Final Project

1. Scenario

15th Jan, 2022, the Georgetown University student Lisa Chen was shocked by the news of a volcanic eruption that happened in Tonga. She was very worried about the situation there, since one of her best friends Fangu Fukofuka is from Tonga. Lisa texted to Fangu right after she saw the news in the morning. 6 hours passed, Fangu still did not text back. And the whole world had trouble connecting with Tonga as well. Lisa desperately wanted to know how devastating the effect of the volcanic eruption could be, and how possibly it can be related to other natural disasters, including earthquakes, tsunami, etc. She then opened the website “VolcaKnows”……

2. Introduction

Volcano eruptions and the natural disasters coming along with them could be catastrophic. Although it could happen at any time and anywhere around the world, some regions are more vulnerable and sensitive. Those activities can severely destroy people’s lives and livelihoods, for example, the volcanic eruption that happened in Hawaii 2018 destroyed 700 homes. Furthermore, the losses occurring might even be the major obstacles to human’s sustainable development. Our group used datasets above to tell a data science story about the eruption of volcanoes and trying to address a bunch of data science questions:

1. How volcanoes are distributed on the planet?

2. What trends could we find from years of volcanic eruption?

3. How can we display the volcanic explosivity index (VEI)?

4. What are the economic effects of the volcanic eruption?

5. How dangerous volcanic eruption can be and how it can be related to earthquakes?

6. Where are the most destructive volcanoes located?

7. What is the networking relationship between volcanic eruptions?

Therefore, we are going to analyze the distributions and activities of volcanoes, explore the VEI, visualize the influence of different geographical locations on volcano movements, the relationships between volcanic eruption and earthquakes, and its impact on economics.

3. Methods

If the reader is interested in the technical details associated with this project then please refer to the following.

4. Results

Figure 1, Volcanoes' Worldwide Distribution

It’s clear to see that most volcano activities happen along the coasts’ side and in islands with several occuring in the middle of oceans. The plot offers a lot of insights that the country of Indonesia, Canada, the USA, and Japan has the most overlapping scatters of volcano activities; this is also helpful for scientists and readers to do prediction and related analysis.

Figure 2, Volcanoes' Number in Different Era and Countries

According to this bubble plot, the volcano's active area is relatively stable. The first five most active areas are Indonesia, Japan, the United States, Russia and Chile.And according to the axis range of the scatter plot, we can also find that earth has become more and more inactive. In other words, most of the volcanoes are going to go dormant.

Figure 3.1 & Figure 3.2, The Area of Activity where Volcanic Eruptions Last Between 1-2 Years

According to the pattern shown in this Multiview plot, it can be seen that if we limit the eruption's duration days to a year or two, most volcanoes have moderate relative explosivity status. Since the upper figure presents all the volcanoes’ performance, Japan and Indonesia have more active volcanoes than other geographic areas. And volcanic eruptions are more likely to last one year to one year and a half. Volcanoes in Indonesia and Papua New Guinea have the least duration days; Volcanoes in Indonesia have the longest durations between 1-2 years. After calculating the duration days in one country as the average duration days, the performance of eruptions can be checked from the bottom figure, the majority of volcanoes still have moderate relative explosivity-status, followed by small explosiveness, non-explosive status and large explosiveness. Based on the overall volcanic eruptions, it can be observed that Japan has the shortest durations days while the longest eruptions take place in DR Congo. Besides, the duration of eruptions is evenly distributed over a period of one to two years.

Figure 4, The Volcanic Explosivity Index(VEI) Map After 1800

This plot shows the worldwide distribution of volcanic eruption with special measurement, VEI. Based on the value of longitude and latitude, we can find that active volcanoes focus on and swarm around Southeast Asia (Indonesia and Philippines), East Asia (Japan, East of Russia), the east coast of the Americas and the Mediterranean continent. And mostly are moderate explosive volcanoes.

Figure 5, Comparison Between Volcano Eruption Evidence Categories

From this density plot, we can see that uncertain and confirmed eruptions observed after year 2000 reached a higher level compared to the eruption dated before. However, discredited eruptions observed mostly after year 1960 increased high. The peaks of conformed eruption, discredited eruption and uncertain eruption are around the year 2000, 1960 and 2000.

Figure 6, Active Volcanoes with Number of Eruptions

Volcanoes studied here are all recorded volcanoes with eruptions starting from year 46 to 2019, ending from year 47 to 2020. It is clear that the Etna volcano in Italy has the most active eruptions with 104 times, mount Fournaise, Piton de la which lies in France has eruptions 70 times, followed by volcano Asosan in Japan with 66 times eruptions. And more than half of the 40 most active volcanoes have had less than 20 eruptions.

Figure 7, Volcano Eruption and Earthquake from 2010 to 2018

INNOVATIVE VIEW: This INNOVATIVE interactive plot shows the distribution of volcanic eruptions and earthquakes worldwide from Jan 2010 to April 2018. According to the plot, the red dots depict the location of earthquakes, and the blue dots pinpoint the location of volcanic eruptions. When green dots shows, it means neither earthquake nor volcanic eruption happens in that month. Apart from the animated way of showing both volcanic eruptions and earthquakes on the plot, this visualization is considered as innovative due to the complexity of processing the dataset. Refer to the method section for more details on data processing.

According to the plot, it can be seen that, in a given month, most volcano eruptions are surrounded by earthquakes. What is more, most volcano eruptions happen at places where historically had earthquakes. This observation is matched to the fact that volcanic eruptions and earthquakes are closely related geological movements.

