Upland vegetation mapping using Random Forests with optical and radar satellite data
Citable Link (URL):http://resolver.sub.uni-goettingen.de/purl?gs-1/16313
Journal Article (Published version)
Uplands represent unique landscapes that provide a range of vital benefits tosociety, but are under increasing pressure from the management needs of adiverse number of stakeholders (e.g. farmers, conservationists, foresters, govern-ment agencies and recreational users). Mapping the spatial distribution ofupland vegetation could benefit management and conservation programmesand allow for the impacts of environmental change (natural and anthropogenic)in these areas to be reliably estimated. The aim of this study was to evaluatethe use of medium spatial resolution optical and radar satellite data, togetherwith ancillary soil and topographic data, for identifying and mapping uplandvegetation using the Random Forests (RF) algorithm. Intensive field survey datacollected at three study sites in Ireland as part of the National Parks and Wild-life Service (NPWS) funded survey of upland habitats was used in the calibra-tion and validation of different RF models. Eight different datasets wereanalysed for each site to compare the change in classification accuracy depend-ing on the input variables. The overall accuracy values varied from 59.8% to94.3% across the three study locations and the inclusion of ancillary datasetscontaining information on the soil and elevation further improved the classifi-cation accuracies (between 5 and 27%, depending on the input classificationdataset). The classification results were consistent across the three differentstudy areas, confirming the applicability of the approach under different envi-ronmental contexts.