3.08 Vegetation Structure (LiDAR)M van Leeuwen and M Disney, University College London, London, United Kingdom

© 2018 Elsevier Inc. All rights reserved.

3.08.1 Introduction 1043.08.2 LiDAR: The Technology 1053.08.3 Global Scale 1063.08.4 Regional to Stand Level 1073.08.5 Plot to Tree Level 1083.08.6 Simulation 1103.08.7 Data Fusion 1113.08.8 Conclusion 112References 112Relevant Websites 116

3.08.1 Introduction

Advances in remote sensing technology since the mid-2000s have drastically increased the potential to acquire traditional forestinventory variables as well as information about the effects of forest canopy structure on radiative transfer, and its implicationsfor tree and ecosystem physiology. Light detection and ranging (LiDAR) has provided means to record in high-resolution (cm-scale), the three-dimensional (3D) structure of the canopy, including height and leaf area density and in certain cases even infor-mation related to the branching architecture of trees (Wulder et al., 2012b; Dubayah and Drake, 2000). LiDAR data are currentlybeing acquired from airborne and ground-based platforms with pulse repetition frequencies over 500,000 pulses per second. Themany billions of returned laser pulses that reflect off the ground surface and other objects, such as trees, shrubs, and rocks, capturedby the detector and for which the exact 3D positions are determined based on accurate measurements of the instrument’s positionand orientation, are accessible in easy-to-use tabular data files, and can be easily displayed on modern graphics hardware througha handful of command line expressions or even user-friendly software packages, often available free of charge to academic andnoncommercial users (e.g. SPDlib.org, see Bunting et al., 2013a,b; cloudcompare.org; rapidlasso.com). Over the past decades,the combined effort of a broad community of academic researchers and corporations has delivered a range of algorithms for theprocessing of these data, such as the extraction of terrain models and surface shapes, including stems and tree crowns, or wall-to-wall forest inventory attributes, including but not limited to basal area, merchantable timber volume, biomass, tree stockingdensity, and tree height. An overview of these algorithms and some of the accuracies by which forest attributes can be extractedfrom the data can be found in review articles and in the introductions of many research papers. A query on the Scopus search enginefor academic research papers shows over 2000 hits on the topic of “LiDAR” and “forestry” alone (Disney, 2016), indicating both thegreat success of the technology in acquiring forest inventory information and the perceived value of the technology to the field.

Airborne LiDAR is generally viewed as a cost-effective means to collect forest inventory information over broad scales (e.g.,regional to national scale). The data are often summarized into simple statistical descriptors such as percentile distributions ofthe return heights above the digital terrain model (DTM) and then stored in two-dimensional raster files, similar in representationto the raster files, which are used for passive-optical satellite or airborne data. The percentile distributions can then be used in linearregression models to estimate a range of inventory attributes and the model coefficients can be determined using training data setsof field measurements. This empirical method has been backed by many research articles from which a good sense of estimationaccuracies can be derived for a range of forest types and attributes of interest. The empirical nature of this remote sensing method,however, does not warrant an accurate estimation of attribute values when the correlation models are used outside the calibrationareas, or when other constraints (e.g., time, resources) mean that the homogeneity cannot be assumed or quantified across the studyareas. Hence, these purely empirical approaches lack generality, predictive ability, and, most importantly, physically based param-eters or process understanding.

The exact effect of parameter retrieval uncertainty on the cartographic information being produced is not always known in detailand it is not unusual that maps are short-lived and replaced by the latest insights as new remote sensing technologies advance.Nevertheless, the wealth of information provided by remote sensing enables decision makers and policy makers, more than everbefore, to base their arguments on the latest information available. In many ways, forestry and forest conservation have beencombined commercial and political efforts, where the collection of information over broad spatial scales has been a very challengingone until these recent advances in remote sensing emerged. In the early days of forestry practice, information was gathered throughobservers who would go out to the field and estimate attributes as forest stocking density and tree height by means of visual assess-ment (Kangas and Maltamo, 2006). The results were neither necessarily accurate nor necessarily worse than tape measurements or

104

angle gauges and finger counts. The emergence of remote sensing technology has greatly advanced the consistency and transparencyby which attribute estimates can be acquired and used to guide decision making (Herold and Johns, 2007). A prediction accuracy of60% may sound weak to some, but on the other hand, it can provide valuable insight for a decision maker who is confronted withthe difficult, but necessary task of marshalling finite, scarce resourcesdfor example deciding on areas to keep for recreationalpurposes while selecting others to exploit for resource extraction. Similarly, when timber properties can be predicted only withweak accuracy based on extrinsic observations of tree vigor and stem diameter or lean angle, significant value might arise fromthe allocation of downed trees to specialized sawmills.

The technical accuracy of LiDAR instruments may far exceed our ability to link point cloud data to forest attributes. In this case,a further reduction of measurement error would not greatly improve the estimation of biophysical or biochemical attributes. Thisleaves some questions around the need for repetition in studies on the potential of LiDAR remote sensing in forestry. While beingthe cornerstone of science, the repetition with which the suitability of LiDAR for forest attribute retrieval has been established seemsto some extent excessive, and one might even expect that the explosive drift to conduct and report LiDAR studies that we have seenover the years, has had something to do with the emphasis that has been put on publication metrics. That said, significant researchpotential remains in exploring the use of LiDAR combined with other remote sensing technologies in studies that focus on the inter-action between forest structure and physiological functions. For example, branching architecture can take on a range of angularform, planophile or erectophile or anywhere in between; canopies can be dense or sparse, and consist of mixed species ormono-cultures, and can be multilayered vertically. All of these traits have an impact on the radiation budget and hence, the photo-synthetic and respiratory functioning of the canopy. While standard passive optical, multispectral, and hyperspectral remote sensingcan provide coarse information about canopy structure, it can only generally do so for the upper layer(s) of dense canopies andrequires empirical or process-based models to do so. LiDAR (and to some extent radar) is the only method able to providenear-direct measures of structural information that are easy to interpret and work with.

A significant source of high-resolution (cm to mm-scale) point cloud data is terrestrial laser scanning (TLS), including mobile orhand-held laser scanning (Calders et al., 2015a; Disney, 2016). TLS allows trees to be scanned from multiple directions to increasethe fidelity of the acquired data while reducing the effects of occlusion where multiple objects shadow one another. TLS is charac-terized by a much higher point sampling density and smaller laser footprint that is often of the order of millimeters. This allowsa near-continuous sampling of surface geometry from which solid geometry models can be derived through subsequent processing.The increase in the utility of TLS facilitated the production of lower-cost instruments in the range of $10 K to $100 K, where cheapermodels are often limited to recording pulses up to 30 m in range (Kelbe et al., 2015) and the more expensive scanners allow rangesof (up to) 100s m (Calders et al., 2016; Jupp et al., 2009). This distinction can be important when the emphasis is placed on anaccurate acquisition of canopy structure. However, for plot-level characterization of stem attributes or for the autonomous, unat-tended acquisition of data that reveals structural change over time (Eitel et al., 2013) faster and more lightweight scanning devicesprove highly efficient. Great advances in TLS processing have been achieved since the early 2010s, and include the extraction ofbranching architecture based on an application of graph network theory (Bucksch et al., 2010; Raumonen et al., 2013; Verroustand Lazarus, 2000; Runions et al., 2007; Côté et al., 2009, 2011, 2012; Hackenberg et al., 2015; Newnham et al., 2015). Also,the use of some high-resolution point cloud data for the assessment of leaf orientation angles (Zheng and Moskal, 2012) andmarker-free coregistration of scans that exploits patterns in the positions of tree stems (Kelbe et al., 2016, 2017) are among themany examples that have advanced the state of art of TLS in forestry. Ongoing efforts, originating from the 1990s, include theuse of returned laser intensity information to approximate the apparent surface reflectance (Strahler et al., 2008; Jupp et al.,2009), which in the light of dual-wavelength laser scanners (Douglas et al., 2012; Danson et al., 2015) introduces exciting newpotential to derive 3D, geometrically explicit maps of vegetation indices that may reveal vital information about attributes suchas leaf moisture (Gaulton et al., 2013).

The remainder of this chapter will explore the various uses of LiDAR remote sensing and reviews some of the potentials that itholds for forestry and tree physiology. First, we discuss temporal and spatial domains in which LiDAR remote sensing offers newopportunities, ranging from the global scale to the stand and plot levels and further down to the scale of individual trees and treecrowns. Following this, we discuss matters of structural assessment as well as how canopy structural information can be used tostudy the radiation budget and impacts on physiological functions. Finally, we present a brief review of potential synergies arisingfrom a combination of LiDAR technology and other remote sensing technologies.

3.08.2 LiDAR: The Technology

The development of LiDAR scanning has resulted from the confluence of a number of technologies, including lasers, inertial navi-gation, global positioning, as well as timing electronics and signal digitization at very high sampling speeds. The effective use ofLiDAR in forestry may date back to the beginnings of the 1980s (Lim et al., 2003). Some initial efforts have focused on the acqui-sition of LiDAR data from satellite instrumentation for monitoring of the Earth surface from space. The NASA IceSat/GeoscienceLaser Altimeter System (GLAS) project collected surface topographic data in addition to cloud height information from January2003 until its failure in 2010. However, data collection was always intermittent due to issues around the powering of the laser

Vegetation Structure (LiDAR) 105

source (Breon et al., 2005). The GLAS laser footprint was approximately 60 m on the ground, with each footprint separated alongthe track by around 1 km, and with no side scanning. At this resolution, extraction of attributes such as tree height can be chal-lenging when the underlying terrain is sloped (Sun et al., 2008; Harding and Carabajal, 2005). GLAS was not designed for forestapplications, but proved extremely useful for mapping canopy height, in flat, tropical, and boreal regions (Saatchi et al., 2011;Baccini et al., 2012). Perhaps more convincing evidence of the power of LiDAR for canopy applications has come from airbornedata collection campaigns. Flown at around 1000 m above ground, the laser footprint may be somewhere around 30 cm or less,with points collected at scan rates of 10s of kHz, resulting in dense point cloud data that are visually pleasing and immediatelyspeaks to the imagination, even of those with an untrained eye: Terrain and build-up structures such as roofs, roads, and treesare immediately recognized.

Some of the more recent contributions include the collection of LiDAR data from unmanned airborne systems (UAS) and hand-held (and carried) devices where an operator simply sweeps or walks the instrument around to collect data in areas that wouldotherwise be hard to cover with conventional remote sensing platforms (e.g. Lin et al., 2011; Bosse et al., 2012; Parker et al.,2004). Much is also to be expected from the automotive industry where self-driving vehicles require real-time acquisition of thesurrounding 3D world. These demands are likely to drive down both the cost as well as the size of laser scanning equipment.With equipment becoming more compact and cheaper, it is likely that operational use of the technology extends beyond thesampling of traditional forest inventory attributes, and in some cases, might even move towards the collection of a comprehensivecensus for select sites, with important implications for the efficiency of forest resource utilization or the study of energy budgets andbiogenic fluxes.

Some evolution in LiDAR scanners can also be discerned in the technique by which individual returns are registered, althoughdifferent techniques may continue to coexist due to trade-offs in cost and utility value. Initially, commercial scanners were equippedto digitize the incoming waveform of laser intensity as a few discrete pulses. Different techniques for registering discrete returns exist,including time-of-flight and phase-based scanning, where the latter is faster and often cheaper but suffers the problem of ‘ghosting’when targets overlap along the photon path. Many “research-grade” LiDAR instruments are equipped to record the entire waveform.The full-waveform scanners are more expensive due to data storage requirements. Discrete scanners only record a range distancewhen the returned laser intensity exceeds a set threshold. Upon the registration of a return, the instrument electronics requiretime to process the returned signal and, as a result, the discrete-return LiDAR scanners suffer from an instrument-dead timewhen two targets are within short range of one another (Armston et al., 2013). The full-waveform data are more challenging toprocess and analyze, typically requiring a deconvolution step where the entire waveform is described as a superposition of Gaussiandistributions. The full-waveform scanners have the added benefit that by integrating waveform-intensity values over a range, an esti-mate of target reflectance can be made given assumptions about the backscattering cross-section and normal angle of the interceptedtargets (e.g. Jupp et al., 2009; Cawse-Nicholson et al., 2014). It is worthwhile mentioning that most commercial airborne scannersnow have full-waveform recording capabilities.

A newer type of LiDAR scanner comprises the single photon LiDAR that registers incoming light more efficiently and, hence,can be used to acquire denser point clouds than the more traditional discrete-return or full-waveform scanners offer. Singlephoton LiDARs can also operate at a higher pulse repetition frequencies and are said to bear some potential to penetrateground fog and thin cloud layers. The precise significance of the single photon LiDAR technology to forestry applicationsremains a topic of ongoing research and efforts are required to filter these data from noise induced by solar radiation (Swatantranet al., 2016).

3.08.3 Global Scale

Mapping forest canopy characteristics at a global scale is beneficial to sustainable resource management as well as climate science.Notably, one instrument, the GLAS instrument aboard the ICESat satellite, has been used to infer tree height information world-wide. Due to its discontinuous sampling pattern along tracks (i.e., spaced 170 m along the track and several kilometers across track),studies have generally focused on the fusion of GLAS data with other Earth observation data. For example, Lefsky (2010) used theMedium Resolution Imaging Spectroradiometer (MODIS) to extrapolate tree height estimates that were derived from GLAS foot-print data. The MODIS data were segmented based on their spectral and textural attributes and a regression model was calibratedto relate spectral and textural indices derived from the image segments to the waveform-derived height estimates. The latter esti-mates were obtained from coincident GLAS footprints and tree height measurements from select field sites. Simard et al. (2011)produced a similar map based on extrapolated GLAS tree height estimates using 1-km MODIS data and ancillary climatologydata (Baldocchi et al., 2001). Field mensuration data was gathered from 66 FLUXNET sites that are distributed globally for thecollection of micro-meteorological data. Simard et al. found that the former map by Lefsky generally produced taller estimatescompared to theirs for regions that are characterized by a sloping terrain, whereas on flat terrain the latter map produced generallytaller estimates. A cross-validation where Lefsky’s tree height estimates were compared against the 66 validation sites used by Simardet al., and the authors found no correlation (r2¼0.01) at all, shedding doubt on the map quality. Authors have emphasized thatestimation errors may be large over sloping terrain and for complex or dense forest canopies and processing methodologies mayrequire sophisticated techniques to decompose the waveform (Wang et al., 2016).

106 Vegetation Structure (LiDAR)

Despite the large uncertainties around global height estimates, some authors have advocated the use of such maps with inferencetechniques to estimate global biomass (Hese et al., 2005; Hu et al., 2016). For example, Hu et al. (2016) used random forest clas-sifiers and ancillary EO data to extrapolate GLAS footprints and produce a global map of waveform parameters. Using a MonteCarlo simulation (n¼100), regression models were used to link biomass reports from literature studies to the spatially continuouswaveform-parameter and other EO data. This produced 100 aboveground biomass maps fromwhich mean and standard deviationswere derived. Similar ensemble methods have often been demonstrated to be successful in remote sensing literature; However,cross-validation results often indicate regression coefficients around 50% and it is not always clear whether such global productsare rightfully characterized as “globally consistent” (e.g. Healy et al., 2012) or whether further research into the overall map qualityis required before such predicate can be assigned. For example, the study by Hu et al. (2016), reported an R2 and RMSE from selectedvalidation sites of 56% and 87.53 Mg/ha (with a mean global above-ground biomass estimate of 210.09 Mg/ha), while at the eco-zone level the R2 was even reduced to 0.38%.

Much more effort into the development and accessibility of LiDAR instruments and data seems to be needed to deliver globaltree height and biomass information with an accuracy that is relevant in the light of monitoring and reporting commitments such aswithin the framework of the United Nations’ Reduction of Emissions from Degradation and Deforestation (UN REDDþ). Uncer-tainty also exists around the extrapolation of allometric relationships that is required to produce global biomass estimates. Due topractical reasons, allometric relationships have predominantly been derived among smaller trees and only from partial treemeasurements, adding great uncertainty around the applicability of these relationships for taller trees (Disney et al., 2016).Proposed missions include the ICESat-2 and the Global Ecosystem Dynamics Investigation (GEDI) that both provide promisingvenues for global forest mapping efforts (Qi and Dubayah, 2016; Moussavi et al., 2014).

3.08.4 Regional to Stand Level

LiDAR acquisitions at regional levels currently fall entirely within the domain of airborne remote sensing; historical acquisitions ofLiDAR data from orbit are, of course, still available from the IceSat/GLAS instrument. Regional, wall-to-wall acquisitions are oftenconducted with an interannual regularity as part of a national effort to monitor the change of the visible landscape, including land-use change, tectonic movements, and the integrity of natural or man-made structures. Monitoring efforts may also include theregular safeguarding of the integrity of power lines or ensuring that tax is paid at household level for structures such as shedsand garages that were built as add-ons to existing real estate. It is within these overarching monitoring programs that informationabout the land surface can be collected in a cost-effective manner and bemade available to the various domains of investigation. It isthrough the same collaborative effort that goes into the collection of the data, that researchers have been able to learn about themeans to process LiDAR data to useful end products. One of the most elementary of approaches is to aggregate LiDAR returnsin percentile distributions for use in linear regression models to predict the forest inventory attributes of interest (see e.g. Naesset,2014, and references therein). Of these percentiles, the median, 75%, 95%, and 99% percentiles have been proven most useful.While the median is strongly correlated with the transparency of the canopy medium to light propagation and relates to inventoryattributes such as biomass or stocking density, the upper percentiles are used to estimate tree height or Lorey’s mean height, definedas the average height of all trees in the stand weighted by their basal area. The upper percentiles remove false returns from theairspace above the trees that may result from the interception of laser light by avifauna or an otherwise undesired behavior ofthe digitizing electronics inside the LiDAR scanner.

Accurate retrieval of tree height, naturally, also requires that terrain elevation can be modeled with a suitable fidelity. Radar-derived terrain maps such as the Shuttle Radar Topography Mission (SRTM; Farr et al., 2007) provides terrain topology ata 30 m or 90 m resolution, which may not be adequate for many forestry applications. Besides, detailed information about theterrain underneath the forest canopy can often be extracted from LiDAR data. Perhaps the most challenging step in producingDEMs from LiDAR data is the separation of returns into ground and nonground hits (Liu, 2008); however, in order to speed upprocessing or to facilitate processing on inexpensive computer hardware, reduction of data dimensionality is of critical importance(Isenburg et al., 2006). Effects of forest structure and acquisition parameters on DEM accuracy have been reported in a number ofstudies. For example, Reutebuch et al. (2014) reported DEM errors of around 20–30 cm in heavily forested areas. Khosravipour et al.(2015) investigated effects of terrain slope on treetop detection and found that treetops were best identified on the raw (i.e., notnormalized with respect to terrain elevation) LiDAR data. Bater and Coops (2009) investigated a number of different DEM extrac-tion techniques and found that natural neighbor interpolation produced superior results against linear, quintic, spline-based andfinite difference methods, while Su and Bork (2006) found inverse distance weighing to produce superior results over kriging andspline-based interpolation for a number of different vegetation types in Alberta, Canada.

This statistical approach to LiDAR analysis has proved widely useful and much of the literature to this end focuses on thesampling designs needed to balance estimation accuracy with the costs related to field-data collection. These issues are discussedin Maltamo et al. (2014) and Van Leeuwen and Nieuwenhuis (2010) among others. Significant early collections from researchinstruments include the NASA campaigns conducted with the Land Vegetation and Ice Sensor (LVIS; Blair et al., 1999), and the Scan-ning LiDAR Imager of Canopies by Echo Recovery (SLICER; Harding et al., 2001). A low-cost profiling instrument, the PortableAirborne Laser System (PALS), was designed and employed by Nelson et al. (2003). Commercial scanners evolved rapidly as a resultof the growing demand for high-resolution data. Europe proved an important consumer of such high-resolution data and some ofits smaller countries were the first to adopt nation-wide, interannual, wall-to-wall LiDAR surveys, for a range of applications, not just

Vegetation Structure (LiDAR) 107

forestry. For example, between 1997 and 2003, the Netherlands started a program to acquire national LiDAR data with pointsampling densities around 1 point per 16 square meter, which around 2009 was being replaced by a second survey with over 10points per square meter (http://www.ahn.nl/). Initiated by the national office for dike safety and water affairs (Rijkswaterstaat),these data were made available to a wide group of users including landscape designers and nature conservationists. The acquisitionof interannual LiDAR data facilitates the modeling of land surface change. However, initial attempts with early commercial instru-mentation were not without problems. The low point sampling density in combination with the often unknown or proprietarycharacteristics of the individual scanners used (Disney et al., 2010; Armston et al., 2013) meant that temporal changes in thedata could not be attributed exclusively to land surface changes, but often to differences between scanners and even survey charac-teristics. These latter effects can be mitigated by using higher ground-sampling densities, such that details in terrain elevation can bemore clearly distilled from the point cloud data. A more insightful account of the effects of aging on the structure of forests might bederived when the same instrument is flown across a natural gradient of forests of different ages and developmental stages or whenforest growth is evaluated against a gradient of environmental conditions. An example of this type of campaign has been providedbyWulder et al. (2012a) where a 25,000 km trajectory was flown across the Canadian boreal shield, providing an excellent source ofdata to study gradients in forest structure across the landscape and how these gradients may relate to other variables of interest suchas primary productivity and stand recovery after a disturbance event (Bolton et al., 2013).

