Describir: Hybridization of process-based models, remote sensing, and machine learning for enhanced spatial predictions of wheat yield and quality