X/C-HH InSAR and L-PolInSAR over Lodgepine Forest
M. Lorraine Tighe(1), Doug King(2) and Heiko Baltzer(3)
(1) Intermap Technologies, 8310 South Valley Highway, Denver CO 80111, United States
(2) Carleton University, 1125 Colonel By Drive, Ottawa Ontario K1S5B6, United States
(3) University of Leicester, Bennett Building, Leicester LE17RH, United Kingdom
One of the most vital environmental issues of our time is the sustainable management of the world's forest resources. Forests are an important global ecosystem and play a key role in its economic and environmental state (Leckie et al., 1998). Furthermore, demand for forest products is increasing worldwide concurrently ad the forested surfaces continue to diminish (Le Toan et al., 1992). Canopy height and vegetation cover mapping are key parameters in forestry measurements and are important ecological parameters, as they are strongly correlated to timber volume, fire models and biomass and hence carbon stocks. Forest canopy height alone is a critical parameter in better quantifying the terrestrial carbon cycle, forest inventory programs and as an input to many biomass models. However, collection of such forest parameters manually is time-consuming and costly to measure in the field using conventional techniques. The ability to derive forest canopy heights and vegetation maps remotely would therefore be of great benefit. Significant advances in remote sensing technologies have led to a new era of global topographic observations, where reliable forest measurements are becoming a possibility (Homer et al., 2007; Maune, 2007). Foremost among these technologies are interferometric synthetic aperture radar known as InSAR and polarmetric InSAR (PolInSAR) (Madsen et al., 1993; Hensley et al., 2001; Mercer, 2004; Walker et al., 2007). Phased-based InSAR techniques exploit the interference patterns of two electromagnetic waves to remove the random component of phase of the single complex radar signal to estimate the first surface feature heights given by the scattering phase center height (hspc). The hspc of forested canopy may include elevations of the tree top, tree leaves, twigs and branches, tree trunks, undergrowth, the edge of a tree canopy, and/or a ground contribution (Treuhaft and Siqueira, 2000; Hodgson et al., 2003; Kellndorfer et al., 2004a; Balzter et al., 2007; Walker et al., 2007; Andersen et al., 2008). Where the hspc lies within a forest canopy and subsequent digital surface model (DSM) is dependent on the wavelength, DSM processing techniques (e.g. elevation posting) and to a lesser extent the SAR polarization, system parameters, vegetation density and slope (Izzawari et al., 2006; Dall, 2007; Anderson et al., 2008; Becek, 2008, Viergever et al., 2008). Motivated by the need to remotely derive vegetation cover type and true vegetation canopy heights and the advances in and availability of InSAR/PolInSAR technologies for collecting nation-wide surface and bare earth elevation models, this research will focus on the analysis of mean vegetation canopy heights derived from single-pass InSAR from X/C-HH InSAR and L-PolInSAR elevation data over Lodgepine vegetation in Burnstick, Alberta, Canada. The primary objective of this study is to assess the relation between HH-polarized multi-frequency InSAR scattering phase center heights (hspc), L-band PolInSAR tree heights and in-situ to describe the link. To test whether this relation is a reasonable approximation X/C-HH InSAR hspc posted at 5 and 30, and L-band PolInSAR data posted at 5 m are evaluated against in-situ tree heights and a LiDAR tree height model. Preliminary results indicate that the X-band InSAR, C-band InSAR and PolInSAR underestimate the true tree heights by 20%, 30% and 20% respectively when compared to in-situ and LiDAR.