Airborne LiDAR remote sensing in forest research is not limited to the characterization of traditional inventory attributes orgrowth rates. Studies have investigated the potential to also predict species or species groups on the basis of laser intensity data(Korpela et al., 2010; Ørka et al., 2009). Perhaps not surprisingly does availability of leaf-on and leaf-off LiDAR intensity data(i.e., relating to different seasons) improve the capacity to discriminate between evergreen and deciduous species, even thougheffects of scans angle and flying altitude may have to be minimized through calibration procedures (Kim et al., 2009). Perhapsgreater use is derived from merging spectral information from imaging spectroscopy with additional information on the tree crownshape, either to delineate individual tree crowns (Baldeck et al., 2015) or to provide information about branching heights andcrown radii (Brandtberg et al., 2003; Holmgren and Persson, 2004). Fusion of LiDAR with radar data has also successfully beendemonstrated as a means to estimate vegetation biomass. For example, Mitchard et al. (2012) used radar-derived products to clas-sify a 5000 km2 area in Gabon into vegetation classes for which average biomass estimates were assigned using LiDAR and fielddata, while Tsui et al. (2013) used kriging, cokriging, and regression kriging to predict aboveground biomass across a 25 km2

area in British Columbia with combined LiDAR transects and radar datasets, and found synergistic value arising from the frequencyof observations and all-weather capacity of the radar data, with potential to reduce cost of flight campaigns or the need fora complete wall-to-wall coverage of LiDAR data.

LiDAR has also been used to predict wood quality attributes (e.g. Hilker et al., 2013). The rationale behind this endeavor is thelinkage between stem attributes, visible on the outside of the bark or from the social status of trees within the stand, and intrinsicwood attributes such as ring width, microfibril angle or lignin content, all of which affect the mechanical properties of the wood.While many of these relations have long been known in silviculture, many studies have also reported contradicting results (VanLeeuwen et al., 2011). Some of these disagreements potentially arise from the limitations in the sampling design or the fact thatit is simply not feasible to compile a comprehensive database of wood fiber properties large enough to accurately cover the variationwithin trees as well as throughout the landscape. The wide-scale acquisition of LiDAR data offers a unique potential to characterizestands and even individual tree structural properties. By linking these data to quality measurements taken at the receiving mills, itmay become feasible to greatly enhance the study of relations between wood fiber traits and stand location.

3.08.5 Plot to Tree Level

An important aspect of the success of LiDAR for vegetation applications is the range of spatial scales it covers. The technologyprovides for wide-area coverage, but at the same time, each laser return represents a footprint area that can be as small as a fewcentimeters in diameter. As a result, contours of individual tree crowns in airborne LiDAR data sets can often be matched by fieldobservations, including from TLS. Fig. 1 illustrates the match than can be seen between two independent acquisitions of point clouddata: an airborne LiDAR point cloud and its derived canopy height model (CHM; in blue) and a terrestrial laser scan and its derivedCHM (in green). The match between the collocations is clearly visible and any discrepancy between the crown outlines is, besidesmis-registration, at least partially due to the movement of the crowns in the wind, which can amount to several meters dependingon weather conditions.

Traditionally, ground-truth data are required to calibrate and validate models that relate remotely sensed reflectance (LiDAR,optical) or back-scattered signals (radar) to the inventory attributes of interest. TLS has drastically increased the possibilities for cali-bration and validation of remote sensing data products by facilitating the mapping of 3D forest structure. A good overview of TLSapplications is provided by Liang et al. (2016) which includes works related to forest resource management as well as ecology. Someof the earliest uses of TLS in forestry were demonstrated by Lichti et al. (2002) and Hopkinson et al. (2004). Popular applicationsincluded the retrieval of stem diameter and taper and this has since remained an area of interest for its significance to the industry aswell as biomass modeling (Bienert et al., 2006; Gorte andWinterhalder, 2004; Pueschel et al., 2013; Kelbe et al., 2015; Calders et al.,2015b). Recent extensions of this work include an intercomparison of stem retrieval algorithms, led by the Finnish GeodeticResearch Institute (http://www.eurosdr.net/). A major challenge in developing these algorithms is the need for a robust treatmentof outlier detection, due to the potentially large amounts of occlusion and scattering by leaves, needles, and branches encroaching

108 Vegetation Structure (LiDAR)

into the region of interest. For this, hypothesis-generating-and-testing approaches are often utilized, such as the Hough transform(Hough, 1962) and Random Sample Consensus (RANSAC; Fischler and Bolles, 1981). These approaches rely on a certain imple-mentation of a voting procedure where the individual data samples, or laser returns, cast a vote over the plausibility of a previouslygenerated hypothesis or set of hypotheses. Least squares approaches may be used to fit stem diameters after outliers have beenremoved (Liang et al., 2012). A reliable retrieval of the complete branching structure proved difficult with early scanners as a resultof the low point sampling densities. Côté et al. (2009) demonstrate the reconstruction of branching topology based on Dijkstra’salgorithm. Given a root node, the algorithm searches for the shortest path between points in the point cloud and the root node. Bybinning the paths together, the authors were able to derive a branching topology. Especially for fine twigs, the algorithm does notnecessarily find the correct branching topology, but one that is demonstrated to be close to the real topology. Others, includingLivny et al. (2010) and Raumonen et al. (2013), demonstrated similar techniques based on graph network algorithms. Otherthan for simple inventory studies, the recovery of branching topology requires high-resolution scans and, hence, a denser spacingof scanner locations to acquire the data. Since the point cloud data is often too sparse to reliably detect fine branches, the recon-struction of tree branching structures is typically characterized by two stages: one where the major branches are detected anda second step in which small twigs are added based on a space colonization algorithm (Runions et al., 2007) or simply by imputingcrown structures that were derived with the use of tree modeling software (Morton et al., 2016; Van Leeuwen et al., 2013; Weber andPenn, 1995; Pradal et al., 2009). Space colonization follows and extends the work of Prusinkiewicz and Lindenmayer (2004) whoinvestigated the occurrence of fractals, or self-similar mathematical patterns, in plant architectures. The reconstruction of individualtrees has grown out to the reconstruction of plots of increasingly larger size. Perhaps the largest to date has been achieved inWythamWoods, Oxford, UK (Calders et al., 2016) where an area of 300 m by 200 m has been scanned and reconstructed. Scan locationswere laid out in a 20 m by 20 m regular grid and were taken with a RIEGL VZ-400 at an angular interval of 0.04� in both azimuthand zenith directions (www.riegl.com). Sites of this kind are crucial for the validation of theories and models around canopy radi-ation transfer and provide a way for scaling between leaf-level physiology and stand-level, top-of-canopy reflectance signals as theyare observed by remote sensing platforms. This is achieved using rendering applications such as ray tracing and path tracing (Disneyet al., 2000; Govaerts and Verstaete, 1998; Goodenough and Brown, 2012). Sites scanned in detailed 3D using TLS like this facilitatean end-to-end traceability chain of radiative and physiological quantities. The reconstructed 3D canopy architecture can be useddirectly in radiative transfer schemes and growth models. However, these digitized representations of the real stand are not onlysignificant as surrogate “truth” against which radiative transfer models can be validated; perhaps more important is the potentialto modify the 3D model scene and, hence, the ability to predict changes in top-of-canopy reflectance resulting from the particularstructural or spectral attributes (Yao et al., 2016). After all, and despite the high level of realism that can be achieved using modernscanning and computer graphics techniques, a 3D model derived from laser scanning may never be interpreted as being a copy ofthe real forest. Many structural details must be simplified or even ignored, such as sunfleck dynamics caused by the swaying ofbranches and leaves in the wind. While the significance of these effects on modeling results is unclear, they can potentially be testedand quantified, if required. As a result, uncertainties in the radiative transfer model predictions can be traced explicitly and there isgreat potential to study the validity of various modeling assumptions across a range of different plant architectural types or acrossspatial scales ranging from leaf to stand level.

Fig. 1 Illustration of coregistered point clouds that were acquired from the ECHIDNA® terrestrial laser scanner (green) and an airborne LiDARsystem (blue). Corresponding canopy height models are shown in matching colors. Data was acquired from site DF49, Vancouver Island, BC, 2008.

Vegetation Structure (LiDAR) 109

Early contributions to TLS include also the development of the ECHIDNA� laser scanner, by the Commonwealth Scientific andIndustrial Research Organization (CSIRO), Melbourne, Australia. The ECHIDNA� scanner features a relatively broad-beam laserand full-waveform recording capabilities. The wide laser footprint facilitates capturing multiple returns, a feature principallyused to derive a gap probability distribution as a function of range, from which a foliage profile with canopy depth can becomputed. Given assumptions around the leaf angle orientation distribution and reflective (transmissive) properties, the returnedlight intensity at the specified range that is recorded by the instrument is integrated into a postprocessing stage and is used to derivethe fraction of uncollided beam as a function of range, or the gap probability. This measure is calibrated using laser footprints inopen sky and those fully intercepted by tree stems and the hemispherical sampling pattern of the instrument facilitates that gapprobability as a function of range can be specified for different zenith angles. Since the gap probabilities for different leaf angledistributions are similar around the 57-degree zenith angle (Welles and Norman, 1991), this portion of the scan is often usedto derive a vertical profile of the gap probability and, from this, a vertical foliage profile (Jupp et al., 2009). The hemisphericaldata also provide for the quantification, in angular terms, of the gap size distribution that can be related to the foliage clumpingindex (Zhao et al., 2012) that is unity when foliage is dispersed randomly throughout the canopy space and is within the rangezero to one for clumped canopies. The gap size distribution has earlier been used to quantify clumping (Chen and Cihlar,1995) using transect measurements and hemispherical photography (Leblanc et al., 2005), however, the availability of range infor-mation provides, in principle, for analysis of any vertical stratification in foliage clumping.

The DWEL (Dual-Wavelength Echidna LiDAR) is a follow-up of the ECHIDNA instrument and features two laser sources oper-ating at different wavelengths and is principally being developed by Boston University in collaboration with CSIRO. While the firstband is located at 1064 nm the second band is located at 1548 nm and is used to differentiate water absorptions such that woodyand foliage material can more effectively be discriminated (Douglas et al., 2012, 2015). Another dual-wavelength TLS system, theSalford Advanced Laser Canopy Analyser (SALCA) has been under development by Salford University (Danson et al., 2014) andfeatures similarly positioned wavebands (centered at 1063 and 1545 nm, respectively) and studies have demonstrated both the cali-bration of reflectance as well as the detection of water stress from data acquired with the SALCA scanner (Gaulton et al., 2013).

3.08.6 Simulation

The ability to describe individual crown outlines facilitates modeling of intercrown shading and its effects on remote sensing obser-vations. Hilker et al. (2008, 2012) used a hill-shade model derived from a CHM and estimated the shadow fraction within the fieldof view of a tower-mounted revolving spectrometer. This shadow fraction was used to correct for the effect of a changing illumina-tion environment on the state of the photochemical reflectance index that relates to the light-use efficiency of photosynthesis(Gamon et al., 1992). While a hill-shade model does not provide insight into the effects of intracrown shading, these effects canbe included if a suitable tree architectural model were to be used at the respective treetop locations. Instead of using an opaquesurface, the shading would then be derived from a porous model of canopy structure that matches, at a minimum, the topologyof the top of the canopy and the constellation of dominant trees in the stand. This might also include the morphology of individualbranches where such information is available. To illustrate, while a typical workstation might have registered a maximum of 32GBof RAM in 2008, at the time of writing (2017) we find desktop workstations with up to 1 TB RAM. As a result, increasingly complexvector models can be handled with ease to allow a separation of foliage into sunlit and shaded leaves as well as how each of thesefractions are visible to a remote sensor. Examples of such efforts can be found in ray tracing (path tracing) literature, where paths ofindividual rays of light are followed inside virtual scenes that comprise a geometrically explicit representation of the forest canopy,including the position, orientation, and shape of leaves, needles and branches. Upon interaction with objects within a 3D scene,rays are either scattered (reflected, transmitted) or absorbed based on the spectral composition of the illumination and materialproperties. Ray tracing simulations can be used to compute angular dependence in top-of-canopy reflectance (Widlowski et al.,2013, 2015; Lewis, 1999; Gastellu-Etchegorry et al., 2015) as well as canopy transmission (Yao et al., 2016) and interception(Van der Zande et al., 2011) and can be used to calibrate or validate other models that link remote sensing observations to surfacecharacteristics. While the virtual, 3D scenes may not represent the true structure and spectral composition of any real forest, theyhave become increasingly accurate and sophisticated. These models provide excellent means to study issues of scale in radiativetransfer theory and have therefore primarily been used for the validation and calibration of other models using a series of well-documented benchmark scenes (Widlowski et al., 2013, 2015). Fig. 2 shows an example of the simulation of surface reflectancefrom a vector model of a conifer and illustrates how shadow effects appear. Path tracing differs from ray tracing in the way splittingof the scattering photon packets is achieved. While ray tracing implements an actual branching of the photon packet in multipledirections, a path tracer samples an individual direction, which minimizes the number of secondary rays (i.e., those that propagateafter a first intersection with the scene) that need to be queued for evaluation and this proves beneficial in simulations in the near-infrared where scattering is significant.

The ability to model shadow fractions across multiple scales simultaneously opens new opportunities for the design and param-eterization of vegetation models. Multiresolution models reduce the need for clumped parameters, hence the effect of differentmorphological traits on physiological functioning can be studied in greater detail. This provides for the study of effects of shootmorphology on the radiation budget or for a decomposition of stand-level productivity into the contributions made by differentcanopy height strata (e.g. Van Leeuwen et al., 2015) but this remains an area of ongoing research. The virtual scenes as such act aslaboratories where parameter sensitivities can be traced across different spatial scales (Lewis, 1999). However, parameter

110 Vegetation Structure (LiDAR)

uncertainties derived from these virtual models are to be interpreted with care and the high level of realism of these virtual scenescan perhaps encourage overconfidence in the verisimilitude of these scenes; validation is ever more important. Since these virtualscenes can be developed from TLS point clouds that were already acquired with state-of-the-art laser scanning equipment, a rigorousmeans to validate these scenes is often not available. Moreover, many scattering characteristics may be excluded, such as specularscattering effects, fine surface texture and scattering within leaves or needles. While ray tracing (or path tracing) does not necessarilyneglect such effects, computational tractability often requires that certain details about scattering characteristics must be omitted.This raises questions as to how confidence intervals around retrieved model parameters ought to be interpreted.

When models are used for inference, a model state serves as an estimate of reality (i.e., system state), but of course is in itself notan attribute of reality. Similarly, a confidence interval should be regarded as an attribute of the model; provided that the modelingassumptions are valid, this confidence interval would coincide with the true distribution of (real-world) system states. However, it isinconsistent to assume that both a 1D layered representation as well as a detailed multiscale 3D model bears the same relation toreality as may be implicit in the underlying modeling assumptions. Thus, while laser scanning has drastically increased the level ofrealism by which radiative transfer models can be parameterized, more work remains in assessing the accuracy of scene represen-tation relative to the real forest stands they resemble. Validation efforts may often have to resort to quantifying morphological andgeometric discrepancies between the algorithmic reconstructions from point cloud data versus manual reconstructions conductedby experienced researchers (Boudon et al., 2014).

3.08.7 Data Fusion

Considerable effort has also been paid to the fusion of LiDAR remote sensing data with other remotely sensed data, notably hyper-spectral. For example, the Carnegie Airborne Observatory (CAO) team lead by Gregory Asner has demonstrated significant potentialto simultaneously acquire biochemical and structural information about the forest landscape (Baldeck et al., 2015; Asner et al.,2015). Other recent efforts in the development of multisensor airborne mapping systems that include LiDAR have been madeby the US-based National Ecological Observation Network (NEON) (Kampe et al., 2010); the National Aeronautics and SpaceAdministration (NASA) with the G-LiHT instrument that combines a hyperspectral, a LiDAR, and a thermal imager (Cook et al.,2013); and the UK-based Natural Environment Research Council (NERC) with the Airborne Research Facility (ARF) (www.bas.ac.uk). The simultaneous access to structural and spectral information allows modelers to place further constraints on radiativetransfer parameterizations. This information can be used, for example, to discriminate between ground-level, understory, and over-story reflectance which may be impossible based on either hyperspectral or LiDAR data alone (Koetz et al., 2006, 2007). Other

Fig. 2 Showing renderings acquired with a path tracer code and a conifer tree model. The scene was built from a repetitive instantiation of shootinstants.

Vegetation Structure (LiDAR) 111

applications may be to improve classification using information about stand height and canopy volume profiles (Jones et al., 2010),or improve timber volume estimates (Breidenbach et al., 2010). Much is yet to be learned, however, about techniques to invertradiative transfer models using combined hyperspectral and LiDAR data. The use of multispectral LiDAR scanning systems couldgreatly facilitate biophysical characterization as the spectral information is then stored as attribute of the point cloud (or waveform)data, such that effects of spectral mixing at the pixel level are better understood (Woodhouse et al., 2011; Morsdorf et al., 2009;Hakala et al., 2012). The broad-scale coverage of fine-resolution data facilitates the estimation of surface characteristics witha marked multiscale dependence. This may be beneficial for highly heterogeneous landscapes or regions with very high biodiversity(Ter Steege et al., 2013; Anderson et al., 2004). Climate-related research into vegetation response requires a characterization ofbiogenic emissions from the land surface. However, nitrous oxide (NOx) and methane fluxes remain challenging to estimatewith a desirable accuracy since both are strongly dependent on gradients in terrain elevation and soil ecology and hydrology (Bridg-ham et al., 2013). The use of spectral and topographic information, in tandem, could possibly be used to infer spatial and seasonalpatterns in soil hydrology, with relevance to the modeling of fine-scale biogenic fluxes (Mohammed, 2015). Similarly, when remotesensing observations over forest canopies are aggregated to a coarse resolution (e.g., >100 m) and then used for the estimation ofprimary productivity, the underlying physiological processes of photosynthesis and respiration may not be understood in detaileven if the computed estimates match validation data closely. The reason is that any model, whether process-based or empirical,has at least some potential to emulate a set of observations for reasons other than a correct understanding of the underlying physicalprocessesdthe right answer but for the wrong reasons. Indeed, many detailed physical and physiological underpinnings are neces-sarily avoided in models that are to be evaluated at regional to global scales. A proper estimation framework may therefore requirenot only a coupling of models of various scale, such as demonstrated in many process-based models of stand productivity (e.g.Ibrom et al., 2006; Medlyn et al., 2002; Wang and Jarvis, 1990) but also a translation of these process-based models to empirical“drop-in” functions that are able to mimic the overall model behavior at a lower computational cost (Gomez-Dans et al., 2016).This allows the complexity of the process-based models to be distilled, to reduce the number of model parameters needed, andoften speeds up computations by several orders of magnitude, while the emulated model is approximated with a known tolerance.Efforts to this aim seem increasingly beneficial for the assimilation of knowledge and observations from a diverse set of disciplinesand it is by combining knowledge from these various disciplines that progress in understanding the Earth system can be achieved.

3.08.8 Conclusion

This chapter summarizes research in LiDAR remote sensing of vegetation structure. The breadth of work conducted in this area hasgrown significantly since the early 2010s, indicating both the development and expansion of remote sensing technology as well asthe challenge in applying it. As a direct measurement technique, LiDAR data are more accessible and easier to handle than manyother remote sensing products. LiDAR allows for new development and application of both empirical and process-based models.The latter have the benefit that site-specific calibrations may not be needed, possibly reducing costs associated with the data acqui-sition campaigns. In addition to providing near-direct measurements of vegetation structure, LiDAR has also been used to provideindirect measures related to biodiversity, energy balance, and biogenic fluxes. The coming years will most certainly see continuingprogress in the use of LiDAR data as newer systems become available, including multiwavelength and photon-counting LiDARsystems. Future datasets are likely to provide ever-increasing point densities, allowing new research questions to be addressed.Many of the future opportunities and applications are likely to involve data fusion and integration, as well as the improved cali-bration and validation of other remote sensing data products. Much of the knowledge we gained over the last decades includesprocesses at microscopic and macroscopic scale, whereas the intermediate range has been paid comparatively little attention(Passioura, 1979). With the technologies that have become available to the field of remote sensing, an ever-increasing potentialis anticipated to further integrate the knowledge that has been gained from both microscopic and macroscopic domains, hopefullythrowing light (literally and metaphorically) on this intermediate region.

See also: 1.15. Lidar Sensors From Space.

References

Anderson, M. C., Norman, J. M., Mecikalski, J. R., Torn, R. D., Kustas, W. P., & Basara, J. B. (2004). A multiscale remote sensing model for disaggregating regional fluxes tomicrometeorological scales. Journal of Hydrometeorology, 5, 343–363.

Armston, J., Disney, M., Lewis, P., Scarth, P., Phinn, S., Lucas, R., Bunting, P., & Goodwin, N. (2013). Direct retrieval of canopy gap probability using airborne waveform LiDAR.Remote Sensing of Environment, 134, 24–38.

Asner, G., Ustin, S., Townsend, P., Martin, R., & Chadwick, K. (2015). Forest biophysical and biochemical properties from hyperspectral and LiDAR remote sensing. InP. S. Tenkabail (Ed.), Land resources monitoring, modeling, and mapping with remote sensing (pp. 429–448). Boca Raton: CRC Press.

Baccini, A., Goetz, S. J., Walker, W. S., Laporte, N. T., Sun, M., Sulla-Menashe, D., Hackler, J., Beck, P. S. A., Dubayah, R., Friedl, M. A., Samanta, S., & Houghton, R. A. (2012).Estimated carbon dioxide emissions from tropical deforestation improved by carbon-density maps. Nature Climate Change, 2, 182–185.

Baldeck, C. A., Asner, G. P., Martin, R. E., Anderson, C. B., Knapp, D. E., Kellner, J. R., & Wright, S. J. (2015). Operational tree species mapping in a diverse tropical forest withairborne imaging spectroscopy. PLoS One, 10, e0118403.

112 Vegetation Structure (LiDAR)

Baldocchi, D., Falge, E., Gu, L., Olson, R., Hollinger, D., Running, S., Anthoni, P., Bernhofer, C., & Davis, K. (2001). FLUXNET: A new tool to study the temporal and spatial variabilityof ecosystem-scale carbon dioxide, water vapor, and energy flux densities. Bulletin of the American Meteorological Society, 82, 2415–2434.

Bater, C., & Coops, N. (2009). Evaluating error associated with LiDAR-derived DEM interpolation. Computers and Geosciences, 35, 289–300.Bienert, A., Maas, H.-G., & Scheller, S. (2006). Analysis of the information content of terrestrial laserscanner point clouds for the automatic determination of forest inventory

parameters. In ISPRS WG VIII/11 & EARSeL Joint Conference ‘3D Remote Sensing in Forestry’. Vienna.Blair, J. B., Rabine, D. L., & Hofton, M. A. (1999). The laser vegetation imaging sensor: A medium-altitude, digitisation-only, airborne laser altimeter for mapping vegetation and

topography. ISPRS Journal of Photogrammetry and Remote Sensing, 54, 115–122.Bolton, D. K., Coops, N. C., & Wulder, M. A. (2013). Measuring forest structure along productivity gradients in the Canadian boreal with small-footprint LiDAR. Environmental

Monitoring and Assessment, 185, 6617–6634.Bosse, M., Zlot, R., & Flick, P. (2012). Zebedee: Design of a spring-mounted 3-D range sensor with applications to mobile mapping. IEEE Transactions on Robotics, 28,

1104–1119.Boudon, F., Preuksakarn, C., Ferraro, P., Diener, J., Nacry, P., Nikinmaa, E., & Godin, C. (2014). Quantitative assessment of automatic reconstructions of branching systems

obtained from laser scanning. Annals of Botany, 114, 853–863.Brandtberg, T., Warner, T., Landenberger, R. E., & McGraw, J. B. (2003). Detection and analysis of individual leaf-off tree crowns in small footprint, high sampling density LiDAR

data from the eastern deciduous forest in North America. Remote Sensing of Environment, 85, 290–303.Breidenbach, J., Næsset, E., Lien, V., Gobakken, T., & Solberg, S. (2010). Prediction of species specific forest inventory attributes using a nonparametric semi-individual tree crown

approach based on fused airborne laser scanning and multispectral data. Remote Sensing of Environment, 114, 911–924.Breon, F. M., O’Brien, D. M., & Spinhirne, J. D. (2005). Scattering layer statistics from space borne GLAS observations. Geophysical Research Letters, 32, L22802. http://dx.doi.org/

10.1029/2005GL023825.Bridgham, S. D., Cadillo-Quiroz, H., Keller, J. K., & Zhuang, Q. (2013). Methane emissions from wetlands: Biochemical, microbial, and modeling perspectives from local to global

scales. Global Change Biology, 19, 1325–1346. http://dx.doi.org/10.1111/gcb.12131.Bucksch, A., Lindenberg, R., & Menenti, M. (2010). SkelTredRobust skeleton extraction from imperfect point clouds. The Visual Computer, 26, 1283–1300.Bunting, P., Armston, J., Clewley, D., & Lucas, R. M. (2013a). Sorted pulse data (SPD) library. Part II: A processing framework for LiDAR data from pulsed laser systems in terrestrial

environments. Computers and Geosciences, 56, 207–215. http://dx.doi.org/10.1016/j.cageo.2013.01.010.Bunting, P., Armston, J., Lucas, R. M., & Clewley, D. (2013b). Sorted pulse data (SPD) library. Part I: A generic file format for LiDAR data from pulsed laser systems in terrestrial

environments. Computers and Geosciences, 56, 197–206. http://dx.doi.org/10.1016/j.cageo.2013.01.019.Calders, K., Newnham, G., Burt, A., Murphy, S., Raumonen, P., Herold, M., Culvenor, D., Avitabile, V., Disney, M., Armston, J., & Kaasalainen, M. (2015a). Nondestructive estimates

of above-ground biomass using terrestrial laser scanning. Ecology and Evolution, 6, 198–208. http://dx.doi.org/10.1111/2041-210X.12301.Calders K, Newnham G, Armston J, Disney MI, Schaaf C, and Paynter I (2015a) Terrestrial LiDAR for forest monitoring, Chapter 2.10 in GOFC-GOLD (2015) A sourcebook of

methods and procedures for monitoring and reporting anthropogenic greenhouse gas emissions and removals associated with deforestation, gains and losses of carbon stocks inforests remaining forests, and forestation. GOFC-GOLD Report version COP21–1, (GOFC-GOLD Land Cover Project Office, Wageningen University, The Netherlands).

Calders, K., Burt, A., Origo, N., Disney, M., Nightingale, J., Raumonen, P., & Lewis, P. (2016). Large-area virtual forests from terrestrial laser scanning data. In 2016 IEEEInternational Geoscience and Remote Sensing Symposium (IGARSS). In , IEEE International Geoscience and Remote Sensing Symposium, 1 JanuaryIEEE International Geoscienceand Remote Sensing Symposium Proceedings, IEEE (pp. 1765–1767). http://dx.doi.org/10.1109/IGARSS.2016.7729452.

Cawse-Nicholson, K., Van Aardt, J., Hagstrom, S., Romanczyk, P., Schaaf, C., Strahler, A., Li, Z., & Krause, K. (2014). Improving waveform LiDAR processing toward robustdeconvolution of signals for improved structural assessments. In M. D. Turner, G. W. Kamerman, L. M. Wasiczko-Thomas, & E. J. Spillar (Eds.), Laser radar technology andapplications XIX; and atmospheric propagation XI. Baltimore: SPIE, 277 pp.

Chen, J. M., & Cihlar, J. (1995). Quantifying the effect of canopy architecture on optical measurements of leaf area index using two gap size analysis methods. IEEE Transactions onGeoscience and Remote Sensing, 33, 777–787.

Cook, B. D., Corp, L. W., Nelson, R. F., Middleton, E. M., Morton, D. C., McCorkel, J. T., Masek, J. G., Ranson, K. J., Ly, V., & Montesano, P. M. (2013). NASA Goddard’s LiDAR,Hyperspectral and Thermal (G-LiHT) airborne imager. Remote Sensing, 5, 4045–4066. http://dx.doi.org/10.3390/rs5084045.

Côté, J.-F., Widlowski, J.-L., Fournier, R. A., & Verstraete, M. M. (2009). The structural and radiative consistency of three-dimensional tree reconstructions from terrestrial LiDAR.Remote Sensing of Environment, 113, 1067–1081. http://dx.doi.org/10.1016/j.rse.2009.01.017.

Côté, J.-F., Fournier, R. A., & Egli, R. (2011). An architectural model of trees to estimate forest structural attributes using terrestrial LiDAR. Environmental Modelling & Software, 26,761–777. http://dx.doi.org/10.1016/j.envsoft.2010.12.008.

Côté, J.-F., Fournier, R. A., Frazer, G. W., & Niemann, K. O. (2012). A fine-scale architectural model of trees to enhance LiDAR-derived measurements of forest canopy structure.Agricultural and Forest Meteorology, 166–167, 72–85. http://dx.doi.org/10.1016/j.agrformet.2012.06.007.

Danson, F. M., Gaulton, R., Armitage, R. P., Disney, M., Gunawan, O., Lewis, P., Pearson, G., & Ramirez, A. F. (2014). Developing a dual-wavelength full-waveform terrestrial laserscanner to characterize forest canopy structure. Agricultural and Forest Meteorology, 198–199, 7–14.

Disney, M. I. (2016). Remote sensing of vegetation: Potentials, limitations, developments and applications. In K. Hikosaka, U. Niinemets, & N. P. R. Anten (Eds.), Advances inphotosynthesis and respirationCanopy photosynthesis: From basics to applications (pp. 289–331). Berlin: Springer, ISBN 978-94-017-7290-7. http://dx.doi.org/10.1007/978-94-017-7291-4.

Disney, M. I., Lewis, P., & North, P. R. J. (2000). Monte Carlo ray tracing in optical canopy reflectance modelling. Remote Sensing Reviews, 18, 163–196. http://dx.doi.org/10.1080/02757250009532389.

Disney, M. I., Kalogirou, V., Lewis, P. E., Prieto-Blanco, A., Hancock, S., & Pfeifer, M. (2010). Simulating the impact of discrete-return LiDAR system and survey characteristics overyoung conifer and broadleaf forests. Remote Sensing of Environment, 114, 1545–1560.

Disney, M., Burt, A., Calders, K., Raumonen, P. A., Herold, M., Lewis, P., Boni Vicari, M., Rowland, L., Meir, P., & Mitchard, E. (2016). 3D Measurements of Tropical ForestStructure for BIOMASS, Morphology and Calibration and Validation of Satellite Observations. In Paper presented at Living Planet Symposium 2016. Czech Republic: Prague, 9/05/16 – 13/05/16.

Douglas ES, Strahler A, Martel J, Cook T, Mendillo C, Marshall R, Chakrabarti S, Schaaf C, Woodcock C, Li Z, Yang X, Culvenor D, Jupp D, Newnham, G and Lovell J (2012) DWEL:A dual-wavelength echidna LiDAR for ground-based forest scanning. In: IEEE International Geoscience and Remote Sensing Symposium. Munich.

Douglas, E. S., Martel, J., Howe, G., Hewawasam, K., Marshall, R. A., Schaaf, C. L., Cook, T. A., Newnham, G. J., Strahler, A., & Chakrabarti, S. (2015). Finding leaves in the forest:The Dual-Wavelength Echidna Lidar. IEEE Geoscience and Remote Sensing Letters, 12(4), 776–780. http://dx.doi.org/10.1109/LGRS.2014.2361812.

Dubayah, R., & Drake, J. B. (2000). LiDAR remote sensing for forestry applications. Journal of Forestry, 98, 44–46.Eitel, J., Vierling, L., & Magney, T. (2013). A lightweight, low cost autonomously operating terrestrial laser scanner for quantifying and monitoring ecosystem structural dynamics.

Agricultural and Forest Meteorology, 180, 86–96.Farr, T. G., Rosen, P. A., Caro, E., Crippen, R., Duren, R., Hensley, S., Kobrick, M., Paller, M., Rodriguez, E., Roth, L., Seal, D., Shaffer, S., Shimada, J., Umland, J., Werner, M.,

Oskin, M., Burbank, D., & Alsdorf, D. (2007). The shuttle radar topography mission. Review of Geophysics, 45. RG2004.Fischler, M., & Bolles, R. C. (1981). Random Sample Consensus: A paradigm for model fitting with applications to image analysis and automated cartography. Communications of

the ACM, 24, 381–395.Gamon, A., Penuelas, J., & Field, C. B. (1992). A narrow-waveband spectral index that tracks diurnal changes in photosynthetic efficiency. Remote Sensing of Environment, 41,

35–44.

Vegetation Structure (LiDAR) 113

Gastellu-Etchegorry, J.-P., Yin, T., Lauret, N., Cajgfinger, T., Gregoire, T., Grau, E., Feret, J.-B., Lopes, M., Guilleux, J., Dedieu, G., Malenovský, Z., Cook, B., Morton, D., Rubio, J.,Durrieu, S., Cazanave, G., Martin, E., & Ristorcelli, T. (2015). Discrete anisotropic radiative transfer (dart 5) for modeling airborne and satellite spectroradiometer and LiDARacquisitions of natural and urban landscapes. Remote Sensing, 7, 1667–1701.

Gaulton, R., Danson, F. M., Ramirez, F. A., & Gunawan, O. (2013). The potential of dual-wavelength laser scanning for estimating vegetation moisture content. Remote Sensing ofEnvironment, 132, 32–39.

Gomez-Dans, J. L., Lewis, P. E., & Disney, M. (2016). Efficient emulation of radiative transfer codes using Gaussian processes and application to land surface parameter inferences.Remote Sensing, 8, 119.

Goodenough, A. and Brown, S. (2012). DIRSIG 5: Core design and implementation. In: Proceedings of the SPIE 8390, Algorithms and Technologies for Multispectral, Hyperspectral,and Ultraspectral Imagery XVIII. doi:10.1117/12.919321.

Gorte, B., & Winterhalder, D. (2004). Reconstruction of laser-scanned trees using filter operations in the 3D raster domain. International Archives Photogrammetry, Remote Sensingand Spatial Information Science, XXXVI(Part 8/W2), 39–44.

Govaerts, Y. M., & Verstaete, M. M. (1998). Raytran: A Monte Carlo ray-tracing model to compute light scattering in three-dimensional heterogeneous media. IEEE Transactions onGeoscience and Remote Sensing, 38, 493–505.

Hackenberg, J., Spiecker, H., Calder, K., Disney, M., & Raumonen, P. (2015). SimpleTreedAn efficient open source tool to build tree models from TLS clouds. Forests, 6,4245–4294.

Hakala, T., Suomalainen, J., Kaasalainen, S., & Chen, Y. (2012). Full waveform hyperspectral LiDAR for terrestrial laser scanning. Optics Express, 20, 7119–7127.Harding, D. J., & Carabajal, C. C. (2005). ICESat waveform measurements of within-footprint topographic relief and vegetation vertical structure. Geophysical Research Letters, 32,

L21s10.Harding, D. J., Lefsky, M. A., Parker, G. G., & Blair, J. B. (2001). Laser altimeter canopy height profiles: Methods and validation for closed-canopy, broadleaf forests. Remote

Sensing of Environment, 76, 283–297.Healy, S. P., Patterson, P. L., Saatchi, S., Lefsky, M. A., Lister, A. J., & Freeman, E. A. (2012). A sample design for globally consistent biomass estimation using lidar data from the

Geoscience Laser Altimeter System (GLAS). Carbon Balance and Management, 7, 10.Herold, M., & Johns, T. (2007). Linking requirements with capabilities for deforestation monitoring in the context of the UNFCCC-REDD process. Environmental Research Letters, 2,

045025. http://dx.doi.org/10.1088/1748-9326/2/4/045025.Hese, S., Lucht, W., Schmullius, C., Barnsley, M., Dubayah, R., Knorr, D., Neumann, K., Riedel, T., & Schroeter, K. (2005). Global biomass mapping for an improved understanding

of the CO2 balance - the Earth observation mission Carbon-3D. Remote Sensing of Environment, 94, 94–105.Hilker, T., Coops, N. C., Wulder, M. A., Black, T. A., & Guy, R. D. (2008). The use of remote sensing in light use efficiency based models of gross primary production: A review of

current status and future requirements. The Science of the Total Environment, 404, 411–423. http://dx.doi.org/10.1016/j.scitotenv.2007.11.007.Hilker, T., Hall, F. G., Tucker, C. J., Coops, N. C., Black, T. A., Nichol, C. J., Sellers, P. J., Barr, A., Hollinger, D. Y., & Munger, J. W. (2012). Data assimilation of photosynthetic

light-use efficiency using multi-angular satellite data: II Model implementation and validation. Remote Sensing of Environment, 121, 287–300. http://dx.doi.org/10.1016/j.rse.2012.02.008.

Hilker, T., Frazer, G. W., Coops, N. C., Wulder, M. A., Newnham, G. J., Stewart, J. D., Van Leeuwen, M., & Culvenor, D. S. (2013). Prediction of wood fiber attributes from LiDAR-derived forest canopy indicators. Forest Science, 59, 231–242.

Holmgren, J., & Persson, A. (2004). Identifying species of individual trees using airborne laser scanner. Remote Sensing of Environment, 90, 415–423.Hopkinson, C., Chasmer, L., Young-Pow, C., & Treitz, P. (2004). Assessing forest metrics with a ground-based scanning LiDAR. Canadian Journal of Forest Research, 34, 573–583.Hough, P. V. C. (1962). Method and Means for Recognizing Complex Patterns. US Patent 3,069,654.Hu, T., Su, Y., Xue, B., Liu, J., Zhao, X., Fang, J., & Guo, Q. (2016). Mapping global forest aboveground biomass with spaceborne lidar optical imagery, and forest inventory data.

Remote Sensing, 8, 565.Ibrom, A., Jarvis, P. G., Clement, R., Morgenstern, K., Oltchev, A., Medlyn, B. E., Wang, Y. P., Wingate, L., Moncrieff, J. B., & Gravenhorst, G. A. (2006). Comparative analysis of

simulated and observed photosynthetic CO2 uptake in two coniferous forest canopies. Tree Physiology, 26, 845–864.Isenburg, M., Liu, Y., Shewchuk, J., Snoeyink, J., & Thirion, T. (2006). Generating raster DEM from mass points via TIN streaming. In M. Raubal, H. J. Miller, A. U. Frank, &

M. F. Goodchild (Eds.), Lecture notes in computer science: 4197. Geographic information science. GIScience 2006. Berlin and Heidelberg: Springer.Jones, T. G., Coops, N. C., & Sharma, T. (2010). Assessing the utility of airborne hyperspectral and LiDAR data for species distribution mapping in the coastal Pacific Northwest,

Canada. Remote Sensing of Environment, 114, 2841–2852.Jupp, D. L. B., Culvenor, D. S., Lovell, J. L., Newnham, G. J., Strahler, A. H., & Woodcock, C. E. (2009). Estimating forest LAI profiles and structural parameters using a ground-

based laser called “Echidna”. Tree Physiology, 29, 171–181. http://dx.doi.org/10.1093/treephys/tpn022.Kampe, T., Johnson, B., Kuester, M., & Keller, M. (2010). NEON: The first continental-scale ecological observatory with airborne remote sensing of vegetation canopy biochemistry

and structure. Journal of Applied Remote Sensing, 4, 043510.Kangas, A., & Maltamo, M. (2006). Forest inventory: Methodology and applications. Dordrecht: Springer, 362 pp.Kelbe, D., Van Aardt, J., Romanczyk, P., Cawse-Nicholson, K., & Van Leeuwen, M. (2015). Single-scan stem reconstruction using sparse terrestrial laser scanner data. IEEE Journal

of Selected Topics in Applied Earth Observation and Remote Sensing, 8, 3414–3427. http://dx.doi.org/10.1109/JSTARS.2015.2416001.Kelbe, D., Van Aardt, J., Romanczyk, P., Van Leeuwen, M., & Cawse-Nicholson, K. (2016). Marker-free registration of forest terrestrial laser scanner data pairs with embedded

confidence metrics. IEEE Transactions on Geoscience and Remote Sensing, 99, 1–17. http://dx.doi.org/10.1109/TGRS.2016.2539219.Kelbe, D., Van Aardt, J., Romanczyk, P., Van Leeuwen, M., & Cawse-Nicholson, K. (2017). Multiview marker-free registration of forest terrestrial laser scanner data with embedded

confidence metrics. IEEE Transactions on Geoscience and Remote Sensing, 55, 729–741.Khosravipour, A., Skidmore, A. K., Wang, T., Isenburg, M., & Khoshelham, K. (2015). Effect of slope on treetop detection using a LiDAR canopy height model. ISPRS Journal of

Photogrammetry and Remote Sensing, 104, 44–52.Kim, S., McGaughey, R. J., Andersen, H.-E., & Schreuder, G. (2009). Tree species differentiation using intensity data derived from leaf-on and leaf-off airborne laser scanner data.

