Parameter tuning for CLM6

[wwieder]

Here's a place to discuss calibration efforts for CLM6.
Notable 1 degree simulations are:

• _#69 CLM5.3 baseline with initial PPE informed calibration
• _#91 Above with Tom's Nfix scheme and additional parameter modifications to address regional / PFT biases
• _#100 CLM5.4 surface data & CAM requested changes for snow
• _#116 & _#118 CLM5.4 surface data & PPEcal115 parameter set

[wwieder]

Quick look at survivability and maximum annual LAI for #89 below:

For reference the same plot for #69 is here:

[linniahawkins]

Thanks for sharing these! For reference, here is a map of the PFT fractional coverage per gridcell in a 2-degree CLM-SP run with the 5.3 surface datasets (probably should remake this with a 1-degree SP). Some of the pink areas where we have low survivability correspond to areas where the PFT occupies less than 1% of the gridcell. Just something to keep in mind.

[linniahawkins]

Couple initial thoughts

• We reduced the high LAI bias in BDBS and C3 arctic Grass without killing it (yay)
• We increased C4 grass survivability (success)
• We increased C3 non-arctic grass survivability, but now the LAI is crazy high!
• NDBT and BDTT LAI look much more reasonable.
• The combination of my tuning and turning on Tom's Nfix functions killed off NEBT and BDBT in high latitudes. (oops)

[wwieder]

I find the trajectories of land C somewhat surprising, with nearly identical initial veg C pools, but greater veg C losses from #89, I'm assuming this difference in land use signal reflects a shift in vegetation C towards the tropics (and away from boreal forests) with the latest parameterization.

[wwieder]

A few thoughts on our next round of calibration (based on land only and preliminary coupled results):

• increase survival of pft2 (boreal forests)
• relax down regulation of tropical pfts (4 & 6) to increase survival in savanna regions (and maybe increase amazon LAI in coupled runs)
• maybe mess around w/ c3 grasses, tough nut here since it's going crazy in wetter regions and barely surviving in dry regions. Is there a way to flatten the precip-lai curve for this pft? Maybe try this with nutrients (instead of leafCN/SLA).