Illinois
Fertilizer Conference Proceedings
January 26-28, 1998
| Home | 1998 Index | Search |
| Home | 1998 Index | Search |
J. W. Hummel and S. J. Birrell1
Site-Specific Crop Management (SSCM) was an exciting concept in Illinois in 1929, when Linsley and Bauer noted that, "The soils of this state, often within a single field, vary widely in their need for limestone," and "It is important, therefore, that systematic and detailed tests be made of the field so that limestone may be applied according to the need for it." Use of the concept waned over the years, as larger fields and larger, faster equipment made varying application rates within a field impractical. Today, SSCM is an area of expanding opportunities for a broad spectrum of the agricultural community. Low-cost, powerful computers, real-time controllers, variable-rate application hardware, and accurate location systems have combined to again make SSCM an exciting concept.
A basic question in SSCM is, "What soil attributes need to be sensed on a site-specific basis?" Commercial fertilizer suppliers started offering grid sampling and analysis services for assessing phosphorus and potassium levels, so that fertilizer could be spatially applied. But others have maintained that drainage, slope, and inclination are more important, and correlate better with yield maps. There are many factors that contribute to growth and yield of a crop, and we'll discuss some of the sensors under development. Of course, knowing which soil attributes are most important for a particular locale or soil series will not be readily apparent until research agronomists and crop scientists have been able to collect site-specific data for a range of soil attributes over a number of years.
Discussions with farmers attempting to utilize SSCM suggest that water retention, runoff, ponding, etc. are important factors in SSCM. Slope and aspect may be valuable data layers, not only because of surface runoff and erosion considerations, but also because of solar radiation interception and soil warming considerations. In addition, changes in slope and elevation may correlate closely with changes in soil type, soil organic matter content, and other soil physical and chemical attributes that affect crop growth, pest infestation, etc. In the central U.S. Corn Belt, surface topography is relatively flat and the economical topographic maps that are obtainable from photogrammetry do not provide the needed accuracy for SSCM. Conventional engineering survey techniques can produce precise maps, but the cost is high.
Fortunately, Global Positioning System (GPS) receivers have recently become commercially available with high accuracy capability. Some real-time kinematic GPS systems are advertised to provide position precision of 1 in. ± 1 ppm. The ‘ppm' refers to the distance between the rover receiver and the base station; 1 ppm means that the positional error will increase about 0.6 in. for every 10 mile increase in distance between the rover and the base station. These accuracies refer to horizontal precision; and vertical errors are typically increased by a factor of 2-3. Clark (1996) investigated the potential of precision GPS instruments, operated in several data collection modes, to produce highly accurate topographic maps for use in SSCM. He concluded that elevation error of only 1.2 in. could be maintained using a static collection technique, and increased by 0-2.4 in. for the kinematic collection technique as compared to a static collection technique. These error levels were achieved in fields of less than 10 acres, with a tripod-mounted base station located within the field. These data suggest that commercially available kinematic GPS units could provide the accurate topographic maps needed for SSCM.
Soil physical characteristics are extremely important for crop production. Soil texture affects many aspects of crop production systems, e.g., aeration, infiltration, soil nutrient availability, erosion, tile spacing and depth, and herbicide adsorption. Soil structure or degree of aggregation influences root growth, hydraulic conductivity, and aeration due to the size of the pores present in the soil matrix. The structure of the soil surface, or seedbed, is important in providing good soil-seed contact for germination. The presence and location of a restrictive layer, such as a hardpan, plowpan, or claypan, can be a major factor in crop growth. In a more general sense, the soil "physical condition" can be viewed as those physical characteristics which are important for plant growth and are influenced by tillage operations. The ability to precisely manipulate a soil from an initial to a desired final physical condition implies the need for sensors to assess this property (Schafer et al., 1981).
