Documente Academic
Documente Profesional
Documente Cultură
analysis system.
Prepared by
Chris Ryan
www.toolkit.net.au/catchmentsim
USER GUIDE
Document History
Date
01/06/2005
Author
Chris Ryan
Revision
1.00
Description of Change
Initial Document
Copyright Notice
CRC for Catchment Hydrology, Australia 2005
Legal Information
To the extend permitted by law, the CRC for Catchment Hydrology (including its employees and
consultants) accepts no responsibility and excludes all liability whatsoever in respect of any persons use or
reliance on this publication or any part of it
CatchmentSIM
CONTENTS
1
Introduction ....................................................10
1.1
1.2
Related documents........................................................................................ 11
1.3
1.3.2
1.4
Product components...................................................................................... 12
1.5
Installation......................................................13
2.1
2.2
2.3
Folders ........................................................................................................ 13
2.4
Installation ................................................................................................... 14
2.5
2.4.1
Stand-alone PC .......................................................................... 14
2.4.2
2.4.3
3.1
Introduction.................................................................................................. 16
3.2
3.3
3.4
3.3.1
3.3.2
3.4.2
3.4.3
3.5
3.5.2
3.6
Filling Algorithm.......................................................................... 29
3.6.2
3.7
3.8
3.9
3.8.1
Theory ....................................................................................... 35
3.8.2
3.10
3.10.2
3.11
3.12
Urbanisation Tools........................................................................................ 45
3.12.1
3.12.2
3.12.3
3.12.4
3.13
3.14
3.15
Summary...................................................................................................... 52
Program Reference......................................... 54
4.1
File Menu................................................................................... 54
New Project (Ctrl+N) .................................................................. 55
Project Information ...................................................................... 55
Projection ................................................................................... 56
Design Plane .............................................................................. 57
ii
iii
Scaling Factors ........................................................................... 58
Open Project (Ctrl+0) ................................................................. 58
Project Options........................................................................... 59
Projection Properties.................................................................... 65
Import Mapping Layer ................................................................. 65
Import Raster Layer...................................................................... 65
Save Project................................................................................ 65
Save Project As ........................................................................... 65
Print........................................................................................... 66
Exit Program (Ctrl+Alt+Q) .......................................................... 66
4.1.2
4.1.3
iii
4.1.5
iv
v
Deleting Hydraulic Controls ......................................................... 95
Draw Hydraulic Control Submenu ................................................ 96
Draw Channel ............................................................................ 96
Draw Gutter ............................................................................... 96
Delete Hydraulic Control ............................................................. 96
View Hydraulic Breaches.............................................................. 97
Inspect Pixel................................................................................ 98
4.1.6
4.1.7
4.1.8
5.1
Introduction................................................................................................ 127
5.2
5.3
5.3.2
5.4
5.5
5.6
5.5.2
5.6.2
6.1
6.2
vi
6.1.1
6.1.2
vii
6.2.1
6.2.2
6.2.3
6.2.4
6.3
References ....................................................154
vii
TABLE OF FIGURES
Figure 3-1 : Vector to Raster Conversion of All Underlying Cells .......................................................... 21
Figure 3-2 : Selective Vector to Raster Conversion.............................................................................. 21
Figure 3-3 : Stream Processing Sequencing ....................................................................................... 22
Figure 3-4 : CatchmentSIM after Watercourse Integration ................................................................... 23
Figure 3-5 : Adherence of Calculated Flow Paths to Stream Network ................................................... 23
Figure 3-6 : HUD Digitising of Contour Lines..................................................................................... 24
Figure 3-7 : HUD Digitising of Streams ............................................................................................. 25
Figure 3-8 : HUD Digitising of ITLs ................................................................................................... 25
Figure 3-9 : Interpolation of Digital Elevation Model .......................................................................... 26
Figure 3-10 : Flat Cross-Section Discounting Scenario........................................................................ 26
Figure 3-11 : Increasing Resolution of DEM Interpolation.................................................................... 27
