Conference PaperPDF Available

Cartographic Rendering of Elevation Surfaces using Procedural Contour Lines


Abstract and Figures

We describe a method to render a height field given on a triangular mesh in a manner that is similar to well-known cartographic map views. This approach utilizes hypsometric tints and creates topographic contour lines procedurally as part of the rendering process. We discuss essential graphical enhancement techniques such as line smoothing, gradient compensation, contour line visibility fading, nonlinear elevation mapping, and shininess variation. The method is exemplified upon various numerical data sets from application sciences, demonstrating its ability to visually allow quantitative and qualitative precise assessment of the data values.
Content may be subject to copyright.
Cartographic Rendering of Elevation Surfaces using
Procedural Contour Lines
Neha Manya
Louisiana State University
Center for Computation &
Department of Computer Science
and Engineering Division
3127 P. F Taylor Hall
USA, LA-70803, Louisiana
Isaac Ayyala
Louisiana State University
Center for Computation &
Department of Electrical and
Computer Engineering
3101 P. F. Taylor Hall
USA, LA-70803, Louisiana
Werner Benger
Louisiana State University
Center for Computation &
213 Johnston Hall
USA, LA-70803, Louisiana
We describe a method to render a height field given on a triangular mesh in a manner that is similar to well-
known cartographic map views. This approach utilizes hypsometric tints and creates topographic contour lines
procedurally as part of the rendering process. We discuss essential graphical enhancement techniques such as
line smoothing, gradient compensation, contour line visibility fading, nonlinear elevation mapping, and shininess
variation. The method is exemplified upon various numerical data sets from application sciences, demonstrating
its ability to visually allow quantitative and qualitative precise assessment of the data values.
Contour lines, hypsometric tints, scientific visualization, bathymetry, coastal modeling, terrain visualization.
The challenge of rendering bathymetric data is
emphasizing features over a large scale while
sustaining accurate data representation. Cartography
is an ancient field of interest to draw upon for this
purpose. The use of colors to represent ranges of
elevation, known as hypsometric tints, is said to have
been invented by Leonardo da Vinci [Vin03].
Contour lines complement hypsometric tints to depict
more detailed information quantitatively. Both
methods are familiar and intuitive. A similar
approach is therefore highly desirable for application
upon numerical datasets where interactive
exploration is mandatory for data analysis, especially
when exploring complex geomorphology.
Deltaic estuaries in the northern Gulf of Mexico are
good examples of complex geomorphological
systems. These estuaries are characterized by
numerous channels and wetland networks on the
order of 20m 100m in size, and have a highly
complex bathymetry within a fairly narrow depth
range of 0m 5m. Implementation of 3-dimensional
hydrodynamic models is particularly challenging in
these systems because numerical grids need to
resolve both the complex bathymetry and intricate
wetland features. Consequently, bathymetric
rendering and visualization can provide invaluable
information for the generation of 3D numerical grids.
Related Work
The Collaborative Ocean Visualization Environment
(COVE [Gro10]) has been designed with
oceanographers to provide a set of elaborated tools
for bathymetric data exploration, including
hypsometric tinting and display of simple contour
lines. In our approach we seek to achieve even higher
accuracy in emphasis of finer variations in elevation,
limited only by numerical precision. In [Gul10] an
algorithm is presented to extract contour lines from
scanned color topographic maps. Topography maps
already have the contour lines in them, but they
might be broken or overlapping. They provide
techniques to filter noise, remove bifurcations and
connect lines properly. In contrast, here we intend to
create contour lines from the data itself. [Du04]
discusses the limitations of contour line generation.
They found that useful contour lines need to be
joined, but should only be extracted if they were not
too fragmented and enough spatial information was
available. The quality of procedural contour lines
such as in our approach will of course depend on the
Permission to make digital or hard copies of all or part of
this work for personal or classroom use is granted without
fee provided that copies are not made or distributed for
profit or commercial advantage and that copies bear this
notice and the full citation on the first page. To copy
otherwise, or republish, to post on servers or to
redistribute to lists, requires prior specific permission
and/or a fee.
21st International Conference on Computer Graphics, Visualization and Computer Vision 2013
Full papers proceedings
ISBN 978-80-86943-74-9
quality of available spatial information as well, but
joining and fragmentation of lines is not a hindrance.
