FoRTE team meeting -- 2018-10-16

[ashiklom]

Alexey's updates:

• Mostly working on setting up and debugging PEcAn on local machine and PNNL PIC node. Because it's Docker, should be easy to set up anywhere else moving forward.
• Currently in a position to do ED2 site-level runs, though will probably hit more PEcAn bugs moving forward. Currently working on doing some test runs (without thinking too much about parameters, initial conditions, etc.) at UMBS.
• Jotting down some paper ideas in this repository. Basic idea for first paper -- can ED (and other models) get the initial state right (i.e. as opposed to the disturbance response). I.e. can models reproduce the observed composition and fluxes at UMBS running from "bare ground" circa 1900.

Alexey's plans:

• For AGU, I need to do a bunch of ED runs (short 1-2 week runs at 55 sites, plus 3-5 multi-decadal runs from bare ground).
• Before spring field work, will try to have full ED2 (and probably other models) parameter sensitivity and uncertainty analysis to help guide measurement priorities. First paper goes with this.

[bpbond]

AS has spent a lot of first 6-8 weeks working on setting up PEcAn here and more general infrastructure improvements. CG notes that we should write in @ashiklom 's work on PECaN as a broader impact!

CG: Summer was spent setting up plots, thinking about capturing first wave of disturbance response (probably physiological versus structural). Dendro bands on 90 trees - fast physiology responses.

@atkinsjeff says he's soaking in the science.

AS: how bare ground was it in 1900? CG: harvest followed by fire, pretty bare. BBL notes this planned first-year paper would be a pre-disturbance analysis (not considering FoRTE at all). Which of these models can simulate time/carbon asymptote correctly?

[bpbond]

CG: Ellen is writing a parallel paper on empirical side. Lots of empirical data available for comparison. Do we have to 'nudge' (in DA sense) models once, or continuously, to get them to sort-of-match observations? What combination of model structural characteristics do you need to get this right?

Have set up a pretty good compute node on PNNL's HPC virtual cloud.

JA: November visit to UMBS, Nov 11-Thanksgiving, Lisa/Kayla/Liz/Jeff. Survey remaining trees, esp those targeted for kill; collect litter; Kayla soil resp (2 days); dendro band readings. Small window!

[bpbond]

Some question about UMBS disturbance timing (late April-mid May) and exactly how the disturbance treatments will be implemented, how they'll play out, on the ground.

AS suggests (but does not promise) we may try to do some disturbance model runs before the actual treatment starts. Like preregistration!

JA: FoRTE lidar data sitting there at the moment; working on another paper; missed one plot this summer -> let's make it a control, problem solved.

BBL: we own a drone!

CG: reached out to Jane re synthesis work, haven't heard back. She might not have time until spring, though.