ENV_ACT troubles

[Neha Jaggeshar]

Hi all,

I have a case of LU emissions, with a “no-harvest” technology that moves land into a stock and I attach emissions using env_act~. The emission commodities (CO₂, CH₄, N₂O) are defined with type = ENV.

When I include only CO₂ (env_act~CRPSINKCO2) the CO₂ flow appears in results (VAR_FOut). 
But when adding values for CH₄ and/or N₂O, the CRPLANDNOHARVEST activity is set to zero in the solution and ofc none of those emissions are reported.

I tried a workaround in an earlier test: adding dummy sink processes consuming the CH₄/N₂O made emissions show up (emissions set to COMM_OUT instead of env_act). 
In this case, that dummy approach doesn’t change anything either.

Could this be a commodity type or reporting issue...? Any suggestions on what to check (or why CO₂ works but CH₄/N₂O cause the activity to disappear) would be appreciated.


Thank you!

Neha. 

[AKanudia]

Can you post the items view of this process - such that it shows the pcg etc as well?

[Neha Jaggeshar]

Hello Amit, 
I did think that it might be related to the PCG, and added CRPLAND_Noharvest as primary comm for the process. But, it did not make any difference as far i remember. 
These are the other tables for your reference: 

[Antti-L]

You have a negative ENV_ACT(CRPSINKCO2).  That might be causing your problem. 
Try changing it to be a positive flow.  You used positive *SINK flows earlier, and I think that approach would be fine (for the final net balance, you would need to combine the emissions and sinks anyway).  Also, recall that negative emissions require allowing such explicitly (see: Negative emissions)

[Neha Jaggeshar]

Hi Antti, 

I will try a positive flow and see if there's any change in the output. 

My sink flows were always negative (Just to reiterate, when adding only the (-ve) coefficient for CRPSINKCO2, i have my activity anf corresponding -ve emissions in the results)

Also, i did define the negative emissions (as attached). Should the issue still persists in that case? 

[Antti-L]

> Should the issue still persists in that case?

No, not in that case.  So, my guess was not correct, sorry, I did not know you had it already in your model.  
But I just checked, in your earlier model your *SINK* flows were indeed positive...

[Neha Jaggeshar]

> But I just checked, in your earlier model your *SINK* flows were indeed positive...

You mean in the previous version i was working, yes. I'm sorry, i thought you meant in the current set up. 


I tried changing the sink flow to be positive as you suggested but i still observe the same behaviour: 
With only one emission (CO₂ sink, whether positive or negative), the CRPLANDNOHARVEST activity runs normally, and the emission is reported.
As soon as I add additional emissions (CH₄ and/or N₂O), the activity goes to zero and none of the emissions are reported.

It seems to occur irrespective of the sign convention used 

[Antti-L]

Yes, very puzzling. It seems I have no way to tell why it is so without seeing the full model, but you still do have the option to post the model files (*.DD, *.RUN), like before.

[Neha Jaggeshar]

I have only env_act~CRPSINKCO2 included for the process in this one.

[Antti-L]

I ran the model and it solved fine.  The objective was:
**** OBJECTIVE VALUE          384055.3097

Then I tested by putting the ENV_ACT(AGRCRPCH4), and I did get also the AGRCRPCH4 emissions reported in the results. And yes, now I also see why you obtained zero activity and zero emissions:

There was, in fact, no demand for the activity or main output of CRPLANDNOHARVEST, which is CRPLAND_Stock. I verified that by putting a zero upper bound for the activity of that process, and solving again. The model solved fine and the objective value was the same as before!
**** OBJECTIVE VALUE          384055.3097

Hence, I have now verified that the results that you obtained (zero activity, zero emissions) were indeed quite correct. The AGRCRPCH4 emissions are thus also being correctly reported.

Your problem is that because you have neither any demand for CRPLAND_Stock, nor any implied supply of CRPLAND_Noharvest, and the process CRPLANDNOHARVEST has no costs at all, your model may choose freely between either producing CRPLAND_Stock or not producing any of it, without any impact on the objective function.  Therefore, you can randomly get results where the process has some activity and emissions, or results where the activity and emissions are zero. The activity of CROLAND_Mgt is changing accordingly, producing either smaller or larger amounts of CRPLAND_Noharvest (the price of which is zero).

I hope that can explain the model behavior to you?

[Neha Jaggeshar]

Certainly, it helps a lot. 
Thank you so much, Antti!