In-Paddock Variability of Plant Available Water
PhD Candidate: Peter Weir
With the adoption of Precision Agriculture (PA) and implementation of site-specific crop management (SSCM) by farmers, there is a better understanding of both spatial and temporal variabilities in cropping fields and to a lesser extent, pastures. Large amounts of data have been acquired by farm businesses by sampling and analysis, and from monitors on harvesters, with the aim of improving crop and soil management. These ‘big data’ sources have the potential to improve the management and efficiency of farm enterprises.
The utilisation of disparate legacy spatio-temporal datasets to determine in-paddock variability of soil attributes associated with crop production and pasture management is to be investigated, focussing on assembling, interrogating, and cross analysing the data that has already been collected by farmers.
Development of a conceptual model, using existing spatio-temporal data as the basis for predictions of within-paddock variability in plant-available water (PAW), which will consider local constraints, and uncertainty in the data. Assuming a credible and robust fine-scale PAW model can be achieved, the variation in annual yields against the variations in seasonal PAW could be correlated to elicit information about the paddocks and soils that can feed into predictive models. Predictions will be based on a range of publicly available data sets including satellite imagery, landscape metrics, soil moisture measurements, digital soil data, and meteorological data. Multi-variate statistical analysis and machine learning techniques will be used to develop the prediction algorithms.
Soil moisture is one of the most common requirements in decision support systems used by farmers, agronomists, and agricultural services. These outputs will contribute to the decision support systems running “under-the-hood”, contributing to the “dashboard” display being developed as part of other research projects currently underway.