With all the discussion about zone sampling versus grid sampling between advocates of both systems, I think it's absolutely essential to know the benefits and shortcomings of each system.  

There is no doubt that zone sampling saves money versus grid sampling. With grid sampling, in order to ensure an acceptable degree of accuracy, a lot of samples need to be pulled and analyzed which drives up costs. Conversely, zone sampling reduces costs because it is assumed that given areas within a field, as one poster in a thread below so eloquently put it, are homogeneous. In other words, the variance in pH and nutrient levels are minimal so these zones can be sampled as a composite, thus reducing costs. But is this true? Are zones homogeneous enough that we can in fact reduce sampling frequency in order to reduce costs? I have found that the answer to that question is usually a resounding no. Here's why.....

Over a period of several years, I have had the opportunity to grid sample a number of fields on 1.1 acre grids. This has provided me with the ability to view nutrient variability on a scale not typically seen in the industry. With this density of samples, I not only get an excellent look at field scale variability, but I can also measure the variability within differing soil types within the field. What I consistently find is that with respect to P & K levels, they are not homogeneous within these zones - although pH is a different story.

Here is a field showing sample points on 1.1 acre grids and also the different soil types:

The next image is the field broken into different sampling zones based on the soil types seen above:

Now that the field is broken down into sampling zones based on soil types, an analysis of the samples that fall within each zone can be done.

After breaking down the samples within their respective zones, it becomes quite apparent that there is a lot of variability within the zones themselves. In order to get an idea on how large the variability actually is within the zones, I did a standard deviation calculation on each zone. Keep in mind that the standard deviation is the average variation from the zone average. For P & K (actually P2O5 & K2O), it's easy to see that the variation is quite large, and for pH, the variation is somewhat smaller. But what does all this mean? And how does it affect the amount of fertilizer that is recommended to be applied?

The next step is to calculate the average amount of error as compared to the zone average:

This chart shows the percent error of the zone average when compared to the individual samples (the chart is labeled incorrectly). In other words, the zone average over/underestimates the nutrient level by the listed percentage. Clearly we can see that for P & K, using the average for the zones results in a large amount of error. A person needs to keep in mind that over/underestimating the level of fertility directly affects the amount of fertilizer being applied by roughly the same percentage. So in effect, if I were to use the zone average as a basis for a fertilizer recommendation, I would be overapplying or underapplying P by approximately 40% and K by 20%. In my opinion, this is unacceptable.

I do think anything under 10% is acceptable, so for pH and the application of limestone, using a composite average by zone is a viable alternative. But not for P & K.

I know someone is going to bring up the fact that this is only one field and I am cherry-picking the data. I have run many of these analysis on different fields it it almost always comes up the same. Very seldom does a composite sample by soil type accurately reflect the P & K fertility within that zone. Again, pH is a different story. In the majority of fields, pH correlates fairly well with soil type, so a zone composite is usually acceptable. But who is going to sample pH by zone and P & K by grid? It's simply not feasible so grid sampling is the best option and far superior to zone sampling.

I will add one caveat however. Zone sampling for mobile nutrients such as nitrogen may produce acceptable results, but since we do not sample for nitrogen in Illinois, I have not done a similar analysis.

http://agadvocate.wordpress.com/2008/11/01/grid-sampling-vs-zone-sampling/#more-45