Accuracy of aerial sketch-mapping estimates of spruce budworm defoliation in New Brunswick

Abstract

The accuracy of aerial sketch-mapping estimates of spruce budworm (Choristoneurafumiferana (Clem.)) defoliation was evaluated from 1984 to 1993 in 222–325 sample plots in spruce (Picea sp.)–balsam fir (Abiesbalsamea (L.) Mill.) stands in New Brunswick. Operational aerial defoliation estimates were used, wherein all productive forest in known budworm infestation zones was surveyed each year from small aircraft with flight lines 2–5 km apart, and rated in classes of nil (0–10%), light (11–30%), moderate (31–70%), and severe (71–100%). Aerial defoliation estimates were compared with ground-based binocular estimates of current defoliation for an average of 10 trees/plot (range 5–20). Overall, 56% of plots were correctly rated by aerial sketch mapping in four classes (nil, light, moderate, and severe), with 37% of the plots underestimated and 7% overestimated. The predominant error (26% of plots) was rating defoliation as nil (0–10%) from the air when it was actually light (11–30%). This error was deemed not important in terms of predicting tree response, since data from the literature indicated that defoliation less than 30% did not cause tree mortality, although if continued, it would reduce growth. Using three defoliation classes (by combining nil and light, 0–30%), 82% of the plots were correctly classified by aerial sketch mapping. The probability of correct aerial classification of defoliation was significantly affected by defoliation class, weather conditions prior to and during observation flights, and the defoliation class × weather interaction. It was concluded that aerial sketch mapping of spruce budworm defoliation is a viable technique that can be used for both surveys and decision support systems that estimate forest response to budworm outbreaks and management activities.