improve/make new figures for revisions

[ValentineHerr]

@teixeirak,

I like Ben's suggestion about the Whittaker plot.

Looking at sites with some data submited to IPCC, quite a few of them are missing map or mat (or both). (in image below the numbers under "TRUE" are the number of sites with missing value

FYI, the 130 missing mat are distributed in these different types of missing values:

NA  NAC  NI 
 20 105   5 

We can probably acquire values for some.

For sites that have both info, this is what the Whittaker plot would look like:

[teixeirak]

This looks nice! Is the code for the bigger background image easy to implement? The advantage of that is that it gives some sense of how common those climates are. This would do better at getting at the reviewer's desire for a better analysis of how representative these data are.

[ValentineHerr]

shouldn't be too hard. I'll work on that tomorrow.

[teixeirak]

@ValentineHerr , I knew this was super familiar... you've already made it! :-)

https://github.com/forc-db/ForC/blob/master/figures/Climate_of_forC_sites.png

[teixeirak]

For this paper, let's make the following changes:

• [ ] try to make the symbols small enough that its possible to see the underlying climate space
• [x] same symbols for all site regardless of whether the climate data came from ForC or WorldClim (which we can use to fill the missing climate values)
• [ ] more emphatic symbols for data sent to EFDB
• [x] the colors on Ben's version are nice, if easy to change.
• [x] remove the A in the x-axis units
• [x] as in the map, limit to records relevant to EFBD (not all of ForC)

Also, I'd skip the one with the Whittaker biomes. That's confusing because EFDB uses the Koppen/FAO climate zones.

[ValentineHerr]

That sounds good. I'll work on that asap. Yesterday I was trying to see if I could color Ben's version by different scales of colors based on biome but sounds like it not worth it. Plus it was not really working. This is what I had left it at:

[teixeirak]

Oh, that's a super cool idea! But I agree its best to drop it since its not working. If you want to do color coding, I'd use the FAO ecozones and drop the Whittaker biome lines.

[teixeirak]

@ValentineHerr , and as in the map, we should limit to records relevant to EFBD (not all of ForC).

[ValentineHerr]

How is this looking? I made tiny dots for ForC and triangles for EFDB sites.

[teixeirak]

I love it!

[teixeirak]

I really don't see anything I'd change.

[ValentineHerr]

I pushed the figure on the ForC repo (see commit above)

[teixeirak]

Thanks! Does this use the same data set as Ben's figure (WorldClim 2, Fick & Hijmans, 2017)?

[teixeirak]

Actually, could you please change the legend as follows: (dot) - ForC site with EFDB-relevant data (triangle) - site with ~some~ data submitted to EFDB

I don't know if a dot will show up well enough. You might just write out "(dot)"

[teixeirak]

Also, can you please confirm that this is limited to terrestrial area?

[teixeirak]

Also, please confirm spatial resolution of the data (0.5° in Ben's figure).

[ValentineHerr]

Thanks! Does this use the same data set as Ben's figure (WorldClim 2, Fick & Hijmans, 2017)? Also, can you please confirm that this is limited to terrestrial area?

It uses CRU data. I have in my notes: "summary of CRU data for 1990-2014 given by Ben Bond-Lamberty through GitHub on 8/23/2017".

I am assuming he gave us terrestrial data back then...

Also, please confirm spatial resolution of the data (0.5° in Ben's figure).

It is 0.5 resolution