Forecasting Earthquakes
This data is from the National Earthquake Information Center NEIC which is the US archive for earthquake information. This forecast solely involves data from 1981 and 2016. This closereading is meant for the general public
Section 1:Forecasting Earthquakes
It is commonly known that sismic activity is unpredictable and there is not enough current evidence to establish that earthquakes happen at a specific place, climate or time. We currently only predict their long-term occurence and the general area of sismic activity leaving us without much preparation for the eventually of a dangerous earthquake.
I will be attempting to forecast the magnitude and prevalence of these earthquakes in the US with data from the NEIC by modeling the shape and direction of the data.
Mortality: Earthquakes kill around 360 people every year and injure thousands.
Section 2: Earthquakes Across Time
According to the data, Earthquakes have been occurring pretty regularly since 1981, with a relatively stable monthly magnitude on average.
However, there seems to be no common knowledge and limited literature on the predictable trends or seasonal patterns of earthquakes. Such knowledge would help us take proactive measures to maximise safety and potentially decrease the US earthquake mortality int he future.
Section 3: “Earthquake prevalence: Total monthly US Earthquakes from 1981 to 2016
The monthly occurence of US earthquakes is pretty stable as well, except for some irregularly large spikes in monthly earthquake count with:
83 in November of the year 2000, 110 earthquakes in December 2004 and 226 earthquakes in March 2011.
These peaks seem to be getting taller across the years so perhaps it isworth gathering more data to support or refute this overall trend.
These three prevalent sysmic magnitudes correspond with Caltech’s data for November of 2000, the 2004 Yellowstone Earthquake Swarm towards the end of end of March and start of April, as well as the great Tohoku earthquake that took place in Japan on 2011. This large-scale earthquake greatly impacted US sismic activity in regions like California and Oregon.
Section 4: Is there such a thing as Earthquake season?
While we do not currently have evidence to establish that earthquakes happen in any specific place, time or conditions, recent research is exploring some potentially significant inrcreases in the magnitude and prevalence of earthquakes in areas of the Northern hemisphere like the US.
Our US data supports this notion with seasonal patterns shown in the subseries plot below, dsiplaying higher earthquake magnitudes in Spring months like April and May, as well as Autum months like October and November.
Section 5: Forecast
Statistical analysis,we were able to effectively use our current data as a way to capture the possible “earthquake season.” These monthly patterns help us create a model that would estimate -with some certainty- the average monthly earthquake magnitudes in the US over the next four years.
Our forecasts are statistically appropriate and tested for relative accuracy. The implications of building upon this data and reinforcing the our current research on findings could potentially lead to the foundations of monthly earthquake forecasts that rely on the seasonal patterns of earthquakes. With future data, more solid forecasts can be produced out of this prospective geological season.
Sources:
We can only predict where earthquakes are most likely to occur
We can predict earthquake long term occurrence
Earthquake Commentary for November, 2000
Yellowstone Earthquake Swarm 2004
The great Tohoku Earthquake impacted US seismic activity