Developing a new National Radon Risk Map

Size: px
Start display at page:

Download "Developing a new National Radon Risk Map"

Transcription

1 Developing a new National Radon Risk Map 2014 J. A. Hodgson, S. Carey and R. Scanlon Geological Survey of Ireland 2014

2 Developing a new national radon risk map Geological Survey of Ireland The Geological Survey of Ireland is responsible for providing geological advice and information, and for the acquisition of data for this purpose. GSI produces a range of products including maps, reports and databases and acts as a knowledge centre and project partner in all aspects of Irish geology. GSI is a division of the Department of Communications, Energy & Natural Resources (DCENR). Geological Survey of Ireland Beggars Bush Haddington Road Dublin 4 Ireland Tel Fax tellusborder@gsi.ie Web: Disclaimer Although every effort has been made to ensure the accuracy of the material contained in this report, complete accuracy cannot be guaranteed. Neither the Geological Survey of Ireland, the Geological Survey of Northern Ireland nor the authors accept any responsibility whatsoever for loss or damage occasioned, or claimed to have been occasioned, in part or in full as a consequence of any person acting or refraining from acting, as a result of a matter contained in this report. Copyright Government of Ireland. Basemaps Ordnance Survey Ireland Licence No. EN ii

3 Developing a new national radon risk map Document Information Written by: Position Date Jim Hodgson Geophysicist 20 th May 2014 Shane Carey GIS and Data Manager Ray Scanlon Principle Geologist, GSI Reviewed by: Mairéad Glennon Assistant Project Manager Vincent Gallagher Assistant Geochemist Change record Date Author Version Details 17 th June 2014 Jim Hodgson Final edits 31 st June 2014 Jim Hodgson Executive Summary iii

4 Developing a new national radon risk map Executive Summary National geological datasets, including 1:100,000 bedrock and aquifer-type maps as well as groundwater recharge-coefficient data and depth-to-bedrock data, have been modelled in conjunction with available country-wide indoor radon measurements to produce a national radon risk map. The new radon risk map is modelled at a 1 km grid scale compared to the 10 km grid scale of the existing radon map of Ireland. Data is modelled based on multivariate linear regression. Individual datasets were assessed to determine their suitability for the model and the derived model has also been evaluated against existing radon maps produced by the Radiological Protection Institute of Ireland (RPII). Variations due to underlying bedrock geology and rural or urban effects were also investigated. Aquifer type and the movement and the concentration of radon gas within groundwater seem to be very significant in the transport of radon. This is demonstrated by the strong relationship between regionally important karstified aquifers, particularly those with conduit flow, and radon highs. Model data was compared with models produced in regions where airborne geophysical data is available and has been used to assist in predicted radon distributions. Where possible airborne geophysical data, in particular uranium measurements, should be used to help model radon distribution, as the percentage variance can be significantly decreased when it is incorporated into models. The new map shows generally good agreement with previous mapping and with detailed investigations carried out in certain areas such as Tralee and Castleisland in Co. Kerry. Radon highs, i.e. areas with greater than 10% chance of a dwelling exceeding the reference level of 200 Bq/m 3, occur predominantly over (1) karstified limestones in the west, in counties Clare, Galway, Mayo, Roscommon and Sligo, and (2) in the southeast, over the Leinster Granite and Lower Palaeozoic succession, including acid volcanic rocks, in counties Waterford, Wexford, Carlow and Wicklow. Other notable highs are found around Castleisland in County Kerry, southwest of Limerick city and in northern County Cork. iv

5 Developing a new national radon risk map Model summaries and data statistics are contained in Appendices 1 to 3 while a detailed account of how all data was used and manipulated within GIS and R language environments is contained in Appendix 4. Acknowledgements The Radiological Protection Institute of Ireland (RPII) provided all the indoor radon data used in the modelling and its staff, were generous with their time and support for the project. Stephanie Long, in particular, made significant contributions to the work. v

6 Table of Contents Executive Summary... iv Acknowledgements... v 1. Introduction Existing Radon mapping and models New Models and Data Indoor Radon data New Datasets New Grid Radon reference level Data Evaluation Available data Linear regression Further Evaluation Bedrock Terrain Rural and Urban Airborne Geophysical Data Grid Size and Spatial distribution Conditional dependence Final Model Model Evaluation Comparison with previous models Test Area Tralee-Castleisland Conclusions and Recommendations References Appendix 1: Data Assessment Summary by 1km Grid Appendix 2: Indoor radon mean values per bedrock unit Appendix 3: Regression Summary Appendix 4: A guide to Data preparation for Radon Modelling A1 Introduction A1.1 Datasets Assessed A2. Data Preparation Recommendations

7 List of Tables TABLE 1: NUMBER OF GRID SQUARES WITH INCREASING RADON POINTS TABLE 2: DATASETS TESTED FOR MODELLING TABLE 3: NUMBER OF GRID SQUARES FOR EACH DATASET SUBSET BY AIRBORNE NUMBER OF RADON DATA POINT. TB REFERS TO TELLUS BORDER PROJECT WITH COUNTIES DONEGAL, LEITRIM, SLIGO, CAVAN, MONAGHAN AND LOUTH

8 List of Figures FIGURE 1: RADON RISK MAP PRODUCED BY RPII BASED ON 10 KM GRID SQUARES FIGURE 2: NATIONAL RADON MAP PRODUCED BY HSE FOR HEALTH ATLAS OF IRELAND BASED ON 29,660 GEO-REFERENCED DATA POINTS FIGURE 3: OUTPUT FROM TELLUS BORDER MODEL DERIVED FROM AIRBORNE URANIUM MEASUREMENT, GROUNDWATER RECHARGE AND KARST VALUES AS PART OF THE TELLUS BORDER PROJECT (HODGSON AND CAREY 2013) FIGURE 4: DISTRIBUTION OF INDOOR RADON MEASUREMENTS USED IN STUDY. PROVIDED BY RPII FIGURE 5: 1KM REFERENCE GRID BASED ON CSO GRID FIGURE 6: AQUIFER CLASSIFICATION WITH LOCATIONS OF INDOOR RADON MEASUREMENTS THAT EXCEED 200 BQ/M FIGURE 7: R 2 VALUES FOR THE FIT OF GWRC AGAINST %RL VERSUS THE NUMBER OF RADON MEASUREMENTS PER GRID SQUARE FIGURE 8: MODEL OF RADON RISK BASED ON AIRBORNE URANIUM MEASUREMENTS, AVERAGE RADON PER BEDROCK UNIT, GWRC, DTB AND AQUIFER VALUES FIGURE 9: CORRELATION MATRIX OF DIFFERENT DATASETS FOR NATIONAL MODEL, NUMBER IS THE R 2 VALUE FIGURE 10: GRID SQUARES WITH 7 OR MORE RADON READINGS SHOWN IN RED, GRID SQUARES WITH 1-6 READINGS IN GREEN. GRID SQUARES WITH LESS THAN ONE RADON READING IN WHITE.. 33 FIGURE 11: MODEL OF THE RISK OF DWELLING EXCEEDING RADON REFERENCE LEVEL OF 200BQ/M FIGURE 12: COMPARISON OF RADON DISTRIBUTION MAPS PRODUCE BY NEW MODEL (LEFT) AND INDOOR MEASUREMENTS RPII FIGURE 13: COMPARISON OF RADON RISK MODELS FOR A) NEW NATIONAL MODEL FROM FIG. 11 AND B) MODEL OF TELLUS BORDER REGION USING AIRBORNE URANIUM DATA FIGURE 14: RADON POTENTIAL MAP OF THE TRALEE-CASTLEISLAND AREA PRODUCED USING THE AIRBORNE MEASUREMENTS OF URANIUM (EU) AND PERMEABILITY (P) LINEAR REGRESSION MODEL (URBAN AREAS OUTLINED IN WHITE). FROM APPLETON ET AL. (2011) FIGURE 15: NEW NATIONAL MODEL OF CASTLEISLAND AREA (2014) FIGURE 16: MODEL BASED ON AIRBORNE URANIUM, AVERAGE RADON PER BEDROCK UNIT, AQUIFER, DTB AND GWRC VALUES. MODEL SHOWS SIMILAR OUTPUT AS TO APPLETON MODEL USING AIRBORNE URANIUM AND PERMEABILITY DATA (FIGURE 14) FIGURE 17: RPII 10KM GRID RADON MAP OF THE TRALEE-CASTLEISLAND AREA (MAIN URBAN AREAS OUTLINED IN WHITE, REDRAWN FROM THE DATA IN FENNELL ET AL. (2002)) FIGURE 18: PERCENTAGE OF BEDROCK TYPE OVER WHOLE COUNTRY OF IRELAND BASED ON1: 1 MILLION BEDROCK MAP AND 1 KM GRID SQUARES FIGURE 19: PERCENTAGE OF QUATERNARY SEDIMENTS ACROSS THE WHOLE COUNTRY BASED ON NEW QUATERNARY MAP GSI FIGURE 20 CSO CENSUS KM GRID DATASET FIGURE 21 1:100,000 (100K) BEDROCK GEOLOGY FIGURE 22 TABULATE INTERSECTION FROM THE STATISTICS TOOLBOX IN ARCTOOLBOX FIGURE 23 SORT THE DATA USING DATA MANAGEMENT TOOLBOX, SORT FIGURE 24 USE THE SUMMARY STATISTICS TOOLBOX FROM ANALYSIS TOOLS TO OBTAIN STATISTICS FOR EACH GRID SQUARE FIGURE 25 TABULATE INTERSECTION FROM THE STATISTICS TOOLBOX IN ARCTOOLBOX FIGURE 26 JOIN BY SPATIAL LOCATION FIGURE 27 ZONAL STATISTICS TOOLBOX USED TO CALCULATE GRID SQUARE MEAN ELEVATION FIGURE 28 TABULATE INTERSECTION FROM THE STATISTICS TOOLBOX IN ARCTOOLBOX USED TO CALCULATE PERCENTAGE AREA OF GRID SQUARES THAT FALL WITHIN A TOWN POLYGON

9 1. Introduction Previous investigations (Tellus Border Radon Report 2013, Appelton et al., 2011, Appleton and Miles 2008) have studied the distribution of radon gas using airborne geophysical data and geological parameters for areas within Ireland and Northern Ireland. This project continues this work and endeavours to develop a model of radon risk for the entire country of Ireland. The existing maps of radon distribution in Ireland are based on indoor radon measurements carried out by the Radiological Protection Institute of Ireland (RPII) and were mapped on a 10 km grid scale. These maps although definitive are of low resolution due to the large grid size used. Therefore local variations in radon gas values are not apparent. Models produced during the Tellus Border Project for the border region (Hodgson and Carey 2013) and for Northern Ireland (Appleton and Miles 2008) were developed on a 1 km grid scale based on airborne measurement of uranium. They incorporated geological data on soil permeability, groundwater recharge and karst distribution. These models greatly improved the predictive distribution of radon gas. To improve the resolution of national radon maps a 1km grid scale was adopted. However, any new model at this scale would have to be based on data which is available across the whole country. Therefore, airborne geophysical data was not included in the new national radon model. Instead national geological datasets were tested to determine their relationship with the distribution of indoor radon measurements. 1.1 Existing Radon mapping and models Currently, radon maps for Ireland are produced by the RPII. These maps are derived from indoor measurements from monitors placed in people s homes. Indoor data are reported in Becquerel per cubic meter (Bq/m 3 ) and have been mapped at a 10 km grid scale (Figure 1). The most recent national radon map for Ireland (Figure 1) subdivides the country into a 10 km 2 grid and provides a prediction of the percentage of homes within each grid square that exceed the reference level of 200 Bq/m 3. This map was generated using 11,319 measurements collected as part of the National Radon Survey carried out by the RPII between 1990 and 1999 (RPII 2002). Indoor measurements have continued since 1999 and a total of up to 40,000 indoor measurements have now been carried out across the country. In 9

10 November 2013 the HSE produced a new map of radon distribution based on 29,660 individually geo-referenced indoor measurements collected by the RPII (Figure 2). Figure 1: Radon risk map produced by RPII based on 10 km grid squares. 10

11 Figure 2: National radon map produced by HSE for Health Atlas of Ireland based on 29,660 geo-referenced data points. This new national radon map provided greater accuracy in the location of individual indoor radon measurements but the resolution of the map is still limited to a 10 km grid scale. The 11

12 bulk of indoor measurements were made in the vicinity of cities and towns. Coverage of rural areas is still relatively sparse and the density of measurements is too low to allow improvements to the map resolution. As part of the Tellus Border Project a model based on airborne geophysical data and geological parameters was produced at a 1km grid scale (Hodgson and Carey 2013; Figure 3), based on previous work by Appleton et al. (2011) and Appleton and Miles (2008). This work used multivariate regression analysis to model the distribution of indoor radon data with different geological and geophysical parameters. Figure 3: Output from Tellus Border model derived from airborne uranium measurement, groundwater recharge and karst values as part of the Tellus Border Project (Hodgson and Carey 2013). One of the possible limitations of the 2013 model based on airborne uranium measurement, groundwater recharge and karst values, is that these parameters may be terrain-specific. The model was based on a limited number of grid squares where the underlying bedrock is predominantly Carboniferous in age. In non-carboniferous areas modelled radon values may be over-estimated based on relatively high airborne uranium levels. Studies in non- 12

13 Carboniferous terrains (Appleton et al. 2011) suggest that, where modelling is based on airborne uranium measurement in combination with geological parameters, specific models may have to be developed for specific geological terrains in order to map radon distributions accurately. A national model would have to be applicable to all areas and therefore terms used would have to be applicable to all geological regions or if not would involve the combination of different models. Also, airborne data is not available across the whole country and therefore a new national model would have to be developed. The new national model of radon risk developed here builds on the methodology adopted for these previous maps (Appleton et al. 2011, Hodgson and Carey 2013). However, as with any new high-resolution national map of radon risk it needs to be tested to determine its suitability in all areas. 13

14 2. New Models and Data 2.1 Indoor Radon data The RPII is the main depository of indoor radon measurements and has carried out both national and local campaigns. It has records of over 40,000 measurements carried out throughout the country but not all of these measurements have been geo-referenced. Work carried out during the Tellus Border Project identified accurate locations for approximately 5000 of these measurements in the border region. The RPII provided its datasets to the Health Services Executive (HSE) which accurately geo-referenced 29,660 indoor radon measurements made across the whole country. Together with the geo-referenced data from the border region and some data for Local Authority housing, a total of 32,108 georeferenced indoor radon measurements are available for the country, after removal of duplicates. This dataset represents all available geo-referenced data from different sources. The inclusion of local authority housing data along with data for private houses may increase the variability of the data as previous studies have shown that radon measurements in local authority houses are typically lower than those for private houses within the same area owing to better construction (Hodgson and Carey 2013). Data was checked and seasonally adjusted based on criteria determined by RPII (RPII 2010). Any data with no documentation to show if any corrections had been applied were removed from the dataset. An outside radon value correction of 5.6 Bq/m 3 was then deducted from the new geo-referenced indoor radon database (32,108 records). This correction was based on outdoor radon measurements made at 18 meteorological stations (Gunning et al. 2014,). Figure 4 shows the location of all corrected indoor radon points used in this study. 14

15 Radon Modelling for a National Map of Ireland 06/08/2014 Figure 4: Distribution of Indoor radon measurements used in study. Provided by RPII

