Problems requiring remedial action will inevitably arise during the production of a crop. Sampling is the activity required to obtain information about problems or conditions that will reduce yield so that proper corrective action can be taken.
There are two types of sampling. The first is predictive sampling, or sampling before the desired crop is planted. Sampling for nematodes in the fall following harvest is an example. The second is diagnostic sampling, or sampling to determine the cause of a problem after it occurs. Sampling for disease and insect damage during the growing season is an example.
There are three interconnected parts to sampling. The first is recognizing the need to sample. This is followed by selecting and properly applying the correct sampling technique(s). Finally, data collected from sampling must be correctly assessed and analyzed so that proper and effective action can be taken.
Grid sampling is a widely used method for site specific soil and crop management. Despite its popularity and use, there are deficiencies with this method. Grid sampling may result in data being collected across dissimilar soil types and thus being inappropriately averaged. Also, defining the appropriate grid size is arbitrary. Uniform grids may be subject to errors associated with field topography, drainage, and other physical factors, as well as management history.
Sampling within defined field areas can be used in lieu of grid sampling. With this method of directed sampling, the target area or field is divided into units based on the variability within the field. Prior knowledge of the area or field based on experience, aerial crop imagery, soil series maps, and maps of past yield, sample, and management history can be used to define sample units.
Sampling involves several components. The first is selecting or defining the proper sample unit. This may involve dividing a field into its soil series components or sampling field sections based on crop performance history. The objective is to ensure that the targeted problem or occurrence is sampled in units that are environmentally and agronomically uniform in order to reduce variability not associated with the sample target.
The second and third components involve determining how large a sample unit should be and how many samples should be taken from each unit to accurately assess the situation. Samples should be large enough to capture the intended target, but small enough to manage for analysis. Several samples from each defined unit are usually required to determine the variability of the sample target within the unit.
Fourth, it should be determined if sampling should be repeated at some later date to further assess the problem or the effect of an applied remedy. Treatment of some problems is only effective for a short time, and re-sampling may be needed to determine if the problem recurs at a treatable level. Also, re-sampling may be the only way to determine if an applied remedy was effective.
Fifth, in the interest of economy, the sampling protocol should be designated. Often, there are defined procedures for collecting crop and soil samples. It is too expensive to collect data that are not needed to assess a perceived problem. On the other hand, it is also too expensive to collect samples and not use them to their fullest benefit. This is true for soil samples taken and assessed for fertility needs. For example, knowledge of the complete nutrient status of a soil should be determined from soil samples in order to address possible interacting nutrient deficiencies.
From this discussion, there are three major points. First, know or define why you are sampling and ensure that your sampling procedure will accomplish the intended goal or detect the intended target.
Second, realize that variability exists in all fields and plant communities. Do not combine or composite samples unless there is evidence that the entity you are sampling is uniform within the defined sample unit. Compositing can mask problems by diluting occurrences that occur in isolated but severe levels.
Third, keep a record of results from all samplings in a field within and across years and build a map of occurrences that the samples detect. Often, problems that affect crop performance in a given field or sample unit interact with each other or change with time. This can be determined by evaluating sampling records that cover the same area over a period of time.
Application of proper sampling methodology can improve problem-solving and crop management. On the other hand, application of poor sampling methods may produce false data. This can result in erroneous conclusions about the problem or occurrence, leading to the application of improper remedies and the waste of time and resources.
For a more detailed treatment of sampling and sampling techniques, I encourage you to read the chapter entitled "Sampling tips and analytical techniques for soybean production" authored by Willers, Hergert, and Gerard that appears in the book Soybean Production in the Midsouth. This chapter presents many useful sampling ideas and procedures that can be used to more accurately determine the extent of field problems and how to assess their impact.