Nugget from variogram.fit vs GLS
Hi Zev, it seems that gls returns the relative nugget rather than the absolute returned by gstat:
1.30231/(1.413246+1.30231)
[1] 0.479574 Could that be? I'm quite surprised results are that similar, doesn't gls use REML to fit variograms? Best wishes from a vivid geostat 2010,
On 07/01/2010 04:53 PM, Zev Ross wrote:
Hi All, I'm verifying coefficients generated with GSTAT kriging with external drift using GLS. The coefficients match absolutely perfectly. The range also matches perfectly. But there is a big difference in what GLS gives as the output for the nugget. Can anyone explain why this is? Zev ###### KED # here's where I fit the variogram and the fit output v.fit = fit.variogram(v, vgm(1.5, "Exp", 20, 0.5))
v.fit
model psill range
1 Nug 1.302310 0.0000000
2 Exp 1.413246 0.8098587
# the fit is used in a KED model
g = gstat(formula=formula(MODEL), data = so2, model = v.fit)
# here I get the coefs (these coefs match those from GLS below perfectly)
UKcoef<-predict(g, lattice, BLUE=c(TRUE,TRUE))
###### GLS
# use the same parameters as above and fix them
nug<-v.fit[1,2] # this is 1.302
sill<-v.fit[1,2]+v.fit[2,2]
range<-v.fit[2,3]
theGLS<-gls(formula(so2FinalAll), data=so2, na.action=na.omit,
correlation=corExp(c(range, nug/sill ),
form=~X+Y, nugget=TRUE,fixed=TRUE), method="ML")
summary(theGLS)
Generalized least squares fit by maximum likelihood
Model: formula(so2FinalAll)
Data: so2
AIC BIC logLik
578.6551 593.7083 -284.3275
Correlation Structure: Exponential spatial correlation
Formula: ~X + Y
Parameter estimate(s):
range nugget
0.8098587 0.4795740
# note the nugget of 0.479 vs the 1.302 above.
Edzer Pebesma Institute for Geoinformatics (ifgi), University of M?nster Weseler Stra?e 253, 48151 M?nster, Germany. Phone: +49 251 8333081, Fax: +49 251 8339763 http://ifgi.uni-muenster.de http://www.52north.org/geostatistics e.pebesma at wwu.de