16 2.2 New Datasets The radon distribution model produced during the Tellus Border Project (Figure 3) used airborne measurements of uranium, groundwater recharge coefficient data and a parameter that reflecting the degree of karstification based on 100,000 scale bedrock data. Other parameters were also assessed during the project including airborne measurements of potassium and thorium and soil geochemistry analyses of uranium, thorium, potassium, yttrium and iron. However, most of this geophysical and geochemical data is not available on a national basis and therefore several other national datasets were tested to see if they were useful in helping to determine the distribution of radon. These included previously existing datasets and others that were generated specifically for this project, including: depth to bedrock; aquifer type; elevation; subsoil permeability; mean radon value per bedrock unit; mean radon value per quaternary sediment unit; and degree of faulting. A detailed assessment and evaluation of the datasets employed is presented in Section New Grid A new 1 km grid, based on that used by the Central Statistics Office (CSO), was used for initial modelling. This will facilitate the addition of data in the future, allowing population data and other potentially useful national datasets to be incorporated into the model. This new 1 km grid was constructed in ARCGIS 10.1 and contains a total of 73,230 individual grid squares (Figure 5). All available data were then assigned to the new grid. Indoor radon readings were available for 12,928 of the grid squares. Only 18% of grid squares contained radon data with greater concentrations in urban areas or areas perceived to have radon issues. This relatively uneven distribution of indoor radon data may affect the accuracy of mapping in certain areas. 16

17 Radon Modelling for a National Map of Ireland Figure 5: 1km Reference Grid based on CSO grid 17 06/08/2014

18 2.3 Radon reference level Individual datasets were assessed to determine the strength of their correlation with the estimated percentage reference level (%RL). The %RL is derived from indoor radon readings and indicates the probability that a dwelling within a grid square has a radon value greater than the national reference level of 200 Bq/m 3. This is the parameter adopted by RPII for use in its National Radon Survey (RPII 2002) and its use in any new developed model allows for easy comparison with existing maps. The %RL is based on the calculated geometric mean (GM) and geometric standard deviation (GSD) for indoor radon data within each grid square. Data analysis of indoor radon values (RPII, 2002; Commission of European Communities, 1987) has shown that indoor radon values are log-normally distributed at various scales, from country scale to kilometre scale. The GM and GSD of indoor radon values were calculated for each grid square using the statistical software package R. A value of K can be defined for each grid square from the following equation [1], where K is a transform threshold value for use with a standard normal distribution: K= ln(200)-ln(gm) / ln(gsd) [1] This allows a %RL value to be read from statistical tables relating to the area under a standard normal curve or K value. The value of %RL can then be assigned to each relevant grid square. This methodology was employed by the RPII for the Irish National Radon Survey (RPII, 2002). To obtain a %RL value, a minimum of 3 indoor measurements are necessary for each grid square. The more readings used per grid square will improve the estimate for the %RL. Of the 12,928 grid squares for which indoor radon data are available, 2318 had three or more readings from which a %RL value could be determined. This will subsequently control the number of grids squares that can be used within the model. Using grid squares with a %RL based on a higher number of radon points should produce better fits although these are less in number and hence less well distributed across the country. Table 1 shows the number of grid squares available for modelling using regression analysis by the minimum number of radon measurements per grid square. 18

19 Table 1: Number of grid squares with increasing radon points No of Radon readings No of grid squares 3 or more or more or more or more or more or more or more or more or more or more or more 70 The more measurements per grid square, the more reliable the estimate of %RL should be. However, as can be seen from Table 1 the number of grid squares available for evaluation decreases as the number of indoor radon measurements within them increases. The geographical distribution also becomes more restricted as a significant proportion of grid squares containing a high number of radon measurements are located within urban areas. 19

20 3. Data Evaluation 3.1 Available data The initial model produced using Tellus Border data (Figure 3) employed a combination of airborne measurements of uranium, the groundwater recharge coefficient (GWRC) and a derived karst term based on the likely degree of karstification as a function of the 1:100,000 bedrock map. However, the main objective of this project is to create a national radon model and unfortunately airborne data is only available for part of the country. Therefore as outlined in Section 2.2 only national datasets were considered and evaluated. Several Geological Survey of Ireland (GSI) datasets (Table 2) were evaluated to determine their suitability for modelling radon distribution across the country. These datasets were considered to be potentially useful in helping to model either the source or pathway of radon gas. Table 2: Datasets tested for modelling 1. 1:100,000 bedrock geology 2. 1:1 Million bedrock geology 3. 1:50,000 quaternary geology 4. Aquifer type 5. Karstification classification 6. Groundwater recharge coefficient (GWRC) 7. Depth to bedrock (DTB) 8. Degree of faulting 9. Permeability 10. Elevation For evaluation, data from the various datasets were assigned to each 1 km grid square. Where data for a given dataset varied within the boundaries of a grid square, the data were merged to generate an average value for the grid square. A full account of how all data were prepared for evaluation and construction of the model can be found in Appendix 4. 20

21 For datasets 1 3 (Table 2), the indoor radon point data were joined to corresponding geological units on a country-wide basis. An average radon value was then calculated for each geological unit based on all the points falling within that unit. This average value was then assigned to all grid squares overlying this geological unit. For some 1:100,000 geological units there were insufficient (less than 20) indoor readings. In these cases, the radon point data were grouped together based on the 1:1 Million bedrock map and the average value determined from these larger groupings. The same process was applied to the quaternary geology units. The aquifer type and karstification classification are based on categories rather than numerical values and therefore a weighting was assigned to each of the different categories. For aquifer type the following weightings were applied: Aquifer type Weighting Regionally important karstified aquifers with conduit flow (Rkc) 4 Other regionally important aquifers (Rkf, Rkd, Rk) 3 Locally important karstified aquifer (Lk) 2.5 Other locally important aquifers (Lm, LL) 2 Poor Aquifers (PU, Pl) 1 For karst terms the following weighting were used Karst classification Weighting Pure-bedded limestone (high degree of karstification) 4 Pure unbedded limestone (moderate degree of karstification) 3 Impure Limestone (low degree of karstification) 2 Non limestone (no karstification) 1 Datasets 6 10 (Table 2) are all expressed numerically. However, in cases where the data varies within an individual grid square an average was calculated. All work was carried out using ARCGIS 10.1 and R software and full details of the methodology used in the data preparation are contained in Appendix 4. 21

22 Following the assignment of all data to individual grid squares, the various datasets could be evaluated by comparing the data for each grid square to the %RL values determined for the same grid squares. 3.2 Linear regression Linear regression was used to determine the strength of the relationship between the percentage reference level (%RL) derived from indoor radon measurements and the varying datasets. For grid squares with values for both %RL and a given dataset a coefficient of determination (R 2 ) could be estimated. The R 2 expresses the amount of variance can be explained by the different factors. A value of 0 indicates no correlation, while a value of 1 indicates a perfect correlation with the different elements accounting for all of the variance. Typically the variance accounted for by individual parameters is relatively low. This approach follows work undertaken by Appleton et al. (2011), Appleton and Ball (2001) and previous radon modelling carried out as part of the Tellus Border Project (Hodgson and Carey 2013). Of the 10 different datasets assessed (Table 2) only five were found to have some correlation with %RL values. These were: 1. Indoor radon mean per bedrock geology unit (combination of 100,000 and 1:1 Million scale bedrock geology maps); 2. aquifer type; 3. degree of karstification; 4. groundwater recharge coefficient (GWRC) and 5. depth to bedrock (DTB). A summary of the regression analysis for all datasets, including different combinations of the datasets by grid squares with increasing minimum indoor radon readings, are listed in Appendix 2. Overall, the calculated R 2 for each of the different datasets was relatively low with generally a small portion (less than 30%) of the variation in %RL explained by the dataset terms. This is not unexpected as the initial indoor radon measurements are highly variable. Indoor radon measurements can vary widely between neighbouring houses owing to the different construction and ventilation of the houses. Therefore the natural (source and pathway) 22

23 component of radon distribution may explain less than two thirds of the measured variation in indoor radon, with the remainder a function of dwelling construction and lifestyle of occupants Radon mean per bedrock unit. Measured indoor radon data was joined to the underlying bedrock unit using the 1:100,000 geological bedrock map. All readings falling within each unit were then averaged to provide a mean value for that unit. This mean was then applied to occurrences of the same bedrock unit. Means were calculated only for bedrock units overlain by 20 or more radon points (214 in total). Where there were insufficient or no radon measurements for units at 1:100,000 scale, the geological units were grouped together into similar types based on the 1:1 Million scale geological map. As expected, radon mean per bedrock unit showed a reasonable correlation with %RL, with R 2 ranging up to 0.3. The fit improved as grid squares with increasing number of indoor readings were used Aquifer Type As radon is soluble in water and can be readily transported away from its source, the type of aquifer and the movement of groundwater are likely to be major influences on the distribution of radon gas. There have been a number of studies that have linked karstified rock with high concentration of radon (Gammage et al., 1992; O Connor et al., 1992). Karstified rocks are widespread throughout Ireland and therefore may play a significant role on the distribution of radon. Figure 6 shows an aquifer classification map superimposed by 3873 locations where measured indoor radon values exceed 200 Bq/m 3. 23

24 Figure 6: Aquifer classification with locations of indoor radon measurements that exceed 200 Bq/m 3. As can be seen from Figure 6 there seems to be a spatial correlation between regionally important aquifers, in particular regionally important karstified aquifers with conduit flow (Rkc), correlate with high radon values. Previous surveys in Ireland (O Connor et al., 1992, 24

25 1993) have shown that radon concentrations in soil gas and in dwellings are in places higher on limestone bedrock adjacent to granite bedrock than on the granite itself, even though granite generally contains a much higher concentration of uranium, and hence radon, than limestone. A calculated R 2 value indicates that aquifer type accounts for up to 10.4% of the overall observed variation in %RL, increasing to up to 17% when only using measurements on Carboniferous bedrock. Carboniferous rocks account for approximately 54% of bedrock in the country Degree of karstification As outlined in Section 3.1 the degree of karstification of the bedrock has been determined based on the properties of the rock type and given a ranking from 1 to 4. In certain areas the karstification classification can account for up to 18% of the variance of %RL data. However, due to the fact that the degree of karstification is imbedded within the aquifer type parameter and the degree of karstification in the many of rocks across the country is zero it was decided to drop the karst term and use aquifer type instead Groundwater recharge coefficient (GWRC) The GWRC is a measure of the ability of rainwater to pass through the subsurface and replenish the aquifer below. GWRC may thus also provide information as to the ease with which radon gas can move from the subsurface to the surface. GWRC is expressed as a percentage and is a national dataset based on the work of Hunter Williams et al. (2011). GWRC is derived using a combination of groundwater vulnerability, incorporating subsoil permeability and thickness values, the ability of the underlying aquifer to accept percolating waters and the degree of saturation of overlying soils. Calculated R 2 values for %RL and GWRC indicate that up to 19% of indoor radon variance can be accounted for by GWRC in certain areas. However, for the whole country GWRC accounts for just 7.2% of %RL variance. Further investigation of the GWRC value using grid squares with an increasing number of radon measurements led to the observation that, after initial improvements with the fit, i.e. increased R 2, the R 2 values started to decrease after using grid squares with 5 or more radon points (Figure 7). This decrease can be explained by the fact that all urban areas and/or 25

26 made ground areas have been assigned a fixed GWRC value of 20%. Grid squares with a greater number of readings are also those with a greater number of houses and hence more likely to be in urban areas or on made ground. Therefore the GWRC is only useful in rural areas. Similar responses showing R 2 values decreasing after initial rises for increasing number of radon points are also found for aquifer type and degree of karstification parameters R 2 values of GWRC based on grid squares with increasing radon points R2 value R2 of Recharge No of radon points per grid square Figure 7: R 2 values for the fit of GWRC against %RL versus the number of radon measurements per grid square Depth to bedrock (DTB) The national depth-to-bedrock (DTB) map is based on groundwater vulnerability mapping, available borehole information and mapping of outcrop. R 2 values indicate up to 20% of the variation in indoor radon measurements can be accounted for by DTB in certain areas. However, for the whole country 6.8% of variation in indoor radon measurements may be explained by DTB. It is interesting to note that DTB appears to have a greater correlation with indoor radon measurements in urban rather than in rural areas, possibly a reflection of the greater number of boreholes carried out in urban areas. This is in contrast to the GWRC value, and is also more significant in non-carboniferous areas than in areas underlain by Carboniferous (primarily limestone) rock. 26

27 4. Further Evaluation 4.1 Bedrock Terrain The model produced during the Tellus Border Project used airborne measurements of uranium in conjunction with GWRC and karst classification values and provided a good overall fit with indoor radon measurements. However, the model seems to favour certain geological terrains. Most of the grid squares used in the model were located within areas of Carboniferous bedrock. When the resultant model equation is applied to the entire grid in the Tellus Border region, the model seems to over-predict radon highs in non-carboniferous areas. Previous work by Appleton et al,. (2011) has shown that different beta coefficients, (variables derived from linear regression equation) are determined depending on the geology and hence it may be necessary to use terrain-specific models. To further evaluate the influence of bedrock geology on the different components of the model, the different datasets used in the model were subdivided by bedrock type and new regression analyses were applied. The datasets were separated into three main groups 1. All data (Carboniferous and non-carboniferous) 2. Carboniferous 3. Non-Carboniferous Results from regression analysis of the different parameters are shown in Appendix 3. The most significant parameters, i.e. those with the best fits to %RL, were radon averages per bedrock unit and aquifer type. However, data for grid squares overlying the non- Carboniferous bedrock show weaker correlations between the different variables and the %RL. In particular the aquifer type has a weaker correlation with %RL. However, DTB shows an increased correlation. It can be concluded that the %RL for all grid squares are more strongly correlated with the presence of Carboniferous bedrock. Grid squares overlying Carboniferous bedrock account for approximately 2/3 of all grid squares used within the model. The Carboniferous bedrock data was further sub-divided in to Visean and non-visean rocks using the 1:1 Million bedrock geology map. Little change in calculated R 2 values was seen 27

28 between the different Carboniferous subgroups suggesting Carboniferous rocks can be treated as a single group. Subdivision of the non-carboniferous rocks into an Ordovician Silurian and Neo-Protozoic grouping again using the 1:1 Million bedrock map shows a slight improvement in the fit of some of the data with aquifer type and GWRC parameters showing greater significance. However, fits are still generally quite poor. 4.2 Rural and Urban It was noted that the geological parameters in the model derived from national datasets, i.e. aquifer type, GWRC and karst classification, show a decline in their fit after an initial rise as the number of radon points per grid square increases. This was suggested to be an effect of urbanisation, i.e. the presence of made ground underlying housing, rather than natural soil / bedrock conditions. This was investigated by subdividing the data into rural and urban datasets. A grid square is classified as urban where more than 50% of its area is within an urban area. Regression analyses carried out for all parameters for both urban and rural groups are summarised in Appendix 3. The analyses shows that for rural data the R 2 shows little decrease with increasing number of radon points per grid square and in the case of aquifer type actually increases. Data from the urban group does show a decrease in R 2 values with increasing radon points per grid square. This would suggest that urban conditions do affect some of the model parameters. 4.3 Airborne Geophysical Data Although airborne geophysical data, in particular airborne measurements of uranium, are unavailable for the whole country, in areas where data are available (i.e. in the Tellus Border region of counties Donegal, Sligo, Leitrim, Cavan, Monaghan and Louth along with the area around Tralee and Castleisland in Co. Kerry) they can be shown to help improve the fit of the modelled %RL. Multivariate regression analysis of %RL data within the border region shows that up to 55% of the variation can be explained by a combination of airborne uranium measurements, radon average per bedrock unit, GWRC, DTB and aquifer type. Figure 8 shows the resultant model using these terms for the border region with predicted high 28

