Temperature Trends in Korea from 1984 to 2024

Research Question: “Are Summer and Winter Getting Hotter in South Korea?”

About the Dataset
Sparkline
Line plot
Moving average
Conclusion

About the Dataset

8 Dataset
6 Columns
Total 595399 rows
Date: Year(1984-2024), Month, Day -Location: Seoul, Incheon, Busan, Daegu, Daejeon, Gwangju, Ulsan, Wonju
Mean Temperature _ Minimum Temperature _ Maximum Temperature

Data sourse: Korean Meteorological Administration

Package load

library(readr)
library(tidyr)
library(ggplot2)
library(ggthemes)
library(dplyr)

## 
## 다음의 패키지를 부착합니다: 'dplyr'

## The following objects are masked from 'package:stats':
## 
##     filter, lag

## The following objects are masked from 'package:base':
## 
##     intersect, setdiff, setequal, union

library(zoo)

## 
## 다음의 패키지를 부착합니다: 'zoo'

## The following objects are masked from 'package:base':
## 
##     as.Date, as.Date.numeric

Read CSV file

Busan <- read_csv("Busan.csv")

## Rows: 14916 Columns: 5
## ── Column specification ────────────────────────────────────────────────────────
## Delimiter: ","
## chr  (1): Location
## dbl  (3): Mean Temperature(°C), Minimum Temperature(°C), Maximum Temperature...
## date (1): Date
## 
## ℹ Use `spec()` to retrieve the full column specification for this data.
## ℹ Specify the column types or set `show_col_types = FALSE` to quiet this message.

Daegu <- read_csv("Daegu.csv")

## Rows: 14916 Columns: 5
## ── Column specification ────────────────────────────────────────────────────────
## Delimiter: ","
## chr  (1): Location
## dbl  (3): Mean Temperature(°C), Minimum Temperature(°C), Maximum Temperature...
## date (1): Date
## 
## ℹ Use `spec()` to retrieve the full column specification for this data.
## ℹ Specify the column types or set `show_col_types = FALSE` to quiet this message.

Daejeon <- read_csv("Daejeon.csv")

## Rows: 14916 Columns: 5
## ── Column specification ────────────────────────────────────────────────────────
## Delimiter: ","
## chr  (1): Location
## dbl  (3): Mean Temperature(°C), Minimum Temperature(°C), Maximum Temperature...
## date (1): Date
## 
## ℹ Use `spec()` to retrieve the full column specification for this data.
## ℹ Specify the column types or set `show_col_types = FALSE` to quiet this message.

Gwangju <- read_csv("Gwangju.csv")

## Rows: 14916 Columns: 5
## ── Column specification ────────────────────────────────────────────────────────
## Delimiter: ","
## chr  (1): Location
## dbl  (3): Mean Temperature(°C), Minimum Temperature(°C), Maximum Temperature...
## date (1): Date
## 
## ℹ Use `spec()` to retrieve the full column specification for this data.
## ℹ Specify the column types or set `show_col_types = FALSE` to quiet this message.

Incheon <- read_csv("Incheon.csv")

## Rows: 14916 Columns: 5
## ── Column specification ────────────────────────────────────────────────────────
## Delimiter: ","
## chr  (1): Location
## dbl  (3): Mean Temperature(°C), Minimum Temperature(°C), Maximum Temperature...
## date (1): Date
## 
## ℹ Use `spec()` to retrieve the full column specification for this data.
## ℹ Specify the column types or set `show_col_types = FALSE` to quiet this message.

Seoul <- read_csv("Seoul.csv")

## Rows: 14916 Columns: 5
## ── Column specification ────────────────────────────────────────────────────────
## Delimiter: ","
## chr  (1): Location
## dbl  (3): Mean Temperature(°C), Minimum Temperature(°C), Maximum Temperature...
## date (1): Date
## 
## ℹ Use `spec()` to retrieve the full column specification for this data.
## ℹ Specify the column types or set `show_col_types = FALSE` to quiet this message.

Ulsan <- read_csv("Ulsan.csv")

## Rows: 14916 Columns: 5
## ── Column specification ────────────────────────────────────────────────────────
## Delimiter: ","
## chr  (1): Location
## dbl  (3): Mean Temperature(°C), Minimum Temperature(°C), Maximum Temperature...
## date (1): Date
## 
## ℹ Use `spec()` to retrieve the full column specification for this data.
## ℹ Specify the column types or set `show_col_types = FALSE` to quiet this message.

