Dear All,
Consider yield as attribute and easting and northing in meters as spatial
coordinates.
I used gls function of nlme package because of spatial dependency of the
residual. I tried to remove large-scale trend by using the easting and
northing as covariates of yield.
model.vc<-gls(yield~easting+northing,method="ML").
Being a spatial data, we are most interested in carrying out variogram
modelling of the spatial structure on the small-scale component i.e. the
residual and equally fit the spatial correlation structure on the residual
as well.
Do I need to obtain the residual from the model.vc above i.e
resid<-residuals(model.vc) and fit the other models as follow
model.sph<-gls(resid~1,corr=CorSpher(form=~easting+northing, nugget=T)) or
model.sph<-update(model.vc,corSpher(form=~easting+northing,nugget=T))
Please which one is correct.
Thank you.
Moshood
[[alternative HTML version deleted]]
_______________________________________________
R-sig-Geo mailing list
R-sig-Geo@r-project.org
https://stat.ethz.ch/mailman/listinfo/r-sig-geo
