The theory behind zone sampling is that because you are sampling related areas within a field - e.g., soil types - you can reduce sample frequency, thus reducing costs. The problem with this approach is that you have 2 variables, soil type and fertility, that more often than not have little correlation with each other.

A case in point: Take one field, sample it on one acre grids, then run a correlation matrix on the variables. It will soon become apparent that in most cases, the accuracy of zone sampling leaves much to be desired.