Assessing the Impact of Severe Weather Events on Public Health and the Economy
An Analysis using NOAA Storm Data
Saurabh Kelkar
2024-09-28
1 Introduction
Severe weather events, such as hurricanes, floods, and droughts, have widespread and far-reaching impacts. Economically, they cause extensive damage to infrastructure and agriculture, leading to business disruptions, increased insurance costs, and significant financial losses [1]. The impact on human health is equally severe, as they lead to injuries, fatalities, and heightened mental health challenges. Furthermore, waterborne and vector-borne diseases often surge in the aftermath of floods. These events also displace populations, especially in marginalized communities that lack the resources to recover quickly.
The summer of 2023 was marked by unprecedented extreme weather events, including record heat, wildfires, and intense storms across North America. This season was noted as the hottest on record for the Northern Hemisphere, exacerbating conditions for severe weather [2]. High temperatures also led to increased rainfall intensity and flooding events in various regions [3].
This project investigates severe weather events in the United States. The primary objective is to determine which types of weather events have the greatest impact on public health and the economy. This may help allocate resources more effectively and enhance preparedness strategies to mitigate future risks. Understanding these patterns is crucial for improving response measures and reducing the overall consequences of severe weather on communities.
2 Data and Methodology
2.1 Region
The analysis focuses on five US states: Colorado, Florida, Idaho, Nevada, and Utah, which US News ranked as the top 5 states of the US Economy sector in 2024 [4]. The rankings take into account various surveys on each state’s business environment, labor market, and overall economic growth conducted from 2020.
2.2 Severe Event Data: NOAA Storm Events Database
The data for the top 5 states, Colorado, Florida, Idaho, Nevada, and
Utah, for all events from June 1 to June 30, 2023, was downloaded in
CSV format from the NOAA Storm Events Database [5].
The NOAA Storm Events Database contains the records on:
The occurrence of storms and other significant weather phenomena having sufficient intensity to cause loss of life, injuries, significant property damage, and disruption to commerce.
Rare, unusual weather phenomena that generate media attention.
Other significant meteorological events, such as record maximum or minimum temperatures or precipitation in connection with another event.
The database contains data from January 1950 to June 2024 and is regularly updated by NOAA’s National Weather Service (NWS). For this analysis, I have focused specifically on June 2023, as it marks the beginning of the summer season, a period when the US typically experiences more severe weather events due to intense weather fronts. Moreover, June 2023 represents the most recent complete data for the start of the US summer season, because the data for June 2024 was still being recorded during the data analysis.
2.3 R Packages for Analysis and Visualization
I will use multiple packages to process and analyze data. For data
handling and cleaning, functions from dplyr,
readr, and VIM will be used. Data
visualization will be carried out with the ggplot2 library.
I will also write some functions to speed up the analysis.
suppressPackageStartupMessages({
suppressWarnings({
library(dplyr)
library(readr)
library(VIM)
library(ggplot2)
})
})2.4 Merging Data
Each state has an individual CSV file, so merging them
will make the data analysis and processing smooth.
