[GRASSLIST:2233] Re: trend surface analysis - New Problem

Hello.

I did as Roger said, and used

base2 <- surf.ls(2, east(G), north (G), isobase)

isobase$isobases.auto not isobase, which is a list that may contain more
than one raster layer.

Roger

but then I get this error msg:

Error in as.double.default(z) : (list) object cannot be coerced to
double

How can fix it?
I tried as.double(), but get the same msg.

Thanks again

Carlos
--
+-------------------------------------------------+
        Carlos Henrique Grohmann - Guano
    Geologist - MSc Student at IGc-USP - Brazil
       Linux User #89721 ICQ: 214752832
+-------------------------------------------------+

Citando Roger Bivand <Roger.Bivand@nhh.no>:

> Hello all
>
> I'm trying to do some trend surface analysis with GRASS data in R,
but I with some problems:
>
> 1 - when I get the raster from GRASS, surf.ls() doesn't accept it,
since its not in x,y,z format:
>
>> str(isobase)
> List of 1
> $ isobases.auto: num [1:241266] 771 770 769 768 767 ...
>>
>
> So, how can I change this to a format that surf.ls() understands?
>

isobase$isobases.auto is the z argument. You can generate the x by
east() of the object returned by gmeta(), and y by north() of this
object. I'm not sure how well surf.ls() scales to this number of
points, though.

> 2 - I'll try several degrees of surfaces, so I'll use surf.ls(),
but
what about residuals? can I get them with residuals()?
>
Yes, residuals() should work. Again, scaling to this number of points
is an issue. On the other hand, the residuals and fitted values will
drop straight into the GRASS location.

Roger

> thanks in advance
>
> Carlos Grohmann
>
> --
> +-------------------------------------------------+
> Carlos Henrique Grohmann - Guano
> Geologist - MSc Student at IGc-USP - Brazil
> Linux User #89721 ICQ: 214752832
> +-------------------------------------------------+

--
Roger Bivand
NHH, Breiviksveien 40, N-5045 Bergen, Norway

--
Roger Bivand
NHH, Breiviksveien 40, N-5045 Bergen, Norway