Quantification of Airborne Lidar Accuracy in
Coastal Dunes (Fire Island, New York)
suty
Abstract
To establish a basis for the utilization of lidar topography
as a data source for coastal geomorphological analyses, this
study generated statistical metrics of lidar error through
the comparison of a June 2014
USGS
collection of airborne
lidar with a concurrently collected high-accuracy
GPS
topo-
graphical survey collected within the beach and dunes of a
portion of Fire Island National Seashore. The examination
of bare earth lidar error within the experiment site revealed
a complex association between accuracy and environment
within the coastal landscape. Accuracy was constrained to
better than 50 cm
RMSE
in areas with vegetated dune topog-
raphy and, overall, a 38.9 cm
RMSE
was measured. Higher
accuracies were achieved in the flat, non-vegetated beach.
A three-dimensional minimization of residuals between the
lidar and
GPS
surveys reduced the total
RMSE
to 25.2 cm,
indicating a correctable systematic offset between the sur-
face generated from the lidar and the true ground surface.
Introduction
Topographical datasets are the foundation of methodolo-
gies created to assess and quantify coastal geomorphological
change (e.g., Brown and Arbogast, 1999; Morton
et al.
, 1993;
Thom, 1991; Woolard and Colby, 2002; Brenner
et al.
, 2017).
Moreover, airborne lidar (light detection and ranging) has
evolved into a valuable source of topographical data. This is
particularly true within the trend towards higher-resolution
data that are the product of technological advancement since
the turn of the 21
st
century (Brock and Purkis, 2009). Con-
sidering the significant socioeconomic importance of coastal
geomorphology and its influence within coupled physical
(Leatherman, 1979; Nordstrom and Psuty, 1980), biological
(Roman and Nordstrom, 1988), and anthropogenic (Carapuço,
2016) systems (Fitzgerald
et al.
, 2008), the collection and ap-
plication of lidar data in the coastal realm provides opportu-
nities to better understand the evolution of coastal systems
and benefit coastal communities. However, with modern
airborne lidar (herein referred to as lidar) systems capable of
collecting a high density of data, upwards of 100,000 points
per second and covering hundreds of square kilometer in
a single day, errors in the data can be reasonably expected
(Leigh
et al.
, 2009). Understanding the accuracy of this source
of coastal topographical data is critical to its utilization in
geomorphological analyses.
Prior research has addressed general error and limitations
of lidar systems (Leigh
et al.
, 2009 and references therein), as
well as the error and utility of lidar under a variety of terrains
and land cover (Hodgson and Bresnahan, 2004; Liu, 2008;
Su and Bork, 2006), including the coastal landscape (Gesch,
2009; Krabill
et al.
, 2000; Mitasova
et al.
, 2009; Nayegandhi
et al.
, 2009; Woolard and Colby, 2002). Aspects of the data ac-
quisition, bare earth processing, geodetic transformations, and
topographical modeling procedures contribute to error in li-
dar-derived elevation datasets. The total error in the dataset is
a sum of these discrete contributions. The relative magnitude
of each of these partial contributions is, in part, a function
of variables within the physical environment, ranging from
satellite geometries to densities of ground vegetation; and,
as a result, the physical environment is the primary factor
contributing to the scale of the error. The error reported along
with most publicly available lidar is a metric that minimizes
a number of contributing factors, such as the presence of
variable slope and land cover characteristics. This approach
makes it a useful measure of error for the lidar data acquisi-
tion in ideal conditions and the lidar system in general, but it
is not a metric of error applicable to data collected in physical
environments with complex configurations.
The coastal environment is a challenging setting to mea-
sure and model topography, in part, due to the spatially vari-
able relief, slope, and vegetation density of the beaches and
dunes. These physical characteristics influence the accuracy
of the ground surface elevations derived from lidar. However,
a comparative analysis and evaluation of the lidar end prod-
ucts to a concurrently collected high accuracy control surface
has not been undertaken within the beach/dune environ-
ment. This site-specific appraisal is of significance because:
(1) coastal geomorphology is expected to constitute a signifi-
cant impact of climate change and global mean sea-level rise
(Wong
et al.
, 2014) with societally important implications
(Fitzgerald
et al.
, 2008; Wong
et al.
, 2014); (2) lidar can be a
powerful data source for the measurement of coastal geomor-
phological change (Brock and Purkis, 2009); and (3) the error
of lidar collected in the beach/dune setting needs to be evalu-
ated to validate examinations of coastal geomorphology using
lidar as a data source. Therefore, a test of lidar error within
the coastal beach/dune system was undertaken through the
comparison of high accuracy ground-based
GPS
survey data to
contemporaneously collected airborne lidar to contribute to
an understanding of the distribution and magnitude of vari-
ables influencing the error.
Background: General Limitations of Lidar Accuracy
The accuracy of a topographical lidar
DEM
is dependent upon
a number of factors resulting from a combination of issues
encountered in data acquisition and processing (Leigh
et al.
,
2009). A measure of accuracy is often provided as metadata
William J. Schmelz is with the Department of Earth and
Planetary Sciences, Rutgers University, 610 Taylor Road,
Piscataway, New Jersey 08854; and the Sandy Hook
Cooperative Research Programs, New Jersey Agricultural
Experiment Station, Rutgers University, 74 Magruder Road,
Highlands, New Jersey 07732(wjs107@eps.rutgers.edu).
Norbert P. Psuty is with the Sandy Hook Cooperative Research
Programs, New Jersey Agricultural Experiment Station, Rutgers
University, 74 Magruder Road, Highlands, New Jersey 07732
Photogrammetric Engineering & Remote Sensing
Vol. 85, No. 2, February 2019, pp. 133–144.
0099-1112/18/133–144
© 2019 American Society for Photogrammetry
and Remote Sensing
doi: 10.14358/PERS.85.2.133
PHOTOGRAMMETRIC ENGINEERING & REMOTE SENSING
February 2019
133
