Illinois Fertilizer
Conference Proceedings
January 24-26, 2005
| Home | 2005 Index | Search |
|
|
| Home | 2005 Index | Search |
R. G. Hoeft, German Bollero, Heather L. Clodfelter, L.C. Gonzini, and Kristen Smith
1
Soil testing is a useful, reliable tool for determination of nutrient needs. The reliability of this process requires that a representative sample be collected in the field, that it be processed correctly at the laboratory, and that the recommendations be based on calibrations conducted on soils similar to those of the area. Nearly all soil testing laboratories that service Illinois producers utilize the recommended soil test procedures for the North Central Region (Brown, 1998). Substantial research has been conducted in Illinois and other North Central states to correlate and calibrate the results obtained from these procedures to crop yield response to supplemental application of nutrients.
A soil testing program begins with the collection of soil samples from a field. This is a critical step because, regardless of analytical precision, the soil test will be of little value if the sample does not represent the field or area of the field. Considerable research has been conducted over the years to ascertain sampling techniques necessary to collect samples that accurately characterize the nutrient status of the field or area within the field. This research has shown that number and field position from which samples are collected will vary depending on management practices and soil variability in the field.
Results reported in the literature do not agree on the size of area that a single sample can accurately represent. Work by Franzen and Peck 1993 has demonstrated substantial spatial variability even on a relatively flat Central Illinois soil. Their work suggested a sample intensity of 1 sample per acre would characterize variation in most fields. Work by Bullock 1999 suggested that one sample per 2.5 acres would accurately characterize the nutrient content of fields with low spatial variability, but that even a 1 acre grid was not accurate to characterize spatial variability in fields with high variability. Spatial variability is often much higher on fields that have high testing soils (Mallarino, 1996). This is fortunate as the risk associated with not accurately predicting the true nutrient content on high testing fields is much less than on low testing fields.
Cameron et al., 1971 reported that an estimate of the mean was within 10% for phosphorus 70% of the time when 20 cores were composited, but on a highly variable field, the same 20 core sample resulted in prediction within 10% of the true mean less than 56% of the time. Band application of immobile nutrients creates a challenge for accurate soil sampling (Kitchen et al., 1990, and Mallarino, 1996). In those situations where the location of the band is evident, it has been suggested that samplers avoid the band area completely or collect one in the band for every 20 when the band spacing has been 30 inches or one sample from band for every 8 between the bands for 12 inch band spacing (Kitchen et al., 1990). When the location of the band is not know, suggestions for the number of samples to collect have varied from 20 cores per composite sample (Shapiro 1988) to over 100 (Hooker, 1976).
The Illinois Agronomy Handbook recommends that samples be collected from a 7–inch depth on all fields (Hoeft R.G. and T.R. Peck, 2002). This recommendation is based on the fact that the research conducted to calibrate soil tests was done at a 7–inch depth. Research has shown that this is especially important with reduced tillage systems as the immobile nutrients such as P and K are located in the top few inches of the profile. If the sample is collected at a shallower depth, the results will over estimate the nutrient value of the sample. On most soils, collection of samples from a depth greater than the 7–inch suggestion will underestimate nutrient value.
While the above research papers do not delineate how the samples were handled after collection, one can only make the assumption that the entire volume of soil was ground and thoroughly mixed before the chemical analysis was conducted. Would the results of these studies have been different if only a portion of the sample had been ground and mixed before chemical analysis? The answer to this question is very important in light of the fact that modern day laboratories in an attempt to become more efficient only grind an amount of soil equivalent to that which will fit into a typical drying box. The remainder of the soil is discarded. In many cases, the samples are not pre-crushed by hand before being poured into the grinder. This change in grinding procedure could unknowingly be inserting increased variability into the soil testing process.
The primary objective of this project is to evaluate the variability associated with the grinding and thorough mixing of an entire sample as compared to the grinding and mixing of a portion of the field sample. In particular, we would compare the differential in variability associated with collection of 3 cores (an amount of soil that will fit into a drying box) with that of grinding a portion of 5 cores.