We intend to render a height field, such as given in
the form of a triangular surface mesh, similar to the
well-known cartographic views known from
topographic maps. Firstly, this implies using a
colorization scheme widely known as hypsometric
tints: it renders bathymetric regions (below ‘sea
level’) using bluish shades and hypsometric regions
(above sea level) in shades from green via yellow to
brown to red and finally gray white colors, in
approximation of terrestrial average colors. Secondly,
we like to draw contour lines to allow quantitative
reading of height levels. Because explicit calculation
of contour lines as required for numerically precise
analysis is computationally expensive (even when
optimized using marching squares as 2D version of
marching cubes [Lor87]), here we want to provide a
quick preview method by implementing an
approximation of contour lines as part of the surface
shading process. A lookup table (LUT) containing
ten entries for bathymetric and ten entries for
hypsometric color values is sufficient to provide a
coarse overview of the data set. The intended purpose
of contour lines is to complement the colorization by
providing more detailed elevation information. While
they may be implemented via a very large, high-
resolution LUT, the availability of programmable
GPUs allows implementing them procedurally. Thus,
the achievable resolution is only limited by numerical
precision. The basic idea is to make use of the
modulo function h mod
:= h floor(h/
) for a
given height h. Depending on a level height (e.g. 1m,
10m, 100m) and a given finite thickness of the
contour line, it determines whether the hypsometric
tint from the LUT or a constant color for the contour
line, e.g. black or red, is used:
 
 
Using this modulo function it is straightforward to
implement multiple such contour lines at the same
time using different parameters. A set of five types of
contour lines is easy to interpret visually as it is
familiar from cartographic displays. These may be
placed as red thick lines at levels of 10 (0m, 10m,
20m, …), thick black lines shifted by 5 (5m, 15m,
25m, …), red thin lines at levels of 1 (1m, 2m, 3m,
4m, 6m, 7m, ..), black thin lines shifted by 0.5 (0.5m,
1.5m, 2.5m, 3.5m, 4.5m, ) and finally gray lines at
distance levels of 0.1 (0.1m, 0.2m, 0.3m, 0.4m, …).
This scheme allows reading off a height range of
1:100 quickly and even 1:1000 if we consider a line
width of 1cm, as demonstrated in Figure 1. While
this approach is pretty straightforward, there are
several aspects that can be, or need to be improved.
Figure 1: Contour lines as integrated part of
surface colorization, utilizing five categories of
contour levels.
Line Smoothing
A direct implementation of the modulo function leads
to strong aliasing (moiré) effects, which disturbs the
visual appearance since for each pixel there will be a
binary yes/no decision on whether it belongs to a
contour line or not. We can address this problem by
replacing the binary decision color(h) with a smooth
transition between the contour line and the ground
color by using a hermite spline interpolation as in
  
where  
  
 
. Such a function is
provided by the GLSL smoothstep() call.
Figure 2: Line smoothing to address aliasing.
We then introduce a smoothing distance , which can
be set to be
/4, for instance, and compute a
blending weight W=B
as product of a bottom
weight B
and top weight T
. The bottom part of the
contour line is calculated as
= S(0,
, (h+
) mod
) .
The top part is calculated via
= ( 1 - S(
/2, (h+
) mod
) ) .
Both functions smooth the line’s edges by
. We
need to consider that the line shall be centered
precisely at height h, extending by
upwards and
21st International Conference on Computer Graphics, Visualization and Computer Vision 2013
Full papers proceedings
ISBN 978-80-86943-74-9
downwards, smoothing out over a distance
. The
blending weight W is 1 for h being an integer
multiple of
such that the final color is given as
color(h) =(1-W)
LUT(h) + W
Contour .
Therefore, by using the above formulae, the edges of
the contour lines are smoothened, implementing
Gradient Compensation
As contour lines are implemented in this approach by
identifying a set of points within the same height
range, we will naturally find more points fitting this
criterion in shallow regions, resulting in a broadening
of the lines depending on the terrain’s slope. The
problem is inherent by design because a contour line
becomes undefined in a completely flat area. While
this broadening of lines is visually displeasing as
compared to the computation of an explicit contour
line, it may also be considered as a depiction of the
contour lines becoming numerically unstable in flat
areas. This would not necessarily be evident from an
explicit contour line. Therefore, this effect cannot be
fully cured. But, it can be compensated to some
extent by taking into account the slope of the terrain
and introducing a dependency of the contour line’s
thickness on the surface normal vector n (|n|=1) via
  
Here, g is an exponent allowing fine-tuning the
‘gradient compensation’. “Mathematically correct”
would be an exponential of 0.5, however in practice
the ability of over-compensation turns out to be
beneficial, as depicted in Figure 3.