Figure 3-12 : Effect of 'Stream Burning' on Flow paths ........................................................................ 28
Figure 3-13 : Interpolation Flat-Spots................................................................................................ 30
Figure 3-14 : Priority First Search Algorithm Methodology ................................................................... 33
Figure 3-15 : Calculation of Flow Direction ....................................................................................... 35
Figure 3-16 : Vector Flow Path......................................................................................................... 36
Figure 3-17 : Path Mapping Capabilities of CatchmentSIM ................................................................. 36
Figure 3-18 : Propagation of D8 Error Downslope ............................................................................. 37
Figure 3-19 : D8 Flow Paths vs CatchmentSIM Algorithm.................................................................... 37
Figure 3-20 : Basin Delineation, D8 Method vs CatchmentSIM............................................................ 38
Figure 3-21 : Strahler's (1957) Revision of Horton Stream Ordering..................................................... 39
Figure 3-22 : Flow Accumulation Jump Break-up Algorithm ................................................................ 41
Figure 3-23 : Horton Catchment Break-up Algorithm ......................................................................... 43
Figure 3-24 : Sample CatchmentSIM Nodal Network Arrangement...................................................... 44
Figure 3-25 : Impervious Areas Database.......................................................................................... 45
Figure 3-26 : Rasterisation of Impervious Area Polygons ..................................................................... 46
Figure 3-27 : Effect of Hard-Coding of Road Crowns on Stream Network............................................. 47
Figure 3-28 : Using a Stream Network as Channel Hydraulic Controls................................................. 48
Figure 3-29 : Hydraulic Representation of Gutters .............................................................................. 49
Figure 3-30 : CatchmentSIM Coupling Methodology.......................................................................... 51
Figure 3-31 : Sample CST Script Generated Report............................................................................ 52
Figure 5-1 : Re-projection of a lat/long DEM into a projected coordinate system................................. 128
Figure 5-2: Interpolate Streams Over Existing DEM Algorithm............................................................ 132
Figure 5-3: Comparison of Horton (top) and Flow Accumulation Analysis (bottom) Breakup
Algorithms .................................................................................................................... 134
Figure 5-4 : CatchmentSIM Charting of Stream Drop Scatter............................................................. 136
Figure 5-5 : CatchmentSIM Charting of SAT vs Students T Value ...................................................... 137
Figure 5-6 : Calculation of Bifurcation Ratio .................................................................................... 138
viii
ix
Figure 5-7 : SAT vs Bifurcation Relationship for Two Subcatchments ...................................................138
Figure 5-8: Channel Hydraulic Control Flow Paths............................................................................140
Figure 5-9: Subcatchments Mapped with/without Channel HC...........................................................140
Figure 5-10: Perpetual Flow Path Example .......................................................................................141
Figure 5-11: Effect of Gutter HC with Low Tolerance.........................................................................142
Figure 5-12: Effect of Gutter HC with high tolerance.........................................................................142
Figure 6-1 : Holland Park GIS Layers...............................................................................................144
Figure 6-2 : Holland Park CatchmentSIM Results ..............................................................................145
Figure 6-3 : Holland Park Results Superimposed over Aerial Photography ...........................................146
Figure 6-4 : Holland Park City Plan GIS Database ............................................................................147
Figure 6-5 : Upper Washita CatchmentSIM Results ...........................................................................149
Figure 6-6 : Upper Washita 3D CatchmentSIM Catchment and Streams .............................................149
Figure 6-7 : Upper Washita Catchment Delineation Differences .........................................................150
Figure 6-8 : Upper Washita Catchment Delineation Differences (3D)..................................................151
Figure 6-9 : Subcatchment Delineation near Outlets .........................................................................152
Figure 6-10 : Upper Washita Stream Network Comparison over DEM ................................................152
Figure 6-11 : Upper Washita Stream Network Comparison ...............................................................153
ix
1 Introduction
CatchmentSIM is a stand-alone GIS based terrain analysis system designed to help setup and
parameterise hydrologic models. It automatically delineates catchments, calculates their
properties and creates run files for other models (eg., WBNM, RORB, RAFTS, URBS, DRAINS,
HEC-HMS etc).
