Nerd without portfolio: engineer, musician, sometime sailor, winemaker, skeptic, cantankerous person.
By: skod , 9:11 PM GMT on April 03, 2014
I decided to take some time and download the historical data from the EPA Radnet database for a handful of cities in the Western US from 2011. My intention was to plot it, just to see what the initial arrival of any hypothetical airborne plumes of debris from the radiological disaster there looked like through the eyes of those monitors.
In addition to the Radnet data, I also downloaded the weather data from the nearest airport from the Weather Underground archives. I was inspired to do this after seeing a very consistent uptick in my own background observations during precipitation events: it occurred to me that some of the positive excursions people have described might be as a result of rainout events, and inclusion of the precip data might prove informative.
I settled upon the interval 1/10/2011 to 7/10/2011 to establish a baseline prior to the event. My thought was that having roughly a 3-month interval prior to the event, as well as roughly 3 months afterwards, would show any longer-term deviations or baseline shifts.
What I found was not at all what I expected. These data indicate that the actual plumes themselves were not very profound, by comparison to other phenomena occurring before and after 3/11/2011.
What is plotted here is total gamma CPM from the Radnet observations (the sum of all 9 individual gamma bands), versus total beta (where available), and a 168hr rolling average of the total gamma value.
All of that data has been normalized and recast as +-N standard deviations about the mean, to allow direct comparison of data from equipment with different sensitivities.
Underneath is a plot of an idealized precipitation rate in .01in/hr increments, capped at .1in/hr, to allow rainout events to be identified.
My first conclusion was that the Radnet data is incredibly noisy and somewhat inconsistent: its overall reliability is not inspiring. My second conclusion is that there are relatively few "dry detections" in the weeks following the disaster: many positive excursions do indeed correspond closely with rain events. And my third conclusion is that the total effect of the plume arrival is in some cases dwarfed by other events- especially some prior to the disaster.
My final conclusion is that I now know less than I knew before I began the exercise. However, it certainly appears from this data that the plume arrival is very hard to detect with any confidence. I have no doubt whatsoever that the debris was and is out there- but plotting its movements via this mechanism is extremely difficult. But someone out there might get some amusement from this exercise...
Here's the data in graphical form:
The views of the author are his/her own and do not necessarily represent the position of The Weather Company or its parent, IBM.
Comments will take a few seconds to appear.