Abstract

The invasive poplar leaf miner Leucoptera sinuella (Lepidoptera: Lyonetiidae), has spread throughout poplar plantations in central Chile. We developed and validated models based on two methodologies of foliar damage estimation, and from different bands and indexes obtained from Sentinel-2 satellite images. We estimated foliar damage with field visual application of an ordinal severity scale, or with a laboratory estimation of leaf mined area with an image software (ImageJ) from a sample of leaves using an ordinal severity scale. We developed four models for the visual estimation on the field using red band and three spectral indexes, while we developed four models for the laboratory image software estimation using near infrared (NIR) band and the same three spectral indexes. Models developed from field visual estimation with red band (R-2 = 0.88) and Normalized Difference Vegetation Index (NDVI) (R-2 = 0.89) produced better results, as well as from image software estimation with NIR band (R-2 = 0.86) and NDVI (R-2 = 0.83). The field visual estimation and red band model got the best validation results, with R-2 of 0.90, mean square error of 0.73, mean absolute error of 0.59, and a slope of 0.91. This model could predict the severity of foliar damage by L. sinuella in poplar plantations, representing a potentially useful monitoring tool for decision-making in the management of the poplar leaf miner in large areas of poplar plantations in central Chile.