Source: VERIS TECHNOLOGIES, INC. submitted to
SOIL CORE ANALYSIS NETWORK FOR IN-FIELD MEASUREMENTS OF NITROGEN AND OTHER SOIL PROPERTIES
Sponsoring Institution
National Institute of Food and Agriculture
Project Status
TERMINATED
Funding Source
Reporting Frequency
Annual
Accession No.
0213562
Grant No.
2008-33610-18915
Project No.
KANK-2008-00380
Proposal No.
2008-00380
Multistate No.
(N/A)
Program Code
8.4
Project Start Date
Jul 1, 2008
Project End Date
Feb 28, 2010
Grant Year
2008
Project Director
Lund, E. D.
Recipient Organization
VERIS TECHNOLOGIES, INC.
601 N. BROADWAY
SALINA,KS 67401
Performing Department
(N/A)
Non Technical Summary
Fertilizer application rates that do not match crop usage pose an economic loss for farmers, and excess applications can result in environmental degradation of water and atmosphere. A significant portion of the nitrogen applied to U.S. fields is not needed, due to the availability of nitrogen from the soil. Soil nutrients, especially nitrogen, vary spatially and temporally, within the field and soil profile. In order to deal with the significant variability challenges, a large number of soil measurements must be taken on each field. Using conventional lab analysis, this is not cost-effective. In-field measurements represent an appealing alternative, yet these must be accurate and affordable. This project will develop the field-deployable test equipment required to evaluate several sensor technologies in a side-by-side comparison. Soil sensing technology used for this application will be electrical, optical, and electro-chemical. The sensors to be used in this project represent the most viable candidates from those categories, and have shown initial feasibility to meet the criteria. However, most have not been widely tested under in-field conditions, nor has a thorough side-by-side comparison been conducted. In order to perform this feasibility comparison test, equipment will be devised to collect and process the soil cores, and bring them into contact with the sensors. Results from in-field sensors will be compared with laboratory-analyzed soil tests.
Animal Health Component
100%
Research Effort Categories
Basic
(N/A)
Applied
100%
Developmental
(N/A)
Classification

Knowledge Area (KA)Subject of Investigation (SOI)Field of Science (FOS)Percent
1020110202060%
1020110206140%
Knowledge Area
102 - Soil, Plant, Water, Nutrient Relationships;

Subject Of Investigation
0110 - Soil;

