Accounting for Soil Variability in Novel GPR Applications of Root Phenotyping and Soil Organic Carbon Quantification

The ability to phenotype roots in situ would provide information for carbon sequestration potential through increased root mass, possible water-seeking strategies by plants, and data generation for plant breeders. However, current phenotyping techniques are often labor intensive and destructive to the observed plant. One potential phenotyping technique that is both rapid and nondestructive is ground penetrating radar (GPR). This technology has been proposed due to its ability to detect fine-scale differences in dielectric permittivity, which is strongly influenced by soil moisture content. To detect small differences in soil moisture caused by root growth, we will need to account for the soil signature in the GPR signal. The objective of this study was to test the feasibility of GPR data to be linked with soil electromagnetic data as a means to detect and visualize a rooting system in different soil textural classes. Additionally, GPR’s potential as a device for quantifying soil organic carbon (SOC) was explored. Similar to current root phenotyping techniques, the agricultural field lacks a tool that can rapidly and non-destructively measure SOC in the field. Like root phenotyping, GPR may be a potential solution due to its ability to detect small scale changes in soil moisture in response to changes in SOC. The root phenotyping portion of this study focused on multiple field locations across Texas and one controlled experiment to simulate in situ and ideal conditions. GPR measurements were taken within each plot, along with multiple measurements of soil moisture to account for soil variability. Measurements were collected throughout the growing season and unique post processing techniques were explored to aide in root detection/visualization. GPR’s ability to distinguish root types across different soils conditions were assessed. The SOC portion of this study focused on three sites across the United States to capture the largest range of SOC levels as possible. GPR data was collected on multiple plots at each location, as well as ancillary soil data. Statistics were developed from these measurements and compared with pre-recorded SOC levels to determine GPR’s ability to detect differences in SOC