St. Petersburg Coastal and Marine Science Center
Effective coastal management plans require the accurate and detailed representation of the horizontal and vertical structure of plant communities, also called canopy structure, in addition to traditional thematic maps of vegetation type. As a full-waveform digitizing lidar, the Experimental Advanced Airborne Research Lidar (EAARL) records backscattered laser energy above a very low photon threshold in 1 ns time intervals, creating a pulse-by-pulse reflection record that is highly sensitive to even minor changes in vegetation structure. The small-footprint EAARL waveforms can be used to generate accurate estimates of vertical canopy structure in coastal vegetation communities.
At the nominal flying altitude of 300 m, a single EAARL laser pulse illuminates a small horizontal sampling area (20-cm-diameter footprint) every 2 – 4 m2. To describe the vertical structure of a vegetation canopy, several individual small-footprint laser pulses are combined to make a composite “large-footprint” waveform that defines a larger horizontal area (fig. 1). The size of this composite footprint is a variable and can be determined in post-flight processing software, unlike the large-footprint lidar systems whose footprint size is determined by optical laser-beam-divergence hardware prior to data acquisition. The size of the composite footprint depends on the density of the lidar data, the nature of the forested terrain (dense forested canopies are difficult to penetrate and may require a larger composite footprint to describe the complete vertical structure), and the desired horizontal resolution of the end product. The vertical-sampling resolution (or vertical bin) of the composite waveform is also defined in post-flight processing software and is typically set to 50 cm, which is equivalent to a digitized vertical-sampling rate of 3.3 nanoseconds in air. Within each vertical bin, the amplitude backscatter for all the individual waveforms constituting the composite waveform are averaged as follows:
where βind is the backscatter count for each individual waveform i, n is the number of waveforms in the vertical bin, and βcomp is the resulting backscatter count for the composite waveform. The resulting composite waveform (fig. 2) represents the vertical structure within a circular cone, similar to a large-footprint lidar system.
Bare-Earth elevations (BEEs) are determined from the range to the last peak in the individual small-footprint waveforms. A ‘trailing-edge’ algorithm is used to detect the range to the last return. This algorithm searches for the location prior to the last return changing its direction along its trailing edge (fig. 2). The EAARL waveform has a tightly focused beam, which minimizes the spreading of the ground return, especially on steep slopes. BEE data can be used to model the hydrology of areas and hence predict vegetation assemblages likely to occur across specific topographies. In coastal communities, knowing the elevation of the land surface above sea level gives a fair idea of relative depths to water tables, which can predict water availability to vegetation on the sandy substrates.
The composite-footprint waveforms are used to derive three metrics: canopy height (CH), canopy reflection ratio (CRR), and height of median energy (HOME). CH is the distance from the first return to the ground (fig. 2). The range to the first return is detected at the zero-crossing of the second derivative. The ground elevation is determined as the average of the BEEs derived from the individual waveforms within the composite footprint. CH is a long-established indicator of site quality in forestry applications.
CRR is the sum of the portion of the waveform return reflected off the canopy (CR) divided by the sum of the portion of the waveform return reflected off the canopy and the ground (CR + GR) (fig. 4). CRR is a relative measure of canopy cover. Independent knowledge of the average reflectance of the canopy and ground surfaces within the footprint is necessary to convert CRR to an absolute measure of canopy cover. CRR can likely provide a remotely sensed attribute through which community type could be differentiated.
HOME is the median height of the entire signal from the waveform, including energy returned from both canopy and ground surfaces (fig. 2). HOME is predicted to be sensitive to changes in both the vertical arrangements of the canopy and the degree of canopy openness (including tree density). HOME has been found to be a good predictor of biomass and structural attributes in tropical forests. Applied in a similar manner to coastal vegetation communities, HOME could be utilized in looking at structural changes across environmental gradients and perhaps in assessments of damage to stands from storms and parasite infestations.