Field Of Science
2020 - Engineering; 2061 - Pedology;
Goals / Objectives
Soil-supplied nitrogen is not accounted for in many regions due to the difficulty in measuring it accurately. Soil nutrients, especially nitrogen, vary spatially and temporally, within the field and soil profile. Using conventional lab analysis, it is not cost-effective to accurately characterize the nitrogen variability on most fields. In-field measurements represent an appealing alternative, yet these must be accurate and affordable. As it is unlikely that a single measurement will be able to account for the complex biological and chemical factors affecting optimal crop nutrient rates, this project will evaluate several sensor technologies in a side-by-side comparison. Soil sensing technology that may ultimately be deployed in-field can be categorized as electrical, optical, and electro-chemical. The sensors planned for this proposal represent the most viable candidates from those categories, and have shown initial feasibility to meet the criteria. However, most have not been widely tested under in-field conditions, nor has a thorough side-by-side comparison been conducted. In order to perform this feasibility comparison test, equipment will be devised to collect and process the soil cores, and bring them into contact with the sensors. Objectives for this project are: Objective #1: Develop equipment that will being soil into contact with the soil sensors, and determine feasibility of producing and using this equipment in various soil and field conditions. Objective #2: Conduct field trials with the sensor devices. Objective #3: Compare results from each sensor with lab-analyzed soil samples. Objective #4: From the results of this Phase I project, establish a clear direction for the development of an in-field sensing system during a Phase II. The thorough investigation of sensor candidates during the Phase I portion will insure that the system is optimized for accuracy, affordability, and timeliness.
Project Methods
The three main areas that offer an opportunity for improved fertilizer efficiency are: 1) measuring profile soil nitrate and nitrogen mineralization potential with adequate spatial resolution, in a cost-effective and timely manner, 2) accompany those measurements with improved measurements of other soil factors such as pH, soil moisture, and compaction that affect mineralization and nitrogen use and loss, and 3) simultaneously measure other soil nutrients such as potassium with increased spatial resolution. In order to optimize a system, a side-by-side comparison of the best candidates will be conducted. While the sensors that will be used in this research have all been the subjects of various research initiatives, and results have been published, much of the work was done in a lab setting with very few sensor comparisons. Obstacles to such a test include the lack of suitable field equipment needed to prepare the samples, and to accommodate the sensors in order measure the samples. The project will address those research gaps in several ways. First, by developing equipment that will collect sensor measurements of the soil profile. This will consist of NIR and Mid-IR spectroscopy, ion-selective electrodes, and soil electrical property measurements. Second, by calibrating sensors to lab measurements. Sensor measurements will be calibrated to laboratory measurements of nitrate N, amino N, total nitrogen, total carbon, P, K, pH, magnesium (Mg), calcium (Ca) and cation exchange capacity (CEC). Calibration methods will include single and multi-variate regression for electrical and electrochemical measurements, chemometric techniques such as partial least squares regression for NIR and Mid-IR, along with wavelet analysis and regression techniques for Mid-IR. Third, by comparing results from all sensors. Validation of results will be accomplished using leave-one-out cross-validation (i.e., predicted R-sq. values), splitting the data set into calibration and validations sets, and/or use of independent data sets.

Progress 07/01/08 to 02/28/10

Outputs
OUTPUTS: The Phase I project evaluated several methods of in-field measurement of soil properties using cores extracted by a hydraulic soil sampler. Cores were extracted from each of 12 barrels inserted in three different fields (sandy, silt loam and silty clay loam) with four different concentrations of ions (zero, low, medium and high N and K fertilizer applied). The order of testing for each location proceeded as follows: 1) measured soil in situ using visible and near-infrared spectroscopy (Vis-NIRS) and soil electrical conductivity (EC) sensor probe, 2) collected 1 meter cores adjacent to sensor probe insertions and scanned intact core with Vis-NIRS, 3) sensed core with miniaturized four electrode Wenner array EC probe, and 4) segmented core, created soil solutions and measured with ion-selective electrodes (ISEs), and EC electrode. Sub-samples of each core segment were sent to a soil testing lab for analysis, and lab results were used to validate each in-field measurement approach. None of the outputs have been disseminated to the public, as they are proprietary and form the basis of a Phase II proposal. PARTICIPANTS: Consultants and sub-contractors on this project include: Dr. Shrinivasa Upadhyaya with University of California-Davis, Dr. Viacheslav Adamchuk with the University of Nebraska-Lincoln, and Dr. Ken Sudduth with USDA-ARS. TARGET AUDIENCES: Not relevant to this project. PROJECT MODIFICATIONS: Nothing significant to report during this reporting period.

Impacts
The results of Phase I provide clear direction for a Phase II project. By comparing on a level playing field the frequently cited candidates for in-field nitrate and potassium measurements, the most viable technology for commercial in-field measurements was identified. One technology showed superior ability to perform high-quality measurements of unprocessed soil cores. These cores were collected from heterogonous soil textures ranging from 62% sand to 84% silt to 36% clay, and other soil properties varied significantly as well. This is encouraging, as it demonstrates certain sensor technologies are able to detect field-typical nitrate and potassium levels against a background of soil variability, just as they will need to do in a commercial system. A rationale for developing an effective automated system has been developed. Issues of soil core grinding and dispersing, sensor accuracy, response time, drift, and durability have been thoroughly examined. None of the outcomes have been disseminated to the public, as they are proprietary and form the basis of a Phase II proposal.