Soil texture is related to soil conductivity, which can be measured using electromagnetic induction (EM). Recent research related EM measurements to the depth of a claypan horizon (Doolittle et al., 1994), and used an automated EM sensing system to map claypan depth over a number of fields (Kitchen et al., 1995). They obtained calibration measurements of claypan depth with a soil probe at a number of locations within a field to remove the effects of temporal variations in soil water content and temperature. Since soil conductivity is influenced by changes in texture and moisture availability, and both vary over the landscape and affect productivity, EM sensing of conductivity shows some promise for interpreting grain yield maps, at least in certain soils (Sudduth et al., 1995; Jaynes et al., 1995a, Kitchen et al., 1995). EM data have also been used to estimate other soil properties related to clay content, including cation exchange capacity (McBride et al., 1990) and atrazine partition coefficients (Jaynes et al., 1995b). Research on analyzing the acoustic emissions emanating from a soil-cutting tool/soil interface was reported by Liu et al. (1993). They showed that the power spectra of the signals generated for the four test soils having similar moisture levels and bulk densities were different, suggesting that the technique might lead to a real-time soil texture sensor.
Soil structure measurement has been attempted in the laboratory using a fiber optic displacement sensor to evaluate air-dry soil aggregate size (Zuo et al., 1995). A commercially-available sensor provided an output signal proportional to the distance from the sensor to the soil surface. The spatial pattern of peaks in the sensor output correlated well with soil aggregate size, and was relatively insensitive to the distance between the sensor and the soil surface. Stafford and Ambler (1990) reported on a computer vision system used to assess seedbed structure. Tests indicated that data from the vision system compared well with the traditional sieve analysis for aggregate size.
Soil physical condition sensing for tillage control was investigated by Young et al. (1988). They detected and analyzed instantaneous draft force and vibration of wedge and blunt vertical chisels in an attempt to sense soil compaction. Mean draft, residual draft, and the autoregression coefficients were sufficient to characterize soil physical condition, but they noted that a major difficulty of their work was the need to specify the desired soil condition based upon agronomic considerations. Smith et al. (1994) built upon this work and investigated coulter draft forces as an indication of soil condition. Roytburg and Chaplin (1995) proposed using soil resistance force, as measured with an extended octagonal ring transducer, as an indicator of the changes in soil condition during tillage. Liu et al. (1996) developed an instrumented chisel to sense soil texture/compaction for mapping spatial variability that might be present in a field. Chukwu and Bowers (1997) reported their work on developing and testing a multi-depth soil mechanical impedance sensor. Three prismatic tip sensors were mounted on a modified shank to sense compaction profiles in the soil. The work was conducted in a laboratory soil bin at very low speeds, but indicated that the technique has potential for locating compacted soil layers. Work is currently underway (Hummel et al., 1995) to interface a real-time soil moisture sensor (Sudduth and Hummel, 1993b) with a soil cone penetrometer to measure soil moisture and penetration resistance simultaneously.
Measurement of soil constituents that affect plant growth is a basic task in SSCM. Considerable time and effort have been expended in trying to optimize the sampling procedure to use to produce an accurate map of the soil nutrient variability that exists within a field. Grid sampling has been shown to produce accurate soil nutrient maps (Franzen and Peck, 1995), but may be cost prohibitive in some instances. Selective soil sampling based on soil characteristics such as soil color, texture, depth, slope and erosion phase as used in past soil testing programs may be used (Francis and Schepers, 1997), unless fertilizer usage has introduced large differences in nutrient levels. All of these techniques are based on sample collection and laboratory analysis to produce soil fertility maps. Frequently, the sampling intensity needed to adequately characterize the spatial variability is dense, resulting in sample collection and analysis costs that are economically prohibitive.
Sensors could potentially reduce the costs while improving the accuracy of the data collection and mapping process. However, most soil constituent sensors are not ready for commercial application because their sensing concepts have not been broadly tested and their designs have not been optimized. Real-time sensors, which can sense a soil and estimate the value of an attribute within a few seconds, are under development at a number of universities, government agencies, and companies.
