Kurt:
Sorry for the delay, but we wanted to try to go back as far as possible in the chain of publication for the book and see where we think things went wrong. I apologize in advance to you and others reading this thread since I will go into lots of detail - for the sake of making sure we can all learn from it, and then we can discuss how to "fix" this in the errata for the book on our website.
In short, I think the most important point to make is that the Table 14.1 suffers from a bad job of categorizing flows of energy. The table makes it appear that energy is being double counted, and even just "appearing" that way double counts it even though that wasnt the intent I think. The example in Chapter 5 is "more correct" in classifying things and avoids this confusion.
Note that the book is actually a re-packaging of various papers our research group has written on EIO-LCA, so I went all the way back to the original paper, published by Ochoa et al in
ASCE Journal of Infrastructure Systems (December 2002, pp. 132-138), which was then updated for the book. Unfortunately we cannot post attachments of these files in our forums (yet) and so I cannot post them here - I have emailed them to you Kurt to see them (others reading this who are curious can e-mail a request for them).
In the original JIS article, the equivalent of Table 14.1 Kurt mentions (I will call it the BOOK TABLE) is Table 2 (JIS TABLE). Also, the JIS paper used the 1997 annual IO model, not the 1997 benchmark which is currently used on the website (the benchmark wasnt available yet) - thus some numbers are different when running through EIO-LCA.
(For those of you willing to look, this old 1997 annual model is buried in the website at
www.eiolca.net/cgi-bin/multimatrix/advindexgerman.pl - it is the 480 sector model for 1997 you can choose to run on that page - but it is old and we no longer really support it)
The most important thing to do is see what the effects of the $89 billion of electricity purchases (1.1 billion kWh) were in 1997. Running that through the old annual 1997 model, you get a TOTAL TJ of energy of 8420000 TJ - found in the same way the new 1997 model does it (adding up primary energy input sources). I note this because running the 1997 benchmark model gives a similar but different number.
The JIS TABLE list the "Use Phase" energy from "Electric services-utilities" as 12 million TJ. If I look at the spreadsheet used to calculate this data point, it adds the 8.4 million TJ from the 1997 Annual model with the 1.1 billion kWh (converted to 3.8 million TJ). The JIS TABLE also then separately lists the 1.1 billion kWh as electricity used (actually as 1.1 million MkWh).
In short, that is where the problem lies (the data point for use phase energy use in the EIO-LCA book - 13 million TJ - looks to have been updated with the 1997 benchmark model results but using the same wrong spreadsheet method of adding the two*). We certainly should not be including the 1.1 billion kWh in both the "electricity" and the "energy summaries" (even if the intent was to separate out electricity from energy, because that is confusing), and more importantly even if we did classify them separately, we should not be adding in the energy value of the 1.1 billion kWh, converted to TJ, in the Energy category, because we're certainly then double counting both the energy needed to make the electricity AND the electricity produced.
I hope this clarifies things. We'll start working on an errata for the text and Table in Chapter 14, and distill this discussion into an errata for the Chapter. I hope you'd be willing to help us double check that errata when we're done Kurt?
* Using the 1997 benchmark EIO-LCA model, producing $89B of electricity uses 10,800,000 TJ of energy across the supply chain