# Get a list of all CSV files in the directory
file_list <- list.files("D:/Mini_projects/Project_2", pattern = "*.csv", full.names = TRUE)
# Read, modify, and combine all CSV files into one data frame
df <- file_list %>%
lapply(read_csv, show_col_types = FALSE) %>% # Read each CSV file and add the state column
bind_rows() # Combine them by rows
# Glimpse the combined data
glimpse(df)## Rows: 976
## Columns: 39
## $ EVENT_ID <dbl> 1106772, 1114089, 1092538, 1114091, 1095326, 11067…
## $ CZ_NAME_STR <chr> "ELBERT CO.", "KIT CARSON CO.", "PITKIN CO.", "YUM…
## $ BEGIN_LOCATION <chr> "FONDIS", "BURLINGTON", "(ASE)ASPEN ARPT", "YUMA",…
## $ BEGIN_DATE <chr> "06/01/2023", "06/01/2023", "06/02/2023", "06/02/2…
## $ BEGIN_TIME <dbl> 1330, 1556, 1500, 1620, 1100, 1543, 1550, 1608, 16…
## $ EVENT_TYPE <chr> "Hail", "Hail", "Flood", "Hail", "Flood", "Hail", …
## $ MAGNITUDE <dbl> 1.00, 0.75, NA, 0.75, NA, 1.00, NA, 1.00, NA, 1.00…
## $ TOR_F_SCALE <chr> NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, "EF0", NA,…
## $ DEATHS_DIRECT <dbl> 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,…
## $ INJURIES_DIRECT <dbl> 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,…
## $ DAMAGE_PROPERTY_NUM <dbl> 0, 0, 5000, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,…
## $ DAMAGE_CROPS_NUM <dbl> 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,…
## $ STATE_ABBR <chr> "CO", "CO", "CO", "CO", "CO", "CO", "CO", "CO", "C…
## $ CZ_TIMEZONE <chr> "MST-7", "MST-7", "MST-7", "MST-7", "MST-7", "MST-…
## $ MAGNITUDE_TYPE <chr> NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA…
## $ EPISODE_ID <dbl> 181742, 182684, 179382, 182685, 179833, 181743, 18…
## $ CZ_TYPE <chr> "C", "C", "C", "C", "C", "C", "C", "C", "C", "C", …
## $ CZ_FIPS <dbl> 39, 63, 97, 125, 37, 35, 89, 5, 89, 5, 123, 101, 4…
## $ WFO <chr> "BOU", "GLD", "GJT", "GLD", "GJT", "BOU", "PUB", "…
## $ INJURIES_INDIRECT <dbl> 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,…
## $ DEATHS_INDIRECT <dbl> 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,…
## $ SOURCE <chr> "Public", "Public", "Law Enforcement", "Trained Sp…
## $ FLOOD_CAUSE <chr> NA, NA, "Heavy Rain / Snow Melt", NA, "Heavy Rain …
## $ TOR_LENGTH <dbl> NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, 0.01, NA, …
## $ TOR_WIDTH <dbl> NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, 20, NA, NA…
## $ BEGIN_RANGE <dbl> 3, 0, 4, 0, 0, 1, 2, 2, 2, 1, 3, 1, 2, 3, 4, 1, 6,…
## $ BEGIN_AZIMUTH <chr> "N", "E", "W", "SSE", "SE", "E", "SW", "NNW", "SW"…
## $ END_RANGE <dbl> 3, 0, 4, 0, 2, 1, 2, 2, 2, 1, 3, 1, 2, 3, 4, 1, 6,…
## $ END_AZIMUTH <chr> "N", "E", "W", "SSE", "NW", "E", "SW", "NNW", "SW"…
## $ END_LOCATION <chr> "FONDIS", "BURLINGTON", "(ASE)ASPEN ARPT", "YUMA",…
## $ END_DATE <chr> "06/01/2023", "06/01/2023", "06/02/2023", "06/02/2…
## $ END_TIME <dbl> 1330, 1556, 1600, 1620, 1100, 1543, 1551, 1608, 19…
## $ BEGIN_LAT <dbl> 39.2600, 39.2990, 39.2107, 40.1191, 39.8653, 39.52…
## $ BEGIN_LON <dbl> -104.4400, -102.2610, -106.9474, -102.7195, -106.8…
## $ END_LAT <dbl> 39.2600, 39.2990, 39.2107, 40.1191, 39.8902, 39.52…
## $ END_LON <dbl> -104.4400, -102.2610, -106.9478, -102.7195, -106.7…
## $ EVENT_NARRATIVE <chr> NA, "Public report of 0.75 inch hail with visibili…
## $ EPISODE_NARRATIVE <chr> "A severe thunderstorm produced hail up to quarter…
## $ ABSOLUTE_ROWNUMBER <dbl> 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15,…2.5 Pre-processing
Before proceeding to the impact estimations, data cleaning and handling missing values must be performed. First, the variables of interest are selected and strings for the names of severe events are re-formatted for consistency.
storm_data_clean <- df |>
# Select the variables of interest
select(STATE_ABBR, EVENT_TYPE, DEATHS_DIRECT, DEATHS_INDIRECT, INJURIES_DIRECT, INJURIES_INDIRECT, DAMAGE_PROPERTY_NUM, DAMAGE_CROPS_NUM) |>
# Convert the names in EVENT_TYPE to uppercase
mutate(EVENT_TYPE = toupper(EVENT_TYPE))Let’s check the missing data using aggr() function from
VIM package.
aggr(storm_data_clean, col = c('navyblue', 'yellow'), numbers = TRUE, sortVars = TRUE,
labels = c("STATE_ABBR", "EVENT_TYPE", "DEATHS_DIRECT", "DEATHS_INDIRECT",
"INJURIES_DIRECT", "INJURIES_INDIRECT", "DAMAGE_PROPERTY_NUM",
"DAMAGE_CROPS_NUM"),
cex.axis = 0.3, gap = 1, ylab = c("Missing Data", "Pattern"))
##
## Variables sorted by number of missings:
## Variable Count
## STATE_ABBR 0
## EVENT_TYPE 0
## DEATHS_DIRECT 0
## DEATHS_INDIRECT 0
## INJURIES_DIRECT 0
## INJURIES_INDIRECT 0
## DAMAGE_PROPERTY_NUM 0
## DAMAGE_CROPS_NUM 0Perfect! There seems to be no missing value.
Note: The aggr() function generates a plot that
provides a comprehensive visual overview of missing data in a dataset.
This plot is composed of two primary components. On the left side, a bar
plot illustrates the percentage of missing values for each variable,
where the height of each bar reflects the proportion of missing data in
that specific variable, with the x-axis displaying the variable names.
On the right, the missing value patterns are shown in a matrix-like
representation. Each row in this matrix corresponds to a unique
combination of missing values across variables, with the x-axis labeling
the variables and the y-axis showing different missingness patterns. The
color intensity of each cell indicates the frequency of that specific
pattern, helping users easily detect and understand missing data
structures within the dataset.
Wait a minute! The data seems to have a lot of 0 as seen
in glimpse(df). Entries with 0 can be
confusing. In some cases, 0 may represent a lack of data or
an event that did not yield any results. However, this can easily be
tangled up with missing values (NA). If 0 is
treated as a missing data, but in reality, if it is meaningful value
(e.g. 0 m/s wind speed), it will obscure the analysis (e.g., mean,
correlations, PDFs). Therefore, handling zeros is crucial to avoid
misinterpretations.
Let’s look at how many 0 there are.
total_zeros <- storm_data_clean |>
# Counting zeros
summarise(across(everything(), ~ sum(. == 0, na.rm = TRUE))) |>
sum()
print(paste("The total number of 0:", total_zeros))## [1] "The total number of 0: 5726"That is a lot of 0. If I treat them as missing data,
this dataset will become unusable. But, it is not like I can label them
as meaningful values either. Therefore, I will only select rows with at
least one non-zero entry to get some meaningful information.
# Filtering data to select rows with at least one non-zero data entry.
storm_data_clean2 <- storm_data_clean |>
filter(rowSums(across(where(is.numeric)) != 0) > 0)
# View the cleaned data
glimpse(storm_data_clean2)## Rows: 123
## Columns: 8
## $ STATE_ABBR <chr> "CO", "CO", "CO", "CO", "CO", "CO", "CO", "CO", "C…
## $ EVENT_TYPE <chr> "FLOOD", "FLOOD", "FLASH FLOOD", "FLASH FLOOD", "F…
## $ DEATHS_DIRECT <dbl> 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0,…
## $ DEATHS_INDIRECT <dbl> 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,…
## $ INJURIES_DIRECT <dbl> 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 90, 0, 0, 0…
## $ INJURIES_INDIRECT <dbl> 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,…
## $ DAMAGE_PROPERTY_NUM <dbl> 5000, 0, 5000, 5000, 50000, 50000, 50000, 50000, 5…
## $ DAMAGE_CROPS_NUM <dbl> 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 25000, 0, 0, 0, 0…3 Result and Discussion
3.1 Assessing the Impact of Severe Events on Population Health
INJURIES (Direct and Indirect) and DEATHS (Direct and Indirect) are the primary variables in the storm event data, which indicate the state of population health during or post-severe event.
