Introduction

METEORITES!

Our dataset included information about ~45,000 recorded meteorites such as:

-the “subclass” of meteorite -the mass(g) -whether it was witnessed as a meteor or discovered after it fell -year it was witnessed/discovered -name of the meteorite -the degree of erosion -latitude and longitude of location

Researching and taking time to understand our dataset was essential in guiding our research questions and analysis. In reading the description of the dataset and doing some other online research about meteorites, we came across the following facts:

  1. The categorization of meteorites consists of three large classes(stony, stony-iron, iron), with many subtypes in each. -> This meant that instead of trying to decipher the 400+ subclasses listed in the dataset, we could write some code to put them all in broader categories, which would simplify the data a lot.

  2. “Iron” class meteorites tend to withstand the friction from falling through the Earth’s atmosphere better than the other classes. ->“Iron” meteorites possibly withstand different atmospheric conditions better than the other classes.

  3. Meteorites should theoretically fall on different places of the Earth with equal probability. -> Can we find circumstances where this isn’t true?

  4. A disproportionate number of meteorites are found in Antarctica due to better “environmental conditions” (no heat, no humidity, they stand out against snow). -> Is this true for all “kinds” of meteorites?

This lead to the following research questions:

  1. How do meteorite classes vary in terms of mass, and what types of classes are observed the most?

  2. Can the density of meteorites merely be explained by population density? That is, are meteorites only being recorded in areas with more people there to see/find them?

  3. What other factors can lead to frequency of sightings, such as those in the atmosphere itself?

  4. Do meteorites from the moon or from Mars (yes, we get those!) tend to fall in unique patterns and places on the Earth? How does the mass of the meteorite correlate?

Setup

Research Question #1: How do meteorite classes vary in terms of mass, and what types of classes are observed the most?

First, we wanted to understand trends regarding the three major meteorite classes by examining changes in mass and discovery type.

##        class  min     Q1 median       mean      Q3     max
## 1       iron 0.60 1288.0  10400 523531.731 44250.0 6.0e+07
## 2      stony 0.00    6.4     27   1844.293   160.0 4.0e+06
## 3 stony.iron 0.17   23.5    245  78636.408  4015.6 4.3e+06

The above graph (facetted by discovery type) shows that stony meteorites on average weigh the least, iron meteorites weigh the most, and stony-iron meteorites between the two. We also show summary statistics of the original mass variables. The iron meteorites has outliers an order of magnitude higher than stony or stony-iron meteorites. This is in line with the composition of the meteorites: iron meteorites are primarily iron (7.873 g/cm^3) and nickel (8.908 g/cm^3), stony meteorites are primarily made of silicate (made of silicon and oxygen) minerals (2.515 g/cm^3), while stony-iron meteorites are made of equal amounts of both.

This next plot shows the discovery type for each class of meteorite. The most surprising thing is that there is an above average amount of fell (observed falls) for iron meteorites. This result ties into the observation from the first graph that iron meteorites are the largest, which may be easier to see falling through the atmosphere. Moving back to the first graph and checking discovery will facet the boxplot by discovery type. We can see that meteorites with observed falls have around the same high values for log mass. However, meteorites that did not have observed falls had greater variance of masses per class. Stony and stony-iron meteorites that were found tended to have lower masses than those that were observed falling.

Research Question #2: Can the density of meteorites merely be explained by population density? That is, are meteorites only being recorded in areas with more people there to see/find them?

We wanted to find out what factors led to the frequency of meteorite, and one of the factors we believed was correlated to meteorite sightings was the population being there to discover those meteorites. Therefore, we mapped data about the population from the US census against the longitude and latitude location of meteorites discovered in the US to determine whether or not population was a factor in meteorite discovery.

From the map above, we can see that the midwest had the highest amount of meteorite observations between 1980 and 2000 despite having a generally lower population density than the coastal areas of the US. This was the opposite of what we predicted! We thought that higher population would mean there would be more people to observe or discover meteorites. The reason we only plotted meteorites from 1980 to 2000 was because we only had population data from the 2000 census which we used to create an estimate of the population density, so it would be less accurate to compare years where the population might be more significantly different.

Research Question #3: What other factors can lead to frequency of sightings, such as those in the atmosphere itself?

According to some online research, higher air pressure can cause meteors to deteriorate more as they fall through the atmosphere (https://qz.com/1154163/meteoroids-how-the-earths-atmosphere-works-to-destroy-them/).

It also just so happens that air pressure in North Africa has increased over the past several decades due to climate change (https://www.nature.com/articles/news030317-6).