Wonju <- read_csv("Wonju.csv")

## Rows: 14916 Columns: 5
## ── Column specification ────────────────────────────────────────────────────────
## Delimiter: ","
## chr  (1): Location
## dbl  (3): Mean Temperature(°C), Minimum Temperature(°C), Maximum Temperature...
## date (1): Date
## 
## ℹ Use `spec()` to retrieve the full column specification for this data.
## ℹ Specify the column types or set `show_col_types = FALSE` to quiet this message.

data processing

Busan <- Busan %>%
  separate(col = Date, into = c("Year", "Month", "Day"), sep="-")

Busan <- drop_na(Busan)
Busan$Year <- as.numeric(as.character(Busan$Year))

Daegu <- Daegu %>%
  separate(col = Date, into = c("Year", "Month", "Day"), sep="-")

Daegu <- drop_na(Daegu)
Daegu$Year <- as.numeric(as.character(Daegu$Year))

Daejeon <- Daejeon %>%
  separate(col = Date, into = c("Year", "Month", "Day"), sep="-")

Daejeon <- drop_na(Daejeon)
Daejeon$Year <- as.numeric(as.character(Daejeon$Year))

Gwangju <- Gwangju %>%
  separate(col = Date, into = c("Year", "Month", "Day"), sep="-")

Gwangju <- drop_na(Gwangju)
Gwangju$Year <- as.numeric(as.character(Gwangju$Year))

Incheon <- Incheon %>%
  separate(col = Date, into = c("Year", "Month", "Day"), sep="-")

Incheon <- drop_na(Incheon)
Incheon$Year <- as.numeric(as.character(Incheon$Year))

Seoul <- Seoul %>%
  separate(col = Date, into = c("Year", "Month", "Day"), sep="-")

Seoul <- drop_na(Seoul)
Seoul$Year <- as.numeric(as.character(Seoul$Year))

Ulsan <- Ulsan %>%
  separate(col = Date, into = c("Year", "Month", "Day"), sep="-")

Ulsan <- drop_na(Ulsan)
Ulsan$Year <- as.numeric(as.character(Ulsan$Year))

Wonju <- Wonju %>%
  separate(col = Date, into = c("Year", "Month", "Day"), sep="-")

Wonju <- drop_na(Wonju)
Wonju$Year <- as.numeric(as.character(Wonju$Year))

Combined data

Korea <- rbind(Busan, Daegu, Daejeon, Gwangju, Incheon, Seoul, Ulsan, Wonju)

Korea <- Korea %>%
  rename('maxT' = 'Maximum Temperature(°C)',
         'miniT' = 'Minimum Temperature(°C)',
         'meanT' = 'Mean Temperature(°C)')

Max Temperature

HW_df <- Korea %>%
  select("Location", "Year", "Month", "Day", "maxT")

Min Temperature

CW_df <- Korea %>%
  select("Location", "Year", "Month", "Day", "miniT")

Sparkline

August Mean Max Temperature

Aug_HW <- HW_df %>%
  filter(Month == "08")

Aug_HWL <- Aug_HW %>%
  group_by(Year, Location) %>%
  summarise(mean(maxT))

## `summarise()` has grouped output by 'Year'. You can override using the
## `.groups` argument.

Aug_HWT <- Aug_HW %>%
  group_by(Year) %>%
  summarise(mean(maxT))

Aug_HWT <- Aug_HWT %>%
  mutate(Location = "Total") %>%
  select("Year", "Location", "mean(maxT)")

Aug_Hmean <- rbind(Aug_HWL, Aug_HWT)

Aug_Hmean <- Aug_Hmean %>%
  rename('mmaxT' = 'mean(maxT)')

Aug_Hmean$'mmaxT' <- round(Aug_Hmean$'mmaxT', 1)

Aug_Hmean <- Aug_Hmean %>%
  mutate(Location = factor(Location, levels = c("Total", setdiff(unique(Location), "Total"))))

Aug_Mins <- group_by(Aug_Hmean, Location) %>% slice(which.min(mmaxT))
Aug_Maxs <- group_by(Aug_Hmean, Location) %>% slice(which.max(mmaxT))
Aug_Ends <- group_by(Aug_Hmean, Location) %>% filter(Year == max(Year))
Aug_quarts <- Aug_Hmean %>% group_by(Location) %>%
  summarize(Aug_quart1 = quantile(mmaxT, 0.25),
            Aug_quart2 = quantile(mmaxT, 0.75)) %>%
  right_join(Aug_Hmean)

