broadleaf_hydrodecid_tropical_tree outcompetes broadleaf_evergreen_tropical_tree over S. American tropics with default parameter file

[jennykowalcz]

Just wanted to document this known issue that came up in discussion last week. With the current default FATES parameter file, the hydrodeciduous PFT outcompetes the evergreen over the S. American tropics. I know people are working on calibrating the default parameter file, so this issue is sure to be fixed, but one easy fix in the meantime is to set fates_leaf_slatop for the hydrodeciduous PFT equal to the value for the evergreen PFT (i.e., lower from 0.03 to 0.012). This allows the evergreen to outcompete the hydrodeciduous.

[rosiealice]

Interesting. In my simulations the DDTs are all very sick and the BETs are very healthy. Would be good to find a halfway house! I will maybe go back and look at the defaults to see where I went wrong...

[mpaiao]

Indeed a halfway is needed @rosiealice! Though I would expect the coexistence between evergreens and deciduous to be heavily skewed towards one or the other depending on climate.

I've been mostly absent from FATES this semester as I am adjusting to the new job, but hope to get back to it soon. One trade-off that I started to explore before all the transitions was that hydro-deciduous should have higher Vcmax but shallower roots—if plants are going to lose leaves during the dry season anyway, they do not need to access deep soil water. But that triggered hydraulic failure mortality in hydro-deciduous, which led me to make all those pull requests linked to #1385.

The other trade-off that I have not tested yet, but has some observational support, is nutrients. Being hydro-deciduousness implies high maintenance costs. So if the environment is nutrient-limited, deciduousness may not be a very good strategy, unless the climate is semi-arid or arid.

[jennykowalcz]

That's interesting @rosiealice and @mpaiao! I note that having SLA approximately equal for these two PFTs is supported by your trait data analysis, @mpaiao. But clearly other parameter changes are needed too, because you can see in the maps I shared above, on the right with the edited SLA for DDT, the gradient of BET/DDT coexistence goes the opposite way we would expect with the rainfall gradient -- lower fraction of BET in the wettest parts of the region. 🤔