Figure 8.1 & 8.2, Total Deaths in Each Country from 2010 to 2018

Figure 8-1: This interactive visualization shows the descending order of the number of total deaths from volcano eruptions from 2010 to 2018, with an extremely large number of total deaths (367) in Indonesia.

Figure 8-2: This interactive visualization shows the descending order of the number of total deaths from volcano eruptions from 2010 to 2018, with the removal of a large number of total deaths in Indonesia. Indonesia , Japan and Philipines have had relatively higher frequency of volcanic eruptions in the last several years. Indonesia , Japan and Eritrea have the top three highest number of total deaths among these countries.

Figure 9, Networking Plot for Volcanic Eruptions

This is a networking plot made by the support rule, which represents the popular level of the items. People could use the dropdown menu to check different traits‘ occurrence probabilities and their networking relationships under different rules. It could be concluded from a world view that volcanic eruptions usually have a short period and the eruptions are not uniform every time. And most of them are in historical status and stratovolcanoes (built up by many layers of hardened lava and tephra.) These are three of the most important traits of volcanic eruptions and they usually occur together.

5. Discussion

According to the results section, it shows that in a worldwide range, the Volcanic belt around Philippine Plate has the largest number of volcanoes. As we know, the volcanic belt can be explained as a geographical region which contains very high levels of volcanic activity. Although the total number of activities in the North and South American Plate is very large (around 200), the number of volcanoes in the Philippine Plate is almost three times as large as American plates. In contrast, the least number of active volcanoes are mostly in the Antarctic Plate of only 4 intotal.These findings can be used to research different geographical features and better serve scientists and governments to protect residents. It also could be concluded that the main area for volcanic eruptions does not change over time. All in all, the areas mentioned above would act as very important roles in volcanic activities research.

Volcanic explosivity index (VEI), commonly used in the study of volcanism, indicates a relative measure of the explosiveness of volcanic eruptions. Since the data points are numeric, which is hard for non-technical personnel to understand, here we add a column called “vei_status”, which stands for the measurement of volcanic explosiveness status, from non-explosiveness to large explosiveness. The majority of volcanoes have moderate explosiveness status, followed by small explosive eruptions. Volcanic explosivity index and status are utilized as size and color labels to generate map and scatter plots to see the activity of volcanoes.

For the time series trends of volcanic eruptions, it suggests that nowadays, the volcanic activities are not the same as the Holocene era anymore. Since magma from the Earth's mantle can no longer reach the volcanoes, they do not have enough energy to erupt. Earth has become more and more inactive and some of the volcanoes are going to go dormant. Although volcanic eruptions still have great influence on human life, the struggle between humans and volcanoes is no longer as fierce as it used to be.

Although the effects of volcanic eruption are more controllable, it still brings lots of economic and human impacts. These effects could be quantified as the number of missing people, number of injuries, damage in million dollars, number of houses destroyed and number of total deaths. These measures help to determine the dangerous level and destructivity for volcanic eruptions. One of the most directed measures of the dangerous level is the total number of deaths from each eruption event. Visualizations show that Indonesia has a higher number of total deaths than other countries from the last 10 years span, with the highest number of 367. Then, Japan ranks second on the scales. It makes sense because Indonesia and Japan have much higher frequency of volcanic eruptions than other countries; and they have a much wider duration of days range. From a duration and VEI point of view, since 1800, a large proportion of volcanic eruptions happened in Indonesia have been labeled as small or moderate explosiveness, and those that happened in Japan have been labeled as small, moderate and even large explosiveness.

What’s more, volcanic eruptions and earthquakes are closely related to geological movements. Although the visualization shows that earthquakes occur much more frequently than volcanic eruptions, both events are very likely to occur along plate boundaries. A place with more earth Earthquakes could be a trigger for a volcanic eruption. The countries we mentioned earlier, Indonesia and Japan, are located at the plate boundaries. In those places, when earthquakes occur, it is reasonable to be concerned about and prepare for volcanic eruptions. The specific relationship should be further studied.

What is the reason that confirmed and uncertain eruption observations become much more after the year 2000? Our guess is that we have more sensors out there in recent times, so we're finding more potential volcanos. Another assumption is that volcanic activities are closely related to the movement of tectonic plates, which can be influenced by environmental parameters, such as humidity and surface temperature.

6. Conclusion:

In conclusion, this project has answered most of the data science questions raised earlier. In a worldwide range, the Volcanic belt around Philippine Plate has the largest number of volcanoes. Active volcanoes swarm around Southeast Asia, East Asia, the east coast of the Americas and the Mediterranean continent and most volcanoes are at a moderate explosiveness level. Volcanic eruptions usually do not have a long period, they tend to erupt for a few days or a few weeks, but not for a few years. And volcanic eruptions are closely related to geological movements like earthquakes.

Indonesia and Japan have the most frequent volcanic activities in the last several years, and they have a higher number of total deaths than other countries. What’s more, those two countries have a wider range of duration. The duration of eruptions is evenly distributed over a period of one to two years. The number of total deaths implies that Indonesia and Japan have some of the most destructive volcanoes, though the specific destructivity should be studied based on more measurements.

Although there are still volcanic activities and more confirmed as well as uncertain eruptions have been observed after the year of 2000, earth has become more and more inactive and some of the volcanoes are going to go dormant.

Appendix: