Quantification of spatial and temporal changes in forest cover is an essential component offorest monitoring programs. Due to its cloud free capability, Synthetic Aperture Radar(SAR) is an ideal source of information on forest dynamics in countries with near-constantcloud-cover. However, few studies have investigated the use of SAR for forest cover estimationin landscapes with highly sparse and fragmented forest cover. In this study, thepotential use of L-band SAR for forest cover estimation in two regions (Longford and Sligo)in Ireland is investigated and compared to forest cover estimates derived from three national(Forestry2010, Prime2, National Forest Inventory), one pan-European (Forest Map 2006)and one global forest cover (Global Forest Change) product. Two machine-learningapproaches (Random Forests and Extremely Randomised Trees) are evaluated. Both RandomForests and Extremely Randomised Trees classification accuracies were high (98.1–98.5%), with differences between the two classifiers being minimal (<0.5%). Increasing levelsof post classification filtering led to a decrease in estimated forest area and an increasein overall accuracy of SAR-derived forest cover maps. All forest cover products were evaluatedusing an independent validation dataset. For the Longford region, the highest overallaccuracy was recorded with the Forestry2010 dataset (97.42%) whereas in Sligo, highestoverall accuracy was obtained for the Prime2 dataset (97.43%), although accuracies ofSAR-derived forest maps were comparable. Our findings indicate that spaceborne radarcould aid inventories in regions with low levels of forest cover in fragmented landscapes.The reduced accuracies observed for the global and pan-continental forest cover maps incomparison to national and SAR-derived forest maps indicate that caution should be exercisedwhen applying these datasets for national reporting.
การแปล กรุณารอสักครู่..