Principal Investigator: David Hooker and Nicole Rabe
Research Institution: University of Guelph and Ontario Ministry of Agriculture, Food and Rural Affairs (OMAFRA)
Timeline: March 2014 – March 2015
- Develop a robust dataset of gamma and soil sample measurements for calibration across variable farm landscapes and long-term experiments in Ontario with inherent variability of various soil parameters.
- Determine which soil parameters are well correlated with gamma measurements towards the development of high-resolution soil maps for identification of management zones for precision agriculture in Ontario (i.e., identify strengths of the technology).
- Determine which soil parameters are poorly predicted using gamma measurements in Ontario (i.e., identify weaknesses and limitations).
- Determine any temporal variability between gamma and soil sampling measurements across robust datasets.
- Characterizing the spatial variability of fields enables the potential for site-specific management and optimizing inputs depending on spatial position in the field. Fast and inexpensive methods to estimate soil parameters would be useful.
- Third-party validation of gamma sensing is limited.
The deployment of precision agriculture has the potential to improve efficiencies of crop production across a spatially variable landscape. Through precision agriculture techniques, nutrients can be applied variably across fields to ensure the correct timing and placements at the rate required in a specific management zone. Successful implementation of precision agricultural technology requires the identification of zones across farm fields that respond differentially to management. However, site- or zone-specific management cannot be deployed until zones can be identified and characterized to show temporal stability in predictions (i.e., year-to-year) with differences in crop management prescriptions. Mapping fields for potential management zones using conventional soil sampling and laboratory analysis is laborious, time consuming, and expensive. Characterizing the spatial variability of soil using sensors would be much more cost efficient and convenient, with much higher resolution of soil data information. One such sensor measures gamma radiation that is naturally emitted from the soil. The gamma radiation spectra have been used to characterize various soil physical and chemical properties within the tillage layer, such as soil texture, soil nutrients, soil organic matter, bulk density, and pH.
The project aimed to develop and enhance gamma ray sensor technology to determine the accuracy and sensitivity of predicted values to soil parameters as they vary spatial across each field for generating high resolution soil maps which could then be used to identify management zones across farm fields in Ontario. The 8 field locations chosen for this study represented a wide range of soil parameters within each field location, which should represent many farm fields across Ontario. Preliminary analysis from only a one‐year dataset shows that the gamma ray sensor predicted some soil parameters with “reasonable accuracy” across most field locations. A reasonable accuracy from this sensor allows for very fast data acquisition, which would then produce high resolution maps that will improve fertilizer use efficiencies and other management strategies. However, some soil parameters were not accurately predicted, and some parameters were predicted better on some sites than others. With this new knowledge, algorithms that were used to predict certain soil parameters may need to be recalibrated using Ontario soils. According to the data generated in this study among the field locations, of the soil textural components, the % sand and % clay within each field at either the 0‐15 cm or 0‐30 cm depths were predicted with reasonable accuracy. Soil organic matter in the surface 0‐15 cm was predicted with reasonable accuracy, whereas the predictions of SOM in the surface 0‐30 cm were not as accurate. Predictions were fair‐ poor for the 0‐15 and 0‐30 cm depths for soil pH and soil test K, and predictions were the poorest for soil test P at either the 0‐15 or 0‐30 cm depths.
Ontario Ministry of Agriculture, Food and Rural Affairs.
Project Related Publications: