[R-sig-Geo] Error fitting separable space-time variogram model

Benedikt Gräler ben.graeler at uni-muenster.de
Sat Mar 8 12:16:38 CET 2014


The error occurs when the internally used optim function tries a non-positive range. You can either try to use different start values for your model or provide boundary arguments as in optim to fit.StVariogram. These are passed on to optim via ..., see the help page of optim for details. The demo stkrige provides as well examples on how to limit the range of possible parameter values. 

Hth,

 Ben

On March 8, 2014 10:15:55 AM CET, d.g.rossiter at utwente.nl wrote:
>I am developing a tutorial on space-time geostatistics, expanding on
>the nice st.pdf vignette. I can successfully fit a metric variogram
>model to an s-t empirical variogram, but when I try to fit a separable
>model I get an error which I can't solve nor interpret:
>
>"Error in vgm(1 - par[4], as.character(model$time$model[2]), par[3],
>par[4],  : 
>   range should be positive"
>
>A reproducible example follows. I updated all packages immediately
>before running this example. Thanks for your help.
>
>System info:
>
>> version
>platform       x86_64-apple-darwin10.8.0   
>arch           x86_64                      
>os             darwin10.8.0                
>system         x86_64, darwin10.8.0        
>status                                     
>major          3                           
>minor          0.2                         
>year           2013                        
>month          09                          
>day            25                          
>svn rev        63987                       
>language       R                           
>version.string R version 3.0.2 (2013-09-25)
>nickname       Frisbee Sailing             
>> help(package="sp")$info[[1]][2]
>[1] "Version:            1.0-14"
>> help(package="spacetime")$info[[1]][2]
>[1] "Version:            1.0-9"
>> help(package="xts")$info[[1]][4]
>[1] "Version:            0.9-7"
>> help(package="gstat")$info[[1]][2]
>[1] "Version:            1.0-18"
>
>
>## condensed and adapted from st.pdf, 13-Feb-2013
>require(sp); require(xts); require(spacetime); require(gstat)
>## Germany & neighbours air quality
>data(air)
>rr <- rural[, "2005::2010"]
>na.stations <- which(apply(as(rr, "xts"), 2, function(x)
>all(is.na(x))))
>r5to10 <- rr[-na.stations,]
>vst <- variogram(PM10 ~ 1, r5to10[,1:200])
>plot(vst, map=FALSE)
>## code from p. 8 of st.pdf
>vgm.sep <- vgmST(stModel="separable",
>                 space=vgm(0.9,"Exp", 123,0.1),
>                 time=vgm(0.9,"Exp", 2.9, 0.1),
>                 sill=60)
>vgmf.sep <- fit.StVariogram(vst, vgm.sep, method = "L-BFGS-B") ## gives
>an error
>## "Error in vgm(1 - par[4], as.character(model$time$model[2]), par[3],
>par[4],  : 
>##  range should be positive"
>
>## my system and package info
>version
>help(package="sp")$info[[1]][2]
>help(package="spacetime")$info[[1]][2]
>help(package="xts")$info[[1]][4]
>help(package="gstat")$info[[1]][2]
>
>
>------------------------------------------------------------------------
>
>_______________________________________________
>R-sig-Geo mailing list
>R-sig-Geo at r-project.org
>https://stat.ethz.ch/mailman/listinfo/r-sig-geo

-- 
Benedikt Gräler

Institute for Geoinformatics
University of Münster

http://www.ifgi.de/graeler

+49 251 83 33082



More information about the R-sig-Geo mailing list