29 radon zones shown in dark brown. The model used grid squares with 5 or more radon points and was calculated for all grid squares with airborne data. These terms accounted for 55% of the variance of the percentage reference level. The linear regression analysis produced the following equation for %RL when airborne data was included.: %RL = U air AveRadon_bedrock GWRC Aquifer DTB A comparison of models derived with and without airborne data is discussed in Section6. Figure 8: Model of radon risk based on airborne uranium measurements, average radon per bedrock unit, GWRC, DTB and aquifer values. 4.4 Grid Size and Spatial distribution Previous mapping of radon data has been carried out on a 10 km grid size. One of the main objectives of this project is to produce a modelled map of radon distribution at a greater resolution. A 1 km grid size was used as, apart from improving resolution by a factor of ten, it corresponds with the grid used by the CSO for its mapping of national statistical data and 29

30 also with previous (Appleton et al., 2008, 2011) studies carried out in Ireland and Northern Ireland. However, there are too few indoor radon measurement locations to calculate a %RL value for most of the 1 km grid squares and hence the modelling is based on a relatively limited number of grid squares. A 4 km grid size was also tested to see if this provided any improvements in the model as you would expect to have more radon points per grid square. For this all data was assigned to the new 4 km grid squares, in the same manner as for the 1 km grid outlined above, and regression analysis carried out and summarised in Appendix 3. The radon-per-bedrock-unit and aquifer-type terms show slight improvements in their correlation with %RL when compared to the 1 km grid square correlations but the GWRC term shows zero correlation with %RL. The lack of correlation between GWRC and %RL in the 4 km grid square model, in contrast to the correlation observed for the 1 km grid square model, may be due to the fact that the GWRC national database is based on information from soil, aquifer and vulnerability data compiled at a scale of 1:25,000 to 1:100,000. This relatively high mapping resolution is not mirrored in the modelling because the data are averaged over the entire 4 km grid square. Therefore the subtlety of the GWRC term is lost when compared to the 1 km grid square modelling. In conclusion, increasing the grid square size from 1 km to 4 km does not lead to a significant improvement in correlation between %RL for the various parameters modelled and in the case of GWRC it actually leads to a decrease in correlation. Using 4km grid squares also leads to a loss of resolution by a factor of 4 compared to using 1km grid squares. Hence the 1km grid was used in order to produce a model at a greater resolution. 4.5 Conditional dependence Figure 9 shows a correlation matrix for the different datasets assessed for the final model. The results show that there is no significant inter-dependence between the different terms and that any resultant model would be relatively stable. 30

31 Figure 9: Correlation matrix of different datasets for national model, number is the R 2 value 31

32 5. Final Model Based on the evaluation of the different datasets and assessment of the model over the entire area as well as over subsets of the area it was decided to use the following terms in the multivariate linear regression equation: radon mean per bedrock unit; aquifer type; GWRC and DTB. Although the correlation between DTB and %RL in some cases was poor and the GWRC value was not significant in urban areas, it was decided to use both as they appeared to complement each other. In the urban areas where GWRC became less relevant the DTB term increased in significance and vice versa in rural areas. The regression was carried out based on a 1 km grid with the best fit found where there was a minimum of 7 readings per grid square. A total of 759 grid squares (approx. 1% of total) each with a minimum of 7 indoor radon readings were used for the model (distribution of used grid squares shown in Figure 10). This produced an adjusted R 2 value of 0.321, i.e. 32% of the variance of the radon distribution for these 759 grid squares can be explained by these four terms. The distribution of grid squares with 7 or more readings is mostly limited to urban areas, although a reasonable geographical spread is found although there is little representation within the midlands region. Linear regression analysis produced the following equation for %RL: %RL = AveRadon_bedrock GWRC Aquifer 0.5 DTB This equation could then be applied to all 72,668 grid squares to produce a modelled response of radon risk across the country. 32

33 Figure 10: Grid squares with 7 or more radon readings shown in red, grid squares with 1-6 readings in green. Grid squares with less than one radon reading in white. To directly compare with the existing national radon map the data was normalised to remove any negative values and each grid square was then assigned to one of the following five categories based on published maps by the RPII; 33

34 i) less than 1%, ii) ii) 1 5 %, iii) iii) 5 10%, iv) iv) 10 20% v) v) greater than 20%. These five categories represent the percentage risk of exceeding the reference level of 200 Bq/m 3. The distribution of the modelled data to these five categories was done to correspond with the distribution of radon data as reported in the Fennell et al., With; 40.5% of modelled data assigned to %RL of less than 1% 59% of modelled data assigned to %RL of less than 5% 68% of modelled data assigned to %RL less than 10 81% of modelled data assigned to %RL less than 20 Figure 11 shows a map of the final model. Areas of high radon (i.e. areas with greater than 10% chance of a dwelling exceeding the reference level of 200 Bq/m 3 ), dominate over Carboniferous limestones in the west of the country in counties Clare, Galway, Mayo, Roscommon and Sligo and also in the southeast of the country over the Leinster Granite and Lower Palaeozoic succession, including acid volcanic rocks, in counties Waterford, Wexford, Carlow and Wicklow. Other notable highs are found around Castleisland in County Kerry, southwest of Limerick city and in northern County Cork. 34

35 Radon Modelling for a National Map of Ireland 06/08/2014 Figure 11: Model of the risk of dwelling exceeding Radon reference level of 200Bq/m3. 35

36 6. Model Evaluation 6.1 Comparison with previous models The modelled map of radon distribution generally agrees well with the existing 10 km grid map of radon in Irish dwellings produced by the RPII. Zones of high radon are found in broadly similar areas to the west, northwest and southeast, although these are not exactly the same. The new model is strongly controlled by the local geology through the radonmean-per-bedrock-unit term and aquifer-type term and it is thus applicable to all areas and not just to certain geological terrains. Figure 12: Comparison of radon distribution maps produce by new model (left) and indoor measurements RPII. When the new model is compared with the model for the Tellus Border region developed using airborne uranium measurements (Figure 13), similar zones of high radon are predicted within areas of county Sligo and Leitrim, and a similar linear band running NE through the centre of the region, through counties Leitrim, Cavan and Monaghan. However zones of high radon values predicted by the Tellus Border airborne model (Figure 13B) for the area of north county Louth and the Cooley peninsula are not reproduced by the new model (Figure 13A). These areas may have been over estimated in the Tellus Border model due to the factor, derived from the regression equation, by which airborne uranium values are multiplied. This factor is generally large as most data corresponds to lower uranium values 36

37 over Carboniferous rocks. However, over high uranium values in these non-carboniferous rocks the relatively large factor results in elevated predicted radon values. Separate models incorporating the airborne data which account for different geological settings may need to be adopted. This conflict between the two models may equally be a result of limited availability of indoor data within the area. A B Figure 13: Comparison of radon risk models for A) new national model from Fig. 11 and B) model of Tellus Border region using airborne uranium data. 6.2 Test Area Tralee-Castleisland The area of Tralee and Castleisland in County Kerry has recorded some of the highest indoor radon readings measured not just in Ireland but throughout Europe and consequently has been the subject of considerable investigation. Appleton et al. (2011) produced radon models of the area based on airborne geophysics and geological data using a grid with 1 km squares. This study provides a good comparator for the the national model produced in this study. The Tralee-Castleisland area is underlain by Devonian sandstones, siltstones and conglomerates, Lower Carboniferous limestones and Lower Upper Carboniferous (Namurian) shales and sandstones. These rocks are folded in E W-trending anticlinal and synclinal structures. Soils in the area mostly consist of till containing clasts of Carboniferous and Devonian rocks with some peat deposits. A detailed description of the main bedrock geological units is given in Pracht (1996). Appleton et al. (2011) produced linear regression models using airborne measurements of uranium and data for rock permeability for the Tralee-Castleisland area. The best correlation 37

38 between the %RL and modelled estimates for the six townlands in the Tralee-Castleisland area had a R 2 value of 0.83, although this correlation is based on only 12 data points. The model is reproduced in Figure 14. Figure 14: Radon potential map of the Tralee-Castleisland area produced using the airborne measurements of uranium (eu) and Permeability (P) linear regression model (urban areas outlined in white). From Appleton et al. (2011) Figure 15 shows the same area modelled using the new national radon model produced using radon average per bedrock unit, aquifer type, GWRC and depth to bedrock as discussed above. The same strong radon high running northwest southeast through the centre of the area is again visible. Generally low radon values are present to the west and north of Castleisland. The new national radon model map indicates lower radon values in the southwest of the Castleisland-Tralee survey area as predicted by Appleton et al., (2011). 38

39 Figure 15: New national model of Castleisland area (2014). Figure 16: Model based on airborne uranium, average radon per bedrock unit, aquifer, DTB and GWRC values. Model shows similar output as to Appleton model using airborne uranium and permeability data (Figure 14). 39

40 Overall, reasonably good agreement is seen between the Appleton et al. s 2011 model and the new national model (Figure 15), providing some support for the national model. Some differences do exist and these may reflect the impact of using airborne data as Figure 16 modelled using components of the national model along with airborne uranium values show very similar results to those produced by Appleton et al., (2011). Comparison of the new national model and the RPII s existing radon map in the area (Figure 17) shows similar broad trends although due to the scale of the RPII map (10 km grid squares) any detail is lost and comparison becomes difficult. Both the modelled map of Appleton et al. (2011) and the new national map show a much higher level of spatial detail compared to the published RPII radon map. The model produced using airborne uranium, mean radon per bedrock unit, aquifer and GWRC values (Figure 16) shows a very similar distribution as to that produced by Appleton et al., (Figure 14). Figure 17: RPII 10km grid radon map of the Tralee-Castleisland area (main urban areas outlined in white, redrawn from the data in Fennell et al. (2002)). 40

41 7 Conclusions and Recommendations A multivariate linear regression model based on average indoor radon values per bedrock unit, groundwater recharge coefficient, aquifer type and depth to bedrock has led to a new national map of radon risk. The percentage variance accounted for by the different variables within the linear regression models are relatively low. However, there is good correlation between radon highs and lows identified on this map and those recorded on previous maps and in other detailed investigations. Numerous elements influence the nature of the indoor radon data, including the building type or construction method along with social factors such as how well ventilated is the house is unknown. Therefore the model can only account for the variance explained by natural geological factors. The final model for the derived national radon risk map, used the parameters; radon mean per bedrock unit, aquifer type, Groundwater Recharge Coefficient (GWRC) and depth to bedrock (DTB). All parameters have been shown to be generally independent from one another (Figure 8). It could be argued that the GWRC and DTB terms both include data on overburden thickness. However, the two terms complement each other when applied to rural and urban areas and therefore both terms were included in the final model. The degree of karstification term was not used in the final model (as was part of previous models, (Hodgson & Carey 2013) as this was co-dependent with the aquifer type term. Aquifer type was found to be more significant than degree of karstification and therefore was used in the final model. The overall model is best at explaining natural conditions rather than the made ground found in urban areas. In these urban areas there are typically a large number of indoor radon measurements and therefore these may act as a better guide. However, in more rural areas the new model provides a useful guide to expected radon levels. The final model is based on an equation derived from multivariate regression using 737 grid squares containing 7 or more radon measurements. This was found to be the best fit. When data are split into different areas or groupings, e.g. the border region or by geology, different regression models and fits are produced. However, to produce a national map for all areas of the country it was important to use as many grid squares as possible. The use of the radon mean per bedrock unit has helped removed the requirement to make terrainspecific models, as the variation due to the local geology is included. This is shown by similar beta factors in the model equations derived for the radon mean per bedrock unit term in models within different geological terrains. 41

42 The overall goodness-of-fit of the model based on the percentage variance accounted for by different variables is relatively low at 32%. This is due to the high variability of the indoor data as well as the considerable uncertainty that exists where there are a limited number of indoor measurements. The final model predicts, in much greater detail than previous models, in particular high radon areas are modelled in the west of the country, particularly in parts of counties Clare, Limerick, Galway, Mayo, Roscommon and Sligo. These areas are predominantly categorised by limestone bedrock with considerable karstifation in places. Radon highs are also modelled in the southeast of the county over parts of county Waterford, Wexford, Carlow and Wicklow where granite and acidic volcanic rocks form a major part of the bedrock. Interestingly the model predicts moderate radon values over granite rocks in counties Donegal, Galway and Louth and not the extreme highs associated with many granitic bodies. These granites are more radiogenic than most granites within the country (O Connor et al., 1983) and therefore these lower values may be a result of thicker or less permeable overburden overlying these rocks or simply few available indoor radon data points in these generally more remote areas. Aquifer type and the movement and concentration of radon gas within groundwater seem to be very significant in the transport of radon. This is demonstrated by the strong relationship between regionally important karstified aquifers, particularly those with conduit flow, and radon highs. Even though the limestone rock within which the groundwater flows is generally weakly radioactive with a low source potential for radon gas the fractures within which the groundwater flows provide locations of concentrated radon gas and a mechanism by which they can degas at the surface. The newly derived national model effective reproduces the existing indoor radon data at greater resolution and to geological boundaries. However, it has been shown that modelled airborne data can improve the model fit and potentially identify new radon hotspots previously unmapped by existing radon maps or the newly developed national radon model. 42

43 Recommendations Where possible airborne geophysical data, in particular uranium measurements, should be used to help model radon distribution, as the percentage variance can be significantly decreased when it is incorporated into models. Differences exist between the derived national model and specific models incorporating airborne data where it exists. In particular uranium highs in the northern part of county Louth predict high radon levels from the airborne model but not necessarily from the national model. These differences require further investigation. A focussed campaign to collect new indoor radon measurements along with radon soil gas readings and local mapping are recommended to help investigate these variations. Other factors not considered in this report may control this variability. When new airborne data becomes available this should be tested against existing maps and incorporated into new models. New data may help in trying to understand the relationship between airborne geophysical parameters and variation in radon data over different geological terrains. Further testing of both the national model with models derived from geological and airborne geophysical data should continue. Any newly available in-door radon data should be used to test the viability of the model and then included in the modelling to further improve it. Geographically weighted regression may also be a useful way of modelling the national distribution of radon gas. Any model produced using this approach can then be compared to the latest model produced through multivariate linear regression. 43

44 9. References Appleton, J.D., Doyle, E., Fenton. D. and Organo, C Radon potential mapping of the Tralee-Castleisland and Cavan areas (Ireland) based on airborne gamma-ray spectrometry and geology. Journal of Radiological Protection, 31, Appleton, J.D and Ball, T.K Geological radon potential mapping. In: Bobrowsky, P.T (Ed), Geoenvironmental mapping: Methods, Theory and Practice. Balkema, Rotterdam, pp Appleton J D and Miles J C H 2008 Application of Tellus airborne radiometric and soil geochemical data for radon mapping in Northern Ireland British Geological Survey Commissioned Report CR/08/049 Commission of the European Communities, Exposure to natural radiation in dwellings of the European Communities. Luxemburg, Commission of the European Communities. Fennell, S.G., Mackin, G.M., Madden, J.S., McGarry, A.T., Duffy, J.T., O Colmain M., Colgan, P.A. and Pollard D Radon in dwellings. The Irish national radon survey. Radiological Protection Institute of Ireland. Gammage, R.B.; Dudney, C.S.; Wilson, D.L.; Saultz, R.J.; Bauer, B.C Subterranean transport of radon and elevated indoor radon in hilly karst terrains. Atmospheric Environment Vol.26A, No.12: Geological Survey of Ireland Bedrock Geology of Ireland, 1:1,000,000 Scale (map). Gunning, G.A., Pollard, D. and Finch, E.C An outdoor radon survey and minimizing the uncertainties in low level measurements using CR-39 detectors. Journal of Radiological Protection

