Timber production and dynamic forestry

[abhimishr]

Purpose

This PR intends to bring the timber production functionality in the model via the production of woody biomass from timber plantations and natural vegetation. This PR also brings dynamic forestry sector in the model. More info in the changelog.

Comparison runs

MAgPIE - Baseline : PR187_04_mishra_SSP2-BAU
MAgPIE - Policy : PR187_04_mishra_SSP2-POL
Forestry - Baseline : PR187_04_SSP2_____xDist_Baseline
Forestry - Policy : PR187_04_SSP2_____xDist_Policy

These runs can be loaded using shinyresults::appResults in an R environment and then filtered by results_folder.

Run location : /p/projects/magpie/users/mishra/projects/abhimishr/f_timber_develop/output

Runtime and performance loss

[abhimishr]

@tscheypidi I addressed all the requested changes. Can you have a look at this PR again?

[flohump]

Closely related to it: There is a calculation pm_interest_dev(t_historical,i) = pm_interest_dev("y1995",i) * 0.9; in the forestry module. This calculation needs to be moved to the interest rate module

Done.

Why do we have the factor 0.9 here?

Demand and production for wood and woodfuel should NOT be 0 if the corresponding module is switched off. (and the additional interface caused by the demand=production=0 issue should be removed)

Done. The off realization can now calculate timber demand just like the biomass_mar20 realization and the production is made by paying 0.1 USD for every unit of timber produced. This is similar to using a free variable but with a nominal cost. This helps in fm_trade_balanceflow becoming f21trade_balanceflow again. Code check is successful.

This I don't understand. The question goes back to @tscheypidi . Why should demand and production for wood and woodfuel should NOT be 0 if the timber module is switched off?
The way @abhimishr implemented it now is prone to misinterpretatation because independent from timber on/off there is timber demand and production. This will also be reflected in our reporting variables, without telling us if it's based on a free variable, and therefore meaningless, or based on endogneous land interactions with the Ag sector.

[abhimishr]

Why do we have the factor 0.9 here?

Historical values at some points were reporting a zero value. This was causing a division by zero error. As this was also used in rotation length calculation the historical values are kept at 90% of 1995 values.
Looks like this was unnecessary. I made the changes in following commit.

This I don't understand. The question goes back to @tscheypidi . Why should demand and production for wood and woodfuel should NOT be 0 if the timber module is switched off?
The way @abhimishr implemented it now is prone to misinterpretatation because independent from timber on/off there is timber demand and production. This will also be reflected in our reporting variables, without telling us if it's based on a free variable, and therefore meaningless, or based on endogneous land interactions with the Ag sector.

We also need a consensus on this because going back and forth on zero or non zero demand makes us doing additional tests to confirm no weird dynamics appear as a result of such changes. For now I have pushed the changes according to @tscheypidi 's request.

[tscheypidi]

@abhimishr @flohump We need to discuss the recent changes of the timber/off realization, including the newly introduced interfaces vm_land_fore and pm_land_start_ac. I believe that some further modifications might be necessary here.

When this is solved we are ready to go from my side.

[abhimishr]

@abhimishr @flohump We need to discuss the recent changes of the timber/off realization, including the newly introduced interfaces vm_land_fore and pm_land_start_ac. I believe that some further modifications might be necessary here. When this is solved we are ready to go from my side.

Thanks Jan. We discuss this over other channels and try to reach a consensus.

[abhimishr]

Dynamic forestry in now part of the MAgPIE develop branch :)