search

forc-db / Global_Productivity

Creative Commons Attribution 4.0 International
2 stars 0 forks source link

Figures and tables to follow new paper structure #38

Closed beckybanbury closed 4 years ago

beckybanbury commented 5 years ago

We're going to want to re-jig a lot of the figures now that we have restructured the paper following the layout of the hypothesis table. So far my list of figures is:

Fig 1 - map of all data points used in analysis Fig 2 - stacked plots to show closure of component fluxes with major fluxes Fig 3 - latitude plot (will be updated to simplify and avoid repetition with fig 2)

Fig 4 is currently a plot of the major fluxes against temperature, precipitation, PET, and temperature seasonality. This crosses a lot of different sections of our results, and so I don't know what the neatest way to include it would be. I think it could work to move it later in the results section, e.g. alongside discussion of other climate variables/seasonality. Alternatively it could be split into two sets of graphs, but I do think four presented like this looks neat.

Fig 5 is the plot of MAT*MAP for all carbon fluxes. I'm wondering if it would be better to make this fig 4, included in the section on how productivity relates to MAT and MAP.

Other possible figures to include:

Finally, I currently have a table detailing the R^2^ values of all fluxes for a key set of climate variables. This no longer fits neatly in a section, but is probably still interesting to include. We may want to refine the variables included in this selection according to what we present/discuss: currently it includes latitude, MAT, temp seasonality, PET and MAT*MAP.

@teixeirak

beckybanbury commented 5 years ago

@teixeirak here are two possible simplified latitude plots for figure 3, one without points + one with transparent points (level of transparency can be adjusted to suit!) effect_of_lat

effect_of_lat_transparent

teixeirak commented 5 years ago

I prefer the one with points. It would help to include the symbols. Also, I like your color scheme, but it may be helpful to expand to include greens and blues-- would differentiate the fluxes more.

teixeirak commented 5 years ago

Fig 5 is the plot of MAT*MAP for all carbon fluxes. I'm wondering if it would be better to make this fig 4, included in the section on how productivity relates to MAT and MAP.

I think that makes sense.

beckybanbury commented 5 years ago

I've been using this scheme because it is apparently specially designed to be readable for people with colour blindness - unfortunately all the schemes that I've found that are colour blindness accessible don't really allow good differentiation for as many variables as we have.

Do you think it's better to use a scheme with a wider range of colours but that won't work for colour blindness? I agree that they are hard to distinguish.

teixeirak commented 5 years ago

Thanks for being considerate! Unfortunately I think some of those colors are so close that they are very difficult to distinguish even with normal vision. Perhaps a better solution would be to use two different line types (and show symbols and line types in legend).

teixeirak commented 5 years ago 
* Productivity against length of growing season (linear relationship so not necessarily that interesting)

I agree that could be interesting. Perhaps as a panel in current Fig. 4? (In that case it would make sense to either add another variable or remove one that's currently there.)

teixeirak commented 5 years ago 
* Any of the within-growing-season climate relationships (but weak/NS relationships, so again, not ideal to present)

I'd agree that this wouldn't be interesting to show in a Figure (although maybe worthwhile putting it in the SI)? Just reporting the statistics should be sufficient there.

teixeirak commented 5 years ago

Finally, I currently have a table detailing the R^2^ values of all fluxes for a key set of climate variables. This no longer fits neatly in a section, but is probably still interesting to include. We may want to refine the variables included in this selection according to what we present/discuss: currently it includes latitude, MAT, temp seasonality, PET and MAT*MAP.

I think I'd rather see this as a figure, but either way is okay.

I agree that this doesn't fit neatly in a section, and that's because our current hypothesis table doesn't have anything relating to R2s for climate variables. I hate to propose yet another section in that table, but perhaps it would make sense to add a question/hypothesis on that at the end.

beckybanbury commented 5 years ago

effect_of_lat_transparent Is something like this any better? i think part of the problem is that the lines are all on top of each other, so it will be hard to distinguish regardless. An alternative option would be to remove the scaling factor for this graph (see below), or to present a smaller selection of variables. The stacked plots do contain every variable plotted against latitude. effect_of_lat_transparent_notscaled

beckybanbury commented 5 years ago

combined_plots

This is one option for updating the combined plots: the top two plots refer to our section on MAT and MAP, the second two to the section about other relevant climate variables (solar radiation could be replaced with VPD alternatively), and the last two refer to the section on seasonality/growing season length.

beckybanbury commented 5 years ago

I have replaced solar radiation with VPD, as this is a better predictor of productivity.

teixeirak commented 5 years ago

I like it!