45 Hodgson J.A. and Carey S Radon risk predictive modeling using airborne geophysical data in the border region of Ireland. Geological Survey of Ireland and geological Survey of Northern Ireland Joint report. Hunter Williams N., Misstear B., Daly, D., Johnston, P., Lee, M., Cooney, P. and Hickey, C A National groundwater recharge map for Ireland. National Hydrology Conference Irish working Group on groundwater, gw2 on gwb delineation Miles, J. C. H. and Appleton, J. D Mapping variation in radon potential both between and within geological units. Journal of Radiological Protection, 25: Nazaroff, W.W Radon transport from soil to air. Reviews of Geophysics, Vol 30, Issue 2, pp O'Connor P J, Gallagher V, Van den Boom G, Hagendorf J, Müller R, Madden J S, Duffy J T, McLaughlin J P, Grimley S, McAulay I R, Marsh D 1992 Mapping of 222 Rn and 4 He in soil gas over a karstic limestone-granite boundary correlation of high indoor 222 Rn with zones of enhanced permeability Rad. Prot. Dos O'Connor P J, Gallagher V, Madden J S, Van den Boom G, McLaughlin J P, McAulay I R, Barton K J, Duffy J T, Muller R, Grimley S, Marsh D, Mackin G, Mac Niocaill C Assessment of the geological factors influencing the occurrence of radon hazard areas in a karstic region GSI Report Series RS 93/1 (Environmental Geology) (Dublin: Geological Survey of Ireland) 204pp O Connor, P.J., Long, C.b and Hennessy, J Radioelement geochemistry of the Irish Newer Caledonian Granites. The Geological Survey of Ireland Bulletin. Vol 3. No

46 Pracht M 1996 Geology of Dingle Bay (Dublin: Geological Survey of Ireland). RPII (2010). Protocol for the Measurement of Radon in Homes UNSCEAR Report of the United Nations Scientific Committee on the Effects of Atomic Radiation to the General Assemby. Sources and effects of Ionising radiation. United Nations, New York. UNSCEAR, United Nations Scientific Committee on the Effects of Atomic Radiation (UNSCEAR). UNSCEAR 2006 Report. Annex E. Sources-to-Effects Assessment for Radon in Homes and Workplaces. United Nations, New York WHO, WHO Handbook on Indoor Radon: A Public Health Perspective. World Health Organisation (WHO), Geneva. 46

47 Appendix 1: Data Assessment Summary by 1km Grid There are 72,660 1km grid squares over the whole country with available geological data based on CSA grid. Total land mass of Ireland (excluding Northern Ireland) is 70,273 km 2. Some of the grid squares are not full 1km grid squares. These partial squares are mostly found along the coast. Geology type based on the 1:1Million bedrock map has been assigned to each grid square based on the dominant unit (by percentage area) (Figure 18). The results for the whole country can be summarised as follows; Figure 18: Percentage of bedrock type over whole country of Ireland based on1: 1 Million bedrock map and 1 km grid squares. The carboniferous accounts for nearly 54% of all bedrock within the country however this can be subdivided into; Namurian shale, sandstone, siltstone, coal -(7.8% of total area) Tournaisian limestone - (11.1% of total area) Tournaisian sandstone, mudstone, limestone - (3.8% of total area) Visean limestone & calcacareous shale - (28.4% of total area) Visean sandstone, mudstone, evaporate - (2.1% of total area) 47

48 Westphalian shale, sandstone, siltstone, coal - (0.5% of total area) The Quaternary sediment variation across the country is summarised in Figure 19. Figure 19: Percentage of quaternary sediments across the whole country based on new quaternary map GSI. To assist in modelling of the radon and airborne data the following table shows the number of grid squares with available data. Obviously the number of grid squares available for modelling decreases with increasing number of radon points measured in an individual grid square. Only 12,290 out of 72,660 grid squares have at least one radon measurement made within them. However, only 2327 grid squares have at least 3 or more indoor radon readings needed to calculate the percentage reference level (%RL) required for linear regression modelling. The number of available grid squares and minimum radon readings is also shown by available airborne data as well as within the Tellus Border (TB) and Tralee-Castleisland areas. 48

49 Table 3: Number of grid squares for each dataset subset by airborne number of Radon data point. Tb refers to Tellus Border project with counties Donegal, Leitrim, Sligo, Cavan, Monaghan and Louth. File No Grid Squares All_GRID_DATA All_GRID_DATA_with Airborne data All_GRID_DATA_with Radon data All_DATA_3 or more Radon pts per grid sq 2327 All_DATA_5 or more Radon pts per grid sq 1113 All_DATA_7 or more Radon pts per grid sq 759 Airborne data with3 or more Radon pts per grid sq 503 Airborne data with 5 or more Radon pts per grid sq 234 Airborne data with 7 or more Radon pts per grid sq 148 Airborne_TB_ with 3 or more Radon pts per grid sq 392 Airborne_TB_ with 5or more Radon pts per grid sq 183 Airborne_TB_ with 7 or more Radon pts per grid sq 117 Airborne_Castleisland with 3 or more Radon pts per grid sq 111 Airborne_Castleisland with 5 or more Radon pts per grid sq 51 Airborne_Castleisland with 7 or more Radon pts per grid sq 31 49

50 Appendix 2: Indoor radon mean values per bedrock unit Indoor radon measurements were assigned to the underlying bedrock unit based on 1:100,000 geological Bedrock map, produced by the Geological Survey of Ireland. All measurements falling within each unit are then averaged to produce a mean radon value for each geological unit. The count column indicates how many indoor radon measurements were made within each bedrock unit. Mean and median values are shown in Bq/m ,000 Bedrock Unit Count Mean Median Abbeytown Limestone Achill Head Formation Addergoole River Formation Agglomerate Aghfarrell Formation Aghmacart Formation Aghyaran & Killygordon Ahenny Formation Aille and Barney Fms Aille Limestone Formation Aillwee Member Aillwee member (lower) Aillwee Member (upper) Allenwood Formation Anaffrin Formation Annabella Formation Annagh Division (undiff) Annascaul Formation Appinite suite Ardagh Shale Formation Ardane Formation Ardara Granite G1 variety Ardara Granite G2 variety Ardara Granite G3 variety Ardaturrish Member Ardenagh Formation

51 Ardmore Member Ardnamanagh Member Ardnasillagh Formation Ards Pelite Formation Ards Quartzite Formation Ardvarney Formation Argillaceous limestones Argillaceous Limestones Arklow Head Formation Arthurstown Member Ashleam Bay Formation Ashleam Bridge Graphitic Ashleam Head Formation Askingarran Formation Assaroola Member Athassel Limestone Formation Atlantic Drive Schist Formation Attymass Group (undiff) Aughnanure Oolite Format Avoca Formation Ayle River Formation Aylecotty Member Balbriggan Formation Ballard Member Ballina Limestone Format Ballina Limestone Format Ballinatray Formation Ballindysert Formation Ballinskelligs Sandstone Balliny Member Ballyadams Formation Ballybeg Member Ballybeg Park Limestone

52 Ballybro Formation Ballycar Formation BALLYCOGLY GROUP Ballyconneely Amphibolite Ballyelly Member Ballygeana Formation Ballyglasheen Oolite Formation Ballyhack Member Ballyhest Member Ballyhoge Formation Ballyknock Member Ballylane Formation Ballymartin Formation Ballymore Limestone Formation Ballymore Sandstone Formation Ballynacarrig Member Ballynaclogh Formation Ballynahown Sandstone Formation Ballynakill Schist Formation Ballynamuddagh Granite Ballynash Member Ballyquinn Member Ballyshannon Limestone Formation Ballysteen Formation Ballytrasna Formation Balrickard Formation Banagher Sandstone Formation Bangor/Corslieve Formation Barnanoraun Schist Formation Barney Limestone Formation Basal clastics Basal sandstones Basalt

53 Beenlea Head Member Belcamp Formation Bellagarvaun Formation Bellavally Shale Formation Ben Levy Grit Formation Benbulben Shale Formation Bencorragh Formation Benmore Formation Benmore Formation\ Broad Bennabeola Quartzite For Bere Island Member Bird Hill Formation Birreen Formation Birreencorragh Quartzite Birreencorragh Schist Formation Black Head Member Blackstones Type 2 Equigranular Blackstones Type 2 Micro Bockagh Member Bohoge Member Booley Bay Formation Borrisokane Formation Boston Hill Formation Boultypatrick (Grit) Formation Bouris Formation Boyle Sandstone Formation Boyne Formation Bray Head Formation Bregaun Flagstone Formation Bricklieve Limestone Formation Bricklieve Limestone Lower Bricklieve Limestone Upper Briska Boulder Bed Formation

54 Brittstown Formation Broad Haven Formation Broadford Formation Brownsford Member Brownstown Head Member Buckoogh Formation Bullockpark Bay Member Bulls Head Formation Bunaveela Lough Formation Bundoran Shale Formation Bunmahon Formation Bunnamohaun Siltstone Formation Burren Formation Burrishoole Member Butlersgrove Formation Butter Mountain Formation Cabra Formation Cadamstown Formation Caha Mountain Formation Caherduggan Limestone Formation Callow Formation Callowfinish Granite Calp Campile Formation Capnagower Formation Cappagh Sandstone Formation Cappagh White Sandstone Carna-type Carnes Formation Carnsore Granite Carracastle Formation Carraun Shale Formation Carrawaystick Aplite

55 Carrick OHara Formation Carrickaness Sandstone Formation Carrickateane Formation Carrickatee Formation Carrickleck Sandstone Member Carrighalia Formation Carrighill Formation Carrigmaclea Formation Carrysalia Formation Cashel Schist Formation Castlebar River Fm. / Lo Castlehaven Formation Castlequarter Member Castlerahan Formation Castleshane Formation Central Belt (undiff) Central Clare Group Chert Claragh Sandstone Formation Clare Shale Formation Clashabeema Formation Clashavodig Formation Clashford House Formatio Clatterstown Formation Claudy Formation Clay Gall Sandstone Formation Cloghan Sandstone Format Clogherhead Formation Cloghmore Granodiorite Clogrenan Formation Clonaslee Member Clonmass Limestone Member Clontail Formation

56 Cloonagh Limestone Formation Cloone Flagstone Formation Cloonierin Formation Cloonnamna Formation Collon Formation Cong Canal Formation Cong Limestone Formation Conlanstown Formation Coolbaun Formation Cooldaragh Formation Cootehill Member Copstown Limestone Forma Coranellistrum Formation Cork Red Marble Formation Corn Hill Formation Cornagnoe Formation Cornamona Marble Formation Coronea Formation Corratober Bridge Formation Corraun Quartzite Member Corvock Granite Costello Murvey Granite Coumeenoole Sandstone Formation Coumshingaun Conglomerate Courtmacsherry Formation Courtown Formation Cracoean Reef Member Craggagh Shale Formation Cranford Limestone Formation Creagh Member Creeslough Formation Cregg Limestone Formation Cregganbaun Formation

57 Cregmahon Member Crinoidal limestone Croaghaun Formation Croaghmarhin Formation Croaghubbrid Pelite Formation Croane Formation Croghan Limestone Formation Cross Lake Formation Crossdoney Granite Crosspatrick Formation Crows Point Formation Crufty Formation Cruicetown Group (undiff) Cuilleen-type Culdaff Limestone Formation Cullenstown Formation Cullentra Formation Culmore Formation Cuskinny Member Dargan Limestone Dartry Limestone Formation Deer Park Complex Deer Park Schist Formati Delaney Dome Meta-rhyolite Denhamstown Formation Dergvone Shale Formation Derravaragh Cherts Derryfadda Formation Derrylea Formation Derrymore Formation Derryveeny Formation Devils Glen Formation Dinantian Limestones (undiff)

58 Diorite Dirtoge Limestone Formation Dolerite Dolerite and Gabbro Donard Andesite Member Donore Formation Dooega Head Formation Doonamo Formation Doon-na-Dell Schist Formation Dowery Hill Member Downpatrick Formation Dromkeen Limestone Formation Drumgesh Shale Formation Drumleck Formation Dunabrattin Formation Duncormick Formation Durnish Formation Durrow Formation Early Gabbro Eask Sandstone Formation Edenderry Oolite Member Edergole Formation Elsinore Formation Errisbeg Townland Granite Extrusive Rhyolite Formation Fahan Grit Formation Fahan Slate Formation Falcarragh Limestone Formation Fanad Granite Fanore Member Farnacht Formation Feale Sandstone Formation Fearnaght Formation

59 Feighcullen Formation Feldspar or Quartz Porphrye Ferriters Cove Formation Fiddaun Member Finavarra Member Fingal Group (undiff) Finlough Formation Finnalaghta Formation Galtymore Formation Garraun Member Garryduff Formation Gaugin Quartzite Formation Giants Grave Formation Glandahalin Formation Glascarrig Formation Glashabeg Conglomerate Glaspistol Formation Glen Ding Formation Glen Lodge Formation Glen Pebbly Arkose Formation Glenade Sandstone Formation Glencar Limestone Formation Glencolumbkille Limeston Glencolumbkille Pelite Formation Glencullen River Formation Glencullin River Formation Glenflesk Chloritic Sandstone Glenisland Formation Glenoween Shale Formation Glenummera Formation Gortanimill Formation Graffa More Formation Granite (undiff)

60 Granodiorite and diorite Granophyre Greencastle Green Beds GREENORE POINT GROUP Greyfield Formation Gubroe Quartzite Member Gull Island Formation Gun Point Formation Gyleen Formation Harrylock Formation Hawaiite Lava Hawkhill Member Hazelwood Limestone Formation Herbertstown Formation Herbertstown Limestone Formation Hilltown Formation Hollyford Formation Holmpatrick Formation Horan Formation Hore Abbey Limestone Formation Illaunagappul Formation in Ballyadams Formation in Ballyhack Member in Ballymoyle Formation in Ballynaclogh Formation in Ballyneale Member in Ballysteen Formation in Bray Head Formation in Bricklieve Limestone in Broadford Formation in Bullockpark Bay Member in Burren Formation in Butlersgrove Formation

61 in Campile Formation in Dartry Limestone Formation in Kilmacthomas Formation in Milford Formation in Newtown Head Member in Salterstown Formation in Slieve Bernagh Formation in Visean Limestones (undiff) in Waulsortian Limestone in Wexford Formation Inchacoomb Formation Inishderry Formation Inishkea Division (undiff) Inishowen Head Grits Inniskeen Formation Inshaboy Formation Inver Schist Formation Johnstown Red Marble Formation Kanfinalta Formation Keadew Formation Keeper Hill Formation Kehernaghkilly Formation Kennetstown Formation Kilanena Formation Kilbride Formation Kilbryan Limestone Forma Kilcarry Member Killadangan Formation Killag Formation Killala Oolite Member Killeshin Siltstone Formation Killeter Quartzite Formation Killin Formation

62 Kilmacrea Formation Kilmacthomas Formation Kilmore Formation KILMORE QUAY GROUP Kilmurry Sandstone Formation Kiln Bay Formation Kilnafrehan Conglomerate Kilsheelan Formation Kiltorcan Formation Kingscourt Gypsum Formation Kingscourt Sandstone Fomationr Kinrovar Schist Kinsale Formation Kinturk Member Knock Granite Knockavellish Member Knockerk Formation Knockletteragh Member Knockmaa Formation Knockmealdown Sandstone Knockmore Limestone Member Knockmore Sandstone Formation Knockordan Limestone Formation Knockroe Basalt Lava Flow Knockroe Lithic Tuff Member Knockroe Vitric-Lithic Knockseefin Lava Flow Member Lack Sandstone Formation Lacka Sandstone Formation Lackan Formation Lackantedane Formation Lagganstown Formation Lakes Marble Formation

63 Laragh Formation Layered Gabbro Leckee Quartzitic Formatation Leitrim Group Letterbrock Formation Lettergesh Formation Lickfinn Coal Formation Lifford Volcanic Member Liscarragh Formation Liscarroll Limestone Formation Lisduff Oolite Member Lisgorman Shale Formation Lismaline Micrite Formation Lispatrick Formation Lissylisheen Member Lithologically diverse Little Harbour Formation Little Island Formation Longstone Shale Member Lough Acoose Sandstone Formation Lough Avaghon Formation Lough Brohly Quartzite Formation Lough Carra Member Lough Eske Psammite Formation Lough Foyle Succession Lough Guitane rhyolites* Lough Guitane Volcanicla Lough Gur Formation Lough Kilbride Schist Formation Lough Lurgan Granite Lough Mask Formation Lough Mourne Formation Lough Nacorra Formation

64 Loughros Formation Loughshinny Formation Lower Attymass Formation Lower Crana Quartzite Formation Lower Falcarragh Pelite Lower Limestone Shale Lower Lismoran Formation Lower Sandstone Member Lucan Formation Luggacurren Shale Formation Maam Formation Magma Mixing-Mingling Zone Magoney Bridge Formation Main Donegal Granite Malahide Formation Malin Schist Formation Marble Marginal Porphyritic Granite Maulin Formation Maumcaha Member Maumtrasna Formation Maydown Limestone Formation Meath Formation Meelick Member Meenymore Formation Metadolerite Metagabbro & orthogneiss Metagabbro Milford Formation Milltown Formation Milverton Group (undiff) Mine Head Member Minnaun Sandstone Formation

65 Moathill Formation Moneyteige Member Mornington Formation Mount Partry Formation Moy Sandstone Formation Moyadd Coal Formation Moygara Formation Moyny Point Limestone Member Muckros Sandstone Formation Mudbank limestone Mullaghfin Formation Mullaghmore Sandstone Formation Mullanalt Member Mullyfa and Deele Formation Murvey Granite Mweelrea Formation Namurian (undiff) Narrow Cove Member Naul Formation Navan Beds Nephin Formation Newry Granite Newtown Formation Newtown Member North Carrowgarve Formation Oaklands Formation Oakport Limestone Formation Oghill Formation Old Head Sandstone Formation Old Red Sandstone (undiff) Oldchapel Limestone Formation Oldcourt Cherty Limestone Oldcourt Member

66 Omey Granite Oolitic limestone Oughterard Granite Owenriff Member Ox Mountains Granodiorite Palace Member Paragneiss Migmatite Parsonage & Corgrig Lodge Pigs Cove Member Pipers Gut Formation Platin Formation Pollacappul Formation Pollaphuca Formation Pollareagh Member Polldarrig Formation Porphyritic granophyre Porphyritic-Megacrystic Port Askaig Formation Porters Gate Formation Portnahally Formation Portrane Volcanic Formation Poulgrania Sandstone Formation Quartz Diorite Gneiss Quartz Diorite Gneiss & Schist Quartz porphyry and Felspar Quinagh Formation Rafts in Thorr Granite Rathkeale Formation Rathkenny Formation Rathronan Formation Red Island Formation Red Mans Cove Formation Reelan Formation

67 Reenagough Member Reenydonagan Formation Rickardstown Formation Ridge Point Psammitic Formation Ringmoylan Formation Rinn Point Limestone Formation Riverchapel Formation Rockfield Limestone Formation Rockfield Sdst. Mbr Rockfleet Bay Limestone Rosroe Formation Ross Member Ross Sandstone Formation Rosses Granite G1 variety Rosses Granite G2 variety Rosses Granite G3 variety Roundstone Granite Rush Conglomerate Formation Salrock Formation Salterstown Formation Scotch Port Schist Seamount Formation Serpentinite Sessiagh-Clonmass Formation shale units in Campile Formation Shanagolden Formation Shanmullagh Mill Formation Shannapheasteen Granite Shannon Group Sheeffry Formation Shelmaliere Formation Shercock Formation Sherkin Formation

68 Sheskin Formation Silverspring Formation Skerdagh River Volcanic Skerries Formation Slaheny Sandstone Formation Slate Quarries Formation Slea Head Formation Slevoir Formation Slieve Bernagh Formation Slieve Bernagh Formation Slieve Gamph Igneous Com Slieve Gamph Igneous Com Slieve Gamph Igneous Com Slieve Glah Formation Slieve League Formation Slieve Tooey Quartzite Formation Slievemore Psammitic Formation Slievenagark Member Slievenaglasha Formation Slievenamuck Conglomerate Slievereagh Conglomerate Slishwood Division Creg Slishwood Division Peli Slishwood Division Psam Slishwood Division Semi South Lodge Formation Sraheens Lough Formation Srahlaghy Quartzite Formation Srahmore Lodge Dolomite Srahmore Quartzite Member St. Finans Sandstone Formation St. Marys Basalt Stackallan Member

69 Streamstown Schist Formation Suir Limestone Formation Swan Sandstone Member Syenite Taghart Mountain Formation Tawnaghmore Formation Tawnyinagh Formation Templetown Formation Termon Formation Termon Granite Terryglass Formation Thorr Granite Tipperkevin Formation Tober Colleen Formation Toe Head Formation Tonweeroe Formation Tourmakeady Formation Trachyte Tramore Limestone Formation Tramore Shale Formation Trawenagh Bay Biotite Granite Trawenagh Bay Biotite-muscovite Tubber Formation Tuff Tullagh Point Granite Tullig Sandstone Tullow Type 2 Equigranular Tullow Type 2 Microcline Tullow Type 2 Sparsely Prophrye Tullyallen Formation Twigspark Formation Type 1 Granite Type 1 granodiorite

70 Type 2e equigranular Type 2p microcline porphyre Type 3 muscovite porphyre Ulster Canal Formation Unassigned metasediment Upper Attymass Formation Upper Crana Quartzite Formation Upper Falcarragh Pelite Upper Lismoran Formation Valentia Slate Formation Visean Limestones (undiff) Volcaniclastic Rocks Walshestown Formation Waterfall Member Waulsortian Limestones Westphalian (undiff) Westport Grit Formation Westport Oolite Wexford Formation White Island Bridge Form White Strand Formation Wicklow Head Formation

71 Appendix 3: Regression Summary The following tables show the percentage variance accounted for by different variables in linear regression models carried out during this study. Data is summarised for all available data across the country but is also subdivided into geological categories and rural urban categories. Urban areas are identified by a grid square of which more than 50% of the area is mapped by Ordnance Survey as urban. 71

72 Regression Summary Ranking Model for all data (R-Sq Results) ALL Data National Rn_values per grid sq 3plus 5plus 7plus 9plus 12plus 15plus 20plus 25plus 30plus 35plus 40plus Total grid Sq's Element Rn_bedrock_Mean Aquifer Recharge Aquifer*Recharge Karst DTB Karst_DTB Bed_Rn+Aqu+Rech+DTB

73 Regression Summary Ranking Model (R-Sq Results) based on geology Carboniferous Non-Carboniferous Ordovician - Silurian Rn_values per grid sq 3plus 5plus 7plus 3plus 5plus 7plus 3plus 5plus 7plus Total grid Sq's Element Rn_bedrock_Mean Aquifer Aquifer*Recharge Recharge DTB Karst_DTB bed_rn+aqi_rech+karst_dtb

74 1km Regression Summary Urban 50% (R-Sq Results) ALL Data National Rn_values per grid sq 3plus 5plus 7plus 9plus 20plus 25plus Total grid Sq's Element Rn_bedrock_Mean Aquifer Aquifer*Recharge Recharge DTB Karst_DTB elevation bed_rn+aqi_rech+karst_dtb km Regression Summary Rural 50% (R-Sq Results) ALL Data (national) Rn_values per grid sq 3plus 5plus 7plus 9plus 20plus 25plus Total grid Sq's Element Rn_bedrock_Mean Aquifer Aquifer*Recharge Recharge DTB Karst_DTB elevation bed_rn+aqi_rech+karst_dtb

75 Regression Summary 4km Model (R-Sq Results) ALL Data National Rn_values per grid sq 3plus 5plus 7plus 20plus 25plus Total grid Sq's Element Rn_bedrock_Mean Aquifer Aquifer*Recharge Recharge DTB Karst_DTB bed_rn+aqi_rech+karst_dtb

76 Regression Analysis: %RL versus Average radon per Bedrock, RECHARGE, AQUIFER, DTB The regression equation is %RL = Ave_Radon_bedrock RECHARGE AQUIFER DTB Predictor Coef SE_Coef T P Constant Ave_Radon_bedrock RECHARGE AQUIFER DTB S = R-Sq = 32.3% R-Sq(adj) = 31.9% Model output from Minitab. P-Values show all data is significant. 76

77 Appendix 4: A guide to Data preparation for Radon Modelling A1 Introduction This is a step by step guide to the tools utilized for organising the datasets used in the radon modelling exercise, by the Geological Survey of Ireland. ESRI ArcGIS Desktop ArcMap 10.1 was used to process the data spatially and the open source programming language, R was used to calculate some of the statistics on the various datasets. A1.1 Datasets Assessed The main controlling dataset for this project was the Central Statistics Office (CSO) 2011 Census grid dataset (Figure 20) which can be downloaded from This is a 1km 2 grid dataset which contains information on population density and population type per grid square. A series of different datasets were assessed although not all were used in the final model. 77

78 Figure 20 CSO Census km grid dataset 78

79 All datasets are fitted to this grid. The datasets fitted to this grid are as follows: Dataset Bedrock Geology 1:100K Bedrock Geology 1:1million Quaternary Sediments Groundwater Recharge Coefficient (%) Depth to Bedrock (m) Degree of Karstification National Bedrock Aquifer Type Airborne Total Count (counts/s) Airborne Potassium (%) Airborne Thorium (ppm) Airborne Uranium (ppm) Indoor Radon (Bq/m 3 ) OSI DEM 10 Meter (m) Towns Faults Type Polygon Polygon Polygon Polygon Polygon Polygon Polygon Point Point Point Point Point Raster Polygon Line The forthcoming sections will describe how each dataset was analysed. A1.1.1 Bedrock Geology 1:100,000 The Bedrock Geology 1:100,000 polygons were overlaid with the CSO grid. Each grid square was assigned a geology type. In the event of two or more geology types in one grid square, the predominant geology for that grid square was assigned to that grid square. For example, let s say a grid square contained 65% Limestone and 35% Sandstone, the grid square is given a geology type Limestone. Figure 21 shows the Bedrock Geology layer, while Figures show the tools used to undertake this analysis. 79

80 Radon Modelling for a National Map of Ireland 06/08/2014 Figure 21 1:100,000 (100K) Bedrock Geology GIS analysis of bedrock geology per grid square was carried out in ArcGIS for Desktop Step one was to use the Tabulate intersection tool from Arctoolbox. 80

81 Figure 22 Tabulate Intersection from the Statistics toolbox in ArcToolbox This tool was used to define what geology type intersected what grid square. A percentage value was returned for each grid square. So, if there was only one geology type for a grid square, that square would be give 100%. If there were two geology types, the break down would be given for each geology type for that grid square: Geology A= 55%, Geology B=45% for example. The data was then sorted per grid square, by max percentage. This ensured that when undertaking summary statistics for each grid square, the max geology percentage for that grid square could be assigned to that grid square. 81

82 Figure 23 Sort the data using Data Management Toolbox, Sort Figure 24 Use the Summary Statistics toolbox from Analysis tools to obtain statistics for each grid square Once summary statistics had been undertaken, the data was joined back to the original CSO grid using OBJECTID_1 as your join field. This exact approach was undertaken for the 1:1million Bedrock Geology dataset, the Quaternary Sediments dataset, the Degree of Karstification dataset and the Groundwater National Aquifer map. 82

83 A1.1.2 Karst The degree of karstification was based on the Karst terms of: Non Limestone, Impure Limestone, Pure Unbedded Limestone Pure Bedded Limestone For each category a numerical value was assigned to each grid square (as in section A1.1.1), the following python code was run to give each category a numerical weighting. 83

84 ############################################################################### # PYTHON CODE FOR Karst WEIGHTING ############################################################################### import arcpy fc="l:\\tellusborder\\radon_2014\\modelling_2014.gdb\\radon_grid_output_v2" fieldlist = arcpy.listfields(fc) # Loop through each field in the list and print the name for field in fieldlist: print field.name rows = arcpy.updatecursor(fc) #this is my feature layer for row in rows: if row.getvalue("karst_weighting") == "Non Limestone": row.setvalue("karst_weighting", "0.1") elif row.getvalue("karst_weighting") == "Impure Limestone": row.setvalue("karst_weighting", "2") elif row.getvalue("karst_weighting") == "Pure Unbedded Lmst": row.setvalue("karst_weighting", "3") elif row.getvalue("karst_weighting") == "Pure Bedded Lmst": row.setvalue("karst_weighting", "4") else: row.setvalue("karst_weighting", "0") rows.updaterow(row) del rows ############################################################################### 84

85 A1.1.3 Groundwater National Aquifer map When the Aquifer classification terms of: Rkc, Rkf, Rkd, Rk, Lk, Lm, Ll, Pl, Pu and Unlcassified were assigned to each grid square (as in section 1.1.1), the following python code was run to give each category a numerical weighting. ############################################################################### # PYTHON CODE FOR AQUIFER TYPE ############################################################################### import arcpy fc="l:\\tellusborder\\radon_2014\\modelling_2014.gdb\\radon_grid_output_v9" fieldlist = arcpy.listfields(fc) # Loop through each field in the list and print the name for field in fieldlist: print field.name rows = arcpy.updatecursor(fc) #this is my feature layer for row in rows: if row.getvalue("aquifer") == "Rkc": row.setvalue("aquifer", "4") 85

86 elif row.getvalue("aquifer") == "Rkf": row.setvalue("aquifer", "3") elif row.getvalue("aquifer") == "Rkd": row.setvalue("aquifer", "3") elif row.getvalue("aquifer") == "Rk": row.setvalue("aquifer", "3") elif row.getvalue("aquifer") == "Lk": row.setvalue("aquifer", "2.5") elif row.getvalue("aquifer") == "Lm": row.setvalue("aquifer", "2") elif row.getvalue("aquifer") == "Ll": row.setvalue("aquifer", "2") elif row.getvalue("aquifer") == "Pl": row.setvalue("aquifer", "1") elif row.getvalue("aquifer") == "Pu": row.setvalue("aquifer", "1") elif row.getvalue("aquifer") == "Unclassified": row.setvalue("aquifer", "0") rows.updaterow(row) del rows ############################################################################### 86

87 A1.1.4 Groundwater Recharge Coefficient (GWRC) and Depth to Bedrock (DTB) Datasets Using the Tabulate Intersection tool from the Analysis toolbox, the Groundwater Recharge Coefficient dataset was tabulated to intersect the CSO grid to find the percentage of each recharge category per grid square. Figure 25 Tabulate Intersection from the Statistics toolbox in ArcToolbox The outputs from this table were then imported into R and an average Recharge for each grid square was calculated. The code used is as follows: ############################################################################### # R CODE FOR GWRC ################################################################################# rm(list=ls(all=true)) setinternet2() #library(xlsreadwrite) library(psych) options( StringsAsFactors=F ) LOAD_DATA_recharge<read.table(file="L:\\TellusBorder\\Geophysics\\Radon\\RadonModelling\\Grid_recharge_coeff_ta ble.txt",header=t,sep=",") 87