So, if we look at North Africa before aand after 1950, we may expect that

  1. Meteorites will have lower masses in the second half of the century from being torn apart by the high-pressure atmosphere, and

  2. There might be a higher proportion of “iron” class meteorites in the second half of the century, since those are more likely to withstand the air pressure conditions. Similarly, other classes might be larger in this half of the century since they had to be larger to withstand the air pressure.

Looking at meteorites from 1900-1950:

Here, we see that the majority of meteorites that fell were “stony”, with a handful being “iron”. The mass of the meteors doesn’t seem to differ much between the classes. There are several countries where no meteorites were reported.

Let’s look at the second half of the century:

Results in this chart largely reflect those in the previous chart. The majority of meteorites are still “stony”. However, the size of the points does seem to differ a bit in that more of the points look larger compared to the previous graph. This seems to support the hypothesis that meteorites would be larger in this time frame because they had to be bigger to not be blown into dust by the higher air pressure.

Here, I took a quick look at the proportions of classes. The results show that “stony” remained the highest proportion in both time periods, and that the proportion or “iron” meteorites actually decreased. Hm.

## # A tibble: 3 x 2
##   class      n.obs
##   <chr>      <int>
## 1 iron          22
## 2 stony        306
## 3 stony.iron     3
## # A tibble: 3 x 2
##   class      n.obs
##   <chr>      <int>
## 1 iron          11
## 2 stony        268
## 3 stony.iron     1

Research Question #4: Do meteorites from the moon or from Mars (yes, we get those!) tend to fall in unique patterns and places on the Earth? How does the mass of the meteorite correlate?

Looking at our dataset, we noticed that there are some meteorites that are classified as being from the moon and some that are from Mars. We were then interested in investigating whether or not lunar meteorites tended to impact with the Earth in different places than martian meteorites given their different positions in the solar system relative to the Earth. We were also interested to know whether these meteorites tended to differ in size. To that end, we first plotted the location of these lunar or martian meteorite landings on a world map with the shape of the points differing by the origin of the meteorite and the shape and color differing by the mass in grams of the meteorites.

From this graph, we notice that in general, Martian meteorites appear to be larger in mass than Lunar meteorites. The largest meteorites seem to be Martian meteorites and there seem to be many more Martian meteorites clustered in the upper ranges of mass compared to the Lunar meteorites which seem to be clustered further down in the scale. Additionally, there seems to be two regions of the world with big clusters of meteorite landings, one around North Africa and one more loosely spread out in Antarctica. In the North Africa region there seems to be mostly large Martian meteorites, while the Antarctica region seems to have more Lunar meteorites. Since this larger world map is a little difficult to read, we will focus on these two aforementioned regions in more detail below.

In this graph, we plot the same information as above, but focusing on the continent of Antarctica in greater detail. In fact, although this data set is relaively incomplete in this aspect, the majority of meteorite landings on Earth occur in Antarctica, which makes this region interesting to study. Here we can see that there seem to be 3 clusters of meteorite landings. The cluster in the bottom right seems to consist primarily of Lunar meteorites, the top right cluster seems to consist mainly of Martian meteorites and the last cluster seems to be a mixture of the two. Again, because our data is incomplete here we might predict that this clustering is more due to where people are in Antarctica rather than where meteorite landings tend to naturally cluster. Like the trend we noticed in the world map, in this graph we also notice that the Martian meteorites seem to be larger than the Lunar meteorites.

In this graph, we focus on the region of North Africa in greater detail. In this region, there seem to be many more Martian meteorite landings than Lunar meteorite landings. Again, like we noted earlier, Martian meteorites generally seem to be larger in mass than Lunar meteorites. One interesting observation from this graph is that there are many meteorite landings in what seems like the same spot in Libya which is pretty surprising considering how large the Earth is. Additionally, in this particular region, it appears that there seem to also be a lot of meteorite landings on or near the coastline.

Conclusion

Overall, some of our findings were just as we expected while others surprised us. For example, we expected meteorite classes to be different in terms of mass due to the difference in material (e.g. iron vs stone) and we found that indeed meteorites with the most iron were heaviest. Based on our research, we also found that meteorites apparently have an equal chance of falling anywhere around the globe, but the varying densities of meteorites discovered was due to the difference in people observing them. Therefore, we wanted to see if more densely populated areas had more meteorite discoveries, but our graph shows that it was the opposite which was surprising. We also wanted to see if other factors besides population could lead to the frequency of sightings, such as higher air pressure which could possibly lead to smaller meteorites due to being torn apart by high-pressure atmosphere. We found that meteorites in the 2nd half of the 20th century tended to be bigger, which may be due to the fact that meteorites had to be bigger to not be blown into dust by the higher air pressure. Lastly, we wanted to know if meteorites from the moon differed from meteorites from Mars. We found that Martian meteors are generally bigger in mass and that these meteorites seem to be more concentrated in the Africa and Antarctica regions than others.