Forest Sink - possible approach?

[Neha Jaggeshar]

Hello there, 

I have a question about how to best represent the decreasing forest sequestration capacity (ENV_ACT) as forests age. I’ve attached a visual example of the concept.
Would using the STOCK / SHAPE approach be a reasonable method to represent this—where STOCK represents existing forest and SHAPE represents new forest—similar to how we model car stock and new cars retirement?

In a previous trial, I tried using an interpolation set for ENV_ACT, but the decrease appeared too sharp. I’m looking to show a smoother decline.

Any suggestions would be helpful. 


Thank you,
Neha 

[Antti-L]

> Would using the STOCK / SHAPE approach be a reasonable method to represent this—where STOCK represents existing forest and SHAPE represents new forest—similar to how we model car stock and new cars retirement?

I have no practical experience in forest growth modeling, but can just give some technical remarks:

With SHAPE you can indeed shape several parameters according to the age of each process vintage.  With NCAP_CPX you can shape the capacity transfer coefficients, with NCAP_AFX you can shape availability factors, with FLO_FUNCX you can shape efficiencies and other process transformation parameters, such as emission factors (e.g. FLO_EMIS).  Therefore, if any of these looks suitable for your forest sink modeling, yes, you could use them.

However, note that if by STOCK you mean the VEDA Stock parameter (=PRC_RESID), that parameter defines the amount of existing capacity still remaining in a period.  The "real" age of the capacity is not known in this case by TIMES, and the vintage of it is always BOH-1. And so, if you use any SHAPE for a PRC_RESID capacity, the assumed age will aways be 1 in the beginning of the horizon.  Hence, for me it does not look like a good way of modeling the existing forest, because surely, the average age of the existing forest must be something between maybe 30–60 years, whereas by using PRC_RESID for shaping the growth, the age would thus be assumed to be 1 year in the Base year (BOH) and then aging from there. Therefore, as an alternative, I might suggest to use NCAP_PASTI instead, for defining the age structure of the existing forests, by dividing the forests into age classes (in maybe 5 or 10 year age-ranges), and then defining NCAP_PASTI for the corresponding past vintages.

[Neha Jaggeshar]

> With SHAPE you can indeed shape several parameters according to the age of each process vintage.  With NCAP_CPX you can shape the capacity transfer coefficients, with NCAP_AFX you can shape availability factors, with FLO_FUNCX you can shape efficiencies and other process transformation parameters, such as emission factors (e.g. FLO_EMIS).  Therefore, if any of these looks suitable for your forest sink modeling, yes, you could use them.


Thank you for the detailed explanation, Antti — that really helps clarify things.

Just to double check: even if I use SHAPE on FLO_FUNCX (or FLO_EMIS) to represent how forest CO₂ removals change with age, I would still need to define the age structure of existing forests using NCAP_PASTI, is that correct?

[Antti-L]

I am sorry, it is not clear to me how you would like to model the forests and their growth, and therefore it is not easy for me to say what you need to do.  But I have seen some other forest growth & sink models having the forests divided into age classes, and so that seems an often-used approach.  And one way of employing age classes in TIMES is to start with NCAP_PASTI(v) for the existing forests of vintage v. I am sure there are many other possible approaches to modeling the carbon sequestering in forests, but as you were referring to "STOCK", I just wanted to point out what the age would be assumed if you use the VEDA Stock capacity parameter.

[Neha Jaggeshar]

The method I am trying is dividing the existing forests into age classes to reflect their current age structure — mature and old forests. 
I have also defined the shape curve for the existing forests (FOR_Mature_00, FOR_OLD_00) and commodity FOR_NOH*
for_noh* has an emission/removal coefficient which is supposed to decrease with the age-based sequestration curve

[Antti-L]

In case you are interested, a few years ago one of my colleagues implemented a forest dynamics model extension into TIMES (for a single Nordic country): Energy-climate-forest modelling

Because the author used a TIMES extension, she could implement the model equations basically without using TIMES attributes, but I have no doubt they could have been implemented also by using TIMES processes, commodities and attributes. Anyway, the approach she used might give you some additional ideas for your work.