## Joining with `by = join_by(Location)`

ggplot(Aug_Hmean, aes(x = Year, y = mmaxT)) + 
  facet_grid(Location ~ ., scales = "free_y") + 
  geom_ribbon(data = Aug_quarts, aes(ymin = Aug_quart1, ymax = Aug_quart2), fill = 'grey90') +
  geom_line(linewidth = 0.3) +
  geom_point(data = Aug_Mins, col = 'blue') +
  geom_point(data = Aug_Maxs, col = 'red') +
  geom_text(data = Aug_Mins, aes(label = round(mmaxT, 1)), size = 3, vjust = -1) +
  geom_text(data = Aug_Maxs, aes(label = round(mmaxT, 1)), size = 3, vjust = 2.5) +
  geom_text(data = Aug_Ends, aes(label = round(mmaxT, 1)), size = 3, hjust = 0, nudge_x = 1) +
  geom_text(data = Aug_Ends, aes(label = Location), size = 3, hjust = 0, nudge_x = 5) +
  expand_limits(x = max(Aug_Hmean$Year) + (0.25 * (max(Aug_Hmean$Year) - min(Aug_Hmean$Year)))) +
  scale_x_continuous(breaks = seq(min(Aug_Hmean$Year), max(Aug_Hmean$Year), 5)) +
  scale_y_continuous(expand = c(0.1, 0)) +
  theme_tufte(base_size = 12) +
  theme(
    axis.title = element_blank(), 
    axis.text.y = element_blank(), 
    axis.ticks = element_blank(), 
    strip.text = element_blank()
  )

Aug_filtered <- Aug_Hmean[Aug_Hmean$Year %in% c(1984, 2024), ]

Aug_1984 <- Aug_filtered[Aug_filtered$Year == 1984, ]
Aug_2024 <- Aug_filtered[Aug_filtered$Year == 2024, ]

Aug_diff <- merge(Aug_1984, Aug_2024, by = "Location", suffixes = c("_1984", "_2024"))

Aug_diff <- Aug_diff %>%
  mutate(Difference = mmaxT_2024 - mmaxT_1984)

In August, the overall trend shows a steady rise in maximum temperatures from 1984 to 2024
Except for Daegu, all cities reached their highest temperatures after 2010
Comparing 1984 and 2024, Seoul and Incheon showed the largest increases, rising by 2.8°C and 2.7°C

Mean Min Temperature

Jan_CW <- CW_df %>%
  filter(Month == "01")

Jan_CWL <- Jan_CW %>%
  group_by(Year, Location) %>%
  summarise(mean(miniT))

## `summarise()` has grouped output by 'Year'. You can override using the
## `.groups` argument.

Jan_CWT <- Jan_CW %>%
  group_by(Year) %>%
  summarise(mean(miniT))

Jan_CWT <- Jan_CWT %>%
  mutate(Location = "Total") %>%
  select("Year", "Location", "mean(miniT)")

Jan_Cmean <- rbind(Jan_CWL, Jan_CWT)

Jan_Cmean <- Jan_Cmean %>%
  rename('mminiT' = 'mean(miniT)')

Jan_Cmean$'mminiT' <- round(Jan_Cmean$'mminiT', 1)

Jan_Cmean <- Jan_Cmean %>%
  mutate(Location = factor(Location, levels = c("Total", setdiff(unique(Location), "Total"))))

Jan_Mins <- group_by(Jan_Cmean, Location) %>% slice(which.min(mminiT))
Jan_Maxs <- group_by(Jan_Cmean, Location) %>% slice(which.max(mminiT))
Jan_Ends <- group_by(Jan_Cmean, Location) %>% filter(Year == max(Year))
Jan_quarts <- Jan_Cmean %>% group_by(Location) %>%
  summarize(Jan_quart1 = quantile(mminiT, 0.25),
            Jan_quart2 = quantile(mminiT, 0.75)) %>%
  right_join(Jan_Cmean)

## Joining with `by = join_by(Location)`

ggplot(Jan_Cmean, aes(x = Year, y = mminiT)) + 
  facet_grid(Location ~ ., scales = "free_y") + 
  geom_ribbon(data = Jan_quarts, aes(ymin = Jan_quart1, ymax = Jan_quart2), fill = 'grey90') +
  geom_line(linewidth = 0.3) +
  geom_point(data = Jan_Mins, col = 'blue') +
  geom_point(data = Jan_Maxs, col = 'red') +
  geom_text(data = Jan_Mins, aes(label = round(mminiT, 1)), size = 3, vjust = -1) +
  geom_text(data = Jan_Maxs, aes(label = round(mminiT, 1)), size = 3, vjust = 2.5) +
  geom_text(data = Jan_Ends, aes(label = round(mminiT, 1)), size = 3, hjust = 0, nudge_x = 1) +
  geom_text(data = Jan_Ends, aes(label = Location), size = 3, hjust = 0, nudge_x = 5) +
  expand_limits(x = max(Jan_Cmean$Year) + (0.25 * (max(Jan_Cmean$Year) - min(Jan_Cmean$Year)))) +
  scale_x_continuous(breaks = seq(min(Jan_Cmean$Year), max(Jan_Cmean$Year), 5)) +
  scale_y_continuous(expand = c(0.1, 0)) +
  theme_tufte(base_size = 12) +
  theme(
    axis.title = element_blank(), 
    axis.text.y = element_blank(), 
    axis.ticks = element_blank(), 
    strip.text = element_blank()
  )

