Environmental Drivers of Quinoa Germination: Growth Responses and Machine Learning Predictions

Abstract

Climate change is intensifying abiotic stresses such as salinity and temperature fluctuations, posing serious challenges to crop establishment during seed germination. This study assessed the germination responses of five quinoa (Chenopodium quinoa Willd.) genotypes under six salinity levels (0–250 mM NaCl) and six temperature regimes (10–35 °C). A completely randomized design with three replications was used, and germination rate, seedling length, and seedling biomass were evaluated. Data were analyzed through three-way ANOVA, Tukey’s HSD, Pearson correlation, and principal component analysis (PCA). In addition, a Random Forest model was applied to predict germination traits across environments. Temperature, salinity, genotype, and their interactions significantly influenced all traits (p < 0.001). Severe inhibition occurred at salinity ≥ 150 mM and at extreme temperatures (10 and 35 °C). The genotype Atlas exhibited superior tolerance, maintaining higher germination, elongation, and biomass, whereas Q4 Bis and Bear Canyon were highly sensitive under combined stress. PCA highlighted distinct genotype clustering and strong trait correlations. Random Forest predictions closely aligned with observed values and confirmed temperature and salinity as the key predictors of germination performance. These findings provide a robust framework for identifying stress-tolerant quinoa genotypes and offer practical insights for breeding and cultivation strategies in the face of climate change.