You have data on crop yields by farm, along with a set of regressors that include the county (region) in which the farm is located. Explain how you would go about estimating a regression model for this data that accounted for the fact the random part of yield variation is likely not independent across farms that are close to each other spatially