The use of spectral indices such as normalised difference vegetation index
(NDVI) is becoming an important element in the evaluation of physiological
traits in cereal crops. Determining the correlation between spectral
readings taken at different phenological stages of wheat plants and grain
yield (GY) is crucial. This study aimed to demonstrate the effectiveness of
NDVI in combination with the GGE biplot model in identifying high yielding
durum wheat genotypes. Field experiments were conducted in multiple
environments and growth stages to evaluate the relationship between NDVI and
grain yield (GY). Twenty-five durum wheat genotypes were tested for two
years under rainfed and supplementary irrigation conditions using a split
plot experimental design. The results showed significant positive
correlations between NDVI and GY, especially at the generative stages under
rain-fed conditions. In contrast, NDVI at the grain-filling stage showed a
weaker relationship with GY under supplemental irrigation. Genotypes
exhibiting high NDVI values up to the stem elongation stage tended to have
lower grain yields, emphasizing the importance of considering growth stage
dynamics. GGE biplot analysis provided visual information on genotype-trait
relationships, helping to identify genotypes with consistent performance in
different environments. This study highlights that NDVI can be used to
predict yield potential and guide selection in durum wheat breeding programs
and the GGE biplot model serves as a valuable tool for genotype evaluation
and selection.