[R-sig-eco] Problem with "fun=predict.se" for stand. errors from "Tps"

[Robert J. Hijmans]

Dear Jaime,

The example in the help file for interpolate does exactly that, and it
works for me. My guess is that you need to update the 'raster' package
to a more recent version, as the 'fun=' argument is a relatively
recent addition. Perhaps you ran the script on a different computer
with a different version of 'raster'?

As for question #2, this is wrong:
Raster2 <- interpolate(Elevacion, Slope, Eva2, xyOnly=FALSE)

it should be something like:

s <- stack(Elevacion, Slope)
Raster2 <- interpolate(s, Eva2, xyOnly=FALSE)

Robert

2010/8/18 Jaime Burbano Girón <jaimebg27 at gmail.com>:
> Hi to everybody. Few weeks ago I performed successfully a Thin plate spline
> regression ("Tps" function from fields package) following instructions given
> in this forum by professor Robert J. Hijmans.
>
> In that moment I calculated successfully the standard errors for the Tps
> predictions using the option fun=predict.se from "interpolate" function in
> raster package. I ran this procedure following this code:
>
> Datos <-
> read.table("F:/Especies_Invasoras/Evaporacion/evaporacion_sabana.txt",
> header=T, sep="\t", na.strings="NA", dec=".", strip.white=TRUE)
> Eva2 <- Tps(Datos[,17:19], Datos$Anual) # X,Y and Elevation as independent
> variables for predict annual evaporation values
> Elevacion <- raster("F:/Especies_Invasoras/Evaporacion/altura.asc") # ascii
> grid of elevation values projected onto the desired area
> Raster2 <- interpolate(Elevacion, Eva2, xyOnly=FALSE)
> RasterSe2 <- interpolate(Elevacion, Eva2, xyOnly=FALSE, fun=predict.se)
>
> I´ve tried to repeat this procedure in this moment but I can´t obtain the
> standard errors from predictions. When I run the last line code I obtain the
> same values for Tps predictions ran in penultimate line code. It seems to be
> that now the option "fun=predict.se" is not being recognized. I work under
> Rx64 version 2.11.1 and raster package version 1.2-6.
>
> Additionally I have a question about the code when I include an other
> independent variable like slope for example. I think the code should be
> something like:
>
>  Datos <-
> read.table("F:/Especies_Invasoras/Evaporacion/evaporacion_sabana.txt",
> header=T, sep="\t", na.strings="NA", dec=".", strip.white=TRUE)
> Eva2 <- Tps(Datos[,17:20], Datos$Anual) # X,Y, Slope and Elevation as
> independent variables for predict annual evaporation values
> Elevacion <- raster("F:/Especies_Invasoras/Evaporacion/altura.asc") # ascii
> grid of elevation values projected onto the desired area
> Slope <- raster("F:/Especies_Invasoras/Evaporacion/slope.asc") # ascii grid
> of slope values projected onto the desired area
> Raster2 <- interpolate(Elevacion, Slope, Eva2, xyOnly=FALSE)
> RasterSe2 <- interpolate(Elevacion, Slope, Eva2, xyOnly=FALSE, fun=
> predict.se)
>
> but "interpolate" function presents this error: "Error in model$terms : $
> operator not defined for this S4 class"
>
> I appreciate your help. Best regards for every body.
>
> --
> Jaime Burbano Girón
> Departamento de Ecología y Territorio
> Pontificia Universidad Javeriana
> Bogotá - Colombia
> 3208320 Ext. 4839 - 3168573713
>
>        [[alternative HTML version deleted]]
>
>
> _______________________________________________
> R-sig-ecology mailing list
> R-sig-ecology at r-project.org
> https://stat.ethz.ch/mailman/listinfo/r-sig-ecology
>
>