INTEGRATION OF EIT, SOIL STRUCTURE AND PLANT PHYSIOLOGYMODELS FOR 3D IMAGING OF CROP ROOT FUNCTION PHENOTYPE

Abstract

Climate change necessitates breeding programmes for the development of plant varieties that are tolerant to drought and other environmental stresses. Rapid identification of new plant varieties that will thrive in future climates is increasingly important. Presently, above soil features are monitored in industrial greenhouses and field trials. Indicators of preferential genetic traits are often based on loosely related features (phenotypes) such as an extra ear of corn on a maize plant. The root system is critical to plant water uptake, however this cannot easily be assessed without destroying the crop or disturbing the plant/soil matrix through extractive sampling. A new visualisation tool is being developed for seed breeders, providing on-line data for each individual plant in a screening programme. It will be used to indicate how efficiently each plant utilises the water and nutrients available in the surrounding soil. This will facilitate early detection of desirable genetic traits with the aim of increased efficiency in identification and delivery of tomorrow's drought tolerant food crops. Visualisation takes the form of Electrical Impedance Tomography (EIT), a non-destructive and non-intrusive imaging technique. Measurements are to be obtained for individual plants thus allowing water absorption levels for each specimen to be inferred. This document will discuss progress made in the design and implementation of a new model-based 3D reconstruction method for the interpretation of data acquired using the University of Manchester LCT2 ERT data acquisition instrument. There were two main aspects of the project: (1) A multiphysics based forward solver was implemented in COMSOL Multiphysics. This forward model is comprised of a fluid flow model (Richards' equation), coupled to an electrostatic model (Poisson's equation) via a mixture model (Archie's Law). Each individual element was validated experimentally in isolation and the results are presented, finally the integrated model is demonstrated. (2) Two inversion schemes are proposed, these schemes utilise the multiphysics forward solver described above in combination with both traditional inversion techniques implemented in the EIDORS reconstruction suite and Kalman Filter techniques to form a hybrid reconstruction scheme and also a solver based solely on the Kalman Filter to provide an inversion free, iterative parameter estimation scheme. The advantages and disadvantages of each method are discussed and suggestions regarding their implementation are given.

Date of Award	1 Aug 2015
Original language	English
Awarding Institution	The University of Manchester
Supervisor	Oliver Dorn (Supervisor) & Bruce Grieve (Supervisor)