SP weeks of December 21 and 28 and January 4

COMPASS-DOE / rs-synthesis

0 stars 0 forks source link

SP weeks of December 21 and 28 and January 4 #2

Closed bpbond closed 3 years ago

bpbond commented 3 years ago

[x] Please commit and push your code from last week

Let's continue previous exploratory work at a deeper level.

[x] How many observations (rows) are there for each category of the "Irrigated" manipulations? (Forests, grasslands, and agriculture only.) What categories exist?
[x] What about for soil drainage? This should be simpler, with mostly "dry" and "wet"
[x] What measurement variables are most common? I assume Rs_annual will be frequent but we're also interested in "Rh_annual" and "Rs_growingseason" in particular; maybe one of the Q10s too
[x] Try computing the response ratio for each study in the Dry/Wet category, as log(Rwet / Rdry). Graph these, with the RR on the X axis and study on the Y. For example like this (but for now without error bars):

g20041

Please don't hesitate to ping me if something is unclear! Thank you.

bpbond commented 3 years ago

By next Monday the 11th you and I should have both read the Liu et al. paper.

bpbond commented 3 years ago

BBL to do: literature search for relevant 2018-2020 publications that we can add to the analysis.

stephpenn1 commented 3 years ago

@bpbond I used the forcats package to reorder right in ggplot (can probably do this in the pipeline if we want this ordering to be more permanent. I noticed some studies still have multiple response ratios... are there other variables im not capturing that could be causing multiple ratios?

bpbond commented 3 years ago

😍 😍 😍

Woo @stephpenn1 that looks great. Nice job!

The multiple RRs within a line are because there might be an irrigation +50%, irrigation +75%, etc. for a single year. They're each divided by the "None" Rs and produce separate dots.

bpbond commented 3 years ago

Suggestion: just have study author name on the y axis? (Not year.)

bpbond commented 3 years ago

Also ideally the y axis is "Smith 2010" where 2010 is the publication year, not the observation year.

stephpenn1 commented 3 years ago

Like this?? Would you like me to average multiple irrigation levels together or keep them as separate points?

bpbond commented 3 years ago

I would keep as separate points. Eventually we'll want to color the points by irrigation/drought level I think.

bpbond commented 3 years ago

Included your lovely graph in my slides for DOE!