Integrating airborne LiDAR and Landsat ETM+ data for large area assessment of forest canopy height in Amazonia
Authors: Hill, R.A., Boyd, D.S. and Hopkinson, C.
Conference: Silvilaser 2010, the 10th International Conference on LiDAR Applications for Assessing Forest Ecosystems
Dates: 14-17 September 2010
Abstract:Quantifying structure is essential for modelling tropical forest ecosystem processes and services. This paper investigates the ability to assess canopy height of tropical rainforest from Landsat Enhanced Thematic Mapper (ETM+) data by performing a detailed assessment of the within-pixel variation of forest canopy height characteristics and the influence this has on recorded spectral response. Forest canopy height is derived from airborne small-footprint LiDAR data acquired using a Leica ALS50 II system. Forest types studied include regenerating, floodplain and terra firme forest. Spectral reflectance in all optical ETM+ bands was shown to be negatively correlated with the range, maximum, mean and standard deviation of canopy height per pixel. The strongest overall correlation coefficient (-0.83, p < 0.01) occurred between ETM+ Band 4 and maximum canopy height. It was shown that either near or middle infrared (ETM+ Bands 4 and 5) in a simple ratio or normalised index with either visible green or red (ETM+ Bands 2 and 3) also correlated strongly with canopy height (mean and maximum) and could be used to predict height across the ETM+ scene. In particular, the NDVI has been demonstrated as a potential cost-effective tool for large-area assessment of tropical rainforest canopy height appropriate for UN-REDD (Reduced Emissions from Deforestation and Forest Degradation) monitoring.
Source: Manual
Preferred by: Ross Hill