A brief overview of CatchmentSIM features includes:
Catchment delineation using a more accurate flow routing algorithm than the D8
method adopted in most comparable software applications;
A flexible result export macro language enabling creation of text or binary files in any
format. This enables CatchmentSIM to be coupled with any other model with a
published file format or used to generate tabulated results in a standardised report
format.
Macro scripts are currently available for integration with WBNM, RAFTS, RORB,
URBS, DRAINS and HEC-HMS as well as a generic Rich Text File (RTF) report format.
Introduction
1.1
1.2
Related documents
Other documents that are available for CatchmentSIM include:
1.3
Three (3) technical papers: these papers are available from the Toolkit
website.
Data requirements
1.3.1
Input data
CatchmentSIM is based on raster DEMs which can be created in a number of ways including:
11
1.3.2
1.4
Text or binary files in any desired format (using the CSTalk macro language);
3D Visulisations;
Product components
CatchmentSIM is distributed as a single InstallShield execuatable.
1.5
12
Installation
2 Installation
2.1
Technical specifications
CatchmentSIM requires the following hardware:
Type of machine
Minimum CPU
Minimum memory
Minimum available disk space
Operating system
The person installing the software must have local administrator access on the computer.
CatchmentSIM is a memory intensive program. The memory requirements of the software
and the run of key algorithms increase with the size of the DEM being modelled.
2.2
Licence agreement
A licence agreement is part of the installation procedure. You must acknowledge that you
have read, understood and agree to be bound by the CatchmentSIM software licence
agreement to be able to proceed with the installation.
2.3
Folders
During install, you will have the option to choose the installation directory for CatchmentSIM.
The default option is: C:\Program Files\CatchmentSIM . However, if the local user does not
have administrator privlidges then the software should be installed to an unprotected
13
2.4
Installation
2.4.1
Stand-alone PC
2.4.2
File server/network
CatchmentSIM should not be installed on a file server. Project files may be installed on a file
server however the software should reside on the local PC.
2.4.3
Auto-update feature
If your computer is connected to the Internet, it can automatically check for newer versions of
the software on the Toolkit website. This can be done by selecting Help >> Check For New
Version (page 125). You may also be periodically offered an option to check for a later
version.
2.5
Product Registration
CatchmentSIM is currently free software, although each version includes an expiry date (see
Help >> About page 126) and requires registration. Registration is required simply to
keep track of the number of people using the software. There are two options to register
CatchmentSIM as described bellow:
14
Installation
Following registration, CatchmentSIM will be ready to use. You will not need to register
CatchmentSIM again until a new version is installed.
2.5.1
Uninstalling CatchmentSIM
15
3 Using
CatchmentSIMs
Algorithms
3.1
Introduction
This chapter outlines the basic sequence of tasks that might be employed in a typical
CatchmentSIM project. The algorithms are described in detail in this chapter because a keen
understanding of their respective approaches, advantages and limitations is fundamental to
applying them successfully. To get help with a specific menu item, form or function, the
Program Reference information in Chapter 4 may be more appropriate. Hyperlinks and page
references are provided for each section which refer to the Program Reference section which
contains specific instructions for using the particular tool.
3.2
CatchmentSIM Tutorial
The best way to come to grips with using CatchmentSIM is to complete the tutorial that is
included with the software. This can be accessed by the Help >> Tutorials >> 1
CatchmentSIM Introduction menu option. The tutorial will basically take the user through the
following steps using a set of sample data.
3.3
16
3.3.1
Project Projection
The project projection is selected from a drop down box on the New Project form (refer
page 56). No single projection can be recommended for all applications and the
appropriate projection should be based on many factors including size of the catchment,
geographic alignment of the catchment and objectives of the analysis. Conformal projections
conserve angles while Equal-Area projections conserve areas. For catchment studies, EqualArea projections may be more appropriate. CatchmentSIM includes the Albers Equal Area
Conic projection which can be used to conserve area measurements.
You do not need to work in the same projection as your data. CatchmentSIM supports reprojection of both vector and raster data sources upon importing the data provided the
transformation is supported by CatchmentSIM (You will be advised of an unsupported
projection if you select it). For large projects, you can also work in latitude and longitude.
However, this will disable the length and area calculations since your raster grid will not have
a constant pixel size. To overcome this you can export your subsequent subcatchment
boundaries after processing and analyse them in other GIS software.
3.3.2
Project Boundaries
The project boundaries should be large enough to accommodate the catchment of interest
but should not be excessively large and include a large amount of redundant topography.
The project boundaries are used to trim all data that is imported into the project. This allows
the use of large GIS data sets as input data for a project. For example, after a user sets
appropriate project boundaries which are large enough to contain the catchment under
analysis, they may then assign a very large GIS database as the project source data.
CatchmentSIM will only import and store the digital terrain information that is within the
project boundaries. Contour or stream lines that cross project boundaries are clipped at their
point of intersection. Consequently, the same source database may be used for many
projects without manipulating the data with the parent GIS application. In fact, it is not
necessary to own any commercial GIS software, merely, to have access to the appropriate
database in an accepted format. CatchmentSIM provides all the tools required for basic data
manipulation including addition / deletion of contours and watercourse data.
The File >> New Project menu allows a project to be setup as outlined on page 55.
3.4
3.4.1
Importing contour data requires GIS files containing line or polyline contours with non-visual
data attached that contains elevation information (Vector GIS Data >> Import 3D Contour
17
3.4.2
If the source data for the analysis is in the form of an external DEM, this information can
either be imported (Digital Elevation Model >> Import DEM page 82) or sampled
(reading point elevation from the DEM to populate the elevation of blank pixels in a userdefined DEM). Sampling a DEM allows you to reproject the DEM and change the resolution
whereas importing a DEM preserves the pixel dimensions and projection exactly.
Sampling of external DEMs can either be undertaken by Digital Elevation Model >> Assign
DEM Elevations by Sampling External DEM (page 83) or Digital Elevation Model >>
Download DEM Data (page 84) for an online source.
More information regarding reprojecting DEMs can be found in Section 5.2 (page 127).
ASCIIGRID DEMs
The ASCIIGRID DEM format is a widely accepted raster DEM interchange format. The basic
file format of an ASCII DEM is
ncols
2361
nrows
1824
xllcorner
143.52625873129
yllcorner
-43.808990500013
cellsize
0.0025
NODATA_value -9999
-9999 -9999 -9999 -9999 -9999 -9999 -9999 -9999 -9999 -9999
nrows and ncols are the number of rows and columns, xllcorner and yllcorner are the
minimum x and y corner coordinates, cellsize is the pixels size and NODATA_value is the
code for pixels with no associated elevation. The data is stored by row starting from the most
northward row and progressing west to east. Each row in the file must contain an entire row
of data.
18
3.4.3
CatchmentSIM also allows for a wide range of other GIS data sets to be imported and used
in an analysis. These include impervious area polygons (Vector GIS Data >> Impervious
Areas page 75), Hydraulic Controls (Runoff Routing >> Hydraulic Controls page 93) as
well as displaying external vector or raster (image) GIS files.
3.5
3.5.1
The rectangle defining the external boundary of the DEM can be set by the user to be any
size equal to or smaller than the selected project boundaries, and may have any number or
rows and columns. The software will generate a warning if it determines that the DEM
characteristics are outside of the users computational resources. CatchmentSIM does not
require DEMs to have square pixels and can accommodate rectangular grid pixels. However,
an option is available to ensure that the DEM has square pixels. If this option is selected then
the number of DEM columns will be automatically calculated after the user enters the number
of DEM rows to ensure a square pixel DEM is developed. The user can also elect to form a
DEM based on an explicit pixel size and CatchmentSIM will determine the required number
of rows and columns.
