Space-time cross variogram with gstat

[MaximeBeauchamp]

Hi,

Would it be possible to create a space-time cross variogram with the variogram function. Maybe by using a gstat object instead of a formula as a first argument. But maybe it is a too big modification for the gstat library.

In that case, I'd like to try to modify the code of the variogramST function by myself to get the cross-variogram between 2 variables of my spacetime object.

Is it possible ?

Maxime Beauchamp INERIS Geostatistics Modeling Engineer

[edzer]

Hi Maxime, way to go - I would try the same. You may want to make variogramST an S3 method, with instances for formula and gstat. I'd be happy to receive pull requests, or just working code.

[MaximeBeauchamp]

Hi Edzer,

Sorry for the delay of my response (I was in paternity leave) ! Ok I managed to create my crossvariogramST with this kind of header (see the attached file for the code) :

cross_variogramST = function(formula1, formula2, locations, data, ..., tlags = 0:15, cutoff, width = cutoff/15, boundaries=seq(0,cutoff,width), progress = interactive(), pseudo = TRUE, assumeRegular=FALSE, na.omit=FALSE)

For sure it is not as clean as an S3 method but it was easier to implement and to use once gstat is loaded. I guess I have now to find a way to fit an authorized model for this !

Maxime

variogramST_functions.zip

[edzer]

On 20/07/16 16:03, MaximeBeauchamp wrote:

Hi Edzer,

Sorry for the delay of my response (I was in paternity leave) ! Ok I managed to create my crossvariogramST with this kind of header (see the attached file for the code) :

cross_variogramST = function(formula1, formula2, locations, data, ..., tlags = 0:15, cutoff, width = cutoff/15, boundaries=seq(0,cutoff,width), progress = interactive(), pseudo = TRUE, assumeRegular=FALSE, na.omit=FALSE)

For sure it is not as clean as an S3 method but it was easier to implement and to use once gstat is loaded. I guess I have now to find a way to fit an authorized model for this !

If you want me to look at it, I'd need example code.

What do you mean by authorized model?

Best regards,

Maxime

variogramST_functions.zip https://github.com/edzer/gstat/files/373845/variogramST_functions.zip

— You are receiving this because you commented. Reply to this email directly, view it on GitHub https://github.com/edzer/gstat/issues/5#issuecomment-233958537, or mute the thread https://github.com/notifications/unsubscribe-auth/AAfykyWa9klR9hPsz0QjNTXP2WpxTqXiks5qXiqZgaJpZM4ISCJp.

Edzer Pebesma Institute for Geoinformatics (ifgi), University of Münster Heisenbergstraße 2, 48149 Münster, Germany; +49 251 83 33081 Journal of Statistical Software: http://www.jstatsoft.org/ Computers & Geosciences: http://elsevier.com/locate/cageo/ Spatial Statistics Society http://www.spatialstatistics.info

[MaximeBeauchamp]

I attached a zip file with the code of the cross-variogram ST function in my previous response. I guess it works fine on my data (my field of interest is air quality modelling). By "authorized" I mean fit the model. Maybe I'll try the linear model of coregionalisation of De Iaco for the product-sum model as a first try.

Maxime

[edzer]

I can't see the attachment.

[edzer]

@MaximeBeauchamp any news on this one, or time to close this issue?