ESTIMATING BIOMASS IN MANAGED URBAN GREENERY AREAS USING CANOPY COVER PERCENTAGES DERIVED FROM NDVI

摘要

Biomass estimation with remote sensing methods is greatly challenged in urban areas due to highly heterogenic structure of urban vegetation and large amount of man-made objects. In this paper we present a method to estimate the amount of biomass in large metropolitan areas using land use polygon based canopy cover percentages derived from a 10 m resolution SPOT 5 HRG image. The study area covered all managed greenery areas (i.e. park, recreational and commercial/institutional land use) in Singapore encompassing over 42 000 ha. In addition to the SPOT data, seven 0.5 m resolution GeoEye images and 26 field sample plots were used. The results showed a strong correlation (r = 0.83) between canopy cover percentages derived from the very high resolution satellite imagery using manual delineation and the percentages of areas in the SPOT image with NDVI values above 0.57 after grass areas had been masked out of the NDVI map. Subsequently, a correlation of r = 0.61 was found between canopy cover percentages and biomass in the sample plots. Thereby, a biomass estimation procedure for managed urban greenery environment was developed based on canopy cover percentages within land use polygons. The method not only allows estimation of the current biomass in urban greenery areas but also enables monitoring of biomass changes. With this method, the total biomass in urban greenery areas of Singapore (excluding road side trees, undeveloped and nature reserve areas in the island) was estimated to be around 1.7 Mt, equalling to 40 t/ha or about a third of a typical biomass of young secondary forests. These figures reveal the high amount of biomass found urban areas of Singapore and illustrate the potential of biomass accumulation in urbanised environment.