High throughput phenotyping of cotton in multiple irrigation environments

Rapid screening of plant growth can improve selection of cotton (Gossypium hirsutum L.) in breeding and productivity analysis. We tested several automated phenotyping methods, including measurements of plant height, ground cover fraction (GCF), Normalized Difference Vegetation Index (NDVI), and canopy temperature using a ground-based platform mounted on a research sprayer. In addition, we tested the effects of ten irrigation levels on crop growth, in-season stress indicators, final boll distribution, and fiber quality. The system was evaluated on sixteen cotton cultivars grown during the 2011, 2012, and 2013 seasons in Lubbock, Texas. All measurement parameters showed differences among irrigation levels during the study years. Rainfall and weather conditions during the 2013 season resulted in taller plants and higher yields, but fewer differences among irrigation treatments. Lint yield was positively correlated with each in-season growth parameter during the 2012 and 2013 seasons. The results support the following statements: ground based near remote sensing system can be used to phenotype multiple traits rapidly and precisely over multiple irrigation levels to screen cultivar and irrigation to maximize yield; ground cover estimation can be performed using a visible vegetation index; canopy architecture can be differentiated using canopy sensors; boll distribution can be tied to irrigation level over a variety of environmental conditions; and fiber quality has both a genetic and environmental component.