Publications

  • No publications reported this period


Progress 07/01/08 to 06/30/09

Outputs
OUTPUTS: To briefly summarize the results to date: Testing of Vis-NIR spectroscopy and soil EC has shown that a core-scanning/measuring bed is feasible. Results from the scanning bed are correlated with in-situ measurements for soil EC and carbon, which shows the device is performing properly. Whether those specific sensors are appropriate for nitrate and potassium remains to be seen, but the device is ready for those tests. Testing of nitrate and potassium ion-selective electrodes has revealed promising results, with acceptable errors of soil solutions containing nitrate and potassium levels that are typical of production agriculture fields. The spectroscopy consultant on this project, who has authored several papers on using mid-IR spectroscopy to identify nitrate levels in soil solution, has encountered difficulty replicating his earlier work, using soil samples provided for this project. The consultant is taking the necessary steps to be able to complete the mid-IR portion. Various preliminary tests have been conducted using lab-analyzed samples, in order to test concepts and systems. Here is a listing and description of the various tests: 1) Core scanning with NIRS, mini EC probe/fork, and ISE's: Cores were extracted from the soil using a Veris Technologies hydraulic probe, scanned by the NIRS core scanner, probed, segmented, measured with mini-EC electrodes, pH ISEs, and analyzed by a soil testing lab. 2) Soil solution measurements with ISE in soil solution: Initially, soil solutions of various concentrations were used to test electrode response. Ground and dried soil samples from KSU archives, with a field-typical range of potassium and nitrate levels, were used to further evaluate response and measurement criteria. Here are the initial results: 1) Core scanning with NIRS, mini EC probe/fork, and ISE's: 25 samples were scanned by the core scanner with the VIS/NIR spectrometers and the data was calibrated by PLS regression to predict Carbon contents in the soil cores. The results showed R2 = 0.655, RMSE = 0.243, RMSECV = 0.322, and RPD = 1.74. While carbon is not the focus of this project, calibrations were developed for C using previously analyzed samples in order to compare the scanning table vs. in situ probe. In a similar manner, soil EC from the mini-EC fork was compared to in situ probe EC measurements. Also using the core scanner bed, a metalloid antimony pH electrode inserted into the core at multiple locations along the core, generating profile pH measurements. 2) For Nitrate measurement using Ion Selective Electrode, NICO N sensor showed faster response and more reliable than ASI N sensor even though both electrodes have long time response (at least 50 sec). NICO N sensor output was highly correlated to the Lab data for soil core segments with R2 = 0.89 and RMSE = 4.01. For predicting exchangeable Potassium contents in the segments, Lab K data and ASI K ISE data with CEC values were calibrated by multiple linear regression analysis. This approach showed very high exchangeable K prediction with R2 = 0.87 and RMSE = 62.75. In future tests, EC values will be included rather than CEC for multiple linear regression analysis. PARTICIPANTS: Nothing significant to report during this reporting period. TARGET AUDIENCES: Nothing significant to report during this reporting period. PROJECT MODIFICATIONS: Nothing significant to report during this reporting period.

Impacts
As this project is only halfway through the first phase, final outcomes and impacts are difficult to predict. Several technical and commercial hurdles remain before completion. At this point a core-scanning/measuring bed is technically feasible, however its marketability is questionable, unless measurements conducted in the second part of the Phase 1 project reveals higher than expected correlations to nitrate or potassium. Testing of nitrate and potassium ion-selective electrodes has revealed promising results. These have been completed using prepared samples; the second part of the project will use field-moist cores collected directly from the field and inserted into the slurry mechanism. If the accuracy level remains high, those electrodes would likely form the basis of a Phase II proposal. One aspect of this project is to rule out from future projects sensors that are not field-deployable, and do not meet measurement criteria. While the final results are not yet known, early indications are that mid-IR and polypyrrole sensors will not meet the needed field unit parameters.

Publications

  • No publications reported this period