I will explore this data in three cases. To streamline the analysis
and save time, I will create a function, health_impact,
that can be used with different arguments, eliminating the need to
rewrite the same code multiple times.
health_impact <- function(...){
storm_data_clean2 |>
# Group the data according to the argument passed to it
group_by(...) |>
# Calculate total fatalities and total injuries
summarize(total_fatalities = sum(DEATHS_DIRECT) + sum(DEATHS_INDIRECT),
total_injuries = sum(INJURIES_DIRECT) + sum(INJURIES_INDIRECT)) |>
mutate(total_health_impact = total_fatalities + total_injuries) |>
# Arranging the values in descending order
arrange(desc(total_health_impact))
}The ... in the function(...) means I can
pass any argument to it.
First, Let’s check the health impact in each state, which can be
examined by grouping the data by STATE_ABBR.
# Total health impact due to all severe events in each state
impact1 <- health_impact(STATE_ABBR)
head(impact1)## # A tibble: 4 × 4
## STATE_ABBR total_fatalities total_injuries total_health_impact
## <chr> <dbl> <dbl> <dbl>
## 1 CO 4 95 99
## 2 FL 10 4 14
## 3 ID 0 9 9
## 4 NV 5 0 5If we want to estimate the health impact of each of the severe
events, we just need to change the grouping method to
EVENT_TYPE.
# Total health impact due to each severe event from corresponding states
impact2 <- health_impact(EVENT_TYPE)
head(impact2)## # A tibble: 6 × 4
## EVENT_TYPE total_fatalities total_injuries total_health_impact
## <chr> <dbl> <dbl> <dbl>
## 1 HAIL 0 90 90
## 2 RIP CURRENT 10 3 13
## 3 THUNDERSTORM WIND 0 9 9
## 4 FLOOD 3 5 8
## 5 HEAT 5 0 5
## 6 FLASH FLOOD 1 0 1Whoa! That’s a lot of fatalities due to the Hail event. Most of them are just injuries, though. Regardless, it was the most devastating event in our data.
Let’s visualize it.
ggplot(impact2[1:10, ],
aes(x = reorder(EVENT_TYPE, total_health_impact),
y = total_health_impact,
fill = total_health_impact)) +
geom_bar(stat = "identity", color = "black") +
coord_flip() +
scale_fill_viridis_c(option = "viridis", name = "Mortality") +
labs(title = "Top 10 Most Harmful Events to Population Health",
x = "Event Type",
y = "Total Health Impact (Fatalities + Injuries)")
Now, this is more visually appealing than just looking at some
numbers on the screen. There is another way to examine the impact of
severe events: showing the estimated impact with respect to both the
State (STATE_ABBR) and Severe events
(EVENT_TYPE). However, I will not go deep into it. I will
show you how to perform estimations. Since I have written a function, it
is very simple now. Remember, I defined the health_impact
function using ... to let it accept any arguments. Now, I
can pass the STATE_ABBR, EVENT_TYPE argument together.
# Total health impact in each state due to a particular severe event
impact3 <- health_impact(STATE_ABBR, EVENT_TYPE)## `summarise()` has grouped output by 'STATE_ABBR'. You can override using the
## `.groups` argument.head(impact3)## # A tibble: 6 × 5
## # Groups: STATE_ABBR [4]
## STATE_ABBR EVENT_TYPE total_fatalities total_injuries total_health_impact
## <chr> <chr> <dbl> <dbl> <dbl>
## 1 CO HAIL 0 90 90
## 2 FL RIP CURRENT 10 3 13
## 3 ID THUNDERSTORM W… 0 9 9
## 4 CO FLOOD 3 5 8
## 5 NV HEAT 5 0 5
## 6 CO FLASH FLOOD 1 0 1See! It’s very easy! Using defined functions saves a lot of time. This is a part of Functional Programming, about which I will write another article later.
