I have extracted the TEPCO radiation measurement data as of 2011-03-27 and produced a linear graph and a logarithmic graph. The latest graph is a time series broken down by measuring point. I have grouped measurements taken at a point and in the vincinity of a point together.
As of the morning of the 20th, we can see the radiation at the main operations building decreasing and the radiation at the west gate generally holding steady.
We can see 4 major radiation spikes. The first which started a bit after 22:00 on the 14th was the smallest. The second was the largest, peaking around 10:00 on the 15th. The third is centered around 00:00 on the 16th. Finally, the fourth peaked a bit after 12:00 on the same day. Two of the peaks exceeded 10,000 uSv/hr.
It is not clear why there seems to be a twelve hour period to the spikes.
The raw data is provided courtesy of TEPCO. It was extracted by custom software from a series of PDFs.
Close inspection of the PDF data strongly suggests that TEPCO is transcribing figures by hand. There are several errors that a machine is unlikely to make:
- Mixing up of AM and PM on some time stamps
- Wide and narrow digit characters mixed within a number.
It is also clear that different pages of some PDFs have been cut and paste using different methods. Finally some PDFs show ascending time, and some descending.
The PDFs have been marked with DRM that prevents some tools from extracting the data. What purpose does this serve? Why is the data not provided in an easy to analyse format?
It is left as an exercise to the reader to determine if the quality of TEPCO's data is reliable, or indicative of the sort of attention to detail that would be required, say, to operate a nuclear reactor.