The impact of the RapidEye Red Edge band in Mapping Defoliation Symptoms
The presented study is an outcome of the work package "340 - Assessing New Sensor Data" of the FP7- EUFODOS (European Forest Downstream Services) project. EUFODOS is focusing on the development of operational services, in particular in the domains forest disturbances and forest functional parameters. The research component of the project includes the evaluation of sensor data from a selection of operational high resolution earth observation satellites with regard to performance in fulfilling the service specific requirements. In the domain forest disturbances, insect damage detection and mapping is one important issue dealt with. The study investigates the relationship between defoliation intensities in pure pine stands infested by the nun-moth, and spectral variables of the RapidEye sensor. It compares the classification power with respect to the five RapidEye bands, the NDVI, and the NDRE (Normalized Difference Red Edge Index). The study reveals, that the Red band as a single Rapid Eye band is with a determination coefficient of R2=0.62 the strongest in detecting defoliation symptoms, whereas the Red Edge band yields an R2 of 0.06. However, when the bands are combined with the NIR band in a vegetation index, respectively, it shows that the Normalized Difference Red Edge Index [(NIR - Red Edge)/ (NIR + Red Edge)] outperforms the NDVI [(NIR - Red)/ (NIR + Red)] and all other compared bands. Its average standard deviation in the defoliation classes is least and its determination coefficient as an indicator for the strength of the linear relationship with defoliation is highest with a value of R2=0.74. As a conclusion it is suggested that the Red Edge band as a single band is not suitable for mapping defoliation symptoms since it is too sensitive and recording too much feature ''noise''. The band releases its real strength when combined with the NIR band into the Normalized Difference Red Edge Index. This supports the practical benefit and value of the Rapid Eye Red Edge band in satellite remote sensing -based forest health monitoring.