Soil organic matter (SOM) affects the application rate of many soil-applied herbicides required for effective weed control. Soil organic matter content also affects nitrogen fertilizer needs. Many nitrogen recommendation algorithms allow a credit for the nitrogen-supplying power of the soil, which increases with increasing organic matter content. A real-time SOM sensor developed by Shonk, et al. (1991) at Purdue University was patented and licensed for commercial development and used to control the rate of a granular herbicide formulation (McGrath et al., 1990). Exhaustive studies of reflectance techniques (Sudduth and Hummel, 1991), following earlier work (Krishnan et al., 1980, Worner, 1989), led to a sensor incorporating a rugged, portable near-infrared reflectance (NIR) spectrophotometer for SOM and soil moisture prediction. The sensor was developed and field tested by USDA-Agricultural Research Service engineers in Illinois (Sudduth and Hummel, 1993a,b), and patented and licensed for commercial development. The portable spectrophotometer predicted organic matter in the laboratory (Figure 1), across a range of soil types and moisture contents, with a coefficient of determination (r2) of 0.89, and a standard error of prediction (SEP) of 0.40% SOM, approaching that of data obtained on the same soils with a research-grade spectrophotometer. Moisture content was predicted with a SEP of 1.88% (r2 = 0.94) for a data set including soil moisture tensions of 0.033, 0.33, and 1.5 MPa, and air-dry soil (Figure 2). Field operation of the prototype sensor did not yield acceptable results (SEP = 0.91% SOM), due at least in part to errors introduced by the movement of soil past the sensor during the scanning process (Sudduth and Hummel, 1993b). Additional laboratory tests of the NIR sensor with soils obtained from across the continental United States showed that acceptable soil organic matter predictive capability could be maintained with a single calibration equation for soils from the lower U.S. Corn Belt. Calibrations obtained for wider geographic areas suffered from a significant decrease in accuracy. A redesigned prototype NIR sensor provided faster data collection, and improved portability (Sudduth and Hummel, 1993c), but the ability of the sensor to estimate soil organic matter was essentially unchanged from the initial prototype.
In-field soil nutrient testing and sensing has seen some advancement in recent years. Test kits, such as those provided by LaMotte2 and Hach, are available for the major soil nutrients (nitrogen in both nitrate and ammonium forms, potassium, and phosphate), soil pH, and many of the micronutrients. Pocket colorimeters or spectrometers are available, and newer systems use digital test strip analysis to output nutrient concentrations. Ion-selective membrane technology is available in the form of hand-held meters for NO3-N, K, and Na ions, which provide a reading in a matter of minutes. Of course, the labor and time required for soil sample collection has not been significantly reduced, so the need for an automated technique still exists. Adsett and Zoerb (1991) reported on real-time nitrate sensing using ion selective electrodes. Nitrate extraction time and methodology were limiting factors in the system, and additional research was planned to improve the mixing and extraction phases. Thottan et al. (1994) reported on subsequent laboratory work on the effects of different soil:extract ratios and extract clarity on electrode response and electrode response time. They found that there was no significant difference (α =0.05) among different soil:extract ratios (1:15, 1:5, 1:3) and no significant difference among final nitrate concentration indicated in either decanted, filtered, or soil:extract suspension samples. Normalizing the response of the electrode for time showed that 80% of final concentration was consistently indicated within 12 s, 40% within 6 s and 10% within 4 s, which they felt was within the time required for rapid in-field measurements. Ion selective field effect transistors (ISFETs) are being investigated for real-time soil nutrient sensing (Birrell and Hummel, 1997; Viscarra Rossel and McBratney, 1997). ISFETs, which are based on the same chemical principles as ion selective electrodes, have several possible advantages such as small dimensions, low output impedance, high signal-to-noise ratio, fast response and the ability to integrate several sensors on a single electronic chip. However, ISFETs have the disadvantage of greater long-term drift and hysteresis than ion selective electrodes. The use of a dynamic measurement system such as flow injection analysis (FIA) minimizes the effects of drift and hysteresis, and exploits the specific properties of ISFETs. Birrell and Hummel (1993) investigated the use of a multi-ISFET sensor chip to measure soil nitrate in a flow injection analysis (FIA) system using different flowrates (0.04-0.19 ml s-1), injection times (0.25-2 s) and washout times (0.75-2 s). The multi-ISFET/FIA system was successfully used to measure soil nitrates in manually extracted soil extracts (r2>0.9) using a 0.5 s washout time and a 0.75 s injection time (Hummel and Birrell, 1995) (Figure 3). A prototype automated extraction system was tested; however, the extraction system did not consistently provide soil extracts that could be analyzed by the FIA/ISFET system, and requires considerable improvement. The rapid response of the system allowed samples to be analyzed within 1.25 s, and the low sample volumes required by the multi-sensor ISFET/FIA system make it a likely candidate for use in a real-time soil nutrient sensing system. There are ion-selective membranes available for most of the important soil nutrients and other ionic species in the soil (H+, K+, NH4+, NO3-, Na+, Ca2+, Mg2+, Cl-). Researchers have recently reported on the development of effective ion selective phosphate membranes with a very high selectivity for phosphate, good detection limits (10-5 M), and fast response times. Projects are underway at the University of Illinois, funded in part by the Illinois Council on Food and Agriculture (C-FAR) and by the Illinois Fertilizer Research and Education Council (FREC), to develop and test a real-time soil nutrient analysis system, based on ISFETs. These projects concentrate on the development of potassium and phosphate sensors, and complement an ongoing program to develop a real-time nutrient sampling and extraction system.
