|Sudduth, Kenneth - Ken|
Submitted to: Computers and Electronics in Agriculture
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: 4/25/2001
Publication Date: 6/11/2001
Citation: Hummel, J.W., Sudduth, K.A., Hollinger, S.E. 2001. Soil moisture and organic matter prediction of surface and subsurface soils using an NIR soil sensor. Computers and Electronics in Agriculture. 32:149-165.
Interpretive Summary: As agricultural tractors and field equipment get larger, the concern about soil compaction increases. Larger land areas being farmed by the same set of machinery increases the possibility that field operation will occur when soil is easily compacted. Instrumentation is needed to measure when soil is compacted to a level that, without additional tillage, will restrict crop root growth. Furthermore, instrumentation is needed that can acquire data across a range of soil types and soil moisture contents, and provide a measure of compaction that is independent of soil moisture content. Soil cone penetrometers are currently used to measure soil penetration resistance, but penetration force is affected by soil moisture as well as soil compaction. This paper reports on research to show that a USDA-Agri. Res. Ser. designed and patented soil organic matter and soil moisture sensor that can estimate soil moisture in the plow layer, can also be used to estimate soil moisture of soil at greater depths. Previous research has shown how the moisture sensor might be incorporated into the soil cone penetrometer to address this need. The research shows that the sensor can measure organic matter and moisture of soil below the plow layer, and that, especially for soil moisture, a simplified sensor might be sufficient. Commercialization of this concept could significantly increase the usefulness of soil penetration resistance data, and lead to reduced primary tillage when used to measure soil compaction.
Technical Abstract: Quantification of spatial variability of soil parameters is important to the successful implementation of Site-Specific Management (SSM). There is a need for sensors to more accurately characterize within-field variability. Extensive laboratory tests using a range of surface horizon soil types were used in the development of a near infrared (NIR) reflectance soil sensor, which is capable of predicting soil organic matter and soil moisture across a range of surface horizon soil types. This paper reports research documenting the ability of the sensor to predict soil organic matter and soil moisture contents of B-horizon soils. Three soil cores (5.56 cm dia. x 1.4 m long) were collected at each of 16 sites across a 144,000 km2 area of the U.S. Cornbelt. Cores were segmented and frozen prior to subsampling at eight depth increments. Samples were tested at six soil moisture tensions ranging from air-dry to saturated. Spectral reflectance data (1603 nm - 2598 nm) were obtained in the laboratory on undisturbed soil samples, using the same procedures as reported in previous studies. Data were collected on a 6.6 nm spacing with each reflectance value having a 45 nm bandpass. The data were transformed from reflectance to optical density [OD, defined as log10 (1/reflectance)], normalized, and analyzed using Stepwise Multiple Linear Regression. Data for a surface soil layer was removed from the dataset, resulting in improved predictive capability. Soil organic matter was predicted with a Standard Error of Prediction of 0.62 % organic matter, and an RPD of 2.06. Better predictive capability was demonstrated for soil moisture, with a Standard Error of Prediction of 5.31 % soil moisture and an RPD of 2.78.