Hi there, I have climate data pertaining to extreme precipitation, as well as carbon emissions associated with those precipitation values in a dataframe.? The goal of my analysis would be to determine the probability of exceeding specific thresholds of precipitation extremes, as well as showing this graphically (I am imagining this by placing extreme precipitation on the the x-axis and exceedance probabilities on the y-axis). My question is if anyone has an idea how to approach this, or a good starting place? I have looked online, but there is nothing specific to really draw on. Thank you for your time, and I look forward to your response!
Plotting probability exceedance
4 messages · r@i@1290 m@iii@g oii @im@com, Bede-Fazekas Ákos
Hello,
You should know or make assumptions on the distribution of the
precipitation. Let's say it is normally distributed (i.e. bell-shaped).
Then you can calculate the probability of exceeding the quantile /q/ by
pnorm(q, mean, sd, lower.tail = FALSE).
If you have several spatial points and a lot of measurments (stored in
columns of the sf/data.frame) for each of the points, then use
apply(X, MARGIN = 1, FUN = function(measurements) {return(pnorm(q, mean,
sd, lower.tail = FALSE))})
and you can display the probabilities in a map.
HTH,
?kos
__________
?kos Bede-Fazekas
Centre for Ecological Research, Hungary
2023.02.15. 1:28 keltez?ssel, rain1290--- via R-sig-Geo ?rta:
Hi there, I have climate data pertaining to extreme precipitation, as well as carbon emissions associated with those precipitation values in a dataframe. The goal of my analysis would be to determine the probability of exceeding specific thresholds of precipitation extremes, as well as showing this graphically (I am imagining this by placing extreme precipitation on the the x-axis and exceedance probabilities on the y-axis). My question is if anyone has an idea how to approach this, or a good starting place? I have looked online, but there is nothing specific to really draw on. Thank you for your time, and I look forward to your response! [[alternative HTML version deleted]]
_______________________________________________ R-sig-Geo mailing list R-sig-Geo at r-project.org https://stat.ethz.ch/mailman/listinfo/r-sig-geo
Hi Akos,
Thank you so much for this suggestion! Indeed, I have 25 data points in each column, and yes, the data are normally distributed. Using the pnorm function is actually quite useful, as well. To that end, could the values of exceedance probability from pnorm be somehow plotted against their associated precipitation thresholds on an xy plot, for example?
Another idea that came to mind is the use of Probability Density Functions, but can these really be used to graphically show exceedance probabilities?
Thanks, again!
-----Original Message-----
From: Bede-Fazekas ?kos <bfalevlist at gmail.com>
To: r-sig-geo at r-project.org
Sent: Wed, Feb 15, 2023 1:59 am
Subject: Re: [R-sig-Geo] Plotting probability exceedance
Hello,
You should know or make assumptions on the distribution of the
precipitation. Let's say it is normally distributed (i.e. bell-shaped).
Then you can calculate the probability of exceeding the quantile /q/ by
pnorm(q, mean, sd, lower.tail = FALSE).
If you have several spatial points and a lot of measurments (stored in
columns of the sf/data.frame) for each of the points, then use
apply(X, MARGIN = 1, FUN = function(measurements) {return(pnorm(q, mean,
sd, lower.tail = FALSE))})
and you can display the probabilities in a map.
HTH,
?kos
__________
?kos Bede-Fazekas
Centre for Ecological Research, Hungary
2023.02.15. 1:28 keltez?ssel, rain1290--- via R-sig-Geo ?rta:
Hi there, I have climate data pertaining to extreme precipitation, as well as carbon emissions associated with those precipitation values in a dataframe. The goal of my analysis would be to determine the probability of exceeding specific thresholds of precipitation extremes, as well as showing this graphically (I am imagining this by placing extreme precipitation on the the x-axis and exceedance probabilities on the y-axis). My question is if anyone has an idea how to approach this, or a good starting place? I have looked online, but there is nothing specific to really draw on. Thank you for your time, and I look forward to your response! ??? [[alternative HTML version deleted]]
_______________________________________________ R-sig-Geo mailing list R-sig-Geo at r-project.org https://stat.ethz.ch/mailman/listinfo/r-sig-geo
??? [[alternative HTML version deleted]] _______________________________________________ R-sig-Geo mailing list R-sig-Geo at r-project.org https://stat.ethz.ch/mailman/listinfo/r-sig-geo
Hello, As far as I understand, for one point (i.e., one row of the sf/data.frame) you have several precipitation values, but can calculate one exceedance probability from these values. So an xy-plot is not the best choice for visualization of this kind of data. Or you'll have the same y value (probability) for several x values (precipitation). If this is not a problem for you, then I suggest using tidyr::pivot_longer() to transfrom the precipitation values from the several columns to one. Then you can draw the xy-plot by ggplot or base plot or even lattice plot. HTH, ?kos __________ ?kos Bede-Fazekas Centre for Ecological Research, Hungary 2023.02.16. 2:52 keltez?ssel, rain1290 at aim.com ?rta:
Hi Akos,
Thank you so much for this suggestion! Indeed, I have 25 data points
in each column, and yes, the data are normally distributed. Using the
pnorm function is actually quite useful, as well. To that end, could
the values of exceedance probability from pnorm be somehow plotted
against their associated precipitation thresholds on an xy plot, for
example?
Another idea that came to mind is the use of Probability Density
Functions, but can these really be used to graphically show exceedance
probabilities?
Thanks, again!
-----Original Message-----
From: Bede-Fazekas ?kos <bfalevlist at gmail.com>
To: r-sig-geo at r-project.org
Sent: Wed, Feb 15, 2023 1:59 am
Subject: Re: [R-sig-Geo] Plotting probability exceedance
Hello,
You should know or make assumptions on the distribution of the
precipitation. Let's say it is normally distributed (i.e. bell-shaped).
Then you can calculate the probability of exceeding the quantile /q/ by
pnorm(q, mean, sd, lower.tail = FALSE).
If you have several spatial points and a lot of measurments (stored in
columns of the sf/data.frame) for each of the points, then use
apply(X, MARGIN = 1, FUN = function(measurements) {return(pnorm(q, mean,
sd, lower.tail = FALSE))})
and you can display the probabilities in a map.
HTH,
?kos
__________
?kos Bede-Fazekas
Centre for Ecological Research, Hungary
2023.02.15. 1:28 keltez?ssel, rain1290--- via R-sig-Geo ?rta:
Hi there, I have climate data pertaining to extreme precipitation, as well as
carbon emissions associated with those precipitation values in a dataframe.
The goal of my analysis would be to determine the probability of
exceeding specific thresholds of precipitation extremes, as well as showing this graphically (I am imagining this by placing extreme precipitation on the the x-axis and exceedance probabilities on the y-axis).
My question is if anyone has an idea how to approach this, or a good
starting place? I have looked online, but there is nothing specific to really draw on.
Thank you for your time, and I look forward to your response! ??? [[alternative HTML version deleted]]
_______________________________________________ R-sig-Geo mailing list R-sig-Geo at r-project.org https://stat.ethz.ch/mailman/listinfo/r-sig-geo
??? [[alternative HTML version deleted]]
_______________________________________________ R-sig-Geo mailing list R-sig-Geo at r-project.org https://stat.ethz.ch/mailman/listinfo/r-sig-geo