R help - Jan 2021 - Runs of at least 5 days with certain value in a rasterbrick in R

Dear R users,

I am  working with a rasterbrick named X, which has got 6300 time layers.
Those 6300 time layers represent every winter day since 1.1.1950 till
31.12. I dropped all 29th of february so that I had 90 time layers for
every season. Values throughout the whole domain are between 0.5 and 1.5. I
successfully made a function, which creates a rasterlayer, which contains
amount of runs of at least 5 days with values greater than 1.0. So for
every grid point, I have got amount of events I am looking for. The
function is following:

ff<-function(x,na.rm=TRUE){
y<-x > 1
n<- ave(y,cumsum(y == 0), FUN = cumsum)
sum(n >= 5)
}

This works very well, however, there is a problem. After layer which
represents 28.2., 1.12., 2.12. and so on.. follows. It means, that event
starting 28.2. may be detected if the criterion is satisfied throughout the
first four days of following december. Of course I dont wanna detect such
events as those are two distinct seasons. So I tried something like this
(to somehow divide it into 70 seasons and calculate function ff for them):

indices<-rep(1:70,each=90)
XX<-stackApply(X,indices,fun=ff)

At first sight, this works and no error is written. However, results are
miscounted and there is too much events detected, for example 20 events are
detected in the first layer at maximum, which doesnt make any sense
(20x5=100, there is just 90 days per season). Other fact is, that this
works perfectly fine, if I dont divide the whole series and use the
criterion through original brick X without any indices.
I cant see where the problem is, so every help is deeply appreciated, maybe
another way could be more efficient.

Adam Dragula



<https://www.avast.com/sig-email?utm_medium=email&utm_source=link&utm_campaign=sig-email&utm_content=webmail>
Bez
vir?. www.avast.com
<https://www.avast.com/sig-email?utm_medium=email&utm_source=link&utm_campaign=sig-email&utm_content=webmail>
<#DAB4FAD8-2DD7-40BB-A1B8-4E2AA1F9FDF2>

	[[alternative HTML version deleted]]

Hallo

I do not know anything about rasterbrick but what about splitting it to list
according to season. If you have dates starting 1.12. and ending 28.2 as one
season, diff should be 1day for each season. If you manage to correctly split
your data, you can cycle through list or use lapply to get desired result.

Cheers
Petr
> -----Original Message-----
> From: R-help <r-help-bounces at r-project.org> On Behalf Of Adam
Dragula
> Sent: Wednesday, January 20, 2021 10:46 PM
> To: R-help at r-project.org
> Subject: [R] Runs of at least 5 days with certain value in a rasterbrick in
R
> 
> Dear R users,
> 
> I am  working with a rasterbrick named X, which has got 6300 time layers.
> Those 6300 time layers represent every winter day since 1.1.1950 till
31.12. I
> dropped all 29th of february so that I had 90 time layers for every season.
> Values throughout the whole domain are between 0.5 and 1.5. I successfully
> made a function, which creates a rasterlayer, which contains amount of runs
> of at least 5 days with values greater than 1.0. So for every grid point, I
have
> got amount of events I am looking for. The function is following:
> 
> ff<-function(x,na.rm=TRUE){
> y<-x > 1
> n<- ave(y,cumsum(y == 0), FUN = cumsum)
> sum(n >= 5)
> }
> 
> This works very well, however, there is a problem. After layer which
> represents 28.2., 1.12., 2.12. and so on.. follows. It means, that event
starting
> 28.2. may be detected if the criterion is satisfied throughout the first
four
> days of following december. Of course I dont wanna detect such events as
> those are two distinct seasons. So I tried something like this (to somehow
> divide it into 70 seasons and calculate function ff for them):
> 
> indices<-rep(1:70,each=90)
> XX<-stackApply(X,indices,fun=ff)
> 
> At first sight, this works and no error is written. However, results are
> miscounted and there is too much events detected, for example 20 events
> are detected in the first layer at maximum, which doesnt make any sense
> (20x5=100, there is just 90 days per season). Other fact is, that this
works
> perfectly fine, if I dont divide the whole series and use the criterion
through
> original brick X without any indices.
> I cant see where the problem is, so every help is deeply appreciated, maybe
> another way could be more efficient.
> 
> Adam Dragula
> 
> 
> 
> <https://www.avast.com/sig-
> email?utm_medium=email&utm_source=link&utm_campaign=sig-
> email&utm_content=webmail>
> Bez
> vir?. www.avast.com
> <https://www.avast.com/sig-
> email?utm_medium=email&utm_source=link&utm_campaign=sig-
> email&utm_content=webmail>
> <#DAB4FAD8-2DD7-40BB-A1B8-4E2AA1F9FDF2>
> 
> 	[[alternative HTML version deleted]]
> 
> ______________________________________________
> R-help at r-project.org mailing list -- To UNSUBSCRIBE and more, see
> https://stat.ethz.ch/mailman/listinfo/r-help
> PLEASE do read the posting guide http://www.R-project.org/posting-
> guide.html
> and provide commented, minimal, self-contained, reproducible code.