Soil Surveys and Precision Agriculture

This looks to be a long blog and it may come across as tirade, but I hope my USDA friends will take it for what it is.  Keep in mind that this is written by a USDA trained soil scientist who has also tried to use soil surveys over the years.

I often hear and read about trying to use soil survey maps as the basis for developing sampling zones, seed rate zones, or other precision uses.  At the meeting I attended earlier this week, (Kelly Robertson's Summary) One of the topics of discussion was precision agriculture.  It seemed that many were trying to use soil survey maps for management zone boundaries.  When I was an undergraduate I worked closely with soil fertility professor Dr Ted R Peck.  Peck said that soil maps were not accurate enough for soil fertility work.  I think this a generalization, but it has merit. 

Since the 1960's soil survey has tried to use Soil Taxonomy to define soil mapping units.  The number of soil series has at least doubled in that time frame.  In the mid 70's Ted Peck explained soil survey by saying that we already know all the soils in Illinois.  The guys doing the soil survey are just defining the boundaries between them.  Doubling the number of series indicates this might not have been true, but still Ted had a point.  The most important thing the soil survey can do for farmers is to define the boundaries between landscapes and management units. 

Most "modern" soil surveys are mapped on a scale of 1:24000 to 1:12000.  Lots in Illinois were done on a scale of 1:15840 , which is 4 inches to a mile.  Limitations of scale affected the size of map units that could be shown.  The fact that modern GIS allows display on any scale gives an illusion of accuracy that is not there because of limitations of mapping scale.  Other things that affect the original product were use of unrectified photos as base maps and poor quality photos used as base maps.  I hate to say it too, but the individual who made the map affected it's quality.  Some were too concerned about taxonomy and not concerned enough about landscape.   

When considering the accuracy of the soil survey, you should also consider what happened to maps after the original line placement was made on the base map.  Boundary lines were hand traced to make a map unit overlay.  To digitize those lines, they were often scanned and the lines stretched to fit imagery.  How well that was done can affect accuracy.  I am not saying there is no quality control, I am just pointing out the potential for inaccuracies. 

I have found that sometimes small mapping units are "off" from where they should be.  Sometimes boundaries are off 100 feet or more. 20 feet may be tolerable but the more accurate the better.  Sometimes boundaries on the landscape are gradual so a line in a particular place is not a huge deal.  Soil surveys can be a very useful tool to learn about a farm, but may or may not be good for precision agriculture.  Some counties are better than others and some landscapes are better than others.  My opinion is that soil surveys should be evaluated and adjusted by a professional before they are used. 

Later I will discuss specifics about developing management zones for precision agriculture.  Yes soil surveys can be useful.