Using negative CO2 emissions in a constraint

[clenox]

Trying to run a constraint with an upper limit on CO2 emissions.  When I include only the CO2 emissions commodities that have positive numbers for the emission values, everything runs fine.  When I add our commodity for accounting emissions from biomass, for which most have a negative value though some have a positive value (BIOECO2 in the workbook) the results are strange.  A process that has positive BIOECO2 emissions is used enough to totally zero out the negative BIOECO2 emissions.  Only a small fraction of the output of that process is used further on in the model.  I have attached the constraint that we are using and a spreadsheet of the results for BIOECO2.  Please help me to understand how to resolve this.

[Antti-L]

Recall that most variables are by default non-negative in TIMES (including VAR_COMNET and VAR_COMPRD), because in general one cannot allow negative physical flows or demand exceeding supply. Enabling negative emissions thus requires defining the associated VAR_COMNET variables with a negative lower bound, e.g. -INF (usually at least for TOTCO2, probably needed also for total GHG, and in your case apparently also BIOECO2). And, if you also refer to VAR_COMPRD of those emissions, you may need a negative lower bound for that as well. A shortcut for defining -INF (minus infinity) as the lower bound of VAR_COMNET(r,t,c,s) in each period is to set:

 COM_BNDNET(r,'0',c,'ANNUAL','N')=-1;

This is of course only one possible explanation, I apologize if you have already taken care of the VAR_COMNET bounds.

[clenox]

I "think" I tried what you recommended, but it did not work.  We got the same model results.  Please see attached constraint.  Can you tell me what I am missing?  Do we need to declare somewhere else that there can be negative emissions?  And if so, where is that done? Can you provided an example in excel form?  Thank you.

[Antti-L]

I am sorry but you understood me wrong.

Defining a constraint with a negative lower limit does not help. You need to define negative lower bounds for the variables.

VAR_COMNET has a lower bound of zero by default.  Defining a constraint does not change that bound: The total net emissions of BIOECO2 are represented by the variable VAR_COMNET(r,y,'BIOECO2','ANNUAL'), and because that variable has a lower bound of zero, the emissions cannot go negative.

Therefore, please try defining the lower bound for VAR_COMNET(r,y,'BIOECO2','ANNUAL'), as I suggested, using the parameter COM_BNDNET. And, if you happen to refer to the COMPRD as well, define COM_BNDPRD as well:

 COM_BNDNET(r,'0','BIOECO2','ANNUAL','N')=-1;
 COM_BNDPRD(r,'0','BIOECO2','ANNUAL','N')=-1;

You can define it in a TFM_INS table: Put BIOECO2 under Cset_CN, ANNUAL under Timeslice, 0 under Year, N under LimType, and -1 under Allregions.
   

[clenox]

That solved it Antti.  Thank you!