Feasibility of Using Remote Sensing Techniques to Detect Spider Mite Damage in Stone Fruit Orchards

To investigate the feasibility of using remote sensing to detect pest damage in stone fruit orchards, we estimated spider mite damage of 392 tree canopies in 2006 and of 1153 leaves in 2007 in nine peach orchards in California. We used Partial Least Squares regression to correlate the spectral reflectance between 350 and 2500 nm of all trees and leaves with mite damage. We tested nine published indices and normalized difference reflectance indices calculated from the most significant wavebands for correlation with mite damage. Significant bands appeared in the blue/near UV and red parts of the spectrum for both datasets. The most promising band combinations for predicting mite damage were 390 and 651 nm for canopies and 687 and 744 nm for leaves. Substantial scatter in the dataset and significant effects of sampling site and season will have to be overcome, before this technique can be applied in practice.