QUANTIFYING HAIL DAMAGE FOR CROP LOSS ASSESSMENT: TECHNIQUES USING REMOTE SENSING AND GEOGRAPHIC INFORMATION SYSTEMS

Abstract Hail damage results in losses of approximately 3% of total broad-acre crop production every year across the Darling Downs region of Queensland. Current practices of hail damage assessment are both time and labour intensive. The use of remote sensing and geographic information systems can provide alternative techniques for the assessment of crop hail damage. Remote sensing is currently providing valuable data for precision agriculture and land-use / land-cover mapping but rarely has it been used for insurance and loss calculation. Results have indicated the usefulness of Landsat 7 ETM+ imagery in the calculation of area and quantum of damage for loss adjusting purposes. Depending upon the analysis techniques used, an average accuracy of 5% to 30% difference was established from that observed in the field. It was identified that the Normalised Difference Vegetation Index (NDVI) had the strongest correlation to defoliation followed closely by the Tasselled Cap Transformation (Greenness), while the Modified Soil Adjusted Vegetation Index (MSAVI2) showed unexpected low accuracies. Higher accuracies were obtained from the introduction of a reduced boundary to reduce the impact of mixed pixels, with the 2-pixel boundary, resulting in the highest accuracies