Dear all
I want to perform regression kriging in R using GSIF package or gstat manually. I want to use linear regression model fitted with quadratic term and interaction between variables as the base?model for regression kriging.
Here is my final LM model:
LM1 <-? lm(Site_Form ~ DEM + Profile_Curvature + TPI + TRASP + Valley_Depth + ????Vertical_Distance + I(Plan_Curvature^2) + I(Profile_Curvature^2) + ????I(TWI^2) + I(Valley_Depth^2) + I(Vertical_Distance^2) + TRASP:Vertical_Distance + ????DEM:TRASP + DEM:Vertical_Distance, data=Data)
When I want to run fit.variogram I encounter with this Warning message:
In fit.variogram(vr, model = vgm(4, "Sph", 400, 1)) :
? No convergence after 200 iterations: try different initial values?
I do not where the problem is. I changed the nugget, range and sill values. But it still remains.
I would be very grateful if somebody helps me.
Thank you very much
Regards
Jalil
?