Figure 3 Procedural contour lines in a shallow
wetland area, as-is and using a gradient
compensation’ exponent of 16 to narrow lines.
Visibility Fading
Contour lines of constant thickness as in Figure 4
come with the disadvantage that a thickness value
working pleasingly well in the foreground leads to
invisible or strongly aliased lines at further distances.
In turn, a thickness value providing good background
appearance looks ugly in the foreground. We thus
relate both the blending weight W as well as the
contour line thickness
to the apparent size of the
contour line on the screen, such to have lines in the
distance being broader, but less prominent, and lines
close to the observer being smaller but crisper.
However, we cannot simply introduce a dependency
of the surface point on the observer’s location
because this would fail in the case of an orthographic
camera projection. Instead, we need to compute a
visibility measure that is related directly to the
apparent size of the contour line on the screen.
We proceed as follows: given the normal vector n of
the surface, we can get a vector t tangential to the
contour line by taking the cross product with the ‘up’
vector z=(0,0,1) as t = n × z . The tangential vector t
will always be parallel to the horizontal plane, i.e.
=0. However, this expression also applies to a
general underground, such as the curvature of Earth,
by using an arbitrary local ‘up’-vector z. From this
tangential vector we compute a vector s parallel to
the surface but orthogonal to the tangent via s = n ×
t. Now, given a point on the surface P we can
compute the location Q of the upper boundary of our
contour line as Q = P +
s. Both Q and P, given in
world coordinates, are then projected in screen
coordinates my multiplying in 4D with the OpenGL
projection and modelview matrix M, yielding 4D
homogenous screen coordinates
P=M P,
3D affine screen coordinates are computed via
dividing by the 4
). We are
only interested in the x and y components of these
projected points as they are directly related to the size
of the contour line in pixels. To get the actual number
of pixels we would need to multiply x and y with the
width and height of the viewport in pixels. However,
we want a pixel-resolution independent algorithm
and thus only consider normalized screen
coordinates. We define a visibility factor V as
 
based on some adjustment factor V
to fine-tune the
visual appearance. The visibility factor will be small
at large distances of the contour lines from the screen
as well as the contour-line being viewed tangentially,
which would result in a visibly small line as well. To
compensate for aliasing effects and enhance distant
lines, we increase the line width via dividing by the
21st International Conference on Computer Graphics, Visualization and Computer Vision 2013
Full papers proceedings
ISBN 978-80-86943-74-9
visibility factor while at the same time reducing its
blending weight:
/ V , W
= W
V .
Using this effective line thickness
and blending
factor W
eff ,
we achieve visually pleasing crisp lines in
the foreground, fading out into bigger lines
emphasizing coarser structures in the background, as
demonstrated vs. Figure 5.
Figure 4 Contour lines with no distance
dependency, leading to overly thick foreground
lines while distant lines remain hardly visible.
Figure 5 Contour lines with visibility fading,
enhancing details crisper in the foreground as well
as in background.
Non-linear Hypsometric Tints
Classically each entry in the color lookup table
covers the same height range to allow easy
identification of elevation by its color. A non-
equidistant mapping in contrast provides the
advantage of allowing emphasizing finer variations
in elevation while covering the same height range.
Whereas, the combination with contour lines still
allows quantitatively precise read-off. One pretty
simple choice of a non-linear mapping of height to
colors is to utilize the arctangent function, as it maps
an infinite range] of input values to the
range [–π/2, /2]. Mapping this range to the interval
[0,1] for indexing a LUT is straightforward. We can
further introduce a power function to parameterize
the mapping of height h for additional adjustment:
 
 is the LUT indexing value and
is a tunable exponent. Values smaller
than 1 will compress color equipotential lines around
the sea level h=0 and enhance low-elevation details,
as depicted in Figure 1.
Figure 6 : Linear and non-linear colorization.
Shininess Variations
Modification of the shininess parameter as part
Phong shading allows to view the surface and the
below the sea level of the grid differently.
21st International Conference on Computer Graphics, Visualization and Computer Vision 2013
Full papers proceedings
ISBN 978-80-86943-74-9
Figure 7: Variation of surface reflectivity
(shininess) applied to water bodies within a
shallow wetland area.
We use an icosahedron as a very simple example of a
triangular mesh to validate aspects of our technique.
Here, the height values range from -1.618 to +1.618.
