Unraveling Tornado Trends

INFO 526 - Project 1

Author

Aes-thetics-Squad

Abstract

This data visualization project centers on unraveling patterns in tornado occurrences and their impacts across the United States. Tornadoes are violent, rotating columns of air that form severe thunderstorms capable of producing destructive winds. In this project we delve into two critical aspects: tornado magnitude patterns and the tornado related fatalities in each state . Through innovative visualizations like alley maps and horizontal bar plot, we seek to provide a clearer understanding of tornado phenomena, enabling informed decision-making and proactive disaster management.

Introduction

This tornadoes data comes from NOAA’s National Weather Service Storm Prediction Center Severe Weather Maps, Graphics, and Data Page. It is also available on the tidytuesday github page. The data has 68,693 rows and 27 columns, including numerical and categorical variables. It equips us with a treasure trove of information spanning over seven decades that is from 1950 to 2022.

To gain a deeper understanding of these tornadoes, we delve into this data containing variables such as ‘magnitude’- the magnitude of tornadoes , ‘state’ -the state in which the tornado occurred, ‘fatalities’-number of deaths for each tornado, ‘year’- the year in which the tornado occurred, ‘slat’- starting latitude, ‘slon’-starting longitude, ‘elat’- ending latitude, ‘elon’ - ending longtitude.

Question 1: Are there any patterns in the magnitude of tornadoes based on location or time of year?

Introduction

Understanding the variations in tornado intensity is vital for assessing their potential impact and devising effective disaster response plans. In this section, we explore the patterns in tornado magnitude concerning geographical location (states) and how they fluctuate across different times of the year. The variables that are necessary to answer this question are “yr”- year in which the tornado occurred, “st”- the state in which the tornado occurred, “slat”- starting latitude, “slon”- starting longitude, “mag”- magnitude of the tornado.

Analyzing whether certain regions are more prone to producing stronger tornadoes can contribute to a better understanding of tornado dynamics and the factors that influence their intensity. For instance, if there are regions where stronger tornadoes are more likely, it could lead to targeted preparedness efforts and resources being allocated to those areas. This question is therefore of great interest as it has practical implications for both disaster response and mitigation strategies, ultimately aiming to reduce the impact of tornadoes on communities and save lives.

Approach

To investigate potential patterns in the magnitude of tornadoes based on location and time of the year, we will employ three distinct types of plots: two line plots and a grouped bar plot and an animation consisting of density plot. The first line plot will depict the trends in mean tornado magnitude over the years, offering insights into potential shifts in the intensity of tornadoes gradually from 1950 to 2022. The second line plot shows the count of tornadoes over time with a trend line(using geom_smooth) to show the overall trend in the plot. The use of a line plots allows for a clear visualization of any long-term variations in tornado magnitudes.

Additionally, we utilized a grouped/stacked bar plot to illustrate the distribution of tornado magnitudes across the different 12-year intervals. In this plot for each 12 year interval it shows the tornadoes occcurences and the count of the tornadoes for each EF scale rating(from 0 to 5). This plot is suitable for comparing categorical data across multiple groups, as it provides a visual comparison of tornado occurrences within distinct magnitude categories.

By combining these two types of plots, we aim to comprehensively explore potential patterns in tornado magnitude based on both temporal and categorical factors. This approach allows for a nuanced analysis, considering both the annual trends and the broader distribution across specified time intervals.

Then lastly for the animation of contour plots, we first grouped the data state-wise and summarized the standard deviation of magnitudes for each of the US states. Then we selected top 5 states from it and used it for generating contour maps for tornadoes in the selected states (Tennessee, Alabama, Iowa, Kansas, and North Dakota) over distinct periods (1950-1961, 1962-1973, 1974-1985, 1986-1997, 1998-2009, 2010-2022). This allows us to observe how tornado density patterns have evolved over these years. We choose the contour map plot as we have spatial data and contour plots are best to showcase the density patterns.

Analysis

Discussion

The first set of visualizations, that is the line plot, is focused on tornado magnitude trends over the years. We observed fluctuations in the mean tornado magnitude per year, with peaks and valleys throughout the decades. We can see that the intensity of tornadoes decreased as we moved forward from 1950 to 2022 gradually. From the second line plot we can see that the count of tornadoes increased over time slightly. The trend line in the plot shows this increase effectively from 1950 to 2022. Based on these two line plots, it is evident that the late 1900s experienced fewer tornadoes, but of higher intensity, whereas the present day sees a greater number of tornadoes, albeit with lower intensity.

Furthermore, examining tornado magnitudes in different categories (EF scale) across distinct year categories (1950-1961, 1962-1973, 1974-1985, 1986-1997, 1998-2009, 2010-2022) through a stacked bar plot illuminated the distribution of tornado intensities over time. It interprets that as the time progressed the intensity of the tornadoes decreased, but the number of tornadoes increased.

For the contour maps, we can see the variations in the trends of tornado occurences for the states “Alabama” “Iowa”, “Kansas”,“North Dakota”, “Tennessee”. Each of the density plot represents a specific period, and as they progress, you can observe how tornado density and patterns evolve. In the first two density plots(1950-1973), tornado events were less frequent, resulting in areas of low tornado density. In the next two plots(1974-1997), tornado density increases, and the tornado patterns become more widespread, so we can observe greater concentration of tornado events in various regions, indicating a moderate increase in tornado activity. In the last two plots(1998-2022), tornado density significantly rises during these periods, with widespread and more intense tornado events. These density plots show larger and more connected areas of high tornado density, suggesting that tornado occurrences have become more frequent and impactful.

References

  1. https://www.ncei.noaa.gov/access/monitoring/monthly-report/tornadoes/201913
  2. https://www.nssl.noaa.gov/research/tornadoes/