Question about access to source-receptor matrix

[Gabriela Giusti]

Hello everyone, I am running InMAP with several emission sources in Brazil. For my project, it is important that I can identify the contribution of each source in different grid-cells, not just the final concentration in each grid. I am new to atmospheric modeling research, but I believe that if I can access the model's source-receptor matrix, I will have access to these specific contributions of each source in each grid. I tried running the InMAP code by adding the following configuration:

Source-Receptor Matrix (SR) configuration
[SR]
SR = true
OutputFile = "output/source_receptor_matrix.csv"
OutputShapefile = "output/source_receptor_matrix.shp"

However, I have not been successful in seeing the matrix. Can anyone help me understand if it is actually possible to obtain this kind of information (the specific contribution of the sources in each grid-cell of the outputs)? If it is through the matrix, could someone guide me on how to configure the code to access it? (I have already read the blog and other chats in this group about the SR matriz, but it is still unclear to me).

I really appreciate your help!

[Anil Gogebakan]

Hello Gabriela,

I believe, unfortunately, the answer to your question is that SR matrices are not yet available for Brazil. But the good news is you can still run InMAP even though SR is not available, it will take a while, however. I haven't recently checked, but Chris and his team (Thank you so much for providing this great tool, Chris!)  mentioned extending the source-receptor matrices globally. I remember reading somewhere that it would be available by summer 2024 (maybe they have already developed it, but I'm not sure), but I haven't come across global SR. In short, SR is just defined for the US map for now, please look for if it is extended to a global scale and your job would be easier. Otherwise, give some time to your computer and you would still be able to get the results without SR matrices.

Hope it helps!

Cheers,

Anil

[Gabriela Giusti]

Hello Anil,

Thank you very much for your quick response! 

I found the ISRM data for the United States and I'm trying to explore it a bit. 

I was successful in running the global InMAP with data from Brazil, but with the results I obtained, I’m not able to analyze the contribution of each specific source (I have many sources) in each grid. Do you know if there is a way to access these specific contributions through the global InMAP? Or is the only way through the SR matrix?

Best,

Gabriela

[Anil Gogebakan]

Ah, I see! I thought you had problems running InMAP,  please ignore the earlier message. What SR does is that it is pre-set matrices for US-specific sources and receptors which only increases the computational efficiency. Other than that results would be identical with and without the SR matrices in play. These SR matrices are a simulation of many InMAP runs and show the relationship of each location in the US and reduce the whole complex process down to simple matrix operations, and in that way, it provides computational efficiency.

As for your question, please don't bother with SR as it will not provide you with an answer. I think, your question is something to think for yourself in your own setting. If you are interested in the marginal changes in pollution concentration from each source you would want to feed your emissions data from different sources separately and you would understand their contribution, but this would be an exhausting process. As you currently feed all the pollution changes in different locations you get the combined effect for each grid cell. If for some valid reason, you want disaggregated results you might want to group your emission sources. For instance, EGUs altogether are causing these much changes, manufacturing altogether causing that much changes etc. In each case, you can group your sources and feed them into InMAP and save the result for each and then you could compare the results with each other and you would be able to tell something about their spatial distribution. While there should be smarter ways of doing this in your setting, SR would not be useful to find an answer to your question.

Good luck!

Anil

[Gabriela Giusti]

Anil,

Thank you very much for your response. It definitely clarified my doubt about the use of the SR matrix. I understand that in order to achieve my goal, I will indeed need to look for other paths.

Best,

Gabriela 

[Anil Gogebakan]

However, keep in mind that understanding how SR has been created might be super helpful in your case. Because you are trying to obtain basically your own "SR". I don't know the generation process of SR in-depth, but it could be helpful for you if you would like to invest time in that, it might spark some ideas for your case. But, I'd recommend keeping things as simple as possible.

I am glad it helped :)

Cheers,

Anil

[Chris Tessum]

Hi Gabriela,

The main issue is that making a SR matrix for the whole globe at high resolution would be a very large computational undertaking. What we have done so far instead is to calculate source-receptor relationships for large cities around the globe, with data that should be available here: https://www.sciencedirect.com/science/article/pii/S1352231022002990 . We also have a source receptor matrix by country and emissions sector (so not by individual grid cells) that we will be submitting for publication soon, and that will be openly available once the corresponding article is published.

Chris