Changes to the partitioning of biomass to coarse woody debris classes #1008
Conversation
…he burnable classes
|
Adam, if it is not too much work, could you cluster the results on slide 6 and 7 by decade, and plot them on a similar scale? Then if people adopt this change they will be able to understand by how much they could anticipate the change being. |
|
Thanks Zachary for suggesting the decadal-scale figures which can be found on slide 8 of the results here A new insight from these figures is that mean fire line intensity increases in some decades with the proposed changes and in other decades its very similar. Differences are likely due to having more combustible fuel on the landscape. (I've updated the "Test results" section on the PR above to reflect this insight). |
|
Regression testing on cheyenne is complete. All expected tests pass with multiple Folder location: |
Description:
This pull request is associated with discussion 962 which was continued in discussion 989.
The changes proposed here update the partitioning of biomass (struct + sapw) to coarse woody debris classes (cwd; aka cwd fuel classes). Fates currently sends some biomass to all fuel classes regardless of cohort dbh. For example it sends stem biomass to the 1,000 hr fuel class (logs > 7.6 cm in diameter) even when the cohort dbh is much less than 7.6 cm.
The changes proposed here update the cwd partitioning for small cohorts to be based on cohort dbh such that stem biomass is sent to the appropriate cwd class based on the Rothermel fuel class diameter thresholds (see comment from @jkshuman in discussion 962. The proportions of stem biomass sent to each flammable cwd class (twig, small branch, and large branch) is kept constant, where possible, to minimize changes in fuel properties. Partitioning for cohorts larger than 7.6 cm in dbh is not changed.
Changes are applied to all subroutines where stem biomass is partitioned into cwd classes: CWDInput (ED physiology), mortality_litter_fluxes (patch dynamics), fire_litter_fluxes (patch dynamics), logging_litter_fluxes (logging), and SendCohortToLitter (cohort dynamics).
Collaborators:
@lmkueppers @ZacharyRobbins @ckoven @rgknox @jkshuman @jenniferholm
Expectation of Answer Changes:
Changes should produce more flammable coarse woody debris (twig, small branch, and large branch) and less non-flammable cwd (1,000 hr fuel). These changes should be most apparent a) during early forest development when cohorts are < 7.6 cm in diameter and, b) for pfts that generally stay small (i.e. shrubs). Fuel properties such as SAV and bulk density are calculated as weighted averages of the fuel classes where the weights are the fraction of total fuel belonging to each class. Therefore, when the proportion of these classes change on the landscape, the aggregate fuel properties will also change, thereby changing fire behavior.
Checklist:
Test Results:
I have tested how these changes affect the accumulation of cwd, fuel properties, and fire behavior for a lower montane mixed conifer forest site in California's Sierra Nevada.
CTSM (or) E3SM (specify which) test hash-tag:
ctsm5.1.dev115 (commit#17e2acb6a)
CTSM (or) E3SM (specify which) baseline hash-tag:
ctsm5.1.dev115 (commit#17e2acb6a)
FATES baseline hash-tag:
sci.1.61.0_api.25.0.0 (commit#e663a6e6)
Test Output:
These changes result in a larger accumulation of flammable coarse woody debris classes (twig, small branch, and large branch) and less accumulation of non-flammable cwd (1,000 hr fuel) during early forest development (0-20 simulation yrs). Differences attenuate as the simulation progresses when more of the cwd is coming from larger cohorts (which are not affected by these changes). Although there is more potentially burnable fuel, the landscape is less flammable on any given day for the same FDI because SAV ratio and fuel bulk density move towards less flammable values due to a higher large branch to leaf fuel ratio on the landscape (fuel properties are weighted averages of the classes). These changes to fuel properties result in less burned area on average, and changes to fire line intensity and vary through time. Fire line intensity was the same in some decades and increased in some decades, likely due to more build up of flammable fuels (see slide 8 of results).
Results can be found here.