Searching for Rare Earth Elements (REEs) in Tasmania

Introduction

“Demand for rare earth elements (REEs) is estimated to grow by 400-600% over the next decade” (https://news.climate.columbia.edu/2023/04/05/the-energy-transition-will-need-more-rare-earth-elements-can-we-secure-them-sustainably/)

Questions:

Do REE’s occur in Tasmania? If so which REE’s occur and where? Are there other elements that occur with these elusive minerals to aid mineral exploration geologists to find REE’s?

Rare Earth Elements (REE’s) occur in a series of critical minerals required to support the electrification targets around the world in support of the Green Revolution. Potassium is a radiogenic element that can be defined using a geophysical method called Airborne Radiometrica where radiogenic elements including potassium (K), uranium(U), thorium(Th) are captured over large tracts of land for a nominal cost. Having elevated potassium would aid exploration geologists in prioritizing areas with high potential for Rare earth elements.

Setup

Imports data set from data folder, ensures numeric variables are numeric and factors are factors.

Load packages

library(tidyverse)
library(here)
library(ggbeeswarm)
library(RColorBrewer)
library(knitr)

Read in Rare Earth Element csv data

plotREE <- read_csv(here("data", "dolerite2.csv"))

Ensures numeric variables are numeric and factors are factors

plotREE$Lappm <- as.numeric(plotREE$Lappm)
plotREE$Ceppm <- as.numeric(plotREE$Ceppm)
plotREE$Yppm <- as.numeric(plotREE$Yppm)
plotREE$K2Opct <- as.numeric(plotREE$K2Opct)
plotREE$Area <- as.factor(plotREE$Area)

Explore the raw REEs data

Investigate which areas in Tasmania to focus on to find REE’s of interest and assess what other-mineral compounds that might guide geologists to potential areas with REE’s like potassium and iron. Potassium is a radiogenic mineral that can be found using a geophysical survey called Airborne Radiometrics that can cover large tracts of land by plane or helicopter for a nominal fee. If potassium occurs with REE’s than this type of survey could be utilized as an exploration tool.

Plot Lanthanum (La ppm) raw data by area

plotREE %>%
  ggplot(aes(x = Lappm, y = Area, color = Lappm)) + 
  geom_jitter() +
  scale_color_gradient(low = "blue", high = "red") +
  theme(text = element_text(size = 8)) +
  labs(title = "Lanthanum values in Tasmania by area",
       subtitle = "Raw Data", 
       caption = "Data Courtesy of https://www.pdac.ca/convention/programming/short-courses",
       y = "Area",
       x = "Lanthanum raw data values")

Figure 1. Lanthanum raw data distribution by area in Tasmania to identify outliers and clustering in the raw data set. The raw data plot uses blue to purple for lower to middle values and red for higher values.

Plot Cesium (Ceppm) values by area

plotREE %>%
  ggplot(aes(x = Ceppm, y = Area, color = Ceppm)) + 
  geom_jitter() +
  scale_color_gradient(low = "blue", high = "red") +
  theme(text = element_text(size = 8)) +
  labs(title = "Cesium values in Tasmania by area",
       subtitle = "Raw Data", 
       y = "Area",
       x = "Cesium Raw Data Values", 
       fill = "data from https://www.pdac.ca/convention/programming/short-courses")

Figure 2. Cesium raw data distribution by area in Tasmania to identify outliers and clustering in the raw data set. The raw data plot uses blue to purple for lower to middle values and red for higher values.

Plot Yttrium (Yppm) values by area

plotREE %>%
  ggplot(aes(x = Yppm, y = Area, color = Yppm)) + 
  geom_jitter() +
  scale_color_gradient(low = "blue", high = "red") +
  theme(text = element_text(size = 8)) +
  labs(title = "Yttrium values in Tasmania by area",
       subtitle = "Raw Data", 
       y = "Area",
       x = "Yttrium Raw Data Values", 
       fill = "data from https://www.pdac.ca/convention/programming/short-courses")

Figure 3. Yttrium raw data distribution by area in Tasmania to identify outliers and clustering in the raw data set.The raw data plot uses blue to purple for lower to middle values and red for higher values.

