PE&RS May 2016 - page 353

instrument’s full operational time frame, between 2003 and
2009. During operation, GLAS emitted a Gaussian shaped laser
pulse in the 532 nm, and 1064 nm wavebands, approximately
1 m in temporal width (approximately 5 to 6 ns duration) at
the Full Width Half Maximum (FWHM; width of Gaussian at
half amplitude) (Rosette et al., 2008). GLAS utilized one of three
lasers at any one time for 2 to 3, one month long periods per
year, known as laser campaigns. Laser 1 operated continuously
for a total of 38 days before failure due to energy loss, where
energy loss was most notable from day 10 and beyond (Ab-
shire et al., 2005). The mission operating parameters for lasers
2 and 3 were adjusted to extend their lifetimes and mitigate
energy loss; this was more successful for laser 3, the energy of
which decayed considerably more slowly than laser 2.
Pulses were emitted between ±86° latitude at a rate of 40
Hz, resulting in a footprint separation of 172 m on the ground.
The elliptical footprints exhibited mean diameters (semi-
major and semi-minor axes) of 52 m × 95 m for campaigns
L1-L2C, and 47 m × 61 m for other campaigns, equivalent
to a mean circular diameter of 64 m (Abshire et al., 2005).
Between 82 m and 150 m of the returned waveform was
recorded at 0.15 m (equivalent to a 1 ns sampling rate, ac-
counting for two-way distance) intervals near the ground sur-
face (Brenner et al., 2003; NSIDC, 2012). Peaks in the returned
waveforms are a function of intercepted surfaces, and are
decomposed by up to six Gaussians, fitted according to Duong
et al. (2008), the sum of which defines the “model alternate
fit” return pulse, from which canopy height is often derived
(Rosette et al., 2008; Los et al., 2012).
GLA14 land data parameters are employed throughout
this study for filtering and canopy height retrieval purposes;
these parameters are summarized in Table 1. Data without
geographical information, footprint geometry information,
or that exhibit cloud contamination (i.e., GLA14 product
variable i_FRir_qaFlag ≠ 15) are treated as spurious, and have
been removed.
Airborne Laser Scanning Data
The sensitivity of GLAS canopy height estimates relative to ALS
height metrics were tested over three forest sites in Australia
(Figure 2). The sites exhibit diverse canopy characteristics (i.e.,
variable height, gap fraction, and density) and some topograph-
ic variability, ranging from flat terrain to areas of >20° slope.
Table 1. Summary of GLA14 Land Data Products Employed in this Study.
GLA14 Code Description Use
i_rec_ndx Unique waveform batch number Identify unique footprints
i_UTCTime Time code in reference to 01/01/2000 00:00:00 UTC Find date footprint was acquired
i_lat Latitude of waveform centroid Location of GLAS footprint
i_lon Longitude of waveform centroid Location of GLAS footprint
i_SigBegOff Signal begin range increment Signal start of waveform
i_SigEndOff Signal end range increment Signal end of waveform
i_gpCntRngOff Centroid range increment for up to six Gaussian peaks Locate elevation of waveform features
i_FRir_qaFlag Cloud detection flag Filter cloud disrupted data
i_TxNrg Transmitted laser energy Footprints transmission energy
Figure 2. Location of study sites over across the four main forested landscapes of Australia, with ALS extents and enclosed GLAS footprints.
PHOTOGRAMMETRIC ENGINEERING & REMOTE SENSING May 2016 353