**3.1 Characteristics**

The use of airborne laser scanning data in forest applications has attracted increasing interest over the last decade. Nowadays, lidar is perceived to provide a cost-effective and precise assessment of the vertical and horizontal structure of woodland areas and, therefore, a valid alternative or complementary approach to current field methods for inventory (Wulder *et al.*, 2008). Vertical location of points are often reported to 1ns (~15cm) precision whilst horizontal gelocation accuracy may be expected in the region of 20-30cm.

In this section, an overview is given of small footprint airborne laser scanning applications for both stand-level assessment and the estimation of vegetation parameters at an individual tree level. For a comprehensive description of the use of airborne lidar for forestry purposes including forest community structure, growth assessment, tree stability and timber quality, please refer to Suárez, 2010.

Airborne lidar systems provide relatively dense sampling coverage with footprint size in the region of tens of centimetres. However, they are commonly reported to underestimate canopy height as a result of point distribution and density. This is due to some degree of canopy penetration of the signal that varies according to species (Næsset, 2004). In general, this underestimation is less pronounced for cone-shaped trees like spruce or Douglas fir (Pseudotsuga menziesi) than for spherical-shaped trees like many broadleaves or even Scots pine (Pinus sylvestris L.). Conifers normally create more compact shapes with less energy penetration through the canopy than broadleaves. So, energy returns tend to be produced

sampled by the lidar footprints. The continuous spatial coverage of optical or radar data

Similar data fusion techniques have been used to determine biomass distribution by several authors and encompass a wide breadth of vegetation types and have been applied from regional to continental scales, including a focus on Africa's mangrove forest (Fatoyinbo & Simard, in press 2011), Siberia (Nelson *et al.*, 2009), Quebec (Boudreau *et al.*, 2008), and for mapping throughout Africa (Baccini *et al.*, 2008). Work is currently underway by the latter research group at Woods Hole Research Center, led by Josef Kellndorfer, to estimate tropical forest biomass globally (WHRC, 2011). Additionally, research utilising GLAS is in progress as part of NASA's Carbon Monitoring System initiative (NASA, 2010) to determine biomass distribution within the US (as well as to produce higher resolution biomass maps at a

Global vegetation height products derived from GLAS and optical data (Lefsky, 2010; Los *et al.* 2011) or combining GLAS and radar data (Simard, 2011) open possibilities to improve our understanding of global processes (Los *et al.*, 2008) as well as allowing applications for

However, comparability of data and methods must be taken into consideration. These methods rely on the application of regression equations to extend vegetation parameter estimates across large areas. Nelson, 2010 demonstrates how the calculation of biomass is often sensitive to the equation applied and lidar sensor characteristics. These inconsistencies have implications for repeat analysis and monitoring of change due to the effect of model

The use of airborne laser scanning data in forest applications has attracted increasing interest over the last decade. Nowadays, lidar is perceived to provide a cost-effective and precise assessment of the vertical and horizontal structure of woodland areas and, therefore, a valid alternative or complementary approach to current field methods for inventory (Wulder *et al.*, 2008). Vertical location of points are often reported to 1ns (~15cm) precision whilst horizontal gelocation accuracy may be expected in the region of

In this section, an overview is given of small footprint airborne laser scanning applications for both stand-level assessment and the estimation of vegetation parameters at an individual tree level. For a comprehensive description of the use of airborne lidar for forestry purposes including forest community structure, growth assessment, tree stability and timber quality,

Airborne lidar systems provide relatively dense sampling coverage with footprint size in the region of tens of centimetres. However, they are commonly reported to underestimate canopy height as a result of point distribution and density. This is due to some degree of canopy penetration of the signal that varies according to species (Næsset, 2004). In general, this underestimation is less pronounced for cone-shaped trees like spruce or Douglas fir (Pseudotsuga menziesi) than for spherical-shaped trees like many broadleaves or even Scots pine (Pinus sylvestris L.). Conifers normally create more compact shapes with less energy penetration through the canopy than broadleaves. So, energy returns tend to be produced

selection and lidar system evolution on the outcome of biomass assessment.

permit these estimates to be extrapolated.

county level using airborne lidar data).

biomass assessment.

**3. Airborne lidar systems** 

please refer to Suárez, 2010.

**3.1 Characteristics** 

20-30cm.

from the outer layers of the canopy. However, the degree of penetration is ultimately related to a combination of factors such as the sampling density, beam divergence and scanning angle (Suárez *et al.*, 2005).

Possible scenarios may be as illustrated in Figure 4 below. With higher density of laser pulses, these difficulties are reduced but at the cost of higher operating expense and flight duration restrictions limiting spatial coverage.

Fig. 4. Possible scenarios of lidar point cloud interception of the forest canopy. Source: Suárez *et al.*, 2005.


Despite common perception, lidar does not create tomographic images and, therefore, they cannot be considered as 3-D representations in the strict sense. Only the gaps in canopy cover and transmittance through leaves will allow laser energy to be returned from the ground. As with full waveform data, the critical step for the calculation of vegetation height metrics is to distinguish between those points returned from ground and non-ground surfaces.

Since lidar energy penetration through the vegetation canopy will vary with forest structure, density and laser scanning angle in particular, the last return of an emitted lidar pulse may not necessarily be returned from the ground surface. Therefore a means of filtering points is necessary in order to differentiate those returns reaching the ground from those being intercepted at different heights within the canopy (e.g. Kraus & Pfeiffer, 1998; Zhang *et al.*, 2003). A thorough comparison of different approaches and a complete description of filters can be found in Sithole & Vosselman, 2004. This allows the classification of points into ground and vegetation classes (Figure 5). Other algorithms can refine the classification further to additionally identify features such as buildings, electricity cables, etc.

Lidar Remote Sensing for Biomass Assessment 13

estimate, in a second phase, forest stand parameters for all the test plots in the study area,

In this way, the stand-level approach provides a useful estimation of key stand parameters such as top height, canopy cover, tree density, basal area and volume. The established relationships allow forest parameters such as biomass to be directly inferred from the use of lidar metrics and for this assessment to be implemented across large forest areas where there is lidar coverage. Lidar systems provide point-wise anisotropic sampling unlike the full area coverage common in optical systems. As a result, laser data are interpolated in order to convert the same coverage to a continuous surface working-image and allow the

This method does of course have some limitations. It is heavily dependent on abundant field data collection to train empirical relationships between field and lidar data that often are not easily transferable to other study areas. Different relationships may be present for morphologically similar species (Norway spruce, Sitka spruce or larch) and additionally, stand structure is a significant determinant factor. The effects derived from the spatial distribution of gaps, their size and the spatial distribution of standing trees (whether perfectly aligned, growing in a natural stand or thinned at different intensities) on the

However, the real value of this method is demonstrated in its application in regional studies, where the combination of lidar and field measurements can optimise traditional

Variability in characteristics is a feature of natural systems. Even in systems designed to be as uniform as possible, such as planted forests in monocultures, growing differences are inevitable. Understanding the factors controlling variability is important in developing our knowledge of how ecosystems operate and behave such as the way that trees grow and accumulate biomass in response to their environment and close interaction with their neighbours (e.g. competition for space, light and nutrients). Identifying and locating trees using airborne lidar data permits

Canopy delineation algorithms have been developed by several authors and most of these detect single trees using an interpolated canopy height model (section 3.1) by the detection of local maxima for tree location and watershed or pouring algorithms (as well as derivations of these algorithms) for the delineation of single tree crowns (e.g. Popescu *et al.*, 2003). However, latterly, approaches are becoming increasingly based on raw data using

Figure 6 illustrates the delineation of individual tree crowns shown outlined by blue polygons and overlaid on a lidar canopy height model. Lighter grey shades indicate taller

Using known allometric relationships between tree height and crown width (derived from crown area) with stem diameter at breast height, volume can be calculated on an individual tree basis by approximating tree stems as a cone or by using more sophisticated taper functions or a crop form parameter which defines mean taper characteristics in a stand (Edwards & Christie, 1981; Matthews & Mackie, 2006). Individual tree volume can then be converted to biomass (by accounting for specific density) allowing the spatial distribution

known as two–stage procedure for stand inventory (Næsset, 2002).

vertical interception of laser hits are not parameterised in this approach.

surveys, particularly as part of large area forest inventories (Hollaus *et al.*, 2009).

the spatial variability of biomass distribution within a forest to be considered.

heights which are used to identify tree tops to infer height of individual trees.

clustering or blob detection methods (e.g. Morsdorf *et al.*, 2004).

distribution of forest parameters to be mapped.

**3.2.2 Individual tree based inventories** 

and variability of biomass to be mapped.

The use of airborne lidar data in forestry was originally focused on the construction of two cartographic products: Digital Terrain Models (DTMs) and Digital Surface Models (DSMs), which are used to describe the underlying terrain and top of forest surfaces respectively. These products are used to generate canopy height models (CHM) that subsequently provide accurate estimates of important forest parameters such as canopy heights, stand volume, and the vertical structure of the forest canopy. The estimation of canopy heights is performed by the subtraction of bare ground values (DTM) from the canopy layer (DSM). An accurate estimation of a CHM relies heavily on a good approximation of the ground surface underneath.

Fig. 5. Vegetation distribution and structure shown by classification of an airborne lidar point cloud. Data illustrated were provided courtesy of Forest Research, UK. Location: Scottish Highlands, UK.

More recently, increasingly sophisticated means of analysis are being applied for the estimation of important parameters at both stand (number of trees, volume, basal area, top height, percentage of canopy cover and crown layers) and individual tree level (individual tree heights, stem diameters and crown metrics). The potential areas of application span from timber production to biological diversity, carbon sequestration or general environmental protection. Section 3.2 below considers both approaches.