Note:
In order to maximise the precision of the analysis within the available computational
resources, the DEM boundaries should be chosen carefully to ensure the smallest possible
DEM that contains the entire catchment is created. The total number of pixels should be
maintained in a reasonable range. For example, a project with 1 million pixels (eg., 1000
19
3.5.2
DEM Interpolation
The minimum data requirement for interpolation of a DEM is 3D contour data. The accuracy
and fitness for the purpose of such a DEM will be primarily a function of the level of contour
definition of the imported data and required spatial resolution of the model. However, the
DEM quality may be greatly improved by using a vector watercourse layer in conjunction with
the interpolation algorithm. These data layers are utilised by the DEM interpolation algorithm
in a number of sequential steps, namely:
20
Vector to raster conversion of 3D contour lines (Digital Elevation Model >> Develop
DEM >> Interpolate Contour Data page 80);
Interpolation of raster DEM (Digital Elevation Model >> Develop DEM >>
Interpolate DEM page 81); and,
Stream burning (optional) (Digital Elevation Model >> Develop DEM >> Stream
Burning - page 81).
It can be seen that assignment of all underlying pixels has converted a zero-width line into a
two pixel wide terrain segment of constant elevation. A cross-section generated perpendicular
to this contour line would yield a flat section at each imported contour line. To overcome this
problem it is generally accepted that only selected pixels underlying the line segment should
have the line attribute applied. CatchmentSIM uses a well accepted and documented
decision structure to determine which pixels will form the raster representation of the line,
which is governed by the following rule.
Should the vector component exit a pixel and traverse two of the pixel's neighbouring eight
pixels then only the pixel containing the longest portion of the line will be applied the vector
attribute.
The example illustrated in Figure 3-1 has been reproduced in Figure 3-2 but with the
improved vector to raster conversion methodology. It can be seen that the new algorithm is a
better solution than that portrayed in Figure 3-1. This is especially relevant in areas of closely
spaced contours.
All contour lines that have been imported into the project should be rasterised into the DEM
prior to incorporation of watercourse data layers.
21
The process is complicated due to the fractal network nature of watercourse alignments and
the tendency of this data to be provided as thousands of partial watercourse segments that
exist between junctions of two or more segments. As stated previously, these line segments
have no 3D attributes and often the individual segments are not large enough to intersect the
necessary two contour lines to allow linear interpolation. CatchmentSIM processes the
watercourse network to create single polylines that travel from each upstream tributary down
to the sink associated with that tributary. This algorithm can process any stream network (with
no limit on the number of segments in a junction) provided connecting stream segments end /
start at the same coordinate pair or within a user designated distance tolerance (set in the
Project Options form page 59). ArcGIS also has the capability to form directional drainage
networks from watercourse GIS layers but it requires the user to identify the sinks associated
with all tributaries. This is not a problem if the network is fully connected and thus only has
one sink. However, in many catchments, particularly in Australia, streams are often
discontinuous with ephemeral streams and channel termination common. This can cause
manual identification of all sinks within a catchment to be a tedious process. Consequently,
the CatchmentSIM algorithm was designed to be fully automated and automatically identifies
all sinks. The algorithm only requires each tributary line segment to cross at least two nonequal contours in order to assign drainage direction.
After incorporating the watercourse integration algorithm the DEM pixels underlying stream
segments where watercourse integration was possible will be assigned an interpolated
elevation as shown in Figure 3-4.
22
The net effect of the watercourse integration algorithm is a more realistic and hydrologically
suitable DEM that preserves a known watercourse network. In most cases, after the
application of the watercourse integration algorithm, interpolation of the DEM and
hydrologic conditioning, calculated flow paths will follow those of the imported GIS
watercourse layer as shown in Figure 3-5.
23
Interpolation Aids
The degree to which the DEM is closely representative of the real terrain is a function of the
quantity and quality of the source data from which the DEM was interpolated. Limited source
data may impact on the interpolation algorithm's ability to represent hydrologically important
topographic features such as watershed divides and convergent flow paths. The best solution
for any problematic regions is to import additional contour and watercourse data, however,
this is often not available or economical. For this reason, CatchmentSIM incorporates a
range of tools that can be applied in areas of concern to augment and improve the
automated interpolation process.
24
By importing a good coverage of 3D contour lines and 2D watercourse lines, and placement
of some strategic interpolation aids, users are able to quickly create the basis for
interpolation of a hydrologically suited DEM.
25
Once an interpolation ray and its corresponding mirror ray both intersect pixels with assigned
elevations, linear interpolation is applied to determine the pixel elevation for that particular
interpolation and mirror ray combination. The final value for the pixel is based on a weighted
average of all the cross-section interpolations. The individual cross-section weights are based
on the inverse of the distance between the located assigned pixels for the cross-section.
During the algorithm development process it was found that flat cross-sections should be
discounted (smaller weights) since they were over-flattening the topography in certain
situations. An example of this can be seen in Figure 3-10.
As illustrated in Figure 3-10, the non-flat cross-sections shown in red have the longest
lengths and would consequently be given the smallest weights. However, due to the apparent
ridge line evidenced by the contours, these cross-sections are more representative of the
26
WN =
DN * FB
# rays W
Elevation = n =1 n * Z n
WT
CatchmentSIM allows the user to designate the number of interpolation rays (and mirror rays)
that are used to interpolate the pixel elevation. Increasing the number of rays will increase
the accuracy of the interpolated surface as well as the computational demands of the
algorithm. Figure 3-11 illustrates the advantages of increasing the resolution of the DEM
interpolation algorithm.
It can be seen in Figure 3-11 that the 32 ray interpolation algorithm has four non-flat crosssections which will realistically represent the expected ridge line in the area. Ultimately, the
best solution is reached when a cross-section alignment is found that is normal to the
intersected contour lines, thus is approximately equal to local aspect. The 8 ray algorithm
27
Stream Burning
The watercourse interpolation algorithm outlined in Section 3.5.2 (page 20) does not force
flow paths to follow the imported watercourse alignments, rather it simply ensures that pixels
along the watercourse alignments are linearly interpolated between intersected contour lines
prior to interpolation of any surrounding pixels. In most cases, this will ensure the
watercourse alignments are preserved in the drainage network. However, in some areas of
low relief or complex stream paths, it may be found that calculated flow paths depart from
imported watercourse alignments. If this presents a problem then stream burning (Digital
Elevation Model >> Develop DEM >> Stream Burning page 81) can be implemented.
This algorithm will artificially lower pixels that underlie watercourse alignments by a set
increment (set in File >> Project Options | General) to ensure they are represented as flow
paths in the calculated stream network. However, it may slightly bias slope calculations that
are generated later in the analysis.
Figure 3-12 illustrates the effect of application of the stream burning algorithm. It can be
seen that the pre-stream burning flow path (green line) deviates slightly from the imported
watercourse's path (blue line), whereas application of the stream burning algorithm has
'snapped' the flow path to the observed watercourse.
Prior to using the stream burning algorithm it is important to determine if it is necessary, as its
use can slightly bias calculated values for some of the topographic indices that are generated
at a later stage, such as average vectored slope. Furthermore, in Figure 3-12 it could be
argued that the pre-stream burning flow path is a better interpretation of the source contours
(ie., steeper descent) compared to the imported watercourse alignment.
28
3.6
3.6.1
Filling Algorithm
CatchmentSIM's filling algorithm (Digital Elevation Model >> Remove Flat and Pits Using
Filling Algorithm page 86) works by first raising all pit pixels to the elevation of their lowest
neighbouring pixel and then raising the elevation of flat pixels by a set increment in order to
be able to derive a downslope flow direction.
This algorithm is specifically designed to treat drainage anomalies resulting from the
flattening of hill crests within the DEM where contour definition has not been provided at the
crest of a hill. This occurs because all rays of the interpolation algorithm will find the same
contour value, as illustrated in Figure 3-13.
29
In these situations the iterative process implemented by the filling algorithm ensures that pixel
elevations in large flat areas are raised from the outside in, creating a rounded hill crest that
realistically distributes flow down all sides, with the highest elevation pixel located at the hill
crest centroid.
