Alexey-Ben meeting 2018-11-14

[ashiklom]

Last week's goals:

• FoRTE
  • ED-related fixes to PEcAn
  • Natively pass ED binary arguments -- done (PR 2183)
  • Run ED only with PEcAn-specified PFTs -- Mostly done. I have a workaround, but need to clean up and submit as PR to PEcAn.
  • ED bare ground runs at UMBS -- Mostly done. Hundred year run completed without crashing (~6.5 hours). Need to look at results.
  • Prepare for FoRTE meeting -- done (meeting notes)
• ED RTM -- Didn't work on this week
• Spent a lot of time this week--probably too much--on PEcAn Docker API (PR 2192). Happy with the result, broadly useful for the community, and does make it a lot easier to quickly run models...but was not a goal for this week. I need to avoid diversions like this in the future.

This week:

• FoRTE
  • Better-setup ED2 runs at UMBS
  • FoRTE ED2 paper outline -- Based on Monday's team meeting. Use current outline (in this repo) as a starting point, but flesh out ED-specific details (based on below reading)
  • Background reading -- ED2 recent developments and analyses
• ED RTM for AGU (all carryover from last week)
  • Implement soil moisture reflectance model
  • Leaf and wood allometries in R
  • Sensitivity analyses for LAI, WAI, leaf/wood spectra, soil reflectance, leaf angle, clumping factor, and incident direct/diffuse fraction

[bpbond]

Wednesday work for a meeting (if needed)?

[ashiklom]

Sure! My schedule is wide open, 9:00 to 4:30.

[ashiklom]

Meeting notes:

Re: FoRTE ED runs, @ashiklom 's plan for this week has three major components:

• Produce a reasonable, baseline 100-year ED2 run at UMBS
  • Goal is not necessarily to match observations, but rather to get output that we can fully explain/rationalize. In other words, eliminate model errors (e.g. unexpected collapses of dominant cohorts), but OK to keep systematic model biases (e.g. early successional dominance) because these fit the ecological narrative
  • Specific goals:
  • Why the sudden collapses partway through the run? Look at all the outputs to diagnose. @ashiklom 's current best guesses are carbon starvation (look at non-structural carbohydrate pools and general carbon balance -- GPP vs. respiration) and water depletion (look at plant water potential, soil moisture, evapotranspiration flux).
  • Best possible parameterization -- Define PFTs in terms of species present at UMBS (first pass -- early, mid, late deciduous, and northern pine; at some point, consider running each species as its own PFT) and parameterize them using the available data (add public TRY data for these species into the PEcAn database; for some parameters, look to the literature).
  • Inspect the inputs -- Especially the soil. Maybe also compare several different climate drivers.
• Research the ED2 sub-models we are interested in
  • What are the equations?
  • What are the references?
  • What are the FoRTE hypotheses that the schemes map onto?
  • Most likely candidates:
  • Finite canopy radius -- dictates relative competitive advantage of taller cohorts over shorter ones
  • Radiative transfer scheme -- two-stream vs. multiple scatter. How do they affect the distribution of light in the canopy? In turn, how do these differences in the canopy light profile affect the emergent biophyiscal and ecological predictions?
  • Nitrogen controls on photosynthesis and decomposition
• Develop a written analysis plan/outline for moving forward
  • Combination of to-do list and manuscript skeleton. Ideally, something that can be populated incrementally until it eventually becomes the manuscript itself.

Other discussion topics:

• Alexey will be working from home Wed morning before Thanksgiving, then off that afternoon and the following Monday.
• Alexey may also work from home tomorrow if the weather is bad
• Undergraduate mentoring -- Brainstorm independently for now; then, discuss with Ben and potential undergrad mentees in early 2019.

[bpbond]

Thanks @ashiklom - nice summary.