While traditional remote sensing techniques could only detect combined nitrogen and water deficiencies with 40–55% accuracy, the team’s new approach boosted that figure to between 65% and 90%