EIO-LCA Web Forum

underlying calculation of hybrid model
Page 1 of 1

Author:  mbendewald [ Sun Nov 16, 2008 4:01 am ]
Post subject:  underlying calculation of hybrid model

The hybrid model allows users to change the Xdirect vector, or (I+A)y. However, the calculation procedure to obtain final demand uses the matrix A, not Xdirect. I would like to be able to make a sector within a sector, essentially splitting a column in A (the direct matrix, derived from BEA's Use matrix) in two. If I were to do this, I would expect the direct requirements of the sector itself to change as well as the environmental effect vector. This does not seem to be happening when I use the hybrid model. For instance, I will delete the electricity generation sector of commercial and institutional buildings, but the energy use per dollar remains the same for that sector.

When using the hybrid model what am I changing in the formula to calculate final demand?


Author:  mbendewald [ Sat Jan 09, 2010 1:27 pm ]
Post subject:  error in Custom Model default data?

The new site looks great and the new "Interpreting Results" section is very helpful. Also, I saw the addition of a clarifying statement (titled "Curious how the Hybrid feature operates?") in the Create a Custom Model section that helps to address an earlier question of mine.

My current question is regarding a potential error in the Custom Model data.
The clarifying statement I referred to above says that "The economic numbers you're allowed to alter on the next page include the model sector's initial demand (I) and the direct suppliers for this demand (A) [hence, I+A]. "

Given that this is true, I think there may be an error in the data on the next page for at least the Construction sector, Commercial Buildings. The 1997 commodity use table from the BEA indicates that the construction industry consumes no coal. The default values in the Custom Model show a value of 1.76 for coal. Am I wrong to believe that the commodity use table represents the direct suppliers (hence, the A matrix)? If not, the Custom Model should show a value of 0 for coal.

Thanks for the clarification!

Mike Bendewald

Author:  hscottm [ Thu Jul 01, 2010 10:13 am ]
Post subject: 

Mike -

sorry for the delay but tracking down what we believe to be the underlying issue took some time.

In short, your question is well motivated and made it easy for us to work on, but also sort of distracted us. Its good to see you know the use table is effectively the source of our economic data. However the use table is on an industry by commodity basis but the model we use online is on an industry by industry basis.

So when you see that there is no purchase of coal in the use table, that means that the Commercial Buildings industry consumes no coal. But when the industry-by-industry basis matrix is made, there are some re-mappings and translations that happen such that things arent so clearly delineated. Similar sectors get crossed, etc (you could look at the make table at the detailed level to get an idea for this) and everything is put on industry terms.

In the end, what that means is that there is a mixture of the inputs then needed by the resulting industry-by-industry basis. And for one of those mixtures, there ends up being direct purchases of coal.

Hopefully this answer suffices so you realize it isnt an "error", its more of an artifact of the desire to have an industry-by-industry model used.

Author:  mbendewald [ Fri Jul 02, 2010 1:54 pm ]
Post subject: 

Thanks a lot for the explanation. I have one more question.

Since the Industry by Industry matrix is used in the EIO LCA calculator and not Industry by Commodity, is it methodologically correct to use the Hybrid tool in the following way:

I want to estimate the carbon emissions associated with $1m of commercial building construction, but I want the boundary of analysis limited to only the structure of the building. In other words, for a $1m building, what's the carbon footprint of the structure?

To calculate this, I use the Hybrid tool to delete all dollar amounts in the industries not producing structural elements for commercial building construction. Is this methodologically correct? Since its not an Industry by Commodity matrix, it does not seem so. What do you think?

I really appreciate your time!

Best, Mike

Author:  hscottm [ Wed Jul 21, 2010 10:54 pm ]
Post subject: 

Mike: Sorry, lost track of this reply as I went on vacation.

Please note that the "basis" of the IO models used (IxI, IxC, etc) is not a very important aspect of deciding whether to use it. All that these choices mean on our end is our sense of which are relatively more compatible with the data that we have. I can't think of a reason why to use/not use an IO model simply because of its basis.

Fundamentally any of these models are just estimates, whether its an IxI based estimated, or IxC, etc, is quite secondary.

Author:  mbendewald [ Sun Aug 22, 2010 7:59 pm ]
Post subject: 

Thanks. To be clear, you are saying it does not matter if IndxInd or IndxCom is used to create a custom model? That seems counter-intuitive to me.

Here's a quick explanation of why it does not make sense.

First off, here's my understanding of how to "create a custom model."
The eiolca.net recommended procedure is to change direct supply chain economic values within that sector. As I understand it, the direct supply chain values should be modified according to the actual budget to produce the custom product. (Sharrard identifies this as a "bottom up" approach and likens it to Joshi"s Model II method. According to Bilec et al (2006), Joshi Model II is basically a tiered hybrid model.)

Thus, to calculate the environmental impact of a custom building, I would go to "Create custom model" section of the eiolca tool; select the "hybrid product" option; select Commercial Building sector; and enter the budget for my building design and construction, leaving values for unknown sectors unchanged; then calculate. Alternatively, if I only wanted to the know the env. impact of, say, the building's structure and did not know the budget for the structure, I would simply enter 0 for all non-structure-related eiolca sectors (such as 335110 Electric Lamp Bulb and Part manufacturing), then calculate.

It seems counter-intuitive that I'd substitute values from my actual project budget for the IndxInd values, because the IndxInd values include coal, which could clearly never be a part of my project budget. It seems like I'm mixing values that are fundamentally different. IndxCom values seem like the appropriate data for "creating a custom model," because they reflect actual flows of commodities into the industry. But you are saying that it does not matter?

Please let me know if I'm not explaining myself clearly.


Bilec, Melissa, Robert Ries, H. Scott Mathews, Aurora Sharrard. 2006. "Example of a Hybrid Life-Cycle Assessment of Construction Processes"

Page 1 of 1 All times are UTC - 5 hours
Powered by phpBB® Forum Software © phpBB Group