88 DATA_recharge<data.frame(LOAD_DATA_recharge[0:nrow(LOAD_DATA_recharge),c(0:(ncol(LOAD_DATA_recharge) ))],na.rm=true) DATA_names_recharge<-names(DATA_recharge) ################################################################################## RECHARGE_MULTIPLE<data.frame(t(t(with(DATA_recharge,(Recharge_Coefficient_percent/100)*(DATA_recharge$PERCEN TAGE/100))))) RECHARGE_GRID<-data.frame(cbind(DATA_recharge,RECHARGE_MULTIPLE)) colnames(recharge_grid)<c("objectid","objectid_1","recharge_coefficient_percent","area","percentage","reomve","re CHARGE_MULTIPLE") RECHARGE<data.frame(tapply(RECHARGE_GRID$RECHARGE_MULTIPLE,RECHARGE_GRID$OBJECTID_1,FUN=su m)) dataout_recharge<-"l:/tellusborder/radon_2014/radon_recharge.csv" write.csv(recharge,file=dataout_recharge) ################################################################################## The data outputted from this grid was then joined back to the original grid using OBJECTID_1 as the join. The exact same approach was used for the Depth to Bedrock dataset and the code is as follows: ############################################################################### # R CODE FOR DTB ################################################################################## rm(list=ls(all=true)) setinternet2() library(psych) options( StringsAsFactors=F ) LOAD_DATA_DTM<- read.table(file="l:\\tellusborder\\radon_2014\\dtm.txt",header=t,sep=",") DATA_DTM<data.frame(LOAD_DATA_DTM[0:nrow(LOAD_DATA_DTM),c(0:(ncol(LOAD_DATA_DTM)))],na.rm=TR UE) ################################################################################## DATA_names_DTM<-names(DATA_DTM) ################################################################################## 88

89 DTM_GRID<-data.frame(cbind(DATA_DTM,DTM_MULTIPLE)) DTM_MULTIPLE<data.frame(t(t(with(DATA_DTM,(DTB_WEIGHT)*(DATA_DTM$PERCENTAGE/100))))) colnames(dtm_grid)<c("objectid","objectid_1","dtb_weight","area","percentage","na.rm","dtm_multiple") DTM<data.frame(tapply(DTM_GRID$DTM_MULTIPLE,DTM_GRID$OBJECTID_1,FUN=sum)) dataout_dtm<-"l:/tellusborder/radon_2014/dtm_grid.csv" write.csv(dtm,file=dataout_dtm) ################################################################################## A1.1.5 Airborne Geophysics: Total Count, Potassium, Thorium, Uranium Where the point Airborne Geophysics data was available, it was joined to each grid square. Each point falling within that grid square was summed and an average value was then found for each grid square. The join data by spatial location tool was used to undertake this work. This was completed for all four airborne datasets of Total Count, Potassium, Thorium and Uranium. Figure 26 Join by spatial location A1.1.6 Indoor Radon Indoor Radon data was joined to each grid square in the exact same method as described in section However, the arithmetic mean, the geometric mean and geometric standard 89

90 deviation and the transform threshold value needed to be calculated for the radon data. These were calculated in R and the code is as follows: ############################################################################### # R CODE FOR INDOOR RADON DATA ################################################################################### ############################## rm(list=ls(all=true)) setinternet2() library(psych) options( StringsAsFactors=F ) LOAD_DATA_Indoor_RN_AV<- read.table(file="l:\\tellusborder\\radon_2014\\radon_points_ txt",header=t,sep=",") DATA_Indoor_RN_AV<data.frame(cbind(LOAD_DATA_Indoor_RN_AV$OBJECTID_1,LOAD_DATA_Indoor_RN_AV$Corr_Valu e_non_neg)) colnames(data_indoor_rn_av)<-c("oid_","annual_ave") DATA_names_Indoor_RN_AV<-names(DATA_Indoor_RN_AV) ################################################################################### ############################## gmean <- function(data) { log_data <- log(data) gm <- exp(mean(log_data[is.finite(log_data)])) return(gm) } geosd <- function(x, na.rm = FALSE,...) 90

91 { exp(sd(log(x,...), na.rm = na.rm,...)) } dataout_indoor_rn<-"l:\\tellusborder\\radon_2014\\g_mean_indoor_rn_av_v2.csv" write.csv(indoor_rn,file=dataout_indoor_rn) ################################################################################### ############################## Indoor_RN<data.frame(tapply(DATA_Indoor_RN_AV$Annual_Ave,DATA_Indoor_RN_AV$OID_,FUN=gmean)) Indoor_RN_gsd<data.frame(tapply(DATA_Indoor_RN_AV$Annual_Ave,DATA_Indoor_RN_AV$OID_,FUN=geosd)) dataout_indoor_rn_gsd<-"l:\\tellusborder\\radon_2014\\g_sd_indoor_rn_av_v2.csv" write.csv(indoor_rn_gsd,file=dataout_indoor_rn_gsd) ################################################################################### ############################## GMEAN_DATA<- read.csv(file="l:\\tellusborder\\radon_2014\\g_mean_indoor_rn_av_v2.csv") GSD_DATA<- read.csv(file="l:\\tellusborder\\radon_2014\\g_sd_indoor_rn_av_v2.csv") GMEAN_GSD<-data.frame(cbind(GMEAN_DATA$G_mean,GSD_DATA$G_SD)) colnames(gmean_gsd)<-c("g_mean","g_sd") k<-data.frame(t(t((log(200)-log(gmean_gsd$g_mean))/(log(gmean_gsd$g_sd))))) colnames(k)<-c("k") k_out<-cbind(gmean_data$oid,k$k) dataout_k<-"l:\\tellusborder\\radon_2014\\k_v1.csv" write.csv(k_out,file=dataout_k) 91

92 ################################################################################### ############################## A1.1.8 Elevation Spatial Analyst Tools were utilized to calculate the average elevation of each grid square. The Zonal Statistics tool from the Zonal toolbox was used to calculate the mean elevation per grid square. The OSI 10 meter was the input raster dataset used. Figure 27 Zonal statistics toolbox used to calculate grid square mean elevation A1.1.9 Towns The Ordnance Survey Ireland (OSI) 2007 towns shapefile was also analysed to assess the percentage area of a grid square that falls within a town boundary. A new field (TABULATE) was added to the dataset to allow for analysis to be undertaken per grid square. The tabulate intersect tool was used to carry out this analysis. Once the analysis was undertaken, the data was joined back to the existing grid. 92

93 Figure 28 Tabulate Intersection from the Statistics toolbox in ArcToolbox used to calculate percentage area of grid squares that fall within a town polygon A2. Data Preparation Recommendations The work of preparing data for radon modelling for future work should be undertaken in one environment. As ArcGIS is the GIS used in this study, one Python script should be used to undertake this process of organising and calculating statistics on the data. This should be done using the arcpy mapping module. 93

20. Modelling in-house radon potential using Tellus data and geology to supplement inhouse radon measurements

20. Modelling in-house radon potential using Tellus data and geology to supplement inhouse radon measurements 20. Modelling in-house radon potential using Tellus data and geology to supplement inhouse radon measurements Don Appleton 1 and James Hodgson 2 How to cite this chapter: Appleton, J.D. and Hodgson, J.A.,

More information

RADON RISK MAPPING IN IRELAND

RADON RISK MAPPING IN IRELAND Radon in the Living Environment, 137 RADON RISK MAPPING IN IRELAND S.G. Fennell 1, Y. Pawitan 2, G.M. Mackin 1, J.S. Madden 1 and A.T. McGarry 1 1 Radiological Protection Institute of Ireland, 3 Clonskeagh

More information

20. Modelling in-house radon potential using Tellus data and geology to supplement inhouse radon measurements

20. Modelling in-house radon potential using Tellus data and geology to supplement inhouse radon measurements 20. Modelling in-house radon potential using Tellus data and geology to supplement inhouse radon measurements Don Appleton 1 and James Hodgson 2 How to cite this chapter: Appleton, J.D. and Hodgson, J.A.,

More information

3.0 GEOLOGY AND HYDROGEOLOGY

3.0 GEOLOGY AND HYDROGEOLOGY 3.0 GEOLOGY AND HYDROGEOLOGY 3.1 Methodology The Geological Survey of Ireland (GSI) publication Geology of South Cork was consulted to establish the nature of the bedrock lithology and Quaternary sediments.

More information

Water Framework Directive. Groundwater Monitoring Programme. Site Information. Galbally

Water Framework Directive. Groundwater Monitoring Programme. Site Information. Galbally Water Framework Directive Groundwater Monitoring Programme Site Information Galbally ImagePath1: Galbally\Galbally1.jpg Galbally source is comprised of 2 boreholes situated in Devonian Old Red Sandstones

More information

1 st Draft Tullamore GWB Description 6 th January 2004

1 st Draft Tullamore GWB Description 6 th January 2004 Hydrometric Area Local Authority 25 Brosna Offaly & Westmeath Co. Co. s Topography Tullamore GWB: Summary of Initial Characterisation. Associated surface water features Associated terrestrial ecosystem(s)

More information

Radon potential mapping of the Tralee-Castleisland and Cavan areas (Ireland) based on airborne gamma-ray spectrometry and geology.

Radon potential mapping of the Tralee-Castleisland and Cavan areas (Ireland) based on airborne gamma-ray spectrometry and geology. Radon potential mapping of the Tralee-Castleisland and Cavan areas (Ireland) based on airborne gamma-ray spectrometry and geology. Appleton J D 1, 4, Doyle E 2,5, Fenton D 3 and Organo C 3 1 British Geological

More information

Water Framework Directive. Groundwater Monitoring Programme. Site Information. Kiltrough PWS

Water Framework Directive. Groundwater Monitoring Programme. Site Information. Kiltrough PWS Water Framework Directive Groundwater Monitoring Programme Site Information Kiltrough PWS ImagePath1: Kiltrough PWS\17_009_Kiltough_P1_SiteLo cation.jpg This monitoring point is a well that is part of

More information

Oola PWS - Carrigmore BH

Oola PWS - Carrigmore BH Water Framework Directive Groundwater Monitoring Programme Site Information Oola PWS - Carrigmore BH ImagePath1: Oola PWS - Carrigmore BH\OolaPWSCarrigmore1.jpg Oola PWS Carrigmore BH is a borehole used

More information

Dunkerrin - Guilfoyles Well

Dunkerrin - Guilfoyles Well Water Framework Directive Groundwater Monitoring Programme Site Information Dunkerrin - Guilfoyles Well ImagePath1: Dunkerrin - Guilfoyles Well\IE_SH_G_19_005_a_Pump House_(800_x_600).jpg Dunkerrin/ Guilfoyles

More information

Cappog Bridge (PW-3)

Cappog Bridge (PW-3) Water Framework Directive Groundwater Monitoring Programme Site Information Cappog Bridge (PW-3) ImagePath1: Cappog Bridge\IEGBNI_NB_G_012_18_0 01_A_SiteLocation.jpg Cappog Bridge PW-3 is one of 8 boreholes

More information

Industrialised Peat Extraction Scoping Project. Technical Report & Results. from. University College Cork. November 2010

Industrialised Peat Extraction Scoping Project. Technical Report & Results. from. University College Cork. November 2010 Industrialised Peat Extraction Scoping Project Cataloguing of exposed peat soil areas identified on Peatlands of Ireland Mapped from Landsat Imagery (PIMLI) Technical Report & Results from University College

More information

User Guide: RADON POTENTIAL DATASET - England and Wales

User Guide: RADON POTENTIAL DATASET - England and Wales User Guide: RADON POTENTIAL DATASET - England and Wales This document provides information for users of the joint HPA-BGS RADON POTENTIAL DATASET for England and Wales. 1. Background to joint Health Protection

More information

Tellus Border Project Overview. Marie Cowan Ph.D Project Manager

Tellus Border Project Overview. Marie Cowan Ph.D Project Manager Tellus Border Project Overview Marie Cowan Ph.D Project Manager Presentation Outline History Goals Partners Project summary Legislative Framework Data Impacts Tellus History Proposed by GSNI, GSI, and

More information

Tellus Survey Frequently Asked Questions:

Tellus Survey Frequently Asked Questions: Tellus Survey 2017 - Frequently Asked Questions: 1. What is Tellus? Tellus is a ground and airborne geoscience mapping programme, collecting chemical and geophysical data that will inform the management

More information

Andrew Lee BEng (Hons) CEng MIStructE FGS FPWS

Andrew Lee BEng (Hons) CEng MIStructE FGS FPWS Jim Twaddle BSc (Hons) FGS Andrew Lee BEng (Hons) CEng MIStructE FGS FPWS 22 January 2010 ME50231/RE001 !! "# # $ " %! &' ( "# # $ ) # ' %! "# # $ # "# # $ * + # "# # $ "! # "# # $, # "# # $! -( +.!!/

More information

International Atomic Energy Agency Learning programme: Radon gas. Module 4: Developing and Implementing a Representative Indoor Radon Survey

International Atomic Energy Agency Learning programme: Radon gas. Module 4: Developing and Implementing a Representative Indoor Radon Survey International Atomic Energy Agency Learning programme: Radon gas Module 4: Developing and Implementing a Representative Indoor Radon Survey Content Scope of this module Representative radon survey aims

More information

New mapping of natural and manmade radioactivity

New mapping of natural and manmade radioactivity New mapping of natural and manmade radioactivity Cathy Scheib and Dave Jones Acknowledgements: Tellus and JAC teams; Don Appleton (BGS); Jon Miles & Martyn Green (HPA); Robert Lamour (EHSNI); David Sanderson

More information

The Need for Regional Geological Datasets

The Need for Regional Geological Datasets The Need for Regional Geological Datasets Dr. Eibhlín Doyle PGeo Geological Survey of Ireland Regional Data Why What Who Future 2006 Why? Place developments in context Land use decision making Assist policy

More information

the Quarrying Industry Dewatering and the Quarrying Industry the Quarrying Industry

the Quarrying Industry Dewatering and the Quarrying Industry the Quarrying Industry Dewatering and the Quarrying Industry Dewatering and Dewatering and the Quarrying Industry the Quarrying Industry Les Brown Eugene P. Daly John Kelly Objectives 1) To present a summary of water management

More information

Harvey Thorleifson, Director, Minnesota Geological Survey. Status of geological mapping needed for groundwater protection in Minnesota

Harvey Thorleifson, Director, Minnesota Geological Survey. Status of geological mapping needed for groundwater protection in Minnesota Harvey Thorleifson, Director, Minnesota Geological Survey Status of geological mapping needed for groundwater protection in Minnesota Minnesota is located between the Dakotas and Wisconsin, north of Iowa,

More information

INDOOR RADON MAPPING FOR NEW YORK STATE

INDOOR RADON MAPPING FOR NEW YORK STATE INDOOR RADON MAPPING FOR NEW YORK STATE C. Kunz, J. Green, C. Schwenker, E. Rigilski, and M. Kitto NYS Department of Health, Wadsworth Center, Empire State Plaza Albany, NY ABSTRACT The percent of homes

More information

The European Commission s science and knowledge service. Moderated discussion: Progress on European Geogenic Radon Mapping. Joint Research Centre

The European Commission s science and knowledge service. Moderated discussion: Progress on European Geogenic Radon Mapping. Joint Research Centre The European Commission s science and knowledge service Joint Research Centre Moderated discussion: Progress on European Geogenic Radon Mapping 13 th GARRM, Prague, 15 September 2016 1 European Basic Safety

More information

Water Framework Directive. Groundwater Monitoring Programme. Site Information. Tir na League

Water Framework Directive. Groundwater Monitoring Programme. Site Information. Tir na League Water Framework Directive Groundwater Monitoring Programme Site Information Tir na League ImagePath1: Tir na League\IE_NW_G_078_05_005_ A_PumpHouse.jpg Tir na League is an infiltration gallery situated

More information

Water Framework Directive. Groundwater Monitoring Programme. Site Information. Drum Bingahamstown

Water Framework Directive. Groundwater Monitoring Programme. Site Information. Drum Bingahamstown Water Framework Directive Groundwater Monitoring Programme Site Information Drum Bingahamstown ImagePath1: Drum Bingahamstown\IE_WE_G_16_ 1_a_Spring.jpg Drum Binghamstown is a spring that is used for a

More information

Spatial Trends of unpaid caregiving in Ireland

Spatial Trends of unpaid caregiving in Ireland Spatial Trends of unpaid caregiving in Ireland Stamatis Kalogirou 1,*, Ronan Foley 2 1. NCG Affiliate, Thoukididi 20, Drama, 66100, Greece; Tel: +30 6977 476776; Email: skalogirou@gmail.com; Web: http://www.gisc.gr.