Jan_filtered <- Jan_Cmean[Jan_Cmean$Year %in% c(1984, 2024), ]

Jan_1984 <- Jan_filtered[Jan_filtered$Year == 1984, ]
Jan_2024 <- Jan_filtered[Jan_filtered$Year == 2024, ]

Jan_diff <- merge(Jan_1984, Jan_2024, by = "Location", suffixes = c("_1984", "_2024"))

Jan_diff <- Jan_diff %>%
  mutate(Difference = mminiT_2024 - mminiT_1984)

In January, the overall trend shows a steady increase in minimum temperatures from 1984 to 2024, with a sharp drop in 2011 and a significant rise in 2020.
Comparing 1984 and 2024, Wonju showed the largest temperature increase, rising by 15.6°C.
The coldest northern regions of South Korea have also experienced temperature increases, making their minimum temperatures in 2024 similar to those of other regions

Line Plot

August

Total <- Korea %>%
  group_by(Year, Month, Day) %>%
  summarise(mean(maxT), mean(miniT))

## `summarise()` has grouped output by 'Year', 'Month'. You can override using the
## `.groups` argument.

Total <- Total %>%
  mutate(Location = "Total") %>%
  select("Year", "Month", "Day", "Location", "mean(maxT)", "mean(miniT)") %>%
  filter(Year == 1984 | Year == 2024)

Total <- Total %>%
  rename('mmaxT' = 'mean(maxT)',
         'mminiT' = 'mean(miniT)')

Total$'mmaxT' <- round(Total$'mmaxT', 1)
Total$'mminiT' <- round(Total$'mminiT', 1)

Aug_Total <- Total %>%
  filter(Month == "08") %>%
  select("Year", "Month", "Day", "mmaxT")

Aug_Count <- Aug_Total %>%
  group_by(Year) %>%
  summarize(days_over_33 = sum(mmaxT >= 33))

ggplot(Aug_Total, aes(x = Day, y = mmaxT, group = Year, color = as.factor(Year))) +
  geom_hline(yintercept = 33, color = "gray", linetype = "solid", size = 0.7) +
  geom_line(linewidth = 1, alpha = 0.7) +
  scale_color_manual(values = c("1984" = "skyblue", "2024" = "red"), 
                     name = "Year") +
  labs(x = "Day of August", 
       y = "Maximum Temperature (°C)", 
       title = "Daily Maximum Temperatures in August Between 1984 and 2024") +
  theme_minimal() +
  theme_tufte()

## Warning: Using `size` aesthetic for lines was deprecated in ggplot2 3.4.0.
## ℹ Please use `linewidth` instead.
## This warning is displayed once every 8 hours.
## Call `lifecycle::last_lifecycle_warnings()` to see where this warning was
## generated.

August minimum temperatures in 2024 were higher than in 1984 on nearly every day, with more days exceeding the heatwave threshold of 33°C.

January

Jan_Total <- Total %>%
  filter(Month == "01") %>%
  select("Year", "Month", "Day", "mminiT")

Jan_Count <- Jan_Total %>%
  group_by(Year) %>%
  summarize(days_under_0 = sum(mminiT < 0))

ggplot(Jan_Total, aes(x = Day, y = mminiT, group = Year, color = as.factor(Year))) +
  geom_hline(yintercept = 0, color = "gray", linetype = "solid", size = 0.7) +
  geom_line(linewidth = 1, alpha = 0.7) +
  scale_color_manual(values = c("1984" = "blue", "2024" = "orange"), 
                     name = "Year") +
  labs(x = "Day of January", 
       y = "Minimum Temperature (°C)", 
       title = "Daily Minimum Temperatures in January Between 1984 and 2024") +
  theme_minimal() +
  theme_tufte()

January minimum temperatures in 2024 were higher than in 1984 on nearly every day, with temperatures reaching above zero

