Predicting leaf traits in wine grapes with reflectance spectroscopy.

Area of Science:

• Plant science
• Agronomy
• Spectroscopy

Background:

• Estimating crop traits is vital for predicting responses to environmental change and optimizing management.
• Traditional plant trait quantification methods are time-consuming and resource-intensive.
• High-throughput phenotyping, particularly reflectance spectroscopy, offers rapid trait estimation.

Purpose of the Study:

• To assess reflectance spectroscopy's ability to quantify intraspecific variation in 12 leaf traits across 12 wine grape cultivars.
• To determine if spectral data can detect fine-scale variation in resource acquisition and plant-water traits among cultivars.
• To evaluate model accuracy for predicting different types of leaf traits.

Main Methods:

• Utilized wine grapes (V. vinifera subsp. vinifera) from seven varieties and 12 cultivars.
• Applied reflectance spectroscopy to estimate 12 leaf traits.
• Developed predictive models using spectral reflectance data for trait quantification.
• Assessed model performance using R-squared values.

Main Results:

• Significant trait variability was observed across and within cultivars, particularly in gas-exchange and hydraulic traits.
• Cultivars exhibited variation along a resource-conservative-to-resource-acquisitive trait axis.
• Spectroscopy models successfully differentiated and predicted variation for seven plant traits (R2 = 0.12-0.57).
• Models for leaf chemical, physiological, and morphological traits were more accurate than those for leaf water status.

Conclusions:

• Reflectance spectroscopy can capture dimensions of fine-scale trait variation within genetically diverse agroecosystems.
• The method shows promise for high-throughput phenotyping of crop cultivars.
• Spectroscopic estimation of intraspecific variation in leaf water status requires further refinement.