Plot Potassium oxide values (K2Opct) (a non-REE mineral compound) by area

plotREE %>%
  ggplot(aes(x = K2Opct, y = Area, color = K2Opct)) + 
  geom_jitter() +
  scale_color_gradient(low = "blue", high = "red") +
  theme(text = element_text(size = 8)) +
  labs(title = "Potassium oxide values in Tasmania by area",
  subtitle = "Raw Data", 
  y = "Area",
  x = "Potassium Oxide Raw Data Values", 
  fill = "data from https://www.pdac.ca/convention/programming/short-courses")

Figure 4. Potassium oxide (K2O pct) raw data distribution by area in Tasmania to identify outliers and clustering in the raw data set.The raw data plot uses blue to purple for lower to middle values and red for higher values.

Plot Iron Oxide (a non-REE mineral compound) levels by Area

plotREE %>%
  ggplot(aes(x = FeOpct, y = Area, color = FeOpct)) + 
  geom_jitter() +
  scale_color_gradient(low = "blue", high = "red") +
  theme(text = element_text(size = 8)) +
  labs(title = "Iron oxide values in Tasmania by area",
  subtitle = "Raw Data", 
  y = "Area",
  x = "Iron Oxide Raw Data Values", 
  fill = "data from https://www.pdac.ca/convention/programming/short-courses")

Figure 5. Iron oxide (FeO pct) raw data distribution by area in Tasmania to identify outliers and clustering in the raw data set. The raw data plot uses blue to purple for lower to middle values and red for higher values.

By plotting the raw data in a variable point plot we can see patterns in associations in the REEs (e.g., Lanthanum, Cesium, Yttrium) and associated potassium-bearing mineral compounds. A positive radiometric anomaly for radiogenic potassium may suggest high potential for REE’s.

How does Lanthanum relative to potassium compare across all areas?

plotREE %>%
  na.omit() %>%
  ggplot(aes(x = Lappm, y = K2Opct, color = Lappm)) + 
  geom_jitter() + 
  facet_wrap(~ Area) +
  scale_color_gradient(low = "blue", high = "red") +
  theme(text = element_text(size = 8)) +
  labs(title = " Lanthanum values",
    subtitle = "Faceted across all areas", 
    x = "La ppm",
    y = "K2O percent",
  fill = "data from https://www.pdac.ca/convention/programming/short-courses")

Figure 6. Lanthanum versus Potassium (K2O percent) distribution across all areas. Each point represents a single sample point. The red colors show high Lanthanum values associated with high potassium values at several areas.

How does Cesium relative to potassium compare across all areas?

plotREE %>%
  na.omit() %>%
  ggplot(aes(x = Ceppm, y = K2Opct, color = Ceppm)) + 
  geom_jitter() + 
  facet_wrap(~ Area) +
  scale_color_gradient(low = "blue", high = "red") +
  theme(text = element_text(size = 8)) +
  labs(title = " Cesium values in Tasmania raw data by area",
      subtitle = "Faceted across all areas", 
  y = "K2O percent",
  x = "Ce ppm", 
  fill = "data from https://www.pdac.ca/convention/programming/short-courses")

Figure 7. Cesium versus Potassium (K2O percent) distribution across all areas. Each point represents a single sample point. The red colors show high Cesium values associated with high potassium values at the same areas as Lanthanum levels above.

How does Yttrium relative to potassium compare across all areas?

plotREE %>%
  na.omit() %>%
  ggplot(aes(x = Yppm, y = K2Opct, color = Yppm)) + 
  geom_jitter() + 
  facet_wrap(~ Area) +
  scale_color_gradient(low = "blue", high = "red") +
  theme(text = element_text(size = 8)) +
  labs(title = " Yttrium values in Tasmania raw data by area",
       subtitle = "Faceted across all areas", 
  y = "K2O percent",
  x = "Y ppm", 
  fill = "data from https://www.pdac.ca/convention/programming/short-courses")

Figure 8. Yttrium versus Potassium (K2O percent) distribution across all areas. Each point represents a single sample point. The red colors show high Yttrium values associated with high potassium values at the same areas as Lanthanum and Cesium levels above.