More information

GEOPHYSICS GRAVITY DATA COVERAGE

GEOPHYSICS GRAVITY DATA COVERAGE GEOPHYSICS DATA COVERAGE The Mudgee-Gulgong district lies within the Dubbo 1:250,000 sheet area. This area is now covered by high res_olution gravity, magnetic and radiometric data. The aeromagnetic and

More information

Towards a multivariate geogenic radon hazard index

Towards a multivariate geogenic radon hazard index Towards a multivariate geogenic radon hazard index P. Bossew 1, G. Cinelli 2, T. Tollefsen 2, M. DeCort 2 1 German Federal Office for Radiation Protection, Berlin 2 European Commission, Joint Research

More information

18. Mapping the terrestrial gamma radiation dose

18. Mapping the terrestrial gamma radiation dose 18. Mapping the terrestrial gamma radiation dose David Beamish 1 How to cite this chapter: Beamish, D., 2016 Mapping the terrestrial gamma radiation dose in M.E. Young (ed.), Unearthed: impacts of the

More information

Groundwater Vulnerability Mapping Eastern Newfoundland Executive Summary

Groundwater Vulnerability Mapping Eastern Newfoundland Executive Summary Groundwater Vulnerability Mapping Eastern Newfoundland Executive Summary 123102.00 Executive Summary March 2014 ISO 9001 Registered Company Prepared for: Water Resources Management Division Department

More information

Soil uranium, soil gas radon and indoor radon empirical relationships in the UK and other European countries

Soil uranium, soil gas radon and indoor radon empirical relationships in the UK and other European countries Soil uranium, soil gas radon and indoor radon empirical relationships in the UK and other European countries Don Appleton (BGS) and Jon Miles (HPA) EUROPEAN GEOGENIC RADON POTENTIAL MAP Wide range of numerical

More information

Geophysical Survey. Ballymount Co. Dublin. Licence Ref. 02R029. By John Nicholls Margaret Gowen & Co. Ltd. For LRT

Geophysical Survey. Ballymount Co. Dublin. Licence Ref. 02R029. By John Nicholls Margaret Gowen & Co. Ltd. For LRT Geophysical Survey Ballymount Co. Dublin Licence Ref. 02R029 By John Nicholls Margaret Gowen & Co. Ltd. For LRT 4 th April 2002 Illustrations List of Figures Figure 1 Site Location 1: 50000 Figure 2 Survey

More information

Assessing soil wetness with airborne radiometric data

Assessing soil wetness with airborne radiometric data Assessing soil wetness with airborne radiometric data David Beamish British Geological Survey, Keyworth, Nottingham, NG12 5GG, UK Corresponding author (e-mail:dbe@bgs.ac.uk) Extended Abstract 20 th European

More information

Newcastle West GWB: Summary of Initial Characterisation.

Newcastle West GWB: Summary of Initial Characterisation. Newcastle West GWB: Summary of Initial Characterisation. Hydrometric Area Local Authorities 24 - Deel/ Shannon Estuary Limerick Co. Co. Topography Associated surface water features Rivers: Deel, Daar,

More information

Regional Variation of Seasonal Behaviour for Indoor Radon Levels

Regional Variation of Seasonal Behaviour for Indoor Radon Levels Int. Statistical Inst.: Proc. 58th World Statistical Congress, 2011, Dublin (Session CPS006) p.3924 Regional Variation of Seasonal Behaviour for Indoor Radon Levels Burke, Órlaith University of Oxford,

More information

Met Éireann Climatological Note No. 15 Long-term rainfall averages for Ireland,

Met Éireann Climatological Note No. 15 Long-term rainfall averages for Ireland, Met Éireann Climatological Note No. 15 Long-term rainfall averages for Ireland, 1981-2010 Séamus Walsh Glasnevin Hill, Dublin 9 2016 Disclaimer Although every effort has been made to ensure the accuracy

More information

KANSAS GEOLOGICAL SURVEY Open File Report LAND SUBSIDENCE KIOWA COUNTY, KANSAS. May 2, 2007

KANSAS GEOLOGICAL SURVEY Open File Report LAND SUBSIDENCE KIOWA COUNTY, KANSAS. May 2, 2007 KANSAS GEOLOGICAL SURVEY Open File Report 2007-22 LAND SUBSIDENCE KIOWA COUNTY, KANSAS Prepared by Michael T. Dealy L.G., Manager, Wichita Operations SITE LOCATION The site was approximately four miles

More information

Analysis of travel-to-work patterns and the identification and classification of REDZs

Analysis of travel-to-work patterns and the identification and classification of REDZs Analysis of travel-to-work patterns and the identification and classification of REDZs Dr David Meredith, Teagasc, Spatial Analysis Unit, Rural Economy Development Programme, Ashtown, Dublin 15. david.meredith@teagasc.ie

More information

Summary and Implications for Policy

Summary and Implications for Policy Summary and Implications for Policy 1 Introduction This is the report on a background study for the National Spatial Strategy (NSS) regarding the Irish Rural Structure. The main objective of the study

More information

GEO-STABILITY DESKTOP STUDY FOR A PROPOSED FILLING STATION AT ROODEKRANS MOGALE CITY, GAUTENG

GEO-STABILITY DESKTOP STUDY FOR A PROPOSED FILLING STATION AT ROODEKRANS MOGALE CITY, GAUTENG GEO-STABILITY DESKTOP STUDY FOR A PROPOSED FILLING STATION AT ROODEKRANS MOGALE CITY, GAUTENG Page 1 of 11 GEO-STABILITY DESKTOP STUDY FOR A PROPOSED FILLING STATION AT ROODEKRANS MOGALE CITY, GAUTENG

More information

Geo-hazard Potential Mapping Using GIS and Artificial Intelligence

Geo-hazard Potential Mapping Using GIS and Artificial Intelligence Geo-hazard Potential Mapping Using GIS and Artificial Intelligence Theoretical Background and Uses Case from Namibia Andreas Knobloch 1, Dr Andreas Barth 1, Ellen Dickmayer 1, Israel Hasheela 2, Andreas

More information

Ground-Water Exploration in the Worthington Area of Nobles County: Summary of Seismic Data and Recent Test Drilling Results

Ground-Water Exploration in the Worthington Area of Nobles County: Summary of Seismic Data and Recent Test Drilling Results Ground-Water Exploration in the Worthington Area of Nobles County: Summary of Seismic Data and Recent Test Drilling Results Jim Berg and Todd Petersen Geophysicists, DNR Waters January 2000 Table of Contents

More information

Does the SDCP need inputs from geology?

Does the SDCP need inputs from geology? The British Geological Survey s Experience and Expertise in Supporting Projects such as the Sirte Depression Connection Project David Ovadia Director of International Kingsley Dunham Centre Keyworth Nottingham

More information

Hydrogeological Assessment for Part of Lots 2 and 3, Concession 5, Township of Thurlow, County of Hastings 1.0 INTRODUCTION. 1.

Hydrogeological Assessment for Part of Lots 2 and 3, Concession 5, Township of Thurlow, County of Hastings 1.0 INTRODUCTION. 1. February 10,2017 25506400 Ontario Ltd. Foxboro, ON Attention: Brad Newbatt Re: Hydrogeological Assessment for Part of Lots 2 and 3, Concession 5, Township of Thurlow, County of Hastings 1.0 INTRODUCTION

More information

Coimisiún na Scrúduithe Stáit State Examinations Commission

Coimisiún na Scrúduithe Stáit State Examinations Commission M. 24 Coimisiún na Scrúduithe Stáit State Examinations Commission LEAVING CERTIFICATE EXAMINATION 2004 GEOGRAPHY HIGHER LEVEL MONDAY, 14 JUNE, AFTERNOON 1.30 TO 4.50 Four questions to be answered, namely

More information

Placer Potential Map. Dawson L and U se P lan. Jeffrey Bond. Yukon Geological Survey

Placer Potential Map. Dawson L and U se P lan. Jeffrey Bond. Yukon Geological Survey Placer Potential Map Dawson L and U se P lan By Jeffrey Bond Yukon Geological Survey Dawson Land Use Plan Placer Potential Map 1.0 Introduction Placer mining has been an important economic driver within

More information

INTRODUCTION Water resources are vitally important for the future of humankind. Groundwater from karst aquifers is among the most important drinking w

INTRODUCTION Water resources are vitally important for the future of humankind. Groundwater from karst aquifers is among the most important drinking w Vulnerability Mapping for the Protection of Carbonate (Karst) Aquifers (Ramallah-Al Bireh District) Eng. Amjad da Assi House of fw Water and de Environment May, 2007 INTRODUCTION Water resources are vitally

More information

Environmental Scoping Report for the proposed establishment of a New Coal-Fired Power Station in the Lephalale Area, Limpopo Province

Environmental Scoping Report for the proposed establishment of a New Coal-Fired Power Station in the Lephalale Area, Limpopo Province 8. GEOLOGY, SOILS AND AGRICULTURAL POTENTIAL 8.1. Geology 8.1.1. Candidate Site Geology Due to the layered natured of the geology and various structures the candidate sites are underlain by differing geological

More information

The Pine Creek AEM Survey, Northern Territory

The Pine Creek AEM Survey, Northern Territory The Pine Creek AEM Survey, Northern Territory (Data acquisition, processing, delivery and interpretation) M.A. Craig, M.T. Costelloe, S. Liu, S. Jaireth Onshore Energy and Minerals Division, Geoscience

More information

Source Protection Zones. National Dataset User Guide

Source Protection Zones. National Dataset User Guide Source Protection Zones National Dataset User Guide Version 1.1.4 20 th Jan 2006 1 Contents 1.0 Record of amendment...3 2.0 Introduction...4 2.1 Description of the SPZ dataset...4 2.1.1 Definition of the

More information

Procedure for Determining Near-Surface Pollution Sensitivity

Procedure for Determining Near-Surface Pollution Sensitivity Procedure for Determining Near-Surface Pollution Sensitivity Minnesota Department of Natural Resources Division of Ecological and Water Resources County Geologic Atlas Program March 2014 Version 2.1 I.

More information

Compact guides GISCO. Geographic information system of the Commission

Compact guides GISCO. Geographic information system of the Commission Compact guides GISCO Geographic information system of the Commission What is GISCO? GISCO, the Geographic Information System of the COmmission, is a permanent service of Eurostat that fulfils the requirements

More information

USING DOWNSCALED POPULATION IN LOCAL DATA GENERATION

USING DOWNSCALED POPULATION IN LOCAL DATA GENERATION USING DOWNSCALED POPULATION IN LOCAL DATA GENERATION A COUNTRY-LEVEL EXAMINATION CONTENT Research Context and Approach. This part outlines the background to and methodology of the examination of downscaled

More information

A method for three-dimensional mapping, merging geologic interpretation, and GIS computation

A method for three-dimensional mapping, merging geologic interpretation, and GIS computation A method for three-dimensional mapping, merging geologic interpretation, and GIS computation Soller, David R., U.S. Geological Survey, 908 National Center, Reston, VA 20192 and Richard C. Berg, Illinois

More information

Understanding and Measuring Urban Expansion

Understanding and Measuring Urban Expansion VOLUME 1: AREAS AND DENSITIES 21 CHAPTER 3 Understanding and Measuring Urban Expansion THE CLASSIFICATION OF SATELLITE IMAGERY The maps of the urban extent of cities in the global sample were created using

More information

AND HORSEFLY PROSPECTS, N.W. BRITISH COLUMBIA NTS, 103H/ll, 14 FOR ATNA RESOURCES LTD DELTA GEOSCIENCE LTD

AND HORSEFLY PROSPECTS, N.W. BRITISH COLUMBIA NTS, 103H/ll, 14 FOR ATNA RESOURCES LTD DELTA GEOSCIENCE LTD AND HORSEFLY PROSPECTS, N.W. BRITISH COLUMBIA NTS, 103H/ll, 14 FOR ATNA RESOURCES LTD BY DELTA GEOSCIENCE LTD FE T A. HENDRICKSON, P.GEO. 1 . TABLE OF CONTENTS Introduction............. Page 1. Personnel..........

More information

Draft 1 Curlew Mountains 3 rd October 2003

Draft 1 Curlew Mountains 3 rd October 2003 Curlew Mountains Southeast: Summary of Initial Characterisation. OUTSTANDING ISSUES the volcanic rocks within the Keadew Formation/Basalts and other volcanic rocks Andesitic Lava. Is the pure basis of

More information

Towards a European map of indoor radon

Towards a European map of indoor radon Towards a European map of indoor radon Current issues and forthcoming challenges G. Dubois JRC - European Commission IES - Emissions and Health Unit Radioactivity Environmental Monitoring () group Gregoire.dubois@jrc.it

More information

ISO Measurement of radioactivity in the environment Air: radon-222 Part 5: Continuous measurement method of the activity concentration

ISO Measurement of radioactivity in the environment Air: radon-222 Part 5: Continuous measurement method of the activity concentration INTERNATIONAL STANDARD ISO 11665-5 First edition 2012-07-15 Measurement of radioactivity in the environment Air: radon-222 Part 5: Continuous measurement method of the activity concentration Mesurage de

More information

GNS Science, Lower Hutt, New Zealand NZSEE Conference

GNS Science, Lower Hutt, New Zealand NZSEE Conference A Ground Shaking Amplification Map for New Zealand U. Destegul, G. Dellow & D. Heron GNS Science, Lower Hutt, New Zealand. 2008 NZSEE Conference ABSTRACT: A ground shaking amplification map of New Zealand

More information

MRD283-REV METADATA. Acronyms are Used to Identify the Data Set or Information Holding: MRD283-REV

MRD283-REV METADATA. Acronyms are Used to Identify the Data Set or Information Holding: MRD283-REV MRD283-REV METADATA GENERAL INFORMATION Official Name of the Data Set or Information Holding: Ambient Groundwater Geochemistry Data for Southern Ontario, 2007 2014 Acronyms are Used to Identify the Data

More information

Coimisiún na Scrúduithe Stáit State Examinations Commission

Coimisiún na Scrúduithe Stáit State Examinations Commission 2018. S28 WARNING You must return this paper with your answer book. Otherwise you will lose marks. Write your Examination Number here: Coimisiún na Scrúduithe Stáit State Examinations Commission JUNIOR

More information

Radon Soil Gas Associated With the C2 Zone, Millet Brook Uranium District (NTS 21A/16)

Radon Soil Gas Associated With the C2 Zone, Millet Brook Uranium District (NTS 21A/16) Report of Activities 2009 35 Radon Soil Gas Associated With the C2 Zone, Millet Brook Uranium District (NTS 21A/16) T. A. Goodwin, G. A. O Reilly, K. L. Ford 1 and P. W. B. Friske 1 Introduction Radon