Moving Average

August

Aug_Mean <- Korea %>%
  filter(Month == "08")

Aug_Mean <- Aug_Mean %>%
  group_by(Year) %>%
  summarise(mean(meanT))

Aug_Mean <- Aug_Mean %>%
  rename('mmeanT' = 'mean(meanT)')
  
Aug_Mean$'mmeanT' <- round(Aug_Mean$'mmeanT', 1)

Aug_Mean$Aug_Decade <- (Aug_Mean$Year %/% 10) * 10 
Aug_avg <- Aug_Mean %>%
  group_by(Aug_Decade) %>%
  summarize(Decade_Avg = mean(mmeanT, na.rm = TRUE))

Aug_Mean$Aug_7ma <- rollmean(Aug_Mean$mmeanT, 7, fill = NA, align = "center")

ggplot() +
  geom_bar(data = Aug_avg, aes(x = Aug_Decade, y = Decade_Avg, fill = "Decadal Average"), 
           stat = "identity", alpha = 0.8, width = 8) +
  geom_line(data = Aug_Mean, aes(x = Year, y = mmeanT, color = "Yearly Temperature"), size = 1) +
  geom_line(data = Aug_Mean, aes(x = Year, y = Aug_7ma, color = "7-Year Moving Average"), size = 1) +
  labs(
    title = "Change in August Mean Temperature (1984-2024)",
    x = "Year",
    y = "Mean Temperature (°C)",
    color = "",
    fill = ""
  ) +
  scale_color_manual(values = c("Yearly Temperature" = "yellowgreen", "7-Year Moving Average" = "blue")) +
  scale_fill_manual(values = c("Decadal Average" = "orange")) +
  theme_minimal() +
  theme_tufte()

## Warning: Removed 6 rows containing missing values or values outside the scale range
## (`geom_line()`).

January mean temperatures are rising overall, indicating warmer winters
The year 2011 appears to be an outlier.

January

Jan_Mean <- Korea %>%
  filter(Month == "01")

Jan_Mean <- Jan_Mean %>%
  group_by(Year) %>%
  summarise(mean(meanT))

Jan_Mean <- Jan_Mean %>%
  rename('mmeanT' = 'mean(meanT)')

Jan_Mean$'mmeanT' <- round(Jan_Mean$'mmeanT', 1)

Jan_Mean$Jan_Decade <- (Jan_Mean$Year %/% 10) * 10 
Jan_avg <- Jan_Mean %>%
  group_by(Jan_Decade) %>%
  summarize(Decade_Avg = mean(mmeanT, na.rm = TRUE))

Jan_Mean$Jan_7ma <- rollmean(Jan_Mean$mmeanT, 7, fill = NA, align = "center")

ggplot() +
  geom_bar(data = Jan_avg, aes(x = Jan_Decade, y = Decade_Avg, fill = "Decadal Average"), 
           stat = "identity", alpha = 0.7, width = 9) +
  geom_line(data = Jan_Mean, aes(x = Year, y = mmeanT, color = "Yearly Temperature"), size = 1) +
  geom_line(data = Jan_Mean, aes(x = Year, y = Jan_7ma, color = "7-Year Moving Average"), size = 1) +
  labs(
    title = "Change in January Mean Temperature (1984-2024)",
    x = "Year",
    y = "Mean Temperature (°C)",
    color = "",
    fill = ""
  ) +
  scale_color_manual(values = c("Yearly Temperature" = "yellowgreen", "7-Year Moving Average" = "red")) +
  scale_fill_manual(values = c("Decadal Average" = "blue")) +
  theme_minimal() +
  theme_tufte()

## Warning: Removed 6 rows containing missing values or values outside the scale range
## (`geom_line()`).

August mean temperatures are rising overall, indicating hotter summers.

Conclusion

Winter:

Both minimum and average temperatures in January are rising, with extreme cold becoming less frequent.

Winters are gradually getting warmer over time.

Summer:

Both maximum and average temperatures in August have steadily increased since 1984.

Summers are gradually getting hotter with more frequent and prolonged heatwaves.

CDS301 Final Project

Heyeon Kim, Jaehyoung Jang

Dec 18 2024

Temperature Trends in Korea from 1984 to 2024

Research Question: “Are Summer and Winter Getting Hotter in South Korea?”

Contents

About the Dataset

Package load

Read CSV file

data processing

Combined data

Max Temperature

Min Temperature

Sparkline

August Mean Max Temperature

Mean Min Temperature

Line Plot

August

January

Moving Average

August

January

Conclusion

Winter:

Summer: