Increasing applications of combined remote sensing techniques are evidence that integrated approaches are viable options if good accuracies in biomass estimation should be combined with high efficiency. Due to the increasing need for and importance of fusion of multi-modal remote sensing (LiDAR and multispectral imagery) to improve and scale estimates, we present a case study for forest plantations. The selection of an appropriate LiDAR canopy height sample, as representative of the entire study area, was a central focus of this study. The proposed scheme would use LiDAR transects instead of a blanket coverage with the aim of reducing costs and processing time, while maintaining prescribed levels of accuracy and precisions (Hudak et al. 2002). The remainder of the forest area to be inventoried would then be assessed using IKONOS imagery, based on established LiDAR-IKONOS height relationships.

The study was conducted in the Kwazulu-Natal province located in eastern South Africa. The sampled plantation stands are located approximately 50 km south of the town of Pietermaritzburg. The area is known locally as the southern Natal midlands. Rain falls predominantly in the summer months with cold dry winters and warm wet summers.

Mean annual rainfall ranges from 746-1,100 mm (Schulze 1997) and is associ­ated with either frontal systems originating from the south or from thunderstorms generated from convection activity. Temperatures range from 20 ° C in summer to below 10 °C in the winter.

Extreme temperature changes are a function of altitude and proximity to the warm Indian Ocean. Soils in the area are characterized by fine sandy clay and humic topsoils, underlain by yellow or red apedal subsoils. The topography of the study area is flat with undulating hills and is classified by Schulze (1997) as being low mountains. Altitude ranges from 362 m aml to over 1,500 m aml with an average altitude of approximately 874 m aml.