I am going to do a project based on kriging. The targeted area would be as big as United States and my grid would be 10km*10km. I am wondering the differences of efficiencies for the methods implemented in geoR package and ArcGIS. Have anyone done any research on the performance yet. I would prefer geoR if it works well.
I would also suggest to consider the
gstat package (either standalone or the
R package), which -in my experience - is a bit between
geoR and Geostatistical Analyst: you do not have the very advanced models of
geoR but it's more efficient from a computing perspective (its core is in C).
Also, you might be interested in the