Sapwood is artificially low in default parameter sets.

[rgknox]

Up to this point, we have been generating virtually no sapwood in tropical trees. The old allometry had some issues with its units, and there is now an updated allometry formulation, but the updated allometry is essentially the same concept. Sapwood cross-sectional area scales on leaf area, then we multiply by height to get a volume, and then use density to get a mass.

When I applied a value from the literature, ie ~ 1.6 cm2/m2 (see Calvo-Alvarado et al), the forest collapsed, I assume due to sapwood maintenance respiration suddenly being a thing.

For the time being, I think we should leave an artificially low value in the parameter file, but this is something we really need to rectify. How, I'm not exactly sure, but the allometry parameters are straight forward, I think the issue lies more with how our current carbon balance and radiation scheme is parameterized.

[rgknox]

Here is what I was using for a rough estimate of leaf area to sapwood area. If I use the Carapa Guianensis and Pentaclethra Macroloba, and assume a 30m tree height, the ratio is roughly 1.61 cm2 sapwood per m2 of leaf.

https://www.researchgate.net/publication/23231611_Allometric_relationships_predicting_foliar_biomass_and_leaf_area_sapwood_area_ratio_from_tree_height_in_five_Costa_Rican_rain_forest_species

[jenniferholm]

Hi Ryan, I think this is good to bring up this issue. For what's its worth in my sensitivity tests at Manaus, I was increasing the fates_sapwood_ratio parameter from default 0.001 to 0.005 and it did make a difference in slowing down some of the out of control stem growth that was occurring. I'm not sure if 0.005 was still too low, but increasing it made a difference and helped (at least in the old allometry and allocation scheme).

[jkshuman]

Thanks for finding this Ryan. @xuchongang was sapwood part of your testing with Hydro at all?