Estimating the Severity of Defoliation Due to Pine Processionary Moth Using a Combination of Landsat and UAV Imagery

After the recent outbreak of pine processionary moth (Thaumetopoea pityocampa) was observed in 2016, we attempted to estimate the damage severity in Catalonia, Spain. In this study pre-outbreak and post-outbreak images were obtained from Landsat 8 to capture the maximum defoliation period over winter. The difference in vegetation index (dVI) was used as a change detection indicator and was further calibrated with Unmanned Aerial Vehicle (UAV) imagery. Linear regression models between predicted dVI and observed defoliation % by UAV were compared among five selected dVIs for the coefficient of determination. Our results found that the goodness of fit was highest at 0.787 (R 2 ) using Normalized Difference Vegetation Index (NDVI), which may be promising for estimating the severity of defoliation in affected areas where ground-truth data is limited. We concluded with the high potential of using UAVs as an alternative method to ground-truth data for cost-effective forest monitoring.