Data: 07/11/2018 – quarta-feira
Local: Sala A– Prédio Rotunda– INPE/SJC
Palestrante: Wanda Graf – PhD student at the chair of Forest Inventory and Remote Sensing, University of Göttingen, Germany
Título: Long-term Landsat time series – a new opportunity for forest diversity research
Current diversity of the forest herb layer is not only influenced by site conditions but also by historic natural and anthropogenic disturbances and the developmental stage of stands in temperate forests. Landsat time series allow analyzing forest canopy dynamics over long periods, and these dynamics may be related to herb layer plant diversity.
Long-term Landsat time series from 1985 to 2015 were used to investigate differences in the frequency of natural and anthropogenic disturbances as well as stand age between temperate forest sites in Germany, which were related to indicators of herb layer plant species diversity. To identify different levels in management intensity and stand age, we applied a newly introduced date-independent SAX-dissimilarity measure (D-SAX) to the NDVI time series. Based on D-SAX we clustered the plots and tested for significant differences in diversity indices of the herb layer species, as well as a silvicultural management intensity indicator for temperate forests. Cluster stability was assessed by the mean Jaccard’s similarity. Levels of herb layer plant diversity between clusters of pine and beech forests were significantly different and could be explained by the developmental stage of the stands or the frequency of the management actions.
Our study suggests that differences in natural and anthropogenic disturbances and developmental stage, detected from Landsat time series, may improve understanding differences in herb layer diversity between forests. Cluster analysis of the D-SAX dissimilarity measure is a promising approach to investigate the diversity-productivity-disturbance relationship.