Spatially optimizing vegetation indices integrated with sparse partial least squares regression to detect and map the effects of Gonipterus scutellatus on the chlorophyll content of eucalyptus plantations

摘要

Gonipterus scutellatus is a beetle causing severe defoliation to South Africa's eucalyptus plantations. This defoliation induced by the beetle inhibits the eucalypts ability to photosynthesize, by affecting its chlorophyll content. Therefore, this study integrates spatially optimized and the single 0.5 m resolution vegetation indices with sparse partial least squares regression (SPLS-R) and partial least squares regression (PLS-R) to detect and map leaf chlorophyll content of defoliated eucalyptus plantations. The optimized vegetation indices were spatially resampled to resolutions that best paralleled varying levels ofG. scutellatusdefoliation. From the results, the 0.5 m resolution SPLS-R model (R-2 = 0.76; RMSE of 1.50 (2.88% of the mean measured chlorophyll)) outcompeted the 0.5 m resolution PLS-R (R-2 = 0.73; RMSE of 1.54 (2.95% of the mean measured chlorophyll)) model. Furthermore, the spatially optimized SPLS-R (R-2 = 0.81; RMSE of 1.44 (2.76% of the mean measured chlorophyll) model was more superior in detecting and mapping chlorophyll content of defoliated eucalyptus plantations when compared to the 0.5 m resolution SPLS-R model. The most significant variables selected by the optimized SPLS-R model were DMI, ARI, NDRE, GNDVI, and NDVI. In essence, this study has illustrated the significance of the spatial resolution in effectively detecting and mapping chlorophyll content of defoliated eucalyptus plantations.