Grid Sampling

ON THE ROAD TO PRECISION AGRICULTURE
Grid Sampling: Problems, Improvements and Beyond

Grid sampling is a favored method for site specific soil management because it is unbiased, simple, relatively quick and software exists to facilitate it. After the samples have been pulled, georeferenced and analyzed, a map is made by either filling in the grid cells with the soil test value (grid cell method) or assigning the soil test value to a point and then interpolating between points (grid point method). Grid sampling results in either a "checkerboard" map (grid cell method) or a "smoothed" map (grid point method).

Problems of Grid Sampling: Despite the current popularity of grid sampling, it is becoming apparent that there are many deficiencies with this method of within-field soil mapping. These problems include:

	Grid cell sampling results in cells being averaged over dissimilar soil types.
	There is no consensus on an appropriate grid size (or on how to determine one).
	Maps based on grid point samples can be unduly biased by localized irregularities.
	Uniform grids are susceptible to systematic errors such a fertilizer banding, terraces, uneven fertilizer application, continuous one-way row cropping, tile drains and a variety of environmental factors.
	Uniform grids result in both under and over sampling if soil regions vary in size.

Improving Grid Sampling: Simple steps will improve the quality of a grid sampling plan.

	Use an unaligned sample design to minimize systematic errors.
	Pay attention to each individual sample pattern as you do to an overall grid pattern. Increase the sample radius for each point to 25% of the sample distance (80 ft. radius for 2.5 acre grids).
	Increase or decrease grid sampling intensity based on prior knowledge of variability within separate field regions.

When applying fertilizer based on the sample pattern, incorporate strip plots with fertilizer recommendations based on a separate whole field soil sample. This allows an on-site evaluation of the effectiveness of the grid design and fertilizer application.

Beyond Grids—Directed Sampling: Using "directed" sampling avoids many of the problems of grids. In directed sampling, a field is broken up into an appropriate number of regions which are sampled individually. The shape and size of each region are based on the variability present within the field and the intensity of management considered necessary and practical. To formulate an appropriate directed sampling pattern, prior knowledge about the spatial patterns present within a field must exist. Useful sources include:

	Farmer experience
	SCS soil maps
	Past yield maps
	Past grid sample maps
	Topography maps
	Aerial field imagery
	Management history maps

Roger Brook Abstracted from information by Stuart Pocknee

This page was last updated on 10/08/01.