UC Davis

We explored remote sensing methods for their potential to distinguish aphid- (Aphis gossypii Glover) and spider mite-infested (Tetranychus spp.) cotton (Gossypium hirsutum L.) from uninfested cotton. Field plots were established using selective and disruptive pesticides to establish a range of apbid and mite populations over 2 yr. Aerial and satellite remote sensing data in 2003 and 2004 were supplemented with ground-based remote sensing data in 2004 and by ground-truthing of arthropod populations in both years. Mite- and aphid-infested cotton was detected using aerial data in the green and near-infrared (NIB) wavelengths in 2003, with subeconomic threshold aphid population levels. At the time aerial data were collected, mite populations peaked at 95% leaves infested and exceeded treatment threshold levels of 30-50% leaves infested. However, the number of mites per leaf in the treatments was low to moderate (32,9,4,6, and 2 average mites/leaf). Moreover, cotton infested with cotton aphids above economic threshold levels was consistently detected using NIB wavelengths from the satellite data in 2004. Similarly, aphid-infested cotton was detected at both sub- and supraeconomic threshold aphid levels using NIB wavelengths from the ground-based remote sensing data. Finally, accumulated mite-days were linearly correlated with a canopy, false color, and a vegetation index using satellite data in 2004. Wavelengths in the NIR were fair to moderately accurate predictors of aphid- and mite-infested cotton.