Jinshi-Ben-meeting 2020-05-19 - Githubissues

PNNL-TES / TES-mgmt

1 stars 0 forks source link

Jinshi-Ben-meeting 2020-05-19 #39

Open jinshijian opened 4 years ago

jinshijian commented 4 years ago

Landmodel

See issue # 11 in the landmodels repo

SRDB-V5

I'd prefer to use the GitHub Releases mechanism, and point people there. How to do this?
Keep figure S1, and have a similar figure for MAP?
Submit this manuscript this week and discuss the corresponding author?

Time and collar

Choose a term: measuring time, measurement time, survey length, sampling time.
Same conclusion when using cook distance to diagnose the outlier.
Looked like MAP and MAT does not introduce bias to the relationship between annual Rs and collar insertion depth.

Surrogates

Lin strongly disagree with PCA, I will try to use cluster or hierarchy classification analysis.
Using correlation instead of R2?

Plan for this week

Finish Time and collar manuscript discussion
Surrogates revision
Improve landmodel analysis
Submit SRDB-V5 manuscript

jinshijian commented 4 years ago

If the collar insertion causes a big amount of Rs reduction, theoretically, Rroot to Rs ratio of the "Exclusion" method (involving collar insertion) should be smaller than that from other partition methods (not involving collar insertion), however, we see no significant difference among different partition methods.