Taking this range into consideration, a set of contour
lines is shown. Basically, three red thick lines are
shown representing elevation levels of -1, 0, +1.
Figure 8: Linear and non-linear colorization
without smoothening effect.
Figure 9: Non-Linear colorization with power less
than one and greater than one
Barataria Bay
Barataria Bay is a 120 km long estuary located in the
north-central Gulf of Mexico, just to the west of the
Mississippi River Delta (Figure 10). It is bounded by
the Mississippi River to the east and by a former
channel of the Mississippi River, Bayou Lafourche,
to the west. The estuary is connected to the Gulf of
Mexico through four tidal passes and contains several
large lakes and numerous marshes interconnected by
ponds and channels. The average depth is only about
2 m but there are several deeper channels (e.g., ~ 4 m
deep Barataria Bay Waterway) and tidal inlets (e.g.,
~ 20 m deep Barataria Pass).
Figure 10 Barataria Bay estuary: Mississippi river
is at right, New Orleans just above the center.
Here we explore a triangular mesh consisting out of
63190 vertices and 125038 triangle elements,
covering a total area of 3140km
, out of which
is open water. The grid resolution ranges
from 18m in the channels to 1600m over wetlands
and open water bodies. This mesh carrying precise
21st International Conference on Computer Graphics, Visualization and Computer Vision 2013
Full papers proceedings
ISBN 978-80-86943-74-9
bathymetric information is the basis for several
numerical hydrodynamic models [Ino08, Das10,
Das11, Das12].
Figure 11 Barataria Pass
The maximum and minimum elevations are at 8.2m
and -22.4m, respectively. Our contour line algorithm
automatically places contour lines at the most
appropriate power of 10 levels, therefore placing
three thick red lines at levels of 0.0m, -10m and -
20m. Thin red lines are placed at 1m levels,
alternating with thin black lines offset by 0.5m. Gray
lines represent elevation levels of 10cm.
Figure 12 Lake Salvador outlet
Important features in this numerical mesh such as
Barataria Pass, Figure 11, or Lake Salvador outlet,
Figure 12, are clearly emphasized when applying our
cartographic 3D rendering. With proper vertical
scaling, Barataria Bay Waterway that carries a large
portion of the water flow in the area is represented
more prominently when compared with a
conventional 2D cartographic view.
Figure 13 "Grand Canyon of Barataria Bay" -
Barataria Bay Waterway.
Optionally, we may also just display the contour lines
on their own by rendering the surface itself
transparent but only the contour lines opaque. This
effect on the Barataria Bay waterway is demonstrated
in Figure 14, resulting in a fine detailed depiction of
contour levels that are automatically adjusting to
view distance.
Figure 14 Barataria Bay Waterway displaying
procedural contour lines as derived from the
numerical triangular mesh.
Rheinfelden LIDAR Terrain
The “Rheinfelden” data set has been derived from an
airborne light detecting and ranging (LIDAR) laser
scan using a bathymetric green laser system, the
Riegl hydro-graphic laser scanner VQ-820G [Stei10].
This laser system is able to penetrate a water surface
to provide bathymetric data of shallow water
grounds. Here, a river section around a power plant
in Rheinfelden, Germany, was acquired. The scan
21st International Conference on Computer Graphics, Visualization and Computer Vision 2013
Full papers proceedings
ISBN 978-80-86943-74-9
was complemented by additional sonar
measurements. We computed a digital terrain model
(DGM) and a digital water model (DWM), resulting
in a mesh consisting of 1880352 triangle elements
with 944521 vertex nodes.
The ‘sea level’ of this data set has been set as 266.5m
to correspond with the height of the water surface at
power plant’s location. Figure 15 and Figure 16
provide views along the river using identical
visualization parameters.
For comparison, contour lines were computed using a
CPU marching squares algorithm, a simpler 2D
variant of the 3D marching cubes [Lor87]. For each
elevation meter a contour line was created in the
Rheinfelden dataset, operating on the raw data
representation as a point cloud (1.0m resolution). The
computation time of the serial code was 12.5 seconds
Intel-i7-2670QM@ 2.20GHz. Using parallelization
via OpenMP, 8 threads on 8 CPUS reduced the
computation time to 3.5seconds. In contrast, using an
NVIDIA Quadro3000 on the same system, the
triangular surface with procedural contour lines
renders at 9.0 frames per second (fps) or 0.111
seconds. This results in a speed up of 32.1 in
comparison to the OpenMP parallel CPU code for
visual scene exploration or rendering of contour
plots. If contour lines are required as explicit
geometry for further processing, the shader-based
procedural contour lines cannot replace the CPU
Figure 15 LIDAR-acquired bathymetry of a river
power plant in Rheinfelden.
