1 Introduction
1.1 Background
Radon (222Rn) is a noble gas that is colorless, odorless, and tasteless, and has a half-life of 3.83 days. Production of radon in the subsurface is driven by the decay of uranium (238U), which decays into radium (226Ra) and subsequently radon (222Rn) before reaching the stable lead isotope 206Pb. As radon gas is produced underground, it travels upwards through rock and soil by means of pore spaces and carrier fluids before being released to the atmosphere. Additionally, the significant atomic mass of a radon atom means that radon gas will generally settle in lower-lying areas with little ventilation, such as mines, caves, and basements. Radioactive decay of radon and its daughter products (218Po, 214Po) releases energy in the form of alpha particles (4He) which can cause damage to lung tissue, and prolonged exposure to radon gas has been identified as the second leading cause of lung cancer in the United States.
Radon concentrations are highest near regions with surficial and underlying geologic material containing elevated uranium content, typically associated with intermediate to felsic igneous rocks. Southeastern Oregon has some of the highest uranium content in the U.S. in Tertiary volcanic rocks (McDermitt caldera, Steens Mountain, White King/Lucky Lass mines near Lakeview), and northern Oregon (particularly the Portland metropolitan area) is subject to high radon concentrations due to uranium-rich Quaternary deposits from the Missoula floods, 15-13 kya. Due to Oregon’s general predisposition for significant radon hazards, continuous radon monitoring occurs both indoors and outdoors throughout the state by various methods.
1.2 Data collection
The primary dataset I chose for this project is a shapefile
containing spatial data regarding Oregon’s potential radon hazard, which
was published by the Oregon Department of Geology and Mineral Industries
(DOGAMI) and downloaded via the data.gov website. The foremost product
of this dataset is the FIN_RN_RNK column, which is a set of
polygons describing that area’s radon hazard based on a rank of 1, 2, or
3 (low, moderate, or high hazard). The final rank is determined based on
five main factors (each with their own hazard rank), which are given as
the following from lowest to highest importance: geologic material’s
uranium content, mean aerial radiometric measurements, mean indoor air
test measurements, occurrences where mean aerial radiometric
measurements are greater than mean indoor air test results, and uranium
mine locations. Additionally, the dataset also includes data for the
primary source that was used to determine a polygon’s final rank, and a
set of Oregon’s main geologic groups and terranes. To improve the visual
cohesion of plotted maps, shapefiles for the state and county outlines
of Oregon were obtained from the National Weather Service and U.S.
Census Bureau.
2 Preparing the data
2.1 Initialize packages and read in shapefiles
Since the main dataset is in the shapefile format, we will need to
utilize the sf package to read in the data and
ggplot2 to plot it. Additionally, the
grDevices package will help generate unique color palettes
for large categorical datasets.
library(sf)
library(ggplot2)
library(grDevices)Now that we have all the necessary packages, we can read in the data
from the shapefiles using the st_read function.
setwd("C:/Users/Zachary/Documents/School/Winter_2024/GEOG490/data/")
shp <- st_read("Radon_data.shp")
oregon <- st_read("Oregon.shp")
oregon_counties <- st_read("OregonCounties.shp")The attribute table of the radon data shapefile has many various columns, but there are only a few that are relevant to the analysis we will perform, so we can assign the data from these columns to a new object.
data <- shp[, c("FIN_RN_RNK", "TERRANE_GR", "RNK_SOURCE", "RNK_SRCE2", "AREA_KM2")]This code simply defines some color palettes that will be used to make later plots.
colors <- sample(hcl(seq(15, 345, length.out = 66), 40, 70), 66)
fin_rn_rnk_colors <- c("azure", "lightblue", "lightblue4")
rnk_source_colors <- c("lightblue", "salmon", "lightgreen", "plum", "gold")3 Plotting the data
3.1 Radon potential rank map
To start out, we can first look at the final radon rank data to get
an idea of where the most significant radon hazards are in the state of
Oregon. To do this, we can use the geom_sf function to plot
the data on a map, and set the fill color to be the
FIN_RN_RNK data column. Additionally, we can use this same
function to plot the oregon and
oregon_counties data to overlay an outline of the state and
county borders.
fin_rn_rnk <- ggplot() +
geom_sf(data = shp, aes(fill = as.factor(FIN_RN_RNK)), color = "NA") +
scale_fill_manual(values = fin_rn_rnk_colors, name = "Rank") +
ggtitle("Oregon Radon Potential Rank") +
geom_sf(data = oregon_counties, fill = "transparent", lwd=0.05, color = "black") +
geom_sf(data = oregon, fill = "transparent", color = "black") +
theme(legend.position = "bottom", plot.title = element_text(hjust = 0.5))
ggsave("fin_rn_rnk.png", fin_rn_rnk)
Based on this map, we can see that the majority of the radon hazard is present in the southeastern corner of the state, with some high hazards in the Willamette Valley and along the Columbia River. In comparison, radon hazards are generally low in central Oregon and much of the Coast Range.
3.2 Rank source map
Now that we have an idea of the spatial distribution of radon hazards
in Oregon, we can look at a similar map plotting the primary source used
to determine this rank to learn what factors have the most control over
radon hazards throughout the state. The method for producing this map is
fairly similar to the previous map, but using the RNK_SRCE2
column to fill polygon with a color corresponding to a source category.
The five sources used to determine radon hazards are given as follows,
in order of least to most important: geologic material’s uranium
content, mean aerial radiometric measurements, mean indoor air test
measurements, occurrences where mean aerial radiometric measurements are
greater than mean indoor air test results, and uranium mine
locations.
rnk_source <- ggplot() +
geom_sf(data = shp, aes(fill = as.factor(RNK_SRCE2)), color = "NA") +
scale_fill_manual(values = rnk_source_colors, name = "Sources") +
ggtitle("Oregon Radon Potential Rank Sources") +
geom_sf(data = oregon_counties, fill = "transparent", lwd=0.05, color = "black") +
geom_sf(data = oregon, fill = "transparent", color = "black") +
theme(legend.position = "bottom", plot.title = element_text(hjust = 0.5))
ggsave("rnk_source.png", rnk_source)
From this map, we can get a general sense that much of the state’s radon hazard is based on aerial radiometric measurements, with most other sources having a much smaller, roughly equal distribution. This map does provide a decent amount of insight on the distribution of sources, but I’d like to get a more comprehensive view and look at how each of these sources vary by radon potential rank.
3.3 Stacked column chart - rank vs. area by source
To represent the statistic distribution of rank sources, we can make
a stacked column chart with radon potential rank on the x-axis and total
area on the y-axis, with each stack in the column representing a
different rank source. Since each rank source category contains many
polygons, we must use the aggregate function to sum up the
area values from the AREA_KM2 column for each polygon.
Taking this aggregated data and plotting it using the
geom_col function with the fill being set to the
corresponding RNK_SRCE2 category, we can make a stacked
column chart that shows us the distribution of rank sources by area for
each radon potential rank.
rnk_source_agg <- aggregate(AREA_KM2 ~ FIN_RN_RNK + RNK_SRCE2, data, sum)
rnk_source_col <- ggplot(rnk_source_agg, aes(x = FIN_RN_RNK, y = AREA_KM2)) +
geom_col(aes(fill = RNK_SRCE2), color="black", lwd=0.1) +
scale_fill_manual(values = rnk_source_colors, name="Sources") +
labs(x = "Rank", y = "Area (km^2)", title = "Stacked Column Chart - Rank vs. Area by Source") +
theme(legend.position = "bottom", plot.title = element_text(hjust = 0.5)) +
scale_y_continuous(breaks = seq(0, 200000, by = 25000))
ggsave("rnk_source_col.png", rnk_source_col)
Based on this column chart, we can see that areas in Oregon with the lowest radon potential are dominated by aerial radiometric measurements, with a much smaller proportion of geologic material and indoor measurements. Areas with a moderate radon hazard are also dominated by aerial measurements to a lesser degree, with a similar proportion of indoor measurements and a small amount of uranium mines. Areas with the highest radon hazard still have a majority of aerial measurements, with indoor and aerial > indoor measurements making up roughly a third of the area covered.
3.4 Geologic terranes/groups map
The previous plots have given us an idea of how radon ranks vary
spatially throughout the state with rank sources, but I would like to
get an idea of which specific geologic units are responsible for the
most radon hazard. Fortunately, the radon dataset contains a column with
the geologic terrane or group, so we can plot the geology of Oregon
using the same geom_sf method as before.
terranes <- ggplot() +
geom_sf(data = shp, aes(fill = as.factor(TERRANE_GR)), color = "black", lwd=0.01) +
scale_fill_manual(values = colors, name = "Terrane/Group") +
ggtitle("Oregon Terranes and Geologic Groups") +
geom_sf(data = oregon_counties, fill = "transparent", lwd=0.05, color = "black") +
geom_sf(data = oregon, fill = "transparent", color = "black") +
theme(plot.title = element_text(hjust = 0.5),
legend.position = "bottom",
legend.title=element_blank(),
legend.text = element_text(size = 5),
legend.key.size = unit(0.2, "cm"),
legend.key.width = unit(0.2, "cm"))
ggsave("terranes.png", terranes)
3.5 Stacked column chart - rank vs. area by terrane/group
Using this map, I would like to look at the distribution of geologic
terranes/groups by making a stacked column chart similar to the previous
one, but filling the columns with each group instead of rank sources. To
do this, we again must aggregate the data with respect to
AREA_KM2 and plot the aggregated data against the radon
potential ranks using the geom_col function.
terranes_agg <- aggregate(AREA_KM2 ~ FIN_RN_RNK + TERRANE_GR, data, sum)
terranes_col <- ggplot(terranes_agg, aes(x = FIN_RN_RNK, y = AREA_KM2)) +
geom_col(aes(fill = TERRANE_GR), color="black", lwd=0.1) +
scale_fill_manual(values = colors, name="Terrane/Group") +
labs(x = "Rank", y = "Area (km^2)", title = "Stacked Column Chart - Rank vs. Area by Terrane/Group") +
theme(plot.title = element_text(hjust = 0.5),
legend.position = "bottom",
legend.title=element_blank(),
legend.text = element_text(size = 5),
legend.key.size = unit(0.1, "cm"),
legend.key.width = unit(0.1, "cm")) +
scale_y_continuous(breaks = seq(0, 200000, by = 25000))
ggsave("terranes_col.png", terranes_col)
Based on this graph, we can see that the two most dominant geologic groups across all radon hazard ranks are the Columbia River Basalt Group and Quaternary surficial deposits. However, looking more closely at other groups besides these two, we can see that some of the most dominant groups at a high radon hazard rank are the Little Butte Volcanics, Neogene volcanics, High Lava Plains Volcanic Province, and McDermitt Volcanic Field.
4 Discussion
The pattern that we see in the distribution of geologic groups by area aligns with current knowledge regarding the primary sources of radon. Aside from the two largest geologic groups in Oregon (Columbia River Basalt Group and Quaternary sediments), many groups with a high radon hazard rank and significant contribution to total area are composed of intermediate to felsic (high silica) igneous rock, which makes sense based on their typical high uranium content. Even the presence of Quaternary sediments in the high hazard rank category makes sense given that some of these deposits are uranium-rich rocks brought to Oregon by the Missoula floods.
In terms of rank sources, instances where aerial measurements are greater than indoor measurements almost always produce a high hazard rank, suggesting that comparing aerial measurements with indoor measurements is a valuable method to determine where the most significant hazards are present. Interestingly, the presence of uranium mines does not actually correspond to a high hazard rank, and most of these locations are assigned a moderate hazard rank.
5 Conclusion
Overall, mapping radon hazard ranks and the primary sources used to determine them can provide insight into the processes that produce hazardous radon contamination. Comparing the hazard ranks of geologic units allows us to form new interpretations of compositional differences that exhibit control over the release of radon from the subsurface. Improvements that could be made regarding the methods employed in this project include plotting a combined map of hazard rank and rank source data to compare how these factors vary not just by rank but also spatially, and employing statistical methods to provide quantitative measures of each distribution. Moving forward, performing statistical analyses of actual indoor or aerial radiometric measurements may provide a more nuanced perspective compared to just using categorical hazard ranks.
6 References and data sources
Radon data report: Oregon Department of Geology and Mineral Industries (DOGAMI), https://pubs.oregon.gov/dogami/ofr/O-18-01/O-18-01_report.pdf
Radon data shapefile: DOGAMI, https://catalog.data.gov/dataset/radon-data
Oregon outline shapefile: National Weather Service, https://www.weather.gov/gis/USStates
Oregon counties shapefile: U.S. Census Bureau, https://www.census.gov/geographies/mapping-files/time-series/geo/tiger-line-file.html
Sahu, P., Panigrahi, D.C. & Mishra, D.P. A comprehensive review on sources of radon and factors affecting radon concentration in underground uranium mines. Environ Earth Sci 75, 617 (2016). https://doi.org/10.1007/s12665-016-5433-8