In our world’s currently challenging climate, precision agriculture proves to be crucial to maximize crop yields while minimizing negative environmental effects. The goal of precision agriculture is to provide the tools so each plant can have the exact amount of water, fertilizer, pest protection and other resources it requires. Unlike some conventional agriculture practices where the entire field is treated uniformly, treating the entire area the same way. Using watering as an example, inaccurate practices may lead some plants to receive too much water, and some will not receive enough. To implement a precision agriculture system, it is essential to map the variations in soil texture and composition across the entire field.
BARD-funded investigators, Prof. Michael (Iggy) Liator of Migal Galilee Technology Centre and Prof. Raj Khosla of Kansas State University from the department of Agronomy worked jointly on a project aimed on discovering how to quantify the soil spatial variations. There main objectives were to test the feasibility of using a wide-ranging matrix of ancillary data (including ECa, multispectral bare-soil indices and thermal imagery) to delineate management zones for establishing precision agriculture practices in model farms.
They also wished to formulate a cost-effective multi-objective sampling design problem. They used this to investigate, develop and validate algorithms for solving the derived optimization problems using ancillary data extracted from the model farms. This data would help evaluate the usefulness of remote sensing to reduce soil sampling and analysis in terms of crop-available nutrient concentrations particularly Nitrogen.
Finally, they worked to compute digital maps of recommended fertilization inputs to compare crop yield in plots under the recommended variable application rates versus common uniform applications.
The Israeli investigators found the clustering analysis enabled to sample 22 soils for a fair description of the soil spatial heterogeneity. The reduction in soil sampling was one of the main objectives of the research project. Spatial analysis of the N, P, and K using the soil samples revealed that the parts of the study area showed significant K deficiencies, while available P was quite abundant in the soils. Based on this analysis the VRA cost only 1,930 NIS versus 4,450 NIS using the common application.
The US team of investigators introduced a hybrid fuzzy inference system as a new approach to delineate site-specific management zones. In this study, organic matter and pH have shown a significant difference among the zones. Soil samples were analyzed with the ASD hyperspectral sensor to generate reflectance curves. Spectral information divergence was used to conduct spectral matching by filtering out samples that have lower spectral information divergence values.
To address the concern and theoretical questions concerning effective soil sampling using a minimal set of points while aiming for maximal information. A multi-objective sampling design problem was formulated to optimize the number of sampling locations with stratification of feature space and temporal yield variability.
According to the most recent report published by USDA in 2023, approximately 30% of the farmers have adopted precision agriculture (PA) across the US. The USDA report suggests significant more work needs to be accomplished to enhance the pace of adoption. Currently, in many instances precision technologies are ahead of science, we need to continue to investigate scientifically sound solutions that farmers could understand and rely on comfortably while implementing those solutions in their farm fields. Farmers are looking for economically affordable solutions and this study provides them with some meaningful solutions to further enhance the pace of precision agriculture adoption in the USA.
