How to plot delta changes and it signification in the same map (spatialPlot)

[abelcentella]

Hi

I calculated the delta change between the climatologies for history (climH) and future (climF) precip and put the result on a grid (change)
climH <-climatology(subH)
climF <- climatology(subF)
change<-gridArithmetics(climF,climH,climH,c,operator = c("-","/","*"))

As I don't found a way (in climate4R) to estimate the statistical signification of the changes I did this with my own script based on t.test using subH and subF grids. Then I produced a grid (named r) with 1 and 0 (1 significant). After that I applied sig<-easyVeri2grid(r,subH) and then sigpoint<-map.stippling(sig,threshold = 1, condition = "GE",pch=16,col="black",cex=0.5).

Finally, I got a proper plot with the precip changes and points where the changes are significant. I used the following

spatialPlot(change,main=list(c("Precip Changes 2021-2040 vs 1961-1990"),cex=2),
xlab=list("Longitude",cex=1.5),ylab=list("Latitude",cex=1.5),
colorkey = list(space = "right",title="(%)",cex=2),
backdrop.theme = "coastline",scales=list(draw=TRUE,cex=1),
col.regions = colorRampPalette(colraindelta),
set.min = -50,set.max=50,sp.layout=list(sig_points))

The above is just for one season or a month. Thus as I want to make a multipanel plot using spatialPlot with a multigrid I create a multigrid using multg<-makeMultiGrid(changes 1..12,skip.temporal.check = TRUE) and made a list of 12 with the respective sig_points. Then

mons<-c("JAN","FEB","MAR","APR","MAY","JUN","JUL","AUG","SEP","OCT","NOV","DEC")

spatialPlot(change,main=list(c("Precip Changes 2021-2040 vs 1961-1990"),cex=2),
xlab=list("Longitude",cex=1.5),ylab=list("Latitude",cex=1.5),
colorkey = list(space = "right",title="(%)",cex=2),
names.attr=mons,as.table=TRUE,
backdrop.theme = "coastline",scales=list(draw=TRUE,cex=1),
col.regions = colorRampPalette(colraindelta),
set.min = -50,set.max=50,sp.layout=list(sig_points))

I got a multipanel plot with the changes for every month but the stippling is always the same for every panel. I believe that I'm doing something wrong.

Can somebody explain to me what is wrong or if there is another way to plot the change field with the significance?

[jorgebanomedina]

**Hi,

The problem is in the sig_points object, since it is a list of 12 (one per season). I recommend you to loop over the months and add a member of sig_points object one at a time to spatialPlot, such that:**

figs <- lapply(1:12, FUN = function(z) {
# where change_z is the change of month z, e.g., change_1 is the change of month "JAN"
spatialPlot(change_z, backdrop.theme = "coastline",
sp.layout=list(sig_points[[z]]),
.... "your other plotting parameters")
})

Finally you can call the figs object that store the spatial maps, with grid.arrange:
library(gridExtra)
grid.arrange(grobs = figs, ncol = 4)

By the way, climate4R.value do not capsulate the t-test statistic, but you can consider the p-value of the KS-statistic which can be computed by setting measure.code = "ts.ks.pval" in valueMeasure function, as an alternative to the t-test. This is just in case you want to fully depend on climate4R, you approach with the t-test and easyveri2grid seems reasonable for me.

Please let me know if you attain your desired plots.

Cheers,

Jorge

[jbedia]

Hi,
"[...] and made a list of 12 with the respective sig_points".
At this point, you must include the argument which in each element of the list, so the first stippling goes to subpanel 1 (which = 1), the second to subpanel 2 (which = 2) and so on. There is an example (not exactly your case, but it may help) in ?map.stippling. I hope this helps.

[abelcentella]

**Hi,

The problem is in the sig_points object, since it is a list of 12 (one per season). I recommend you to loop over the months and add a member of sig_points object one at a time to spatialPlot, such that:**

figs <- lapply(1:12, FUN = function(z) {

where change_z is the change of month z, e.g., change_1 is the change of month "JAN"

spatialPlot(change_z, backdrop.theme = "coastline",
sp.layout=list(sig_points[[z]]),
.... "your other plotting parameters")
})

Finally you can call the figs object that store the spatial maps, with grid.arrange:
library(gridExtra)
grid.arrange(grobs = figs, ncol = 4)

By the way, climate4R.value do not capsulate the t-test statistic, but you can consider the p-value of the KS-statistic which can be computed by setting measure.code = "ts.ks.pval" in valueMeasure function, as an alternative to the t-test. This is just in case you want to fully depend on climate4R, you approach with the t-test and easyveri2grid seems reasonable for me.

Please let me know if you attain your desired plots.

Cheers,

Jorge

Hi Jorge,

Thanks a lot. I understood the point. However I followed the suggestion fro Bedia which is simplest for my specific case. I works for me and the plots are find now. Nevertheless I will use your approach in other scripts we are building.
Re ts.ks.pval I was trying to install the VALUE packaged as recommended but I got an error.

devtools::install_github("SantanderMetGroup/VALUE")
Downloading GitHub repo SantanderMetGroup/VALUE@HEAD
Error in utils::download.file(url, path, method = method, quiet = quiet, :
download from 'https://api.github.com/repos/SantanderMetGroup/VALUE/tarball/HEAD' failed

Regards

Abel

[abelcentella]

Hi,
"[...] and made a list of 12 with the respective sig_points".
At this point, you must include the argument which in each element of the list, so the first stippling goes to subpanel 1 (which = 1), the second to subpanel 2 (which = 2) and so on. There is an example (not exactly your case, but it may help) in ?map.stippling. I hope this helps.

Hi Joaquin
Thanks you very much, I got the right plots

[jorgebanomedina]

Hi,

Try devtools::install_github("SantanderMetGroup/VALUE@v2.2.1")

Regards,

Jorge

[abelcentella]

Hola Jorge, muchas gracias. Espero que todo me funcione bien. En cualquier caso te mantendré informado. Gracias y saludos Abel El mar, 1 jun 2021 a las 12:00, Jorge Baño-Medina ***@***.***>) escribió:

…

Hi, Try ***@***.***") Regards, Jorge — You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub <#8 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/ALEOW3GVOBK7CVRY7CDYESDTQT7Y5ANCNFSM45VGNZUA> .

[abelcentella]

Hola Joaquin. Muchas gracias. Era bien simple y todo me funcionó perfecto ahora. Agradezco mucho la respuesta/ayuda Saludos Abel El lun, 31 may 2021 a las 10:37, Joaquin Bedia ***@***.***>) escribió:

…

Hi, *"[...] and made a list of 12 with the respective sig_points"*. At this point, you must include the argument which in each element of the list, so the first stippling goes to subpanel 1 (which = 1), the second to subpanel 2 (which = 2) and so on. There is an example (not exactly your case, but it may help) in ?map.stippling. I hope this helps. — You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub <#8 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/ALEOW3FXNU37W6NTJ6BU34DTQONMNANCNFSM45VGNZUA> .