Figure 16 Bathymetry of a river bend at the
Rheinfelden dataset, exposing some canyons.
Baltic LIDAR Terrain
Another LIDAR dataset was acquired by the same
airborne laser scan system as used for the
Rheinfelden area. The raw dataset provides a point
cloud of 385 million elements. It was reduced to a
triangular surface representing the ground surface
(DGM) and a second triangular surface representing
the water surface (DWM). The DGM has 1659958
vertices with 3228653 triangle elements and the
DWM 2670748 vertices with 5271291 triangle
elements. Rendering of the DGM results in a frame
rate of 6.3 fps, rendering of both, DGM and DWM,
in 2.9 fps, measured on an Intel-Xeon-
X5650@2.67Ghz(x6) NVIDIA-Quadro5000 system.
Figure 17: Variation in surface reflectivity. The
coast line is illustrated by the border of the water
surface extracted from the classified point cloud.
Figure 18: Highly sensitive hypsometric tinting
illustrates structures on the ocean ground. The
coast line matches the coast line of Figure 17.
Figure 19: Procedural contour lines emphasize the
sand dunes on the ocean ground. Sand dunes have
a length of about 50 to 800 meters.
21st International Conference on Computer Graphics, Visualization and Computer Vision 2013
Full papers proceedings
ISBN 978-80-86943-74-9
Figure 17 demonstrates the effectiveness of varying
the surface reflectivity to emphasize sea level. It
shows how the coastline, illustrated by the shader
(opaque surface), matches the coast line that was
generated from laser echoes which had been
classified as water surface (transparent surface). This
is, for example, well-illustrated in the harbor region.
Figure 18 provides an alternative view clearly
depicting sub-surface structures on the ocean ground,
such as sand dunes or the excavated harbor entrance.
Figure 19 complements the colorization scheme with
procedural contour lines. This scheme emphasizes
the sand dune structures on the ocean ground.
We found the impact of including procedural contour
line in addition to usual color-coding of surfaces by
height information to be negligible. For the Barataria
Bay the rendering frame rate was found to be at most
four percent slower when compared to rendering
without contour lines. For the larger Rheinfelden
dataset, the rendering frame rate of the grid was only
0.3 percent slower with contour lines than without.
Table 1 summarizes our measurements. Since the
rendering is view-dependent, the view point and
direction influences the rendering time, so the
compared views from the top, the front and the side.
The values for the rendering frame rate are averaged
from repeated execution of the actual OpenGL
rendering call.
of the
Frame rate
ia Bay
Table 1: Rendering time performance of the grid
with contour lines and without contour lines
To calculate the performance, the graphics processor
used was a Quadro FX5600 with memory 1536 MB.
Also, eight CPU’s with model name Intel® Xeon®
CPU with memory 32GB was used.
These measurements merely refer to the performance
of the shader. The triangular surfaces have been
rendered as-is, without dynamic triangle refinement
and level-of-detail. Those mechanisms, common
from terrain rendering in particular when rendering
height data given on a regular grid, are orthogonal to
our shading technique and can be used to
complement the overall performance, but are not part
of our algorithm.
In this article we presented a highly efficient method
to render a height field given on a triangular mesh in
a manner similar to well-known cartographic views,
enhanced by essential or useful features such as line
smoothing, gradient compensation, visibility fading,
non-linear elevation mapping and shininess variation.
The method can be implemented fairly simple on
modern graphics cards to yield real-time renderings
allowing precise quantitative assessment of data
values. The effectiveness of this method has been
demonstrated on a diverse set of geoscientific
datasets from current research projects.
It has been noted [Pat12] that the use of hypsometric
tints may be misleading when representing terrains.
E.g., low areas will always appear ‘lush’ even if they
refer to a desert. Similarly open water above sea level
will look as if it were on a land mass. This problem
may be addressed by using cross-colorization, which
is modifying the colors used in the LUT based on an
additional scalar data field given on the terrain.
Doing such requires a two-dimensional LUT and
faces the challenge of non-repeating color entries to
keep the rendering meaningful. In lieu or
additionally, incorporating blending with RGB
information from aerial photography may be useful
as well. Also, alternative non-linear functions, in
particular mappings based on histograms [Khu12],
may be used to emphasize local details even better.
The visual quality may well be enhanced by adding a
geometry or tessellation shader to insert more
triangles where needed, based on high order
interpolation methods (i.e., dynamic level of detail).
However, from a scientist’s perspective to accurately
identify their data elements, such an enhancement
might not necessarily be desirable. Finally, the
technique presented here is neither limited to
elevation data nor to surfaces. All methods apply as
well to a general scalar field replacing the elevation.
Such may be given on a set of lines or non-planar
surfaces as well, once the gradient of the scalar field
and the base domain is known to allow for gradient
compensation and visibility fading.
21st International Conference on Computer Graphics, Visualization and Computer Vision 2013
Full papers proceedings
ISBN 978-80-86943-74-9
This study was funded in part by BP through the
grants to Louisiana State University and Coastal
Waters Consortium. Special thanks to Frank
Steinbacher for providing the LIDAR datasets.
Thanks to Wolfgang Dobler for providing time
measurements on the data set Rheinfelden. This work
was supported by the Austrian Science Foundation
FWF DK+ project Computational Interdisciplinary
Modeling (W1227), and grant P19300. This research
employed resources of the Center for Computation
and Technology at Louisiana State University, which
is supported by funding from the Louisiana
legislatures Information Technology Initiative.
8. Additional Authors
Marcel Ritter, University of Innsbruck & Airborne
Hydromapping OG, Technikerstrasse 13&21 A-
6020, Innsbruck, Austria; Dubravko Justic,
Department of Oceanography and Coastal Sciences,
School of the Coast and Environment, 2221 Energy
Coast and Environment, Louisiana State University,
Baton Rouge, LA 70803,; Lixia
Wang, Department of Oceanography and Coastal
Sciences, School of the Coast and Environment, 2223
Energy Coast and Environment, Louisiana State
University, Baton Rouge, LA 70803,
[Gul10] Gul, S.; Khan, M.F., "Automatic Extraction
of Contour Lines from Topographic Maps,"
Digital Image Computing: Techniques and
Applications (DICTA), 2010 International
Conference on , vol., no., pp.593,598, 1-3 Dec.
[Du04] Du Jinyang; Zhang Yumei, "Automatic
extraction of contour lines from scanned
topographic map," Geoscience and Remote
Sensing Symposium, 2004. IGARSS '04.
Proceedings. 2004 IEEE International , vol.5, no.,
pp.2886,2888 vol.5, 20-24 Sept. 2004
[Gro10] Grochow, K.; Stoermer, M.; Fogarty, J.;
Lee, C.; Howe, B.; Lazowska, E., "COVE: A
Visual Environment for Multidisciplinary Ocean
Science Collaboration," e-Science (e-Science),
2010 IEEE Sixth International Conference on ,
vol., no., pp.269,276, 7-10 Dec. 2010
[Stei10] F. Steinbacher, M. Pfennigbauer, A. Ulrich,
and M. Aufleger, “Vermessung der
Gewässersohle - aus der Luft durch das
Wasser,” in Wasserbau in Bewegung ... von der
Statik zur Dynamik. Beiträge zum 15.
Gemeinschaftssymposium der Wasserbau
Institute TU München, TU Graz und ETH Zürich,
[Pat12] Patterson, T., & Jenny, B. (2012). The
Development and Rationale of Cross-blended
Hypsometric Tints. Cartographic Perspectives,
0(69), 31-46. Retrieved from
[Ino08] Inoue, M., Park, D., Justic, D., Wiseman, W.
J., Jr. 2008. A High-Resolution Integrated
Hydrology-Hydrodynamic Model of the Barataria
Basin System. Environmental Modelling and
Software 23: 1122-1132.
[Das10] Das, A., Justic, D., Swenson, E. 2010.
Modeling estuarine-shelf exchanges in a deltaic
estuary: Implications for coastal carbon budgets
and hypoxia. Ecological Modelling 221: 978-985.
[Das11] Das, A., Justic, D., Swenson, E., Turner, R.
E., Inoue, M., Park., D. 2011. Coastal land loss
and hypoxia: The 'outwelling' hypothesis
revisited. Environmental Research Letters 6:
025001 (9 pp); doi:10.1088/1748-
[Das12] Das, A., Justic, D., Inoue, M., Hoda, A.,
Huang, H., Park., D. 2012. Impact of Mississippi
River diversions on salinity gradients in a deltaic
Louisiana estuary: Ecological and management
implications. Estuarine, Coastal and Shelf
Science 111: 17-26.
[Vin03] Leonardo da Vinci, A map of central Italy,
1503, Royal Collection by 1690; available from
[Lor87] Lorensen, William and Harvey E. Cline.
Marching Cubes: A High Resolution 3D Surface
Construction Algorithm. Computer Graphics
(SIGGRAPH 87 Proceedings) 21(4) July 1987, p.
[Khu12] S. Khurana, N. Brener, B. Karki, W.
Benger, S. Roy, S. Acharya, M. Ritter, S. Iyengar,
Multi Scale Color Coding of Fluid Flow Mixing
Indicators along Integration Lines, WSCG2012,
Plzen, 2012
21st International Conference on Computer Graphics, Visualization and Computer Vision 2013
Full papers proceedings
ISBN 978-80-86943-74-9
... As you can see from the last row in Table 5, the 8*40,002 points at 0.25 m 1 km 2 visualisation frame rate is below a frame a second and we were unable to gather measurements. We can compare the fps of our implementation to the work of [43]. They generated a LiDAR ground surface model (DGM) consisting of 1,659,958 vertices with 3,228,653 polygons at 6.3 fps on a Intel-XeonX5650 @ 2.67 Ghz(x6). ...
Full-text available
In this paper we investigate the use of games technologies for the research and the development of 3D representations of real environments captured from GIS information and open source map data. Challenges involved in this area concern the large data-sets to be dealt with. Some existing map data include errors and are not complete, which makes the generation of realistic and accurate 3D environments problematic. The domain of application of our work is crisis management which requires very accurate GIS or map information. We believe the use of creating a 3D virtual environment using real map data whilst correcting and completing the missing data, improves the quality and performance of crisis management decision support system to provide a more natural and intuitive interface for crisis managers. Consequently, we present a case study into issues related to combining multiple large datasets to create an accurate representation of a novel, multi-layered, hybrid real-world maps. The hybrid map generation combines LiDAR, Ordnance Survey, and OpenStreetMap data to generate 3D cities spanning 1 km2. Evaluation of initial visualised scenes is presented. Initial tests consist of a 1 km2 landscape map containing up to 16 million vertices' and run at an optimal 51.66 frames per-second.
Full-text available
Large-scale river diversions on the lower Mississippi River are considered to be an important component of wetland restoration efforts in coastal Louisiana. Diversions are used primarily for salinity control but increasingly proposed also as a major way to deliver sediments and nutrients to coastal wetlands impacted by the construction of flood control levees. We used a coupled hydrology–hydrodynamics model of the Barataria estuary, a site of the Davis Pond Diversion – the world's largest river diversion project, to examine salinity variations under different diversion discharge scenarios. Discharge scenarios were selected based on actual freshwater discharges in different years and management alternatives that included a scenario with several new diversions. The model results indicate that river diversions strongly affect salinities only in the middle section of the Barataria estuary. The upper parts of the estuary are fresh most of the time and so the excess fresh water from river diversions has only a minor impact on salinity in this region. Also, the Davis Pond diversion has little impact on salinities in the coastal section of the estuary because of strong marine influence in this area adjacent to the Gulf of Mexico. Interestingly, the predicted salinity differences between different model scenarios can be as high as 10 in some months and places. These differences can be biologically significant depending on the salinity tolerance of different species and could cause a shift in community composition within the affected region.
Full-text available
We present a new algorithm, called marching cubes, that creates triangle models of constant density surfaces from 3D medical data. Using a divide-and-conquer approach to generate inter-slice connectivity, we create a case table that defines triangle topology. The algorithm processes the 3D medical data in scan-line order and calculates triangle vertices using linear interpolation. We find the gradient of the original data, normalize it, and use it as a basis for shading the models. The detail in images produced from the generated surface models is the result of maintaining the inter-slice connectivity, surface data, and gradient information present in the original 3D data. Results from computed tomography (CT), magnetic resonance (MR), and single-photon emission computed tomography (SPECT) illustrate the quality and functionality of marching cubes. We also discuss improvements that decrease processing time and add solid modeling capabilities.
The export of wetland-derived materials to the coastal ocean (i.e., the “Outwelling” hypothesis) has received considerable attention over the past several decades. While a number of studies have shown that estuaries export appreciable amounts of nutrients and carbon, few studies have attempted to estimate the importance of estuarine sources for the coastal carbon budgets in river-dominated coastal ecosystems. A novel tidal prism model was developed to examine estuarine-shelf exchanges in the Barataria estuary, a deltaic estuary located in the north-central Gulf of Mexico. This estuary has been the site of a massive wetland loss, and it has been hypothesized that carbon export from the eroding coastal wetlands supports the development of a large hypoxic zone in the coastal Gulf of Mexico. The model results show that the Barataria estuary receives nitrogen through the tidal passes and releases carbon to the coastal ocean. The mean calculated tidal water discharge of 6930m3s−1 is equivalent to about 43% of the lower Mississippi River discharge. The annual total organic carbon (TOC) export is 109 millionkg, or 57gCm2yr−1 when prorated to the total water area of the estuary. This carbon export is equivalent to a loss of 0.5m of wetland soil horizon over an area of 8.4km2, and accounts for about 34% of the observed annual wetland loss in the estuary between 1978 and 2000. Compared to the lower Mississippi River, the Barataria estuary appears to be a very small source of TOC for the northern Gulf of Mexico (2.7% of riverine TOC), and is unlikely to have a significant influence on the development of the Gulf's hypoxia.
Conference Paper
The automatic extraction of contour lines and generation of digital elevation models (DEMs) from topographic maps is one of the challenging subjects mostly because of aliasing, false colors, closely spaced lines and other features causing intersection or overlapping. In this paper we present an algorithm to extract contour lines from colored images of scanned topographic maps. In our approach, we first segment the color image using adaptive thresholding to extract basic contour structure. Noise in the images is removed utilizing morphological operation. Next, the contour lines are reduced up to unitary thickness using Zhang's thinning algorithm. The bifurcation and holes that result after thinning are removed using different masks. After thinning, end points of the broken contours are identified and best candidate for connection is determined, this is performed by analyzing the Euclidean distance and direction of end points near the gap. Then broken contour lines are joined employing curve fitting technique. The performance of the algorithm is tested on several samples of topographic maps and results show good segmentation of the contour lines. This automatic extraction algorithm for contour lines from topographic maps can save significant amount of time and labor as well as improving the accuracy of the contour line extraction.
Conference Paper
Advances in cyber infrastructure for virtual observatories are poised to allow scientists from disparate fields to conduct experiments together, monitor large collections of instruments, and explore extensive archives of observed and simulated data. Such systems, however, focus on the `plumbing' and frequently ignore the critical importance of rich, 3D interactive visualization, asset management, and collaboration necessary for interdisciplinary communication. The NSF Ocean Observatories Initiative (OOI) is typical of modern observatory-oriented projects-its goal is to transform ocean science from an expeditionary science to an observatory science. This paper explores the design of an interactive tool to support this new way of conducting ocean science. Working directly with teams of scientists, we designed and deployed the Collaborative Ocean Visualization Environment (COVE). We then carried out three field evaluations of COVE: a multi-month deployment with the scientists and engineers of an observatory design team and two deployments at sea as the primary planning and collaboration platform on expeditionary cruises to map observatory sites and study geothermal vents.
Conference Paper
The Barataria Basin, a bar-built estuarine system located directly west of the Mississippi Delta, has been experiencing a significant land loss, especially since the leveeing of the Mississippi River for flood control purposes in early 20th century. Recent efforts to alleviate the land loss problem include the construction of man-made freshwater diversion structure in order to divert river water as well as its associated suspended sediments from the Mississippi River into the Barataria Basin. In order to implement an ecologically friendly management plan of those diversions, a careful examination of the anticipated salinity alterations resulting from the operation of the diversions is required. A high-resolution (O(100m)), integrated hydrology-hydrodynamic model of the Barataria Basin has been developed to simulate the local hydrological cycle over the surrounding drainage basin and hydrodynamics within the basin. The integrated model is forced by observed tides coming from the Gulf of Mexico, local wind, rainfall and evaporation over the model domain, salinity and temperature estimated at the open boundary located offshore of the mouth of the bay. Estimated local precipitation and evaporation over the model domain provide hydrological forcing of the hydrological model, that in turn simulates local runoff into the hydrodynamic model. A novel feature of the hydrodynamic model is its use of a very accurate advection scheme, thus, enabling accurate simulation of salinity variations in response to changes in various hydrological forcing functions. A flood event that took place during the tropical storm Allison in June 2001 resulted in significant sea-level changes especially in the upstream region of the basin. The integrated model appears to be able to capture a significant portion of the observed sea-level variations during the flood. Significant effects on water level and salinity are observed in the multiply connected channels through the marsh in the vicinity of operating diversion structure and in the open waters downstream.