More information

THE STRUCTURE AND THICKNESS OF THE CLINTON AND BEREA FORMATIONS IN THE VICINITY OF WOOSTER, OHIO

THE STRUCTURE AND THICKNESS OF THE CLINTON AND BEREA FORMATIONS IN THE VICINITY OF WOOSTER, OHIO THE STRUCTURE AND THICKNESS OF THE CLINTON AND BEREA FORMATIONS IN THE VICINITY OF WOOSTER, OHIO KARL VER STEEG College of Wooster INTRODUCTION AND ACKNOWLEDGMENTS The data used in the construction of

More information

The Greater Drogheda Area - Drogheda (Co. Louth), Drogheda South (Co. Meath) & the Meath Coast

The Greater Drogheda Area - Drogheda (Co. Louth), Drogheda South (Co. Meath) & the Meath Coast Submission to National Planning Framework The Greater Drogheda Area - Drogheda (Co. Louth), Drogheda South (Co. Meath) & the Meath Coast Date: 31st March 2017 From: Cormac Bohan, Proposal: that Drogheda

More information

=%REPORT RECONNAISSANCE OF CHISHOLM LAKE PROSPECT. October 25, 1977

=%REPORT RECONNAISSANCE OF CHISHOLM LAKE PROSPECT. October 25, 1977 =%REPORT ON FIELD RECONNAISSANCE OF CHISHOLM LAKE PROSPECT October 25, 1977 Bruce D. Vincent Imperial Oil Limited, Minerals - Coal, CALGARY, ALBERTA CHISHOLM LAKE PROSPECT Introduction The Chisholm Lake

More information

SIXTH SCHEDULE REPUBLIC OF SOUTH SUDAN MINISTRY OF PETROLEUM, MINING THE MINING (MINERAL TITLE) REGULATIONS 2015

SIXTH SCHEDULE REPUBLIC OF SOUTH SUDAN MINISTRY OF PETROLEUM, MINING THE MINING (MINERAL TITLE) REGULATIONS 2015 SIXTH SCHEDULE REPUBLIC OF SOUTH SUDAN MINISTRY OF PETROLEUM, MINING THE MINING ACT, 2012 THE MINING (MINERAL TITLE) REGULATIONS 2015 Guidelines should be prepared by the Directorate of Mineral Development

More information

Integrated GIS based approach in mapping the groundwater potential zones in Kota Kinabalu, Sabah, Malaysia

Integrated GIS based approach in mapping the groundwater potential zones in Kota Kinabalu, Sabah, Malaysia Integrated GIS based approach in mapping the groundwater potential zones in Kota Kinabalu, Sabah, Malaysia Zulherry Isnain and Juhari Mat Akhir Faculty of Science and Natural Resources, Universiti Malaysia

More information

Mineral Resources

Mineral Resources Sacramento Local Agency Formation Commission Mineral Resources 3.11 - Mineral Resources 3.11.1 - Introduction This section describes and evaluates potential environmental impacts to mineral resources resulting

More information

Analytical Report. Drought in the Horn of Africa February Executive summary. Geographical context. Likelihood of drought impact (LDI)

Analytical Report. Drought in the Horn of Africa February Executive summary. Geographical context. Likelihood of drought impact (LDI) Executive summary The current drought in the Horn of Africa is affecting especially Somalia, among other countries, in particular the central and southern regions, where most population is located. Overall,

More information

patersongroup Mineral Aggregate Assessment 3119 Carp Road Ottawa, Ontario Prepared For Mr. Greg LeBlanc March 7, 2014 Report: PH2223-REP.

patersongroup Mineral Aggregate Assessment 3119 Carp Road Ottawa, Ontario Prepared For Mr. Greg LeBlanc March 7, 2014 Report: PH2223-REP. Geotechnical Engineering Environmental Engineering group Hydrogeology Geological Engineering Archaeological Studies Materials Testing 3119 Carp Road Prepared For Mr. Greg LeBlanc March 7, 2014 Paterson

More information

704,000 OUNCE MIYABI GOLD PROJECT UPDATE

704,000 OUNCE MIYABI GOLD PROJECT UPDATE 704,000 OUNCE MIYABI GOLD PROJECT UPDATE ASX/ RELEASE 21 April 2016 ASX code RVY Board of Directors: Geoff Gilmour Managing Director Greg Cunnold Technical Director Graeme Clatworthy Non-Executive Director

More information

Quaternary clays alluvial sands of the Shepparton Formation overlie the basement rocks.

Quaternary clays alluvial sands of the Shepparton Formation overlie the basement rocks. NAGAMBIE GOLDFIELD Regional Geological Setting The Nagambie Project is located within the Melbourne Structural Zone of Victoria. The lithologies range in age from the Upper Silurian Broadford Formation

More information

1/28/16. EGM101 Skills Toolbox. Map types. Political Physical Topographic Climate Resource Road. Thematic maps (use one of the above as backdrop)

1/28/16. EGM101 Skills Toolbox. Map types. Political Physical Topographic Climate Resource Road. Thematic maps (use one of the above as backdrop) EGM101 Skills Toolbox Map types Political Physical Topographic Climate Resource Road Thematic maps (use one of the above as backdrop) Map Types Deriving information from maps Hydrographic Geological Soils

More information

Comhairle Contae Mhaigh Eo

Comhairle Contae Mhaigh Eo Comhairle Contae Mhaigh Eo Winter Service Plan 2017/2018 Contents 1. Winter Service Policy 3 2. Winter Service Limitations 4 3. Treatment Routes 4 4. Weather Prediction 5 5. Personnel & Equipment 6 6.

More information

Prospectivity Modelling of Granite-Related Nickel Deposits Throughout Eastern Australia

Prospectivity Modelling of Granite-Related Nickel Deposits Throughout Eastern Australia Prospectivity Modelling of Granite-Related Nickel Deposits Throughout Eastern Australia M P Hill 1 and A McCarthy 2 ABSTRACT Spatial modelling has been used to determine potential locations of granite-related

More information

Predictive Modelling of Ag, Au, U, and Hg Ore Deposits in West Texas Carl R. Stockmeyer. December 5, GEO 327G

Predictive Modelling of Ag, Au, U, and Hg Ore Deposits in West Texas Carl R. Stockmeyer. December 5, GEO 327G Predictive Modelling of Ag, Au, U, and Hg Ore Deposits in West Texas Carl R. Stockmeyer December 5, 2013 - GEO 327G Objectives and Motivations The goal of this project is to use ArcGIS to create models

More information

L.O: HOW GEOLOGISTS SEQUENCE EVENTS IN EARTH'S GEOLOGIC HISTORY IF NOT OVERTURNED, OLDEST ON BOTTOM, YOUNGEST ON TOP

L.O: HOW GEOLOGISTS SEQUENCE EVENTS IN EARTH'S GEOLOGIC HISTORY IF NOT OVERTURNED, OLDEST ON BOTTOM, YOUNGEST ON TOP L.O: HOW GEOLOGISTS SEQUENCE EVENTS IN EARTH'S GEOLOGIC HISTORY IF NOT OVERTURNED, OLDEST ON BOTTOM, YOUNGEST ON TOP 1. Unless a series of sedimentary rock layers has been overturned, the bottom rock layer

More information

Connecticut's Aquifers

Connecticut's Aquifers Page 1 of 5 DEP Search: Connecticut's Aquifers The technical definition of the word "aquifer" is: any geologic formation capable of yielding significant quantities of water to wells. By that definition,

More information

Merging statistics and geospatial information

Merging statistics and geospatial information Merging statistics and geospatial information Demography / Commuting / Spatial planning / Registers Mirosław Migacz Chief GIS Specialist Janusz Dygaszewicz Director Central Statistical Office of Poland

More information

Irish Industrial Wages: An Econometric Analysis Edward J. O Brien - Junior Sophister

Irish Industrial Wages: An Econometric Analysis Edward J. O Brien - Junior Sophister Irish Industrial Wages: An Econometric Analysis Edward J. O Brien - Junior Sophister With pay agreements firmly back on the national agenda, Edward O Brien s topical econometric analysis aims to identify

More information

Lima Project: Seismic Refraction and Resistivity Survey. Alten du Plessis Global Geophysical

Lima Project: Seismic Refraction and Resistivity Survey. Alten du Plessis Global Geophysical Lima Project: Seismic Refraction and Resistivity Survey Alten du Plessis Global Geophysical Report no 0706/2006 18 December 2006 Lima Project: Seismic Refraction and Resistivity Survey by Alten du Plessis

More information

Understanding China Census Data with GIS By Shuming Bao and Susan Haynie China Data Center, University of Michigan

Understanding China Census Data with GIS By Shuming Bao and Susan Haynie China Data Center, University of Michigan Understanding China Census Data with GIS By Shuming Bao and Susan Haynie China Data Center, University of Michigan The Census data for China provides comprehensive demographic and business information

More information

AGGREGATE RESOURCES OF ONTARIO (ARO) METADATA

AGGREGATE RESOURCES OF ONTARIO (ARO) METADATA AGGREGATE RESOURCES OF ONTARIO (ARO) METADATA GENERAL INFORMATION Official Name of the Data Set or Information Holding: Aggregate Resources of Ontario Acronyms are Used to Identify the Data Set or Information

More information

Quality and Coverage of Data Sources

Quality and Coverage of Data Sources Quality and Coverage of Data Sources Objectives Selecting an appropriate source for each item of information to be stored in the GIS database is very important for GIS Data Capture. Selection of quality

More information

Geophysical Exploration in Water Resources Assessment. John Mundell, P.E., L.P.G., P.G. Ryan Brumbaugh, L.P.G. Mundell & Associates, Inc.

Geophysical Exploration in Water Resources Assessment. John Mundell, P.E., L.P.G., P.G. Ryan Brumbaugh, L.P.G. Mundell & Associates, Inc. Geophysical Exploration in Water Resources Assessment John Mundell, P.E., L.P.G., P.G. Ryan Brumbaugh, L.P.G. Mundell & Associates, Inc. Presentation Objective Introduce the use of geophysical survey methods

More information

2-D Resistivity Study: The Horizontal Resolution Improvement by Introducing the Enhancing Horizontal Resolution (EHR) Technique

2-D Resistivity Study: The Horizontal Resolution Improvement by Introducing the Enhancing Horizontal Resolution (EHR) Technique Open Journal of Geology, 213, 3, 1-6 doi:1.4236/ojg.213.32b1 Published Online April 213 (http://www.scirp.org/journal/ojg) 2-D Resistivity Study: The Horizontal Resolution Improvement by Introducing the

More information

For personal use only

For personal use only L4 66 Kings Park Road West Perth WA 6005 P: +61 8 6141 3585 F: +61 8 6141 3599 E: info@drakeresources.com.au ASX Announcement 16 th September 2014 Seimana Gold Project Technical Update and Next Steps Geophysical

More information

A Preliminary Geophysical Reconnaissance Mapping of Emirau Ground Water Resource, Emirau Island, New Ireland Province, PNG

A Preliminary Geophysical Reconnaissance Mapping of Emirau Ground Water Resource, Emirau Island, New Ireland Province, PNG A Preliminary Geophysical Reconnaissance Mapping of Emirau Ground Water Resource, Emirau Island, New Ireland Province, PNG Geological Survey Division of Mineral Resources Authority (MRA) Papua New Guinea

More information

COASTAL QUATERNARY GEOLOGY MAPPING FOR NSW: EXAMPLES AND APPLICATIONS

COASTAL QUATERNARY GEOLOGY MAPPING FOR NSW: EXAMPLES AND APPLICATIONS COASTAL QUATERNARY GEOLOGY MAPPING FOR NSW: EXAMPLES AND APPLICATIONS A Troedson Geological Survey of New South Wales Abstract Detailed geological mapping of the coastal plains of regional NSW was undertaken

More information

Desktop GIS for Geotechnical Engineering

Desktop GIS for Geotechnical Engineering Desktop GIS for Geotechnical Engineering Satya Priya Deputy General Manager (Software) RMSI, A-7, Sector 16 NOIDA 201 301, UP, INDIA Tel: +91-120-2511102 Fax: +91-120-2510963 Email: Satya.Priya@rmsi.com

More information

Groundwater Resources of Missouri. Cynthia Brookshire, R. G.

Groundwater Resources of Missouri. Cynthia Brookshire, R. G. Groundwater Resources of Missouri Cynthia Brookshire, R. G. GROUNDWATER... Water beneath the Earth s surface within a zone of saturation AQUIFER... A geologic formation or group of formations that are

More information

PREDICTING OVERHEATING RISK IN HOMES

PREDICTING OVERHEATING RISK IN HOMES PREDICTING OVERHEATING RISK IN HOMES Susie Diamond Inkling Anastasia Mylona CIBSE Simulation for Health and Wellbeing 27th June 2016 - CIBSE About Inkling Building Physics Consultancy Susie Diamond Claire

More information

TITLE: Pooling, Sharing and Linking: Spatial Representation of Data using Geographical Information Systems

TITLE: Pooling, Sharing and Linking: Spatial Representation of Data using Geographical Information Systems TITLE: Pooling, Sharing and Linking: Spatial Representation of Data using Geographical Information Systems SCIENTIFIC RATIONALE AND RELEVANCE TO THE PROGRAMME (max. 2 pages). With the explosive growth

More information

Geo 327G Semester Project. Landslide Suitability Assessment of Olympic National Park, WA. Fall Shane Lewis

Geo 327G Semester Project. Landslide Suitability Assessment of Olympic National Park, WA. Fall Shane Lewis Geo 327G Semester Project Landslide Suitability Assessment of Olympic National Park, WA Fall 2011 Shane Lewis 1 I. Problem Landslides cause millions of dollars of damage nationally every year, and are

More information

INDOOR RADON ON LOESS DEPOSITS IN WALLONIA

INDOOR RADON ON LOESS DEPOSITS IN WALLONIA INDOOR RADON ON LOESS DEPOSITS IN WALLONIA François TONDEUR, Isabelle GERARDY, Nathalie GERARDY Institut supérieur industriel de Bruxelles, 150 rue Royale, B1000 BRUSSELS, BELGIUM ABSTRACT Pleistocene

More information

Project Appraisal Guidelines

Project Appraisal Guidelines Project Appraisal Guidelines Unit 16.2 Expansion Factors for Short Period Traffic Counts August 2012 Project Appraisal Guidelines Unit 16.2 Expansion Factors for Short Period Traffic Counts Version Date

More information

Refinement of the OECD regional typology: Economic Performance of Remote Rural Regions

Refinement of the OECD regional typology: Economic Performance of Remote Rural Regions [Preliminary draft April 2010] Refinement of the OECD regional typology: Economic Performance of Remote Rural Regions by Lewis Dijkstra* and Vicente Ruiz** Abstract To account for differences among rural

More information

a system for input, storage, manipulation, and output of geographic information. GIS combines software with hardware,

a system for input, storage, manipulation, and output of geographic information. GIS combines software with hardware, Introduction to GIS Dr. Pranjit Kr. Sarma Assistant Professor Department of Geography Mangaldi College Mobile: +91 94357 04398 What is a GIS a system for input, storage, manipulation, and output of geographic

More information

NOA ASSESSMENT HARRIS QUARRY MENDOCINO COUNTY, CALIFORNIA TABLE OF CONTENTS

NOA ASSESSMENT HARRIS QUARRY MENDOCINO COUNTY, CALIFORNIA TABLE OF CONTENTS NOA ASSESSMENT HARRIS QUARRY MENDOCINO COUNTY, CALIFORNIA TABLE OF CONTENTS Introduction... 1 Scope of Services... 1 Project Location and Description... 1 Geologic Setting... 1 Regional Geology... 1 Site

More information