|Hubbard, J - UNIV OF WISCONSIN|
|Flanders, K - AUBURN UNIVERSITY|
Submitted to: Journal of Economic Entomology
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: December 30, 2006
Publication Date: June 1, 2007
Citation: Mankin, R.W., Hubbard, J.L., Flanders, K.L. 2007. Acoustic indicators for mapping infestation probabilities of soil invertebrates. Journal of Economic Entomology. 100:790-800. Interpretive Summary: Insects living in the soil are often important agricultural pests. However, it is difficult to monitor their presence and abundance. Scientists at the Center for Medical, Agricultural, and Veterinary Entomology, Gainesville, FL, University of Wisconsin-Stevens Point, and Auburn University performed digital signal analyses of recordings from four different acoustic surveys of hidden infestations of insects in soil. They found statistically significant relationships between acoustic estimates of soil insect distributions and the numbers of insects present at sampled sites in three of the four surveys. In the one survey where the relationship failed, it was found that the average sound activity at infested sites was below the average background noise level. The results help to clarify conditions under which acoustic methods can augment or substitute for traditional excavation methods reliably in a field survey, and conditions under which they are not likely to be effective.
Technical Abstract: Soil invertebrates were surveyed in forage fields at Auburn and Grove Hill, AL by acoustic monitoring and traditional excavation procedures in 1997 and 1998. For this report, recordings from sampled sites were analyzed by computer and by listeners with previous training in identification of subterranean sounds. The computer analysis discriminated soil invertebrate sounds from background noise by matching individual sound pulses with spectral profiles derived from validated samples from each survey. Sound pulses produced by white grubs and other large soil invertebrates often could not be distinguished from each other; consequently, in comparing acoustic and excavation assessments, all sound-producing organisms were pooled into a single, noisemaker category. Relationships between sound rates and noisemaker counts were analyzed by first transforming the original assessments into categorized (discretized) variables, i.e., low, medium, or high likelihoods of noisemakers, or below- or above-threshold pulse rates. In the Grove Hill-1997 survey, noisemaker pulse rates were not high enough above the background noise level to establish a significant relationship between sound activity and infestation. In the Auburn-1997, Auburn-1998, and Grove Hill-1998 surveys, however, significant relationships were found between noisemaker counts and likelihoods rated by listener- and computer-based acoustic methods. Geostatistical analysis of the Auburn-1998 survey found a significant correlation (R2 = .41) between kriged (spatially interpolated) estimates of above-threshold pulse rates and noisemaker counts. These results clarify initial findings of a significant correlation between log-transformed pulse rates and log-transformed noisemaker counts in the Auburn-1997 survey but neither Grove Hill survey.