How does Iron Oxide relative to potassium compare across all areas?

plotREE %>%
  na.omit() %>%
  ggplot(aes(x = FeOpct, y = K2Opct, color = FeOpct)) + 
  geom_jitter() + 
  facet_wrap(~ Area) +
  scale_color_gradient(low = "blue", high = "red") +
  theme(text = element_text(size = 8)) +
  labs(title = " Iron Oxide values in Tasmania raw data by area",
        subtitle = "Faceted across all areas", 
  y = "K2O percent",
  x = "FeO percent", 
  fill = "data from https://www.pdac.ca/convention/programming/short-courses")

Figure 9. Iron Oxide (FeO pct) versus Potassium (K2O pct) distribution across all areas. Each point represents a single sample point. The red colors show high FeO values associated with low potassium values at the same areas as Lanthanum and Cesium levels above. The very high iron oxide defined by red show the inverse pattern to the La ppm, Ce ppm and Y ppm plots.

Now focus in on three areas with the highest REE’s of interest.

Filter three areas for potassium versus Lanthanum levels

plotREE %>%
  na.omit() %>%
  filter(Area %in% c("Red Hill", "Red Hill Intrusion", "O'Briens Hill")) %>%
  ggplot(aes(x = K2Opct, y = Lappm, color = Area)) +
  geom_jitter() + 
  facet_wrap(~ Area) +
  theme_grey() +
   theme(text = element_text(size = 8)) +
  labs(title = " Lanthanum (La) versus K2O percent",
  subtitle = "Filtered by areas of interest", 
  x = "K2O percent",
  y = "La ppm", 
  fill = "data from https://www.pdac.ca/convention/programming/short-courses")

Figure 10. Lanthanum (La ppm) versus Potassium(K2O pct) distribution across all three areas. Each point represents a single sample point colored by area. There is a high tendency to have higher La ppm levels with higher potassium levels. At the Red Hill Intrusion there is a cluster of K2O pct less than 1 percent with La ppm < 12ppm. This suggests that while an Airborne Radiometric Survey (K-U-Th) would highlight areas elevated in the radiogenic potassium, the K2O anomaly might be weak. The Red Hill has the highest La ppm outlier at 40ppm La.

Filter three areas for potassium versus Cesium

plotREE %>%
  na.omit() %>%
  filter(Area %in% c("Red Hill", "Red Hill Intrusion", "O'Briens Hill")) %>%
  ggplot(aes(x = K2Opct, y = Ceppm, color = Area)) +
  geom_jitter() + 
  facet_wrap(~ Area) +
  theme_grey() +
   theme(text = element_text(size = 8)) +
  labs(title = " Cesium (Ce) versus K2O percent",
       subtitle = "Filtered by areas of interest", 
  x = "K2O percent",
  y = "Ce ppm", 
  fill = "data from https://www.pdac.ca/convention/programming/short-courses")

Figure 11. Lanthanum (Ce ppm) versus Potassium (K2O pct) distribution across all three areas. Each point represents a single sample point colored by area. There is a strong tendency to have high Cesium values with high potassium values. At the Red Hill Intrusion area there is a cluster of samples with K2O pct less than 1.25 percent with Cesium ranging between 25-30ppm. This suggests that an Airborne Radiometric Survey (K-U-Th) could highlight areas elevated in the radiogenic potassium to find potential areas of Cesium. The Red Hill area has the highest outlier at close to 90ppm Ce.