Real-time close-range sensors for soil attributes are under development at a number of university, government, and corporate laboratories, both here in the U.S. and abroad. These sensors are typically being developed by small teams of research scientists and engineers. Information on soil attributes, often obtained today by laboratory analysis of manually collected soil samples, will be enhanced by the use of these and other sensing technologies. Improved data collection capability will enhance our understanding of the interactions among variables influencing crop production and yield; resulting in more economical production practices and reduced environmental impact.
Considerable additional research effort is needed to establish the range of soil types and climatic conditions where the technology can be applied. Closer alliances between the traditional agricultural research sector, the computer and software industries, and the farm and electronics industries are needed to shorten the time from concept to commercialization. Partnering could accelerate the entire sensor development process, and provide early access to sensors for soil attributes for our agricultural producers.
1 J.W. Hummel is an Agricultural Engineer,
USDA Agricultural Research Service, Urbana IL; S.J. Birrell is a Visiting Assistant
Professor, Dept. of Agricultural Engineering, University of Illinois, Urbana
IL.
Doolittle, J.A., K.A. Sudduth, N.R. Kitchen, and S.J. Indorante. 1994. Estimating
depths to claypans using electromagnetic induction methods. J. Soil Water
Cons. 49(6): 572-575.
Birrell, S.J., and J.W. Hummel. 1993. Multi-ISFET sensors for soil nitrate analysis. In: Proc. first workshop on soil specific crop management, p. 349. Robert, Rust, and Larson (eds.) ASA-CSSA-SSSA, Madison WI.
Birrell, S.J. and J.W. Hummel. 1997. Multi-sensor ISFET system for soil analysis. In: Precision Agriculture '97, pp. 459-68. Stafford (ed.), BIOS Scientific Publishers, Oxford UK.
Chukwu, E. and C.G. Bowers. 1997. Instantaneous multiple depth soil mechanical impedance sensing from a moving vehicle. ASAE Paper 97-1077. ASAE, St. Joseph MI.
Clark, R.L. 1996. A comparison of rapid GPS techniques for topographic mapping. In: Proceedings of the 3rd Int'l. Conf. on Precision Agriculture, pp. 785-94. Robert, Rust and Larson (eds.), ASA-CSSA-SSSA, Madison WI.
Francis, D.D. and J.S. Schepers. 1997. Selective soil sampling for site-specific nutrient management. In: Precision Agriculture '97, Vol I., pp. 119-26. Stafford (ed.), BIOS Scientific Publishers, Oxford UK.
Franzen, D.W. and T.R. Peck. 1995. Field soil sampling density for variable rate fertilization. J. Prod. Agric. 8:568-74.
Hummel, J.W., and S.J. Birrell. 1995. Real-time soil nitrate sensing. In: 25th North Central Extension-Industry Soil Fertility Conference Proceedings, pp. 125-36. G. Rehm (ed.), Potash & Phosphate Institute, Manhattan KS.
Hummel, J.W., R.R. Price, S.C. Newman, S.J. Birrell, and C.E. Plattner. 1995. Sensors for site-specific crop management. Poster Abstracts-Information Agriculture Conference, pp. 20-21. June 27-30, 1995, Champaign IL.
Jaynes, D.B., T.S. Colvin, and J. Ambuel. 1995a. Yield mapping by electromagnetic induction. In: Proceedings of Site-specific Management for Agricultural Systems, pp. 383-94. Robert, Rust, and Larson (eds.), ASA-CSSA-SSSA, Madison WI.
Jaynes, D.B., J.M. Novak, T.B. Moorman, and C.A. Cambardella. 1995b. Estimating herbicide partition coefficients from electromagnetic induction measurements. J. Environ. Qual. 24: 36-41.
Kitchen, N.R., K.A. Sudduth, S.J. Birrell, and S.T. Drummond. 1995. Spatial prediction of crop productivity using electromagnetic induction. In: Agronomy Abstracts, p. 299. ASA, Madison WI.
