Permafrost landscapes are one of the earth´s landscapes most affected by climate change. Thus, monitoring these landscapes is necessary. Vegetation cover and vegetation composition often serve as an indicator for the state of the permafrost and thus need to be monitored accordingly. This thesis uses a non-classificatory ordination approach to map species composition in the central Lena delta in Russia. It further describes patterns to be found within the central Lena delta´s species composition as well as which kind of data serves best as a basis for vegetation composition mapping. Partial Components Analysis (PCA) was used to extract species composition and floristic gradients from a dataset collected during the expedition LENA-2018. The extracted values were combined using different kinds of spectral input data, both hyper- and multispectral, as well as aggregated and non-aggregated, to train a Partial Least Square Regressor (PLSR). Species composition was mapped using a Sentinel 2 image from August 2018. It was shown that species composition was dominated by Mosses, Carex chordorrhiza, Salix glauca, Salix pulchra, Eriophorum vaginatum and Poaceae. Each species was linked to a specific type of landform. Hyperspectral, aggregated data performed best as predictor variables, but due to the lack of hyperspectral imagery, the aggregated multispectral data can also be used as it leads to satisfying results. These results indicated that species composition mapping in permafrost landscapes is possible although to a lesser degree of accuracy compared to lower latitudes. These observations could pave the way further studies resulting in a more detailed description of permafrost properties related to species composition like active layer thickness or greenhouse gas content.