GCAM community meeting -- Attended. Got a much better sense of how GCAM works, what people are working on, and how I might be able to contribute down the line.
FoRTE -- Nothing directly paper-related
PEcAn
Worked on doing model runs at UMBS. Several pull requests merged in the process:
Bugfix: Edit PEcAn XML through web interface on Docker container (#2152)
Feature: Add alternative download method to download.CRUNCEP, because our current default (OpenDAP) is fast but unreliable (#2153).
Feature/Bugfix: When downloading CRUNCEP climate data, check against land-sea mask first. Coarse gridded climate products sometimes mask out sites near large water bodies (such as UMBS) as water, so all data gets replaced with NA (#2155).
Updated and configured PEcAn containers on our PNNL OpenStack cloud node. Along the way, realized that Docker (and possibly other things as well) don't work on PNNL's network when running on Ubuntu 18, so had to re-install everything on Ubuntu 16 (install is pretty quick and straightforward, and I have it documented for future reference).
Started a 100 year SIPNET run at UMBS on cloud on my way out the door on Friday, but OpenStack is down for maintenance 10/22 and 10/23 (today), so can't get to results today.
Hector
Didn't really get to any calibration
I did mess around with pulling spatio-temporal patterns out of global NPP data. I found an AVHRR-based global NPP dataset from 1980 to 2000 and performed an EOF analysis on it. I have a basic, incomplete writeup that I can show you and Corinne at our meeting later today.
Goals for this week:
FoRTE
Look at SIPNET UMBS outputs from 100 year run.
Run ED at UMBS, 1901-ish to 2010-ish. May need to do several runs, and tune several knobs to get reasonable output.
Run SIPNET and ED at SERC.
Unlike UMBS historical runs, these are not bare-ground runs, so I will need to get FIA or other composition data for initializing ED.
Goals from last week:
download.CRUNCEP
, because our current default (OpenDAP) is fast but unreliable (#2153).NA
(#2155).Goals for this week: