How to estimate an AR1 x AR1 spatial structure in Mixed or Glimmix

[bprice]

I'm looking to use the spatial procedure outlined by Gilmour et al 1997 ("Accounting for Natural and Extraneous Variation in the Analysis of Field Experiments", Journal of Agricultural, Biological, and Environmental Statistics, 1997; 2(3):269-293). Is this possible? I've tried some preliminary things (random rows and columns), but have convergence issues. Has anyone done this?

TIA

Bill

[Rick_SAS]

Show us what you have tried so far and copy/paste the SAS log that shows the "convergence issues."

That paper also discusses the variogram for the data. You can use PROC VARIOGRAM to analyze the spatial structure for your data by fitting common spatial models.