x404-pagenotfound
We chose to research what factors affect the frequency of volcanic eruptions.
The data we are analyzing comes from the tidytuesdy package, and was collected by the Smithsonian’s Global Volcanism Program. To explore the factors affecting volcanic eruptions, we used 2 data frames, one of which listed every volcano on the planet, giving its position, name, number…etc, while the other data set listed every recorded volcanic eruption. Both dataframes have far too many variables to explore, so we decided to focus on factors like position, country, crust type, and tectonic setting.
Now, the easiest visualization to spot trends by location is obviously a world map, so we will start there. To get an idea of which countries have had the most volcanic eruptions, we created a choropleth map which shows number of eruptions by country. Now, this is helpful, but it is impossible to see any trend, because we can’t see groups of eruptions within the countries. Also, as you can see from the scale, this map only gives a rough idea of where the most eruptions happen.
So to solve this, we create a clustered world map. Now the patterns a becoming a little more apparent. For example, we can now see that most of the large clusters of eruptions are circled around the pacific ocean. Our hypothesis as to why these trends exist had to do with tectonic plates.
So we layered a map of the tectonic borders on to the points, and as you can see, there is an clear pattern. Most of the eruptions took place along a tectonic border, which isn’t surprising because volcanoes are formed by the clashing of two plates. The circle of volcanoes around the pacific that we saw earlier is a result of the pacific plate, and is actually called the Pacific Ring of Fire. So just by looking at the maps, we can see that location has a big effect on the count of eruptions in a given area.
We then wanted to investigate whether the type of crust and tectonic setting influenced the number of eruptions. To clarify these variable’s meaning; the earth’s crust is split into tectonic plates which lie on top of the mantle. The type of plate is given as the crust variable and is either: -Continental (thick land plates) -Oceanic (thinner but heavier sea plates) -Intermediate -And unknown plates. The tectonic setting is the movement of the plates and mantle that causes the volcano to form. It is given as the variables: -Subduction zone (an oceanic plate moves beneath a continental plate) -Rift Zone (two plates move apart) -Intraplate (a hotter area causes increased magma at the surface)
We investigated these as when comparing the maps of distribution it was clear that certain areas of the map had higher frequencies of distinct tectonic settings and crust types in the same areas.
The initial plan was to fit linear models for frequency by crust and tectonic setting separately. However when evaluating both models, the r squared value was very low and the residuals were in distinct clusters by variable category, hence we concluded the linear models weren’t suitable for the data in hand. We decided our research question could be answered by visualization and comparisons.
When reviewing the number of eruptions on each crust type it is clear that continental crust is the crust type with the most eruptions. Comparing the number of eruptions for each tectonic setting, the subduction zone has the most eruptions. When referring back to the distribution maps this connection is shown as all of the subduction volcanoes are on land, whilst most ocean volcanoes appear to be rift volcanoes, a scientific study of why this is the case would be really interesting if future research occurred.
Plot shows the frequency of eruptions in each country, grouped by volcano type. We choose to only show the top 25 groups of country and volcano types to avoid overwhelming so this is only a sample of the data. It can be clearly seen that Japan has the highest number of eruptions, mainly from strato, caldera and complex volcanoes, closely followed by Indonesia, and the composition of the eruptions by volcano type is similar.The US also has a high number of volcanic eruptions, with eruptions almost evenly split between strato and shield volcanoes. If we look at the countries of South America; Mexico,Equador,Costa Rica and Chile, they all have a relatively similar number of eruptions, solely composed of strato volcanoes. This backs up our conclusion that tectonic plate setting, and crust type affect the number of eruptions, and also suggests that these factor also affect the volcano type. If we had more time I think it would be an interesting extension of our project to investigate the relationship between the factors which affect the number of volcano eruptions. The majority of this plot if red which indicates majority of volcanoes with high number of eruptions in our data set are stratovolcanoes. The variation in number of eruptions as primary volcano types varies, may lead us to believe that primary volcano type is a factor which affects the number of eruptions.
