Skip to main content
SLU publication database (SLUpub)

Abstract

Climate change is increasing the frequency and intensity of drought, which hampers wheat productivity from meeting the growing food demand worldwide. Therefore, improvements in yield under drought are urgently needed. This work evaluated a diverse set of 77 winter-wheat lines for two image-based early vigor traits and 15 mature traits of diverse winter-wheat lines. Early and late drought treatments were applied 12 and 65 days after vernalization, respectively. Further, a machine-learning-assisted phenotyping technique was adopted to measure spike area. Old Swedish cultivars showed the lowest early root vigor (4.92 cm) and large root biomass at maturity (5.25 g). No positive correlation was found between root biomass and yield components under the control condition. A high mean of grain yield was obtained in 1RS (9.8 g/plant), 2RL (9.5 g/plant), and cfAD99 (9.5 g/plant) genotypes under control. When including stability across control and two drought treatments, NGB, 1RS, 2RL, and cfAD99 genotypes showed the best performance. Peduncle length, root biomass, and NDVI positively contributed to the grain yield of 2RL genotypes under early drought, while 1000-grain weight and root biomass accounted for the high grain yield of 1RS genotypes under late drought. The image-based spike area measured by a machine-learning model correlated strongly to the yield component grain number (R 2 = 0.70***). Furthermore, combined with yield reduction results, the spike area results suggested increased sterility (empty spikes) as the main cause of drought-induced yield loss instead of changes in spike size. Further integration of traditional measurements, modern phenotyping, and computational image analysis is needed to accelerate evaluations of plant traits under drought conditions. Genes potentially governing drought tolerance can be identified in these diverse lines.

Keywords

drought; early vigor; image-based phenotyping; machine learning model; winter wheat

Published in

Food and energy security
2025, volume: 14, number: 5, article number: e70116
Publisher: WILEY

SLU Authors

UKÄ Subject classification

Artificial Intelligence
Agricultural Science

Publication identifier

  • DOI: https://doi.org/10.1002/fes3.70116

Permanent link to this page (URI)

https://res.slu.se/id/publ/143897