Remote Sensing of Environment, 113, 1575–1586.Koetz, B., Morsdorf, F., Sun, G., Ranson, K. J., Itten, K., & Allgöwer, B. (2006). Inversion of a LiDAR waveform model for forest biophysical parameter estimation. IEEE Geoscience

and Remote Sensing Letters, 3, 49–53.Koetz, B., Sun, G., Morsdorf, F., Ranson, K. J., Kneubühler, M., Itten, K., & Allgöwer, B. (2007). Fusion of imaging spectrometer and LiDAR data over combined radiative transfer

models for forest canopy characterization. Remote Sensing of Environment, 106, 449–459.Korpela, I., Orka, H. O., Maltamo, M., Tokola, T., & Hyyppa, J. (2010). Tree species classification using airborne LiDARdEffects of stand and tree parameters, downsizing of training

set, intensity normalization, and sensor type. Silva Fennica, 44, 319–339.Leblanc, S. G., Chen, J. M., Fernandez, R., Deering, D. W., & Conley, A. (2005). Methodology comparison for canopy structure parameters extraction from digital hemispherical

photography in boreal forests. Agricultural and Forest Meteorology, 129, 187–207.Lefsky, M. A. (2010). A global forest canopy height map from the Moderate Resolution Imaging Spectroradiometer and the Geoscience Laser Altimeter System. Geophysical

Research Letters, 37. L15401.Lewis, P. (1999). Three-dimensional plant modelling for remote sensing simulations studies using the botanical plant modelling system. Agronomie, 19, 185–210.Liang, X., Litkey, P., Hyyppa, J., Kaartinen, H., Vastaranta, M., & Holopainen, M. (2012). Automatic stem mapping using single-scan terrestrial laser scanning. IEEE Transactions on

Geoscience and Remote Sensing, 50, 661–670.

114 Vegetation Structure (LiDAR)

Liang, X., Kankare, V., Hyyppa, J., Wang, Y., Kukko, A., Haggren, H., Kaartinen, H., Jaakkola, A., Guan, F., Holopainen, M., & Vastaranta, M. (2016). Terrestrial laser scanning inforest inventories. ISPRS Journal of Photogrammetry and Remote Sensing, 115, 63–77.

Lichti, D. D., Gordon, S. J., & Stewart, M. P. (2002). Ground-based laser scanners: Operation, systems and applications. Geomatica, 56, 22–33.Lim, K., Treitz, P., Wulder, M., St-Onge, B., & Flood, M. (2003). LiDAR remote sensing of forest structure. Progress in Physical Geography, 27, 88–106.Lin, Y., Hyyppa, J., & Jaakkola, A. (2011). Mini-UAV-borne LiDAR for fine-scale mapping. IEEE Transactions on Geoscience and Remote Sensing Letters, 8, 426–430.Liu, X. (2008). Airborne LiDAR for DEM generation: Some critical issues. Progress in Physical Geography, 32, 31–49.Livny, Y., Yan, F., Olson, M., Chen, B., Zhang, H., & El-Sana, J. (2010). Automatic reconstruction of tree skeletal structures from point clouds. ACM Transactions on Graphics, 29,

151. http://dx.doi.org/10.1145/1882261.1866177.Maltamo, M., Naesset, E., & Vauhkonen, J. (2014). Forestry applications of airborne laser scanning: Concepts and case studies. Dordrecht: Springer, 464 pp.Medlyn, B. E., Dreyer, E., Ellsworth, D., Forstreuter, M., Harley, P. C., Kirschbaum, M. U. F., & Roux, X. L. E. (2002). Temperature response of parameters of a biochemically based

model of photosynthesis. II. A review of experimental data. Plant, Cell and Environment, 25, 1167–1179.Mitchard, E. T. A., Saatchi, S. S., White, L. J. T., Abernethy, K. A., Jeffery, K. J., Lewis, S. L., Collins, M., Lefsky, M. A., Leal, M. E., Woodhouse, I. H., & Meir, P. (2012). Mapping

tropical forest biomass with radar and spaceborne LiDAR in Lopé National Park, Gabon: Overcoming problems of high biomass and persistent cloud. Biogeosciences, 9,179–191.

Mohammed, A. (2015). Scaling up of peatland methane emission hotspots from small to large scale. Doctoral thesis. University of Edinburgh, School of GeoSciences, 228 pp.Morsdorf, F., Nichol, C., Malthus, T., & Woodhouse, I. H. (2009). Assessing forest structural and physiological information content of multi-spectral LiDAR waveforms by radiative

transfer modelling. Remote Sensing of Environment, 113, 2152–2163.Morton, D. C., Rubio, J., Cook, B. D., Gastellu-Etchegorry, J.-P., Longo, M., Choi, H., Hunter, M., & Keller, M. (2016). Amazon forest structure generates diurnal and seasonal

variability in light utilization. Biogeosciences, 13, 2195–2206.Moussavi, M. S., Abdalati, W., Scambos, T., & Neuenschwander, A. (2014). Applicability of an automated surface detection approach to micro-pulse photon-counting LiDAR

altimetry data: Implications for canopy height retrieval from future ICESat-2 data. International Journal of Remote Sensing, 35, 5263–5279.Naesset, E. (2014). Area-based inventory in NorwaydFrom innovation to an operational reality. In: Maltamo, M. (ed.) Forestry applications of airborne laser scanning: Concepts and

case studies. Managing forest ecosystems, Springer, vol. 27. doi:10.1007/978-94-017-8663-8__11.Nelson, R., Parker, G., & Hom, M. (2003). A portable airborne laser system for forest inventory. Photogrammetric Engineering and Remote Sensing, 69, 267–273.Newnham, G. J., Armston, J. D., Calders, K., Disney, M. I., Lovell, J. L., Schaaf, C. B., Strahler, A. H., & Danson, F. M. (2015). Terrestrial laser scanning for plot scale forest

measurement. Current Forestry Reports, 1, 239–251. http://dx.doi.org/10.1007/s40725-015-0025-5.Ørka, H. O., Næsset, E., & Bollandsas, O. M. (2009). Classifying species of individual trees by intensity and structure features derived from airborne laser scanner data. Remote

Sensing of Environment, 113, 1163–1174.Parker, G. G., Harding, D. J., & Berger, M. L. (2004). A portable LiDAR system for rapid determination of forest canopy structure. Journal of Applied Ecology, 41, 755–767.Passioura, J. B. (1979). Accountability, philosophy and plant physiology. Search, 10, 347–350.Pradal, C., Boudon, F., Nouguier, C., Chopard, J., & Godin, C. (2009). PlantGL: A python-based geometric library for 3D plant modelling at different scales. Graphical Models,

71, 1–21. http://dx.doi.org/10.1016/j.gmod.2008.10.001.Prusinkiewicz P and Lindenmayer A (2004) The algorithmic beauty of plants, p. 240. Retrieved from (24 May 2017). http://algorithmicbotany.org/Pueschel, P., Newnham, G., Rock, G., Udelhoven, T., Werner, W., & Hill, J. (2013). The influence of scan mode and circle fitting on tree stem detection, stem diameter and volume

extraction from terrestrial laser scans. ISPRS Journal of Photogrammetry and Remote Sensing, 77, 44–56.Qi, W., & Dubayah, R. O. (2016). Combining Tandem-X InSAR and simulated GEDI LiDAR observations for forest structure mapping. Remote Sensing of Environment, 187,

253–266.Raumonen, P., Kaasalainen, M., Åkerblom, M., Kaasalainen, S., Kaartinen, H., Vastaranta, M., Holopainen, M., Disney, M., & Lewis, P. (2013). Fast automatic precision tree models

from terrestrial laser scanner data. Remote Sensing, 5, 491–520. http://dx.doi.org/10.3390/rs5020491.Reutebuch, S. E., McGaughey, R. J., Andersen, H.-E., & Carson, W. W. (2014). Accuracy of a high-resolution LiDAR terrain model under a conifer forest canopy. Canadian Journal

of Remote Sensing, 29, 527–535.Runions, A., Lane, B., & Prusinkiewicz, P. (2007). Modeling trees with a space colonization algorithm. In Eurographics Workshop on Natural Phenomena (pp. 63–70). Geneva:

Eurographics Association.Saatchi, S. S., Harris, N. L., Brown, S., Lefsky, M., Mitchard, E. T. A., Salas, W., Zutta, B. R., Buermann, W., Lewis, S. L., Hagen, S., Petrova, S., White, L., Silman, M., & Morel, A.

(2011). Benchmark map of forest carbon stocks in tropical regions across three continents. Proceedings of the National Academy of Sciences of the United States of America,108, 9899–9904.

Simard, M., Pinto, N., Fisher, J. B., & Baccini, A. (2011). Mapping forest canopy height globally with spaceborne lidar. Journal of Geophysical Research, 116. G04021.Strahler, A. H., Jupp, D. L. B., Woodcock, C. E., Schaaf, C. B., Yao, T., Zhao, F., Yang, X., Lovell, J., Culvenor, D., Newnham, G., Ni-Meister, W., & Boykin-Morris, W. (2008).

Retrieval of forest structural parameters using a ground-based LiDAR instrument (Echidna ®). Canadian Journal of Remote Sensing, 34, S426–S440. http://dx.doi.org/10.5589/m08-046.

Su, J., & Bork, E. (2006). Influence of vegetation, slope, and LiDAR sampling angle on DEM accuracy. Photogrammetric Engineering and Remote Sensing, 72, 1265–1274.Sun, G., Ranson, K. J., Kimes, D. S., Blair, J. B., & Kovacs, K. (2008). Forest vertical structure from GLAS: An evaluation using LVIS and SRTM data. Remote Sensing of

Environment, 112, 107–117.Swatantran, A., Tang, H., Barrett, T., DeCola, P., & Dubayah, R. (2016). Rapid, high-resolution forest structure and terrain mapping over large areas using single photon LiDAR.

Nature Scientific Reports, 6, 28277. http://dx.doi.org/10.1038/srep28277.ter Steege, H., Pitman, N. C., Sabatier, D., Baraloto, C., Salomão, R. P., Guevara, J. E., Phillips, O. L., Castilho, C. V., Magnusson, W. E., Molino, J. F., Monteagudo, A., Núñez

Vargas, P., Montero, J. C., Feldpausch, T. R., Coronado, E. N., Killeen, T. J., Mostacedo, B., Vasquez, R., Assis, R. L., Terborgh, J., Wittmann, F., Andrade, A., Laurance, W. F.,Laurance, S. G., Marimon, B. S., Marimon, B. H., Jr., Guimarães Vieira, I. C., Amaral, I. L., Brienen, R., Castellanos, H., Cárdenas López, D., Duivenvoorden, J. F.,Mogollón, H. F., Matos, F. D., Dávila, N., García-Villacorta, R., Stevenson Diaz, P. R., Costa, F., Emilio, T., Levis, C., Schietti, J., Souza, P., Alonso, A., Dallmeier, F.,Montoya, A. J., Fernandez Piedade, M. T., Araujo-Murakami, A., Arroyo, L., Gribel, R., Fine, P. V., Peres, C. A., Toledo, M., Aymard, C. G. A., Baker, T. R., Cerón, C., Engel, J.,Henkel, T. W., Maas, P., Petronelli, P., Stropp, J., Zartman, C. E., Daly, D., Neill, D., Silveira, M., Paredes, M. R., Chave, J., Lima Filho Dde, A., Jørgensen, P. M., Fuentes, A.,Schöngart, J., Cornejo Valverde, F., Di Fiore, A., Jimenez, E. M., Peñuela Mora, M. C., Phillips, J. F., Rivas, G., van Andel, T. R., von Hildebrand, P., Hoffman, B., Zent, E. L.,Malhi, Y., Prieto, A., Rudas, A., Ruschell, A. R., Silva, N., Vos, V., Zent, S., Oliveira, A. A., Schutz, A. C., Gonzales, T., Trindade Nascimento, M., Ramirez-Angulo, H., Sierra, R.,Tirado, M., Umaña Medina, M. N., van der Heijden, G., Vela, C. I., Vilanova Torre, E., Vriesendorp, C., Wang, O., Young, K. R., Baider, C., Balslev, H., Ferreira, C., Mesones, I.,Torres-Lezama, A., Urrego Giraldo, L. E., Zagt, R., Alexiades, M. N., Hernandez, L., Huamantupa-Chuquimaco, I., Milliken, W., Palacios Cuenca, W., Pauletto, D., ValderramaSandoval, E., Valenzuela Gamarra, L., Dexter, K. G., Feeley, K., Lopez-Gonzalez, G., & Silman, M. R. (2013). Hyperdominance in the amazonian tree flora. Science, 342. http://dx.doi.org/10.1126/science.1243092.

Tsui, O. W., Coops, N. C., Wulder, M. A., & Marshall, P. L. (2013). Integrating airborne LiDAR and space-borne radar via multivariate kriging to estimate above-ground biomass.Remote Sensing of Environment, 139, 340–352.

Van der Zande, D., Stuckens, J., Verstraeten, W. W., Mereu, S., Muys, B., & Coppin, P. (2011). 3D modeling of light interception in heterogeneous forest canopies using ground-based LiDAR data. International Journal of Applied Earth Observation and Geoinformation, 13, 792–800. http://dx.doi.org/10.1016/j.jag.2011.05.005.

Van Leeuwen, M., & Nieuwenhuis, M. (2010). Retrieval of forest structural parameters using LiDAR remote sensing. European Journal of Forest Research, 129, 749–770.

Vegetation Structure (LiDAR) 115

Van Leeuwen, M., Hilker, T., Coops, N. C., Frazer, G., Wulder, M. A., Newnham, G. J., & Culvenor, D. S. (2011). Assessment of standing wood and fiber quality using ground andairborne laser scanning. Forest Ecology and Management, 261, 1467–1478.

Van Leeuwen, M., Coops, N. C., Hilker, T., Wulder, M. A., Newnham, G. J., & Culvenor, D. S. (2013). Automated reconstruction of tree and canopy structure for modeling theinternal canopy radiation regime. Remote Sensing of Environment, 136, 286–300.

Van Leeuwen, M., Coops, N. C., & Black, T. A. (2015). Using stochastic ray tracing to simulate a dense time series of gross primary productivity. Remote Sensing, 7, 17272–17290.Verroust, A., & Lazarus, F. (2000). Extracting skeletal curves from 3D scattered data. The Visual Computer, 16, 15–25.Wang, Y.-P., & Jarvis, P. G. (1990). Description and validation of an array modeldMAESTRO. Agricultural and Forest Meteorology, 51, 257–280.Wang, Y., Hyyppa, J., Liang, X., Kaartinen, H., Yu, X., Lindberg, E., Holmgren, J., Qin, Y., Mallet, C., Ferraz, A., Torabzadeh, H., Morsdorf, F., Zhu, L., Liu, J., & Alho, P. (2016).

Internal benchmarking of the individual tree detection methods for modeling 3-D canopy structure for silviculture and forest ecology using airborne laser scanning. IEEETransactions on Geoscience and Remote Sensing, 54, 5011–5027.

Weber, J., & Penn, J. (1995). Creation and rendering of realistic trees. In SIGGRAPH 95 (pp. 119–128). New York: ACM.Welles, J. M., & Norman, J. M. (1991). Instrument for indirect measurement of canopy architecture. Agronomy Journal, 83, 818–825.Widlowski, J.-L., Pinty, B., Lopatka, M., Atzberger, C., Buzica, D., Chelle, M., Disney, M., Gastellu-Etchegorry, J.-P., Gerboles, M., Gobron, N., Grau, E., Huang, H., Kalle, A.,

Kobayashi, H., Lewis, P. E., Schlerf, M., Stuckens, J., & Xie, D. (2013). The fourth radiation transfer model intercomparison (RAMI-IV): Proficiency testing of canopy reflectancemodels with ISO-13528. Journal of Geophysical Research. Atmospheres, 118, 6869–6890.

Widlowski, J. L., Mio, C., Disney, M., Adams, J., Andredakis, I., Atzberger, C., Brennan, J., Busetto, L., Chelle, M., Ceccherini, G., Colombo, R., Cote, J.-F., Eenmae, A., Essery, R.,Gastellu-Etchegorry, J.-P., Gobron, N., Grau, E., Haverd, V., Homolova, L., Huang, H., Hunt, L., Kobayashi, H., Koetz, B., Kuusk, A., Kuusk, J., Lang, Mait, Lewis, P. E.,Lovell, J. L., Malenovsky, Z., Meroni, M., Morsdorf, F., Mottus, M., Ni-Meister, W., Pinty, B., Rautianen, M., Schlerf, M., Somers, B., Stuckens, J., Verstraete, M. M., Yang, W.,Zhao, F., & Zenone, T. (2015). The fourth phase of the radiative transfer model intercomparison (RAMI) exercise: Actual canopy scenarios and conformity testing. RemoteSensing of Environment, 169, 418–437.

Woodhouse, I., Nichol, C., Sinclair, P., Jack, J., Morsdorf, F., Malthus, T. J., & Patenaude, G. (2011). A multispectral canopy LiDAR demonstrator project. IEEE Geoscience andRemote Sensing Letters, 8, 839–843.

Wulder, M. A., White, J. C., Bater, C. W., Coops, N. C., Hopkinson, C., & Chen, G. (2012a). LiDAR plotsdA new large-area data collection option: Context, concepts, and casestudy. Canadian Journal of Remote Sensing, 38, 600–618.

Wulder, M. A., White, J. C., Nelson, R. F., Naesset, E., Orka, H. O., Coops, N. C., Hilker, T., Bater, C. W., & Gobakken, T. (2012b). LiDAR sampling for large-scale forestcharacterization: A review. Remote Sensing of Environment, 121, 196–209.

Yao, W., Kelbe, D., Van Leeuwen, M., Romanczyk, P., & Van Aardt, J. (2016). Towards an improved LAI collection protocol via simulated and field-based PAR sensing. Sensors, 16,1092. http://dx.doi.org/10.3390/s16071092.

Zhao, F., Strahler, A. H., Schaaf, C. L., Yao, T., Yang, X., Wang, Z., Schull, M. A., Roman, M. O., Woodcock, C. E., Olofsson, P., Ni-Meister, W., Jupp, D. L. B., Lovell, J. L.,Culvenor, D. S., & Newnham, G. J. (2012). Measuring gap fraction, element clumping index and LAI in Sierra Forest stands using a full-waveform ground-based LiDAR. RemoteSensing of Environment, 125, 73–79.

Zheng, G., & Moskal, L. M. (2012). Leaf orientation retrieval from terrestrial laser scanning (TLS) data. IEEE Transactions on Geoscience and Remote Sensing, 50, 3970–3979.

Relevant Websites

http://ahn.nl/http://algorithmicbotany.org/http://www.bas.ac.uk/http://www.cloudcompare.org/http://www.eurosdr.net/http://www.rapidlasso.com/http://www.riegl.com/http://www.spdlib.org/

116 Vegetation Structure (LiDAR)