Its very important to note that my answer to your question is dependent on using DDM in CMAQv5.2. The interface will be substantially changed in the upcoming release, because CMAQ emissions treatment significantly changed and the following approaches will not work!
For the EMIS term, using the keyword “TOTA” will use all of the emissions. You can specify a subset through using environment variables of the emissions streams. EMIS_1 is probably not the correct environment variable names. I don’t remember the release scripts exactly, but you it should be something like “GRIDDED_EMIS” for the gridded file and something like “PT_EGU” for the point source EGUs (for example). You also need to specify the type of file you are using (this may the problem you ran into).
Here is an example of a sensitivity definition I ran a while ago (again, this will not work in 5.3+):
This example uses 1 (one) gridded file labeled “USANTHRO” and 5 (five) pairs of stack group/stack emissions files. Species NO2 and NO are pulled out of each of these 6 emissions inputs for sensitivity.
The “AMOUNT” can not be used to cut emissions by percentages. It is an actual emissions rate amount. This option is basically never used and there will be much easier ways to accomplish this functionality in the next release, so its probably going away anyways.
The “LAYER” term is also similarly never used, so it will not be upgraded. There will be easier ways to handle these things through the new CMAQ emissions streams upgrade.
I don’t exactly understand what you are trying to accomplish with your first question in the second set. You can, however, calculate sensitivities to emissions not used in the base CMAQ simulations. So, you can specify CMAQ to use FILE_C to calculate concentrations as it normally does. At the same time, you can calculate sensitivity to species in FILE_B. You just need to make sure you define everything in your runscript.