The inherited spatial and temporal variation of soil properties and nutrients is critical for determining VR strategies for crop production, but is typically derived from soil samples data collected at insufficient resolution (typically 1 soil sample per 1-3 ha) to match the underlying variation within the fields. These traditional methods of soil analyses are expensive, time consuming and allow for limited sampling resolution, which is not sufficient to explore within field variation in soil properties. This soil sample actually represents a bulked average of the entire area of the land considered. When an average soil sample is adopted as input information for decision support, within field variability is ignored and poor land management is expected e.g. fertilisation, irrigation, tillage, etc. However, recent advancements on utilising proximal soil sensing for site specific crop management indicate that there are flaws in the current PA technology, in that it does not acquire some key elements in the soil.
Recently new innovative on-line sensors were developed. They are based on visible and near infrared (vis-NIR) spectroscopy, which enables the collection of quantitative values on key soil properties including, organic carbon (OC), total nitrogen (TN), pH, moisture content (MC), phosphorous (P), cation exchange capacity (CEC), magnesium (Mg) and calcium (Ca). The scale of sampling is very fine to allow site specific land management, essential for successful implementation of PA.
Aim of the Project
The project proposes to fuse a set of data on soil and crop together with auxiliary data on topography, land use and weather. Following, management zones will be delineated and utilised for site specific land and crop management including site specific fertilisation, seeding, tillage and irrigation. This functionality will be integrated into a proposed FMIS, which enables end users to access information on crop and soil and recommendations for best crop management practices.