We have some hard-coded, but verbosely defined, parameters in the disturbance code. These parameters specify where the litter ends up when fire or treefall generates a disturbance event.
They are suffixed with "_localization". Here are where we define the parameters:
For instance, when a tree falls and creates a gap, does the litter all end up in the gap (newly disturbed and created land, ie new patch), or does it get retained in the old patch (non-gap), or somewhere in between... Cause trees fall to the side, the don't just crumple down exactly under where their crown used to be. Current assumption is that the litter is retained in the original donor patch (ie outside the new gap).
For fire, this parameter had been leveraged to retain all the litter on the old patches (donor patches where the disturbance was calculated), so that newly disturbed areas were not concentrated with fuel. I think we called it the bonfire effect. As opposed to treefall, this does seem to differ conceptually from how tree-fall would potentially end up, as one would think (I'm not a fire expert) the bulk of the litter following a fire would reside in the newly burned area.
Either way, these parameters, and their effects, should be un-buried and discussed.
Note that we also have one for moving existing litter from old to new patches, but it seems weird to change that value, as existing litter should really just move with the footprint it sits on.
As far as I know, some people have tested these parameters, and they can have some real impact on simulation results.
One potential action of this, could be to add the tree-fall, burn and perhaps landuse-relevant disturbance process localizations to the parameter file.
We have some hard-coded, but verbosely defined, parameters in the disturbance code. These parameters specify where the litter ends up when fire or treefall generates a disturbance event.
They are suffixed with "_localization". Here are where we define the parameters:
https://github.com/NGEET/fates/blob/sci.1.69.0_api.31.0.0/biogeochem/EDPatchDynamicsMod.F90#L146-L148
For instance, when a tree falls and creates a gap, does the litter all end up in the gap (newly disturbed and created land, ie new patch), or does it get retained in the old patch (non-gap), or somewhere in between... Cause trees fall to the side, the don't just crumple down exactly under where their crown used to be. Current assumption is that the litter is retained in the original donor patch (ie outside the new gap).
For fire, this parameter had been leveraged to retain all the litter on the old patches (donor patches where the disturbance was calculated), so that newly disturbed areas were not concentrated with fuel. I think we called it the bonfire effect. As opposed to treefall, this does seem to differ conceptually from how tree-fall would potentially end up, as one would think (I'm not a fire expert) the bulk of the litter following a fire would reside in the newly burned area.
Either way, these parameters, and their effects, should be un-buried and discussed.
Note that we also have one for moving existing litter from old to new patches, but it seems weird to change that value, as existing litter should really just move with the footprint it sits on.
As far as I know, some people have tested these parameters, and they can have some real impact on simulation results.
One potential action of this, could be to add the tree-fall, burn and perhaps landuse-relevant disturbance process localizations to the parameter file.