Krishnan, P., J.D. Alexander, B.J., Butler, and J.W. Hummel. 1980. Reflectance technique for predicting soil organic matter. Soil Sci. Soc. Am. J. 44:1282-85.
Linsley, C.M. and F.C. Bauer. 1929. Test your soil for acidity. Univ. of Ill. Col. of Agri. and Agric. Exp. Stn., Circ. 346, Urbana IL.
Liu, W., L.D. Gaultney, and M.T. Morgan. 1993. Soil texture detection using acoustic methods. ASAE Paper 93-1015. ASAE, St. Joseph MI.
Liu, W., S.K. Upadhyaya, T. Kataoka, and S. Shibusawa. 1996. Development of a texture/soil compaction sensor. In: Proceedings of the 3rd Int'l. Conf. on Precision Agriculture, pp. 617-30. Robert, Rust and Larson (eds.), ASA-CSSA-SSSA, Madison WI.
McBride, R.A., A.M. Gordon, and S.C. Shrive. 1990. Estimating forest soil quality from terrain measurements of apparent electrical conductivity. Soil Sci. Soc. Am. J. 54: 290-293.
McGrath, D.E., J.P. Ellingson and A.O. Leedahl. 1990. Variable application rates based on soil organic matter. ASAE Paper 90-1598. ASAE, St. Joseph MI.
Roytburg, E., and J. Chaplin. 1995. Stochastic modeling of soil condition during tillage. In: Proceedings of Site-specific Management for Agricultural Systems, pp. 581-99. Robert, Rust, and Larson (eds.), ASA-CSSA-SSSA, Madison WI.
Schafer, R.L., S.C. Young, J.G. Hendrick, and C.E. Johnson. 1981. Control concepts for tillage systems. ASAE Paper No. 811601. ASAE, St. Joseph MI.
Shonk, J.L., L.D. Gaultney, D.G. Schulze and G.E. Van Scoyoc. 1991. Spectroscopic sensing of soil organic matter content. Trans. ASAE 32(3):826-9.
Smith, B.E., R.L. Schafer, and C.E. Johnson. 1994. Using coulters to quantify the soil physical condition. ASAE Paper No. 941040. ASAE, St. Joseph MI.
Stafford, J.V., and B. Ambler. 1990. Sensing spatial variability of seedbed structure. ASAE Paper No. 901624. ASAE, St. Joseph MI.
Sudduth, K.A. and J.W. Hummel. 1993a. Portable near infrared spectrophotometer for rapid soil analysis. Trans. ASAE 36(1):187-95.
Sudduth, K.A. and J.W. Hummel. 1993b. Soil organic matter, CEC, and moisture sensing with a portable NIR spectrophotometer. Trans. ASAE 36(6):1571-82.
Sudduth, K.A., J.W. Hummel and S.J. Birrell. 1997. Sensors for site-specific management. In: The State of Site-Specific Management for Agriculture, Chap. 10, pp. 183-210. Pierce and Sadler (eds.), ASA-CSSA-SSSA, Madison WI.
Sudduth, K.A., N.R. Kitchen, D.F. Hughes, and S.T. Drummond. 1995. Electromagnetic induction sensing as an indicator of productivity on claypan soils. In: Proceedings of Site-specific Management for Agricultural Systems, pp. 671-81. Robert, Rust, and Larson (eds.), ASA-CSSA-SSSA, Madison, WI.
Viscarra Rossel, R.A. and A.B. McBratney. 1997. Preliminary experiments towards the evaluation of a suitable soil sensor for continuous, 'on-the-go' field pH measurements. In: Precision Agriculture '97, pp. 493-501. Stafford (ed.), BIOS Scientific Publishers, Oxford UK.
Worner, C.R. 1989. Design and construction of a portable spectrophotometer for realtime analysis of soil reflectance properties. M.S. thesis. University of Illinois at Urbana-Champaign, Urbana IL.
Young, S.C., C.E. Johnson, and R.L. Schafer. 1988. Quantifying soil physical condition for tillage control applications. Trans. ASAE 31(3): 662-667.
Zuo, Y., D.C. Erbach, and S.J. Marley. 1995. Soil structure evaluation by use of fiber-optic sensor. ASAE Paper No. 951317. ASAE, St. Joseph MI.
