Spatio temporal Kriging

Technically speaking it is OK to predict in future (prepare a stfdfObj 
object with spacetime coordinates in future, then run krigeST), but I 
would assume that the prediction variance will be large (well this 
depends on your variogram / temporal nugget and temporal range of 
spatial autocorrelation). Consider for example the prediction models for 
weather in the Netherlands: 
http://www.knmi.nl/waarschuwingen_en_verwachtingen/ensemble.html
As you get further from the measurement horizon, the confidence limits 
become wider and wider.

Also, I do not know how much is krigeST applicable to non-Gaussian data.

PS: Sending more focused questions that refer to some specific line in 
the code / or specific topic of broad interest are usually more 
efficient on R-sig-geo (see: "Do your homework before posting" in 
http://www.r-project.org/posting-guide.html).

T. (Tom) Hengl
Url: http://www.wageningenur.nl/en/Persons/dr.-T-Tom-Hengl.htm
Network: http://profiles.google.com/tom.hengl
Publications: http://scholar.google.com/citations?user=2oYU7S8AAAAJ

Hello

i am exercising (would like) to use spatiotemporal kriging to predict
in the future the rate of populations in neighborhoods in towns. I
have values for population rate in 2011, 2012, 2013, 2014 for many
neighbourhoods and i need to know the corresponding values in 2015 and
2016 with the spatio temporal kriging using R.
I would to know (if possible) how to apply extrapolation (spatio
temporal kriging) with R from the beginning to the end.

i have attached the actual R code with the help of Graeler and data file.

Thank you a lot

Franklin

_______________________________________________
R-sig-Geo mailing list
R-sig-Geo at r-project.org
https://stat.ethz.ch/mailman/listinfo/r-sig-geo