Estimating early season growth and biomass of field pea for selection of divergent ideotypes using proximal sensing

Abstract

The aims of this study were to (i) test ground and aerial-based remote sensing vegetation indices (VIs) for trait-based breeding line selection, (ii) improve our understanding of the association between measured plant traits and readings derived from active and passive sensors and (iii) establish an optimal time for growth assessments in relation to field pea vigour and seed yield. Multispectral sensors were deployed with the handheld Crop Circle (CC) and a sensor mounted on an unmanned aerial vehicle (UAV) to collect data from field trials conducted between 2017 and 2020 at Beulah and Horsham in Victoria and Yenda, Wagga Wagga and Ardlethan in New South Wales in Australia. The result showed that normalised difference vegetation index (NDVI) derived from an aerial-based passive sensor (UAV) was strongly and significantly correlated to NDVI derived from a ground-based active sensor (CC) at both Beulah (R2 = 0.85; n = 1165; p < 0.001) and Horsham (R2 = 0.77; n = 210; p < 0.001). Both methods showed similar NDVI trends in pea genotype rankings. Based on the three seasons of field trial data, NDVI derived from both the CC and UAV sensors were linearly related to biomass production during pre-canopy closure growth. In water limiting environments, seed yield was positively correlated to NDVI measures. Measures calculated from the area under the NDVI curve throughout the growth season, and an additive main effect and multiplicative interaction model (AMMI) identified varieties with high vigour scores (high NDVI). Overall, a high vigour score was correlated to seed yield in lower yielding environments. From these results it appeared that higher vigour helps achieve higher yields in drier environments, however it was correlated with lower yields in better environments.