Site-specific detection of residual soil N levels and the levels of other plant nutrients have been studied during the last twelve years to a greater level, with a greater accumulation of data, than any other nutrient effort in the history of NDSU. The results support the general use of a “zone” approach to soil sampling that could easily be adopted by most growers. With the support of local businesses and their associated crop consultants, growers can choose to map and manage nutrients site-specifically on their own, or with assistance.
The greatest immediately seen economic returns are seen in sugarbeet production, where quality and quantity of the crop are both rewarded by the processor and nitrogen management is a key element in both.
Immediate economic rewards are also seen in other crops, but once zones are addressed and variable-rate fertilizer is applied, the subsequent rewards are more subtle. The tools used in conducting site-specific farming can result in the following benefits-
-An archive of spatial yield data that can be a valuable record, but can also be used to delineate production management zones.
-An understanding of the production of certain soil types under your management, which leads to an understanding of the productivity of neighboring farms under consideration for purchase or for rent.
-A record both in time and space of the application of fertilizers or pesticides, that might be useful in litigation or in environmental enforcement.
My research into site-specific nutrient detection and application began in Illinois as a Ph.D. student under Dr. Ted Peck. Initially the research was conducted to determine the minimum sampling density on two forty-acre fields necessary to delineate soil P, K and pH levels. The results published in the 1993 thesis indicated it would take about one sample per acre to delineate these areas.
Sampling research continued in North Dakota beginning in 1994. The initial results indicated that sampling density of about one sample per acre would be required to delineate soil nitrate and other nutrient levels in North Dakota. However, when the fields were resampled in 1995, similar patterns appeared, suggesting that perhaps if the patterns were related to more easily measured and stable factors, fewer samples would need to be taken. Several papers were written during the next years regarding the use of a zone approach to soil sampling, using topography and other delineation tools to direct soil sampling instead of the use of intensive grid sampling. To date, in 2006, there have been over 50 field-site years of data that support the use of zone sampling to direct soil nitrate sampling in North Dakota. Zone sampling provides at least the same level of delineation accuracy as one-sample-per-acre grid sampling, while using only 3-8 samples per field, depending on the size of the field and the variation found within the field. Commercial samplers have been very successful in their use of the zone approach for soil sampling. Those who are using more than one zone delineation method might find that the stability of their zones is better from year to year than if just one method were used.
Research near St. Thomas, ND from 1997 to 2001 was conducted to determine the reason why sugarbeet quality in this area of the Red River Valley was historically low compared to other areas, and determine methods to overcome this problem. Soil N availability was found to be high, but use of variable-rate N application alone did not solve the problem. The solution was based on J.T. Moraghan and L.J. Smith’s approach of considering the N supplying ability of sugarbeet tops as an N credit to subsequent crops, combined with the zone soil nitrate sampling approach outlined by Franzen and others. The N credit idea was originally established by French scientists Crohain and Rixhon (1967). A review of the development of the sugarbeet top N credit in sugarbeet rotational management can be found in Franzen (2005). The research conducted during this period showed that the combined use of sugarbeet top credits and zone-directed soil sampling and variable-rate application could be used successfully to reduce residual soil N and to increase sugarbeet quality. Residue decomposition studies showed that sugarbeet and potato tops decompose rapidly in the soil, enhancing their nutritional benefit to subsequent crops.
An analysis of soil survey scale showed that the current published scale of Order 2, in which soils are only mapped when they have a spatial extent of more than 2.5 acres in a polygon are not very useful in zone delineation for directing soil sampling. Order 1 soil surveys, which map areas of 1 acre or more in size, would be much more useful in defining zone boundaries for soil sampling.
A three-state, federally supported project (USDA-CSREES-IFAFS) was initiated in 2000 to determine whether similar zone delineation methods could be used across the region, or whether the preferred methods were more localized. Economic and environmental benefits were also studied during the project. The results indicated that across the three states of Montana, North Dakota and Minnesota, zone management of N was effective in delineating areas with relatively homogenous residual N levels. Topography, the use of yield frequency maps, and satellite and aerial imagery were found to be the overall best delineation tools. Soil EC was most useful as a data layer in drier environments, such as Montana, Williston, ND, or in the irrigated sands at Oakes, ND. The study found that a simple method of weighted, classified zone delineation that uses more than one data layer was more effective than more complicated black-box clustering methods. In addition, the study curiously found that assumptions regarding N fertilization were opposite of what the data showed. In both Montana and North Dakota, it was originally assumed that higher rates of N would be appropriate in more productive soils. However, yield and residual nitrate data showed that lower N rates would be most appropriate on productive soils, with the higher rates or perhaps different N timing management would be needed in less productive soils. Economic analysis of the sites showed an edge to variable-rate N application by zone for sugarbeets, but not for other crops in the study. These results were probably realized because several of the sites either suffered from very high residual N levels at deeper depths (Williston), or the variability of the sites was reduced from numerous years of site-specific management in previous work (Oakes, Valley City, Mandan). The study also assumed that the N recommendations published in the current circulars are correct, however, recent review of the N response data from various crops shows that perhaps they are not as correct as they need to be for site-specific N management.
There may be an opportunity with the current state of GIS and site-specific agriculture to develop Order 1 soil maps from terrain modeling and the knowledge of what soils would be expected on certain parts of the landscape. A project funded through USDA-NRCS in North Dakota is researching this possibility, with the assistance of Dr. J. Boettinger, Utah State Univ., who has researched and developed the PURC model for terrain modeling and soil survey enhancement. This project is expected to conclude in the fall of 2007.