Current Projects


Lead Institution: Kansas State University
Carlton Leuschen (KU –PI, Fernando Rodriguez-Morales, Co-I)
Sponsor: USDA/National Institute of Food and Agriculture

Wheat is a globally important grain at the forefront of food security issues. But like rice, maize, and sorghum, wheat is achieving barely 50% of the annual yield progress rate necessary to meet the food needs widely forecast for 2050. For over ten years, researchers have developed an approach melding ecophysiological and quantitative genetic modeling with the potential to accelerate breeding rates of gain. This entails a two-step process that first fits crop models to data and then association maps the resulting parameter values to genetic markers. However, technological limits impede collection of the large amounts of needed plant trait data, especially the geometry of dense plant canopies. Targeting Kansas wheat breeding trials, this project is a proof-of-concept test combining microwave radar sensing with a novel, inversion algorithm to ameliorate the situation. The basic rationale is that (1) it is unnecessary to sense the 3D position, angle, and size of every tiller and leaf in a trial plot - rather one desires the genetic markers and effect sizes associated with these quantities' statistical distributions; (2) models interrelating markers and morphology exist; (3) if radar calculations for plant canopies can be accelerated, then the models in (2) can be inverted to yield genetics in a single-step; and (4) an extension of the Analytical Element Method (AEM) from hydrology to electromagnetic (EM) wave propagation can provide such a speed up.Briefly, the AEM exactly solves the field equations for very simple shapes that are then combined to yield machine accurate-answers for complex geometries. Unlike solvers in common use, the AEM only calculates solutions at the specific points of interest, thus hugely reducing computational loads. Prior work has found AEM solutions for EM waves in two dimensions. This project will extend those solutions to full 3D.Concurrently, an existing wheat model that predicts highly realistic plant shapes will be modified so its outputs are expressed in terms of the AEM basic shapes. A three-layer model will then be built comprising [genetic markers : plant shapes : EM fields] and solved by probabilistic methods. This will yield the genetic markers most associated with the plant shapes sensed by radar. The method will be tested by team members with radar expertise using the facilities of the Center for Remote Sensing of Ice Sheets. Experiments in a large anechoic chamber will compare AEM predictions to actual radar reflectance data for simplified targets. The EM properties of wheat at radar frequencies will also be measured in the chamber using small, movable plots.Based on these data, a prototype field system will be constructed and used to gather plot data in a field trial conducted as part of the on-going Kansas wheat breeding program. Two tests will be performed. First the radar data will be association mapped directly to detect any responses to genetically determined canopy features. If positive results are found, they will be compared to published phenotypic mapping studies and hypotheses developed as to features to which the radar might be responding. The second test will solve the three layer model described above and also compare the results to literature.