Map improves to 12% as against 3% of the P90 Potential Map.

3.3.2. Degree of Identification Regarding the Rates of Exposure to Terrestrial Gamma Radiation

Regarding the rates of exposure to terrestrial gamma radiation, the P90 Potential Map has a greater capacity of identification (57%) as compared to the Calculated Potential Map (32%). However, when analyzing the different ranges in detail, it is found that this reduction in global accuracy is mainly due to low identification in the cells corresponding to the average rates (between 45 nGy/h and 122 nGy/h), because the data on radon concentration in this grid is limited. On the other hand, the increase in the identification of cells with higher rates is notable: in cells with more than 167 nGy/h the level of identification increases to 90%, and in cells between 123 nGy/h and 167 nGy/h it rises to 8%.

#### 3.3.3. Degree of Identification Regarding the Lithostratigraphies

With respect to lithostratigraphies, both maps correctly identify 47% of the cells. The differences emerge when analyzing the different classes associated with concentrations.

In the cells corresponding to lithostratigraphies associated with concentrations of more than 400 Bq/m3 (Class 9), the Calculated Potential Map increases identification to 96% as against 89% of the P90 Potential Map.

It is of particular interest that the Calculated Potential Map correctly identifies 36% of the cells associated with Classes 7 and 8 (lithostratigraphies linked to radon concentrations between 301 Bq/m<sup>3</sup> and 400 Bq/m3), since the CSN map does not have the capacity to identify these areas. Again, Directive 2013/59/Euratom sets the value 300 Bq/m<sup>3</sup> as the reference level for producing National Action Plans against radon gas.

The identification of cells associated with concentrations between 201 Bq/m3 and 300 Bq/m<sup>3</sup> (Classes 4, 5 and 6) is similar in both maps: 27% in the case of the Calculated Potential Map and 31% of the P90 Potential Map.

The identification capacity of the Calculated Potential Map drops to 65% compared to 91% of the P90 Potential Map in Class 2 and 3 cells (lithostratigraphies associated with concentrations between 101 Bq/m<sup>3</sup> and 200 Bq/m3). This is due to the existence of lithostratigraphies that were previously identified with intermediate concentrations, but now have come to be placed in the category of low concentrations: in Class 1 (lithostratigraphies associated with radon concentrations of less than 100 Bq/m3) the Calculated Potential Map has an accuracy of 80% as opposed to the null capacity of the P90 Potential Map.

#### **4. Conclusions**

In conclusion, it has been shown that:


The map also identifies the areas with a probability of finding radon concentrations of between 100 Bq/m<sup>3</sup> and 300 Bq/m<sup>3</sup> more reliably, by homogenizing the weights of the variables. This range of concentrations is also of particular interest, as the WHO designates 100 Bq/m<sup>3</sup> as the recommended reference level to start action plans against radon gas.

**Author Contributions:** Conceptualization, A.F., C.S., and L.Q.; methodology, A.F., C.S., S.C., L.Q., D.R., and I.F.; software, A.F.; validation, A.F.; formal analysis, A.F., C.S., and L.Q.; investigation, A.F., C.S., S.C., L.Q., D.R., and I.F.; writing—original draft preparation, A.F., C.S., and L.Q.; writing review and editing, A.F., C.S., and L.Q.; visualization, A.F. All authors have read and agreed to the published version of the manuscript.

**Funding:** This research received no external funding.

**Institutional Review Board Statement:** Not applicable.

**Informed Consent Statement:** Not applicable.

**Data Availability Statement:** The data presented in this study are available in article.

**Conflicts of Interest:** The authors declare no conflict of interest.

#### **References**

