On The Road to Precision Agriculture
Non-Grid Soil Sampling and Fertilizing Ideas
Grid soil sampling has been used for several years to guide variable-rate fertilization. Recent studies show that soil fertility patterns may be the result of topography or other predictable reasons, so that a high level of information might be obtained with minimal sampling and analytical expense.
Reasons for grid sampling-
Grid sampling is a systematic sampling that uses a high density of soil cores to reveal soil fertility patterns. Grid sampling is especially useful when man-made variations in soil fertility levels occur that are not revealed in plant growth differences. Agronomically, there is nothing wrong with sampling by grids if it is conducted relatively densely, such as one sample per acre. However, few growers are willing to sample densely enough to give adequate information to direct variable-rate application consistently because of the time and expense.
Topography and soil fertility levels
Soil fertility levels may be related to landscape, especially in areas where high levels of fertilizers or manure applications have not been made. Mobile nutrients move in the soil from upland positions to depressions, where water may collect and move into the soil. Water that infiltrates upland positions moves downward until it reaches a discontinuity in texture (coarser over finer, or fine over coarser) which causes the water to flow laterally.
Nitrate may or may not be high in depressions due to the activity of denitrification bacteria. In saturated soils, denitrification bacteria transform nitrate into gaseous forms of N which are not taken up by plants. The nitrogen is therefore lost to plant uptake. In semi-arid areas, the activity of
denitrification bacteria is low, and depressions are often high in nitrate. However, in more humid regions, the activity of denitrification bacteria is relatively high, and low levels of nitrate may be present in depressions, along with high levels of chloride and sulfate.
Non-mobile nutrient levels may also be related to topography. Due to increased plant growth and lower mineralization, depression areas tend to have higher levels of organic matter and clay-sized particles than upland areas. Organic matter accumulates micro-nutrients and P, while clay size particles may contain more K than silt or sand-sized particles. On the other hand, higher crop yields in depressions compared to upland positions may lead to higher P levels on upland positions.
Levels of nutrients are also affected by past erosion. Upland positions are particularly subject to wind erosion in semi-arid environments. In many fields in the central North Dakota glacial till region, hilltops are bare of the original high organic matter surface layer due to the action of wind during the last one hundred years of cropping. The hilltops often contain lower levels of all nutrients because of the loss of organic matter. Depression areas may contain additional clay and organic matter from water erosion and deposition. Or, some intermediate areas within slopes may have shallower organic layers due to accelerated erosion since the onset of cropping. Sulfur availability appears to be particularly sensitive to low organic matter levels, and sulfur deficiency in canola is often seen on eroded slopes, but not in areas on the same slopes with higher organic matter.
Topography sampling is conducted by first measuring elevations. However, dividing the
field into zones of elevation is not appropriate, because hilltops may be the same elevation as depressions in rolling fields. It is the landscape structure, or topographic structure that is important, not the elevation. However, it is difficult to draw lines with confidence around landscape features without other means of support. Other tools can help to define zone boundaries to direct sampling in addition to topography alone.
Soil conductivity sensors
Soil conductivity sensors can be used to identify important zones and for verifying the importance of topography zones. It is important to note that the readings did not correlate directly with nutrient levels. High conductivity may be related to high N, but low conductivity may more often in areas of medium N content, while medium conductivity corresponded with low N levels. These devices reveal patterns and it is the responsibility of the sampling to reveal the significance of these patterns.
Aerial photography and satellite imagery
Satellite imagery can be effective in identifying fertility patterns if the patterns are larger than the resolution of the pictures. In fields where spatial patterns are small, aerial photography might be a better tool.
Soil surveys in North Dakota or Illinois have not been helpful in determining management zones. Sometimes the boundaries are drawn incorrectly, other times, too many soil series are combined into one mapping unit. In all of the fields investigated by this researcher (Dr. Dave Franzen), soil series boundaries were not useful for direct sampling for precision farming.
Yield maps are the result of a number of soil and environmental factors that affect the growth and yield of crops. Yield maps appear to be most useful in defining areas of recurring patterns between years, such as areas with consistently low or no yield. However, those areas of the field that are only slightly different than other areas of the field in yield have not been useful in defining management zones. A great deal of interpretation is essential if yield maps are to included in any management zone development.
Managing Directed Sampling
The concept of directed sampling is exciting because it usually will result in a reduction in sampling expenses, making it more practical to conduct precision farming activities in lower return crops. However, directed sampling may have additional demands on the quality of help necessary to carry out a directed sampling program.
In a directed approach, the sampler must make a decision prior to entering the field on where to define the zones. This requires some field interpretation skills and a higher level of sampler knowledge and experience than generally would be necessary in a grid program. So although directed sampling appears less expensive on the surface, there is also attached to the system a need for more advanced employees.
Some progress is being made to automate the process of developing sampling zones. Differences in topography, conductivity, image values and/or yield levels are used to automate the zone development process. Developing zones automatically is in the beginning stages, but if perfected, could make sampling zones as easy to do as grid sampling is today.
At the Precision Agriculture session during ANR week this year, Dr. Dave Franzen North Dakota State University, discussed this topic with those in attendance.