Small-footprint Lidar Estimations of Sagebrush Canopy Characteristics

The height and shape of shrub canopies are critical measurements for characterizing shrub steppe rangelands. Remote sensing technologies might provide an efficient method to acquire these measurements across large areas. This study compared point-cloud and rasterized lidar data to field-measured sagebrush height and shape to quantify the correlation between field-based and lidar-derived estimates. The results demonstrated that discrete return, small-footprint lidar with high point density (9.46 points/m2) can provide strong predictions of true sagebrush height (R2 of 0.84 to 0.86), but with a consistent underestimation of approximately 30 percent. Our results provided the first successful lidar-based descriptors of sagebrush shape with R2 values of 0.65, 0.74, and 0.78 for respective predictions of shortest canopy diameter, longest canopy diameter, and canopy area. Future studies can extend lidar-derived shrub height and shape measurements to canopy volume, cover, and biomass estimates. Introduction The height and shape of shrub canopies are critical measurements for characterizing shrub steppe landscapes in terms of structure, age, cover, critical wildlife habitat, fuel type, erosion, infiltration, evapotranspiration, disturbance history, and biomass. Aerial biomass estimates for big sagebrush (Artemisia tridentata) have been used to assess fuel loads (Frandsen, 1983), calculate available forage (Wambolt, 1994), delineate wildlife habitat (e.g., Eng and Schladweiler, 1972; Beck, 1977; Davies et al., 2007), and study climate response (Harte and Shaw, 1995; Harte et al., 2006). The traditional, direct method of measuring vegetative biomass by clipping and weighing (Bonham, 1987) is destructive and extremely cost-inefficient (Uresk, 1977; Clark et al., 2008). Allometric equations have therefore been developed for rapid biomass assessments using volume-based metrics derived from height PHOTOGRAMMETRIC ENGINEER ING & REMOTE SENS ING May 2011 1 Jessica J. Mitchell is with the Department of Geosciences, Idaho State University, Boise Center Aerospace Lab, 995 University Blvd., Idaho Falls, ID 83402 ([email protected]). Nancy F. Glenn and Temuulen T. Sankey are with the Department of Geosciences, Idaho State University, Boise Center Aerospace Lab, 322 E Front St., Suite 2240, Boise, ID 83702. DeWayne R. Derryberry is with the Department of Mathematics, Idaho State University, PO Box 8085 Pocatello, ID 83209. Matthew O. Anderson and Ryan C. Hruska are with the Idaho National Lab, PO Box 1625, Idaho Falls, ID 83415. Photogrammetric Engineering & Remote Sensing Vol. 77, No. 5, May 2011, pp. 000–000. 0099-1112/11/7705–0000/$3.00/0 © 2011 American Society for Photogrammetry and Remote Sensing Small-footprint Lidar Estimations of Sagebrush Canopy Characteristics Jessica J. Mitchell, Nancy F. Glenn, Temuulen T. Sankey, DeWayne R. Derryberry, Matthew O. Anderson and Ryan C. Hruska and crown characteristics (e.g., diameter, elliptical area, density, cover) (Harniss and Murray, 1976; Rittenhouse and Sneva, 1977; Uresk, 1977; Dean et al., 1981; Murray and Jacobson, 1982; Frandsen, 1983; Tausch, 1989; Wambolt et al., 1994; Clark 2008; Cleary et al., 2008). Remote sensing technologies offer potential solutions for extending sagebrush measurements collected on the ground to a range of spatial scales in a cost-efficient manner. Multispectral and hyperspectral studies designed to estimate shrub cover and leaf area index in semiarid shrub steppe are limited by multiple scattering, bright soil reflectance, open canopies and spectrally indiscriminate targets (e.g., Smith et al., 1990; Jakubauskas et al., 2001; Okin et al., 2001; Mirik et al., 2007). Small-footprint, discrete return lidar (airborne laser scanning) is not limited by many of these spectral challenges and has the potential for estimating shrub canopy characteristics at a range of scales appropriate for landscape assessments (Ritchie et al., 1992 and 2001; Mundt et al., 2006; Riano et al., 2007; Streutker et al., 2006; Su and Bork 2006, 2007). However, separating lidar returns in low-height rangeland vegetation is difficult because the vegetation canopy returns are often close to ground returns in both space and time. Furthermore, there are fewer vegetation returns in sparsely vegetated semiarid ecosystems than in more foliated ecosystems. A limited number of studies have evaluated the use of lidar in shrub environments (Hopkinson et al. 2005; Streutker and Glenn, 2006; Riano et al., 2007; and Su and Bork, 2007) and these studies have consistently found that small-footprint lidar systems underestimate shrub canopy height. Shrub height underestimation is attributed to the low probability of the laser hitting the top of the canopy, or laser pulses penetrating the canopy, which generates return signals from material within the canopy (Weltz 1994; Næsset and Økland 2002; Gaveau and Hill 2003, Clark et al. 2004). Hopkinson et al. (2005), using a sensor with an average point sampling density of over 3 m 2, determined that their low shrub class ( 2 m) had the highest proportion of height underestimation: 62 percent (52 cm) using the raw lidar point cloud data and 48 percent (39 cm) using a raster. Streutker and Glenn (2006) found that lidar data with an average point density of 1.2 m 2 systematically underestimated mountain sagebrush (Artemisia tridentata subsp. vaseyana) heights ( 2 m) by approximately 50 percent and that underestimation was dependent upon canopy cover, with lower cover having greater error estimates. Riano et al. In Pr es (2007) underestimated shrub height ( 1.68 m) in a dense mixed shrub environment using a lidar sensor with an average point density of 3.5 m 2. Although the magnitude of underestimation was not documented in their study, nearinfrared imagery co-registered to the lidar point-cloud data reduced underestimation. Su and Bork (2007) used smallfootprint lidar with an average point density of 0.54 m 2 to characterize a variety of community types and found a tendency toward shrub height underestimation ( 1.30 m) in open environments. This study reported the magnitude of height underestimation in terms of signed-mean error, which varied from 4 to 16 cm, depending on the species. To our knowledge, only a few lidar studies have estimated the accuracy of shrub canopy shape characteristics, and mixed results were obtained using an aerial cover approach defined by the ratio of vegetation returns to the total number of returns (Ritchie et al., 1992; Weltz et al., 1994; White et al., 2000). Hopkinson et al. (2005) evaluated the accuracy of small-footprint lidar estimates of shrub volume by defining canopy volume as the vertical frequency distribution of the number of vegetation returns, stratified by height quantiles. This study compares lidar point-cloud data to sagebrush canopy characteristics measured in the field with the goal of quantifying prediction errors associated with height and 2D aerial shape measurements (canopy diameter and area) for individual shrubs.