Surface roughness analysis of a conifer forest canopy with airborne and terrestrial laser scanning techniques

• Published in: International journal of applied earth observation and geoinformation 2012-02, Vol.14 (1), p.192-203
• Main Authors: Weligepolage, K., Gieske, A.S.M., Su, Z.
• Format: Article
• Language: English
• Subjects: Airborne laser scanning; Applied geophysics; Conifer forest; Earth sciences; Earth, ocean, space; Exact sciences and technology; Internal geophysics; Momentum roughness; Terrestrial laser scanning
• ISSN: 1569-8432; 1872-826X
• DOI: 10.1016/j.jag.2011.08.014

► The study utilizes laser scanned data to estimate forest aerodynamic roughness. ► The average obstacle density (λ) of the forest was in the range of 0.14 to 0.24. ► The estimated roughness parameters depend very much on the model selected. ► Raupach models with the optimized parameters perform better for a wide range of λ. ► The cumulative area–height modelling proves to be a simple yet a feasible method. Two digital Canopy Height Models (CHMs) were generated using the novel Terrestrial Laser Scanning (TLS) technique combined with Airborne Laser Scanning (ALS) data, acquired over a conifer forest. The CHMs were used to extract cross-sections in order to derive surface geometric parameters. Different morphometric models were applied to estimate aerodynamic roughness parameters: the roughness length (z0) and the displacement height (d0). The CHMs were also used to derive the area–height relationship of the canopy surface. In order to estimate roughness parameters the observed canopy area–height relationship was modelled by uniform roughness elements of paraboloid or conical shape. The estimated average obstacle density varies between 0.14 and 0.24 for both CHMs. The canopy height distribution is approximately Gaussian, with average heights of about 26m and 21m for CHMs generated with data from TLS and ALS respectively. The estimated values of z0 and d0 depend very much on the selected model. It was observed that the Raupach models with parameters tuned to resemble the forest structure of the study area can be applied to a wide range of roughness densities. The cumulative area–height modelling approach also yielded results which are compatible with other models. The results confirm that, to model the upper canopy surface of the conifer forest, both the cone and the paraboloid shapes are fairly appropriate.