Prior to September 2010, earthquakes were a rare event in the Canterbury region. This chart shows how the region's seismic stability has changed over time.
Change the date range in the menu to see how this has since evolved.
A previously unknown and dormant earthquake fault has become active and the region remains seismically active today.
Unselect the 'Magnitude > 3.0' box to see all seismic activity, although note that those less than 3.0 are generally considered to be below the threashold for being felt by people.
Hover over the graph to see date and magnitude (note: dates are in UCT) or use the plotly controls (visible when you hover over the plot) to zoom in on more detail.
One way to comprehend the destructive force of earthquakes is to convert from magnitude to a measure of energy.
To do this we use the Gutenberg / Richter energy-magnitude formula:
Energy(in Joules) = 10^(1.5*Magnitude + 4.8)
Another perhaps useful comparison is to convert this to the 'ton of TNT'. which is generally agreed to be equivalent to 4.184e9 Joules.
The table below shows the calculated monthly amount of energy released by the earthquakes in the selected date range.
For comparison, the atomic bomb dropped on the city of Hiroshima in 1945 had a blast yeild equivalent to 1.50e04 tons of TNT
A previously unknown earthquake fault line existed. Using the earthquake data overlaid on a map of Canterbury and then adding a smoothed trendline gives possible clue to its location.
To see how well this matches the expert assessment, look at the image of the Greendale Fault as plotted by GSN Science contained in the app documentation (on the 'about' tab above.)
The short answer is 'no'.
However, there are a number of prediction methods being researched, some of which use 'seismicity patterns' as an indicator. Plotting magnitude against the frequency in a particular region over time gives the 'seismicity rate' for that area. This is illustrated in the plot below for the Canterbury area.
Changing the date range shows how Canterbury's seismicity rate has changed since 2010.
When the 'log scale' is selected, the resulting correlation is known as the Gutenberg-Richter relationship, which describes the exponential distribution of earthquake magnitudes - usually written as:
N = 10^(a - bM)
Where 'N' is the number of earthquakes with magnitude greater than or equal to 'M', 'a' and 'b' are the intercept and slope respectively, of the fitted trend line. The 'b' value is the 'seismicity rate' for a particular area. Over a long enough period, this can be thought of as the regions 'background' earthquake rate.
Uncheck the 'log scale' option to see a simple relationship of how often earthquakes of certain size happened.
For best results in both case, use magnitudes > 3.0, but for interest sake see what happens when all earthquakes are selected.
Disclaimer I am not a seismologist. This analysis is for education purposes only and should not be used for any other purposes.
New Zealand is a seismically active country, but the distribution of earthquakes is not uniform, rather they are more prevalent along the tectonic boundary separating the Australian Plate from the Pacific Plate.
Prior to 2010, few people believed the city of Christchurch in the South Island's Canterbury region was at risk of earthquake given that the most obvious manifestation of the plate boundary in the South Island is not on the Canterbury Plains, but in the western foothills of the Southern Alps.
But that all changed in September 2010 when the region was struck by a magnitude 7.1 earthquake, luckily causing no fatalities. The region was subsequently struck by another of magnitude 6.3 in February 2011, this time killing 185 people and adding to the damage from the previous event, estimated at $NZ40 billion.
A previously hidden fault line was the cause of these earthquakes and GNS Science experts have plotted its location and named it the Greendale Fault.
This analysis project explores seismic data collected by GeoNet and made publicly available and specifically asks:
Each of these questions is explored on its own tab in the main page.
Controls in the side navigation enable some interactive exploration of the dataset.
You can use the date sliders to change the start and end dates of the earthquake data. This lets you see things like how seismically stable Canterbury was before September 2010 or how long seismic activity lasted after the major earthquakes.
You can use the check boxes to select all earthquakes or only those greater than magnitude 3.0 (a commonly used boundary as humans generally don't notice earthquakes below 3.0) or to use different scales in the earthquake prediction tab.
You can also download the filtered dataset as a .csv file so you can do some analysis of your own.
Have a play.
This app is written in R and Shiny and uses the shinythemes package to change the standard UI to something a little nicer. It also uses a number of R packages as stated below.
The data for this project was sourced from GeoNet (see acknowledgements below). The data catalogue that describes the fields can be found here: Data Catalogue
Data is cleansed to remove non-earthquake data (such as quarry blasting) and to implement the data rounding suggestions located in the GeoNet data catalogue. In addition, the energy released by each earthquake is calculated and added to the dataset.
lubridate and dplyr are used in data cleansing.
Energy release is calculated using the formula:
Energy(in Joules) = 101.5*Magnitude + 4.8
As described by the Gutenberg / Richter energy-magnitude formula. (Note: The constant '4.8' is stated as '4.4' in some articles.)
Conversion of the energy to an equivalent 'tons of TNT' uses 4.184e9 Joules as being equal to one ton of TNT.
As this is an educational project, a range of shiny widgets and controls are used to demonstrate some of shiny's versatility.
Plotly is used with ggplot to make interactive plots.
ggmap is used with ggplot to overlay a Google map over a latitude/longitude plot
GeoNet is sponsored by the New Zealand Government through its agencies: Earthquake Commission (EQC), GNS Science and Land Information New Zealand (LINZ)
I acknowledge the New Zealand GeoNet project and its sponsors EQC, GNS Science and LINZ, for providing the data used in this project.