Filter three areas for potassium versus Yttrium

plotREE %>%
  na.omit() %>%
  filter(Area %in% c("Red Hill", "Red Hill Intrusion", "O'Briens Hill")) %>%
  ggplot(aes(x = K2Opct, y = Yppm, color = Area)) +
  geom_jitter() + 
  facet_wrap(~ Area) +
   theme(text = element_text(size = 8)) +
  labs(title = " Yttrium(Y) versus K2O percent",
  subtitle = "Filtered by areas of interest", 
  x = "K2O percent",
  y = "Y ppm", 
  fill = "data from https://www.pdac.ca/convention/programming/short-courses")

Figure 12. K2O pct versus Yppm. Each dot corresponds to one sample. There is a strong tendency to have high Yttrium values with high potassium values. The potassium values increase with the Yttrium values in support of using an Airborne Radiometric Survey (K-U-Th) to find potential areas of Yttrium.

Filter the areas for potassium versus iron oxide(FeOpct)

plotREE %>%
  na.omit() %>%
  filter(Area %in% c("Red Hill", "Red Hill Intrusion", "O'Briens Hill")) %>%
  ggplot(aes(x = K2Opct, y = FeOpct, color = Area)) +
  geom_jitter() + 
  facet_wrap(~ Area) +
  theme(text = element_text(size = 8)) +
  labs(title = " FeO percent versus K2O percent",
  subtitle = "Filtered by areas of interest", 
  x = "K2O percent",
  y = "FeO percent", 
  fill = "data from https://www.pdac.ca/convention/programming/short-courses")

Figure 13. K2O percent versus FeO percent. Each dot corresponds to one sample. At the Red Hill Intrusion and O’Briens Hill high levels of iron oxide (above 7.75 % FeO) show the opposite pattern to the other REE plots above suggesting that there is a tendency for lower potassium as well as rare earths associated with higher iron oxide values.

Now look in a bit more detail at Lanthanum.

Show Lanthanum by Area using Boxplot

plotREE %>%
  na.omit() %>%
  ggplot(aes(x = Area, y = Lappm)) +
  geom_boxplot() +
  coord_flip() + 
  theme_classic() +
  theme(text = element_text(size = 6)) +
  labs(title = "Lanthanum spread",
     subtitle = "Distribution by Area", 
  x = "Area",
  y = "La ppm", 
  fill = "data from https://www.pdac.ca/convention/programming/short-courses")

Figure 14. The Box plot above of Lanthanum by area shows the highest values and greatest spread of Lanthanum values occur at the Red Hill and O’Brien Hills areas. The mean is shown in the solid black line in each box (mean = 18ppm La at Red Hill area)

Look at correlations between variables

Correlation between Lappm and K20 and colored by Ceppm

plotREE %>%
  na.omit() %>%
  filter(Lappm > 10) %>%
  ggplot(aes(x = K2Opct, y = Lappm, color = Ceppm)) +
  geom_point() +
  geom_smooth() +
  scale_color_gradient(low = "blue", high = "red") +
  theme_classic() +
   theme(text = element_text(size = 8)) +
  labs(title = " Lanthanum values in Tasmania compared to potassium",
       subtitle = "Colored by Cesium and filtered La > 10 ppm",
       caption = "data from https://www.pdac.ca/convention/programming/short-courses")

Figure 15. Lanthanum versus potassium and colored by Cesium correlation. The image shows Cesium in a continuous scheme from blues for poor correlation of Cesium with La ppm and K2O pct or red for good correlation with La ppm and K2O pct. There are only five samples that show a good correlation with Cesium.

Export Figure 15

ggsave(here("output", "Figure 15.png"))

Conclusion

There is rare earth element potential in several areas in Tasmania including Lanthanum, Cesium and Yttrium.
These correlate well with the mineral compound potassium oxide (K2O percent).
An airborne radiometric survey could easily help in identifying high potential areas of high potassium with elevated rare earth elements in Tasmania.

Future work - Calculate and visualize the Correlation Coefficients of a series of REE’s, together with other data including litho-geochemical data as well as and trace element data through a correlogram to provide associations of several variables at the same time using a diverging scale for higher correlation coefficients versus lower values. - Explore All-against-All scatter plot matrix visualizations to see if they are easier to interpret in a series of pair-wise scatter plots.

Explore the output from a typical radiometric survey over western Australia.