Soil samples will be collected from a total of 6 different soil types, each having 2 different tillage systems and 2 different phosphorus levels. Soil types will consist of Drummer, Virden, Fayette, Cisne, Elliot, and Grantsburg. Tillage systems will include 0–till as well as conventional tillage that includes a primary as well as secondary tillage operation at least every other year. Phosphorus levels will be greater than 100 lb⁄acre and less than 40 lb⁄acre. To attain each of these factors, a total of 24 sites will be selected.
At each site, five representative points (replications) will be selected for sampling. A total of 4 types of sampling strategies will be collected, including: 1 sample of 3 cores; 2 samples of 5 cores; and 1 sample of 5 cores that are separately bagged into a 0–3 inch and 3–7 inch samples. All soil contained in the bag of 3 cores will be ground and thoroughly mixed. All soil contained in one of the set of 5 core samples will be ground and thoroughly mixed before analysis, while the other 5 core sample that will be divided 3⁄5 into one sample and 2⁄5 into the other with the individual samples being kept separate and ground and thoroughly mixed prior to analysis. All soil contained in the bags containing the 0–3 inch and the 3–7 inch samples will be kept separate, ground and thoroughly mixed. Cores for all sampling strategies will be collected within a radius of 5 feet of the selected sample point.
After grinding, all samples will be sent to 3 commercial laboratories for analysis for pH, P, and K. The laboratories have agreed to provide their service free of charge for this study. The data will be statistically analyzed to determine differences in variability associated with the different sampling systems and to determine whether tillage, soil type, or inherent soil test value variability is influenced by soil sampling system.
As of mid-December, soil samples were collected from 5 different soil types, each having 2 different tillage systems and 2 different soil phosphorus levels. Soil types sampled included Drummer, Virden, Cisne, Elliot, and Grantsburg. Following collection, the samples were dried and ground according to the protocol and prepared for analysis at commercial laboratories. All samples were split and submitted to three laboratories for independent analysis. Results have not been compiled as of the writing of this paper.
1 R.G. Hoeft is Professor and Interim Head, German Bollero is an Associate Professor, Heather Clodfelter is JBT Scholar, L.C. Gonzini is Senior Research Specialist, and Kristen Smith is Research Specialist, Dept. of Crop Sciences, University of Illinois
Brown, J.R. 1998. Recommended chemical soil test procedures for the North Central region. In J.R. Brown, (ed.). Recommended chemical soil test procedures for the North Central Region. North Central Region Publication 221 (revised). Missouri Agric. Exp. Stn, Missouri, MO.
Bullock, Don. 1999. Soil testing: thoughts about precision. Proceedings Illinois Corn and Soybean Classic. pp. 5–7.
Cameron, D.R., M Nyborg, J.A. Toodgood, and D.H. Laverty. 1971. Accuracy of Field Sampling for Soil Tests. Can. J. Soil Sci. 51: 165–175.
Franzen, David W. and Ted R.Peck 1993. Soil sampling for variable rate fertilization. 1993 Illinois Fertilizer Conference Proceedings (R.G. Hoeft, ed.) pp. 81-90.
Hoeft, R.G. and T.R. Peck. 2003. Soil testing and fertility. In Illinois Agronomy Handbook 23rd Edition. University of Illinois Coop. Ext., University of Illinois, Urbana.
Hooker, M.L. 1976. Sampling intensities required to estimate available N and P in five Nebraska soil types. M.S. thesis. Univ. of Nebraska, Lincoln (Cat. No. LD 3656 H665X 1976).
Kitchen, N.R., J.L. Havlin, and D.G. Westfall, 1990. Soil sampling under no-till banded phosphorus. Soil Sci. Soc. Amer. J. 54:1661–1665.
Mallarino, Antonio P. 1996. Spatial variability patterns of phosphorus and potassium in no–till soils for two sampling scales. Soil Sci. Soc. Am. J. 60: 1473–1481.
Shapiro, C.A. 1988. Soil sampling fields with a history of fertilizer bands. Soil Sci. News. Vol 10 No. 5. Neb. Coop Ext. Serv.
