A retrospective investigation was undertaken of the area around Broken Hill in the Western region of New South Wales, where Australian plague locust (Chortoicetes terminifera) nymphal band control took place during September -November (spring) of 1987. Processed Landsat-5 multispectralscanner data for March 1987 were used to locate the likely spring/summer locust infestation areas using habitat type, condition and soil type as the delineating parameters. The results show that a vegetation index (indicating the degree of vegetation greeness in an area) is useful for predicting where locusts are most likely to aggregate, a classification of vegetation types is needed for determining likely breeding sites, and that there is an additional relationship between soils/geology and the location of locust infestations. These results are discussed in relation to the future operational use of the techniques in the Australian Plague Locust Commission.
SummaryA Maximum Likelihood Classification (MLC) and a Normalised Difference Vegetation Index (NDVI) were used to determine habitat type and condition respectively. A Soil Brightness Transformation (SBT) using bands 6 and 4 was investigated as a method of indicating soil type in the area based on the resultant density sliced 'brightness' values.Thematic maps at 1:250 000 scale were produced for all classification techniques and ground truth surveys showed that the classifications had, in general, a greater than 80% accuracy.The SBT results indicated that 88% of band targets were in areas with brightness values associated with areas defined on geological/soil maps as either floodplains of black and grey clayey silt and sand, flat to gently undulating plains of red and brown clayey sand, loam and lateritic soils, or colluvial deposits of poorly sorted sand. The NDVI analysis showed that 75% of all targets were in areas of green vegetation, and the MLC analysis revealed 70% of targets were in grassed or natural pasture areas.A historical survey of locust data was carried out and known infestations between 1971 and 1987 plotted against the ancilliary geology/soil data. It was found that the majority of infestations occurred on 3 main geological/soil groups: 20% on floodplains of black and grey clayey silt and sand, 42% on flat to gently undulating plains of red and brown clayey sand, loam and lateritic soils and 26% on colluvial