This can be seen more clearly in this bar plot which shows the number of eruptions grouped by primary volcano type, in particular it shows 9 primary volcano types with the top 9 number of eruptions. This very clearly shows that stratovolcanoes are responsible for the greatest number of eruptions, with nearly 6000 eruptions in the dataset being eruptions of stratovolcanoes. This is significantly more eruptions than shield and Caldera volcanoes, which have the second and third highest number of eruptions, with around 1000 eruptions recorded in the dataset. We could conclude that volcano type is a factor which affects the number of eruptions, but I believe to support this conclusion it is important to take into the consideration the number of each volcano type in the dataset.
This bar plot shows the number of volcanoes grouped by primary volcano type. It shows there are 23 different primary volcano types recorded in the dataset. Stratovolcanoes are the most abundant volcano types, therefore it makes sense that they cause the most eruptions, however there are approximately 450 stratovolcanoes, which are responsible for 6000 eruptions in the dataset. Therefore Stratovolcanoes must erupt several times within there lifetime, approximately 13 times each if we were to take a raw average. You can also see if we compare the bar plot of Number of eruptions by volcano type and the bar plot for number of volcanoes by volcano type, that the plot take a similar shape, which may be an indicator that the number of eruptions of each volcano type is proportional to the number of volcanoes of that type.
In conclusion we believe the factors of: -Location of volcano (position on tectonic plate) -Tectonic Plate Setting -Tectonic Plate Type -Crust Type -Primary Volcano Type have an influence on the number of volcano eruptions. Through our investigation we have come across many new potential relationships between these factors, which should further research be conducted with perhaps more expert knowledge on the subject could be interesting.
## ── Attaching packages ─────────────────────────────────────── tidyverse 1.3.0 ──
## ✓ ggplot2 3.3.2 ✓ purrr 0.3.4
## ✓ tibble 3.0.4 ✓ dplyr 1.0.2
## ✓ tidyr 1.1.2 ✓ stringr 1.4.0
## ✓ readr 1.3.1 ✓ forcats 0.5.0
## ── Conflicts ────────────────────────────────────────── tidyverse_conflicts() ──
## x dplyr::filter() masks stats::filter()
## x dplyr::lag() masks stats::lag()
##
## Attaching package: 'janitor'
## The following objects are masked from 'package:stats':
##
## chisq.test, fisher.test
##
## Attaching package: 'glue'
## The following object is masked from 'package:dplyr':
##
## collapse
## Parsed with column specification:
## cols(
## .default = col_character(),
## volcano_number = col_double(),
## latitude = col_double(),
## longitude = col_double(),
## elevation = col_double(),
## population_within_5_km = col_double(),
## population_within_10_km = col_double(),
## population_within_30_km = col_double(),
## population_within_100_km = col_double()
## )
## See spec(...) for full column specifications.
## Parsed with column specification:
## cols(
## volcano_number = col_double(),
## volcano_name = col_character(),
## eruption_number = col_double(),
## eruption_category = col_character(),
## area_of_activity = col_character(),
## vei = col_double(),
## start_year = col_double(),
## start_month = col_double(),
## start_day = col_double(),
## evidence_method_dating = col_character(),
## end_year = col_double(),
## end_month = col_double(),
## end_day = col_double(),
## latitude = col_double(),
## longitude = col_double()
## )
## Parsed with column specification:
## cols(
## volcano_number = col_double(),
## volcano_name = col_character(),
## eruption_number = col_double(),
## eruption_start_year = col_double(),
## event_number = col_double(),
## event_type = col_character(),
## event_remarks = col_character(),
## event_date_year = col_double(),
## event_date_month = col_double(),
## event_date_day = col_double()
## )
Our presentation can be found here.
Mock, T 2020, “Volcano Eruptions”, electronic dataset, Global Volcanism Program/Smithsonian Institution, viewed 4 December 2020, https://github.com/rfordatascience/tidytuesday/blob/master/data/2020/2020-05-12/readme.md