The analysis so far reveals that certain types of severe weather events are particularly detrimental to public health, causing a significant number of fatalities and injuries. The data shows that:
First, considering the five U.S. states examined—Colorado, Florida, Idaho, Nevada, and Utah—Colorado emerges as the state with the highest total impact, recording 99 cases. This was largely due to injuries, whereas Florida, despite having a lower total impact of 14, experienced the highest number of fatalities, with 10 deaths. On the other hand, Nevada and Idaho had relatively lower impacts, with Nevada reporting 5 cases, primarily fatalities, and Idaho recording 9 cases, mostly injuries. Utah, notably, did not report any health impacts during the period analyzed. It’s important to note that this could either imply that severe weather events may not be severe enough to lead to mortality or there were no records for our study period.
From the perspective of different weather event types, hailstorms were responsible for the most significant health impacts, resulting in 90 injuries, though no fatalities were associated with these events. Meanwhile, heat events and flash floods had comparatively minor impacts, resulting in 5 and 1 total cases each, while thunderstorm winds and floods contributed to moderate impacts, with 9 and 8 total cases, respectively. Rip currents, however, caused 10 fatalities and 3 injuries, making them particularly deadly. A rip current is a strong, localized, and narrow flow of water that moves directly away from the shore [6]. It pulls anything in its path out to sea, and attempting to swim against it is often futile, resulting in fatality. It is similar to trying to go against the overwhelming force of a truck that has directly hit you.
Now that the health impacts of severe events are examined, it is time to explore another sector where they can significantly impact: the Economy. We will delve into the damages caused by these events and summarize the overall impact on the Economies of the four states.
3.2 Assessing the Impact of Severe Events on the Economy
From the Storm events database, there are two variables that
represent the damages in USD ($): Property Damage
(DAMAGE_PROPERTY_NUM) and Crop Damage
(DAMAGE_CROPS_NUM). Since, data processing functions are
the same, except the variables, as in the case of
health_impact, I will write a function similar to
health_impact().
Writing a function eco_impact().
economic_imp <- function(...){
storm_data_clean2 |>
group_by(...) |>
summarize(total_prop_damage = sum(DAMAGE_PROPERTY_NUM, na.rm = TRUE),
total_crop_damage = sum(DAMAGE_CROPS_NUM, na.rm = TRUE),
total_damage = (total_prop_damage + total_crop_damage)
) |>
arrange(desc(total_damage))
}Similar to the analysis for public health, I will follow steps to analyze Damage data.
First, let’s check the damages in each state by grouping the data by
STATE_ABBR.
# Total economic impact due to all events in each state
eco_impact1 <- economic_imp(STATE_ABBR)
head(eco_impact1)## # A tibble: 4 × 4
## STATE_ABBR total_prop_damage total_crop_damage total_damage
## <chr> <dbl> <dbl> <dbl>
## 1 ID 14826000 1000000 15826000
## 2 CO 5924000 45000 5969000
## 3 FL 486950 0 486950
## 4 NV 10000 0 10000# Total economic impact due to each severe event from corresponding states
eco_impact2 <- economic_imp(EVENT_TYPE)
head(eco_impact2)## # A tibble: 6 × 4
## EVENT_TYPE total_prop_damage total_crop_damage total_damage
## <chr> <dbl> <dbl> <dbl>
## 1 WILDFIRE 14450000 0 14450000
## 2 FLOOD 3585000 20000 3605000
## 3 FLASH FLOOD 2500000 1025000 3525000
## 4 THUNDERSTORM WIND 549950 0 549950
## 5 TORNADO 100000 0 100000
## 6 HAIL 52000 0 52000# Total economic impact in each state due to a particular severe event
eco_impact3 <- economic_imp(STATE_ABBR, EVENT_TYPE)## `summarise()` has grouped output by 'STATE_ABBR'. You can override using the
## `.groups` argument.head(eco_impact3)## # A tibble: 6 × 5
## # Groups: STATE_ABBR [3]
## STATE_ABBR EVENT_TYPE total_prop_damage total_crop_damage total_damage
## <chr> <chr> <dbl> <dbl> <dbl>
## 1 ID WILDFIRE 9200000 0 9200000
## 2 CO WILDFIRE 5250000 0 5250000
## 3 ID FLOOD 3420000 0 3420000
## 4 ID FLASH FLOOD 2000000 1000000 3000000
## 5 CO FLASH FLOOD 500000 25000 525000
## 6 FL THUNDERSTORM WIND 391950 0 391950While we can visualize all three of the cases, let’s direct our
attention to the visual representation of eco_impact2, the
total economic impact due to each severe event from corresponding
states.
suppressMessages(
ggplot(eco_impact2[1:10, ],
aes(x = reorder(EVENT_TYPE, total_damage),
y = total_damage/1000,
fill = total_damage/1000)) +
geom_bar(stat = "identity", color = "black") +
coord_flip() +
scale_fill_viridis_c(option = "magma", name = "Damage (in Thousand USD)") +
geom_bar(stat = "identity") +
coord_flip() +
labs(title = "Top 10 Events with Greatest Economic Consequences",
x = "Severe Event",
y = "Total Economic Impact (Property and Crop Damage)")
)
Just like the Wildfires, ggplot is on Fire!
The data reveals a comprehensive view of damages caused by severe weather events across different U.S. states and event types. From a state-level perspective, Idaho suffered the highest total damages, with approximately $15.83 million in losses, driven largely by property damage from wildfires. Colorado experienced total damage of around $5.97 million, mostly from wildfires and flash floods, while Florida and Nevada had much smaller impacts, with Florida reporting $486,950 and Nevada a mere $10,000 in damages.
From the perspective of event types, wildfires were the most destructive, causing a total of $14.45 million in damages, all of which were property-related. Floods and flash floods also led to significant losses, with damages totaling $3.61 million and $3.53 million, respectively, though flash floods incurred higher crop damage than regular floods. Thunderstorm winds and tornadoes caused relatively lower damages, with thunderstorms accounting for $549,950 in property damage and tornadoes for $100,000. Once again, Utah reported no damages.
4 Summary
When analyzing the impact of severe events across states according to event types, Colorado stands out as disproportionately affected, particularly by hailstorms, which caused all 90 of the state’s reported injuries. In addition, Colorado’s health burden was exacerbated by floods and flash floods. Similarly, Idaho experienced significant impacts from thunderstorm winds, while Nevada’s fatalities were primarily due to heat. On the other hand, Florida saw all its fatalities linked to rip currents, underscoring its vulnerability to coastal hazards.
In terms of damages, Idaho incurred the highest financial losses, with $9.2 million from wildfires, followed by $3.42 million from floods and $3 million from flash floods, the latter heavily affecting crops. Colorado also suffered significant wildfire-related damages, amounting to $5.25 million, alongside $525,000 in flash flood damages. Florida’s losses were mostly attributed to thunderstorm winds. This data highlights the uneven distribution of weather impacts, particularly from wildfires, and emphasizes state-specific vulnerabilities, such as Idaho’s susceptibility to multiple severe event types.
These findings underscore the need to focus public health resources and emergency response efforts on regions and times of the year most prone to severe weather events.
5 Reference
- NOAA National Severe Storms Laboratory. Severe weather 101. https://www.nssl.noaa.gov/education/svrwx101/
- Berman, N. (2023, August 7). The weather of summer 2023 was the most extreme yet. Council on Foreign Relations. https://www.cfr.org/article/weather-summer-2023-was-most-extreme-yet
- Center for Climate and Energy Solutions. (2024, January 29). Extreme Weather and Climate Change - Center for Climate and Energy Solutions. https://www.c2es.org/content/extreme-weather-and-climate-change/
- Best States of the US Economy sector. https://www.usnews.com/news/best-states/rankings/economy
- NOAA Storm Events Database - https://www.ncdc.noaa.gov/stormevents/
- NOAA’s National Weather Service. RIP Current Science. https://www.weather.gov/safety/ripcurrent-science
Note: This project is written in R Markdown and
kneaded into a document using knitr.