CatchmentSIM's filling algorithm is good at treating drainage anomalies formed in DEMs
interpolated internally or by other ray based approaches. However, for imported remotelysampled DEMs or stubborn flat or pit pixel arrangements that are unable to be resolved by
the filling algorithm, an advanced Priority First Search (PFS) weighted graphed based
breaching algorithm has been included in CatchmentSIM.
3.6.2
The Priority First Search (PFS) algorithm implemented within CatchmentSIM is a breaching
algorithm designed to solve complex arrangements of flat and pit pixels in a DEM. The
algorithm can resolve any flat or pit pixel within a DEM provided a pixel with a lower
elevation exists somewhere within the DEM (subject to minimum downslope gradient criteria).
For each flat or pit pixel, the PFS algorithm searches for a nearby pixel with lower elevation
(outlet pixel) and an optimum drainage path between the two pixels. After finding the outlet
pixel and optimum drainage path, the PFS algorithm will lower the elevation of all pixels
along the optimum drainage path to create a downslope drainage path of consistent
gradient between the original flat or pit pixel and the outlet pixel.
The method used to implement this technique is based on well-documented weighted graph
methodology (Sedgewick, Robert 1988) and has shown promise in hydrologic applications
(Jones, Richard 2002). The algorithm records DEM pixels or 'nodes' in two sets, the priority
tree and the priority queue. As a result, all pixels in the DEM are in one of three states, on the
priority tree, on the priority queue or as yet unseen by the algorithm. Initially, all nondiagonal pixels adjacent to the target flat or pit pixel are added to the priority queue. In turn,
the nodes in the priority queue are examined with reference to a priority function and the
most suitable node in the priority queue is added to the priority tree and removed from the
priority queue. Adjacent pixels to the new node (recently added to the priority tree) are then
added to the priority queue which now consists of the remaining nodes from the previous
iteration and these new nodes. The algorithm continues until a terminating condition is met,
which is triggered when a node in the priority queue satisfies the terminating criteria. In
CatchmentSIM, the terminating criteria requires the node to have a lower elevation than the
starting node and for the resultant downslope gradient between the two points (along the
optimum drainage path) to exceed a user designated minimum gradient threshold.
30
31
a)
b)
c)
d)
e)
f)
g)
33
One of the important capabilities of the PFS algorithm is that once a node is entered in the
priority queue it is not removed until the algorithm is finished or it is transferred to the priority
tree. This means that the priority tree can grow in any direction and will always find the
optimum drainage path.
The PFS algorithm can be either applied to an individual flat or pit pixel (Digital Elevation
Model >> Apply PFS Flat and Pit Removal Algorithm >> Individual Pixel page 87) by the
user or to all flat or pit pixels remaining in the DEM (Digital Elevation Model >> Apply PFS
Flat and Pit Removal Algorithm >> Entire DEM page 87). If the later is applied then the
flat and pit pixels are processed by the PFS algorithm in order of increasing elevation. This
improves the drainage network in flatter terrain and reduces the necessity for multiple
applications of the PFS algorithm.
There are a number of options that a user can set in CatchmentSIM to dictate the properties
of the PFS Algorithm, these are :
Minimum Gradient The minimum gradient that must be found to exist along the optimum
drainage path for the algorithm to accept the outlet pixel. This parameter is designed to
ensure that significant drainage paths are identified and that gradients are not so low as to
produce flat pixels when DEM pixel elevations are rounded to the precision of the DEM
(single or double, see page 19).
No-data Behaviour This parameter dictates how the algorithm will behave if it encounters
pixels which are on the boundaries of the DEM or have not yet been assigned an elevation
value. The algorithm can either terminate leaving the original flat or pit pixel with its initial
elevation, or continue to search ignoring the no-data or boundary pixel.
PFS Break Size This parameter is used to monitor the size of the priority queue and the
algorithm will terminate if the priority queue reaches this size. This is particularly important if
the no-data behaviour parameter is set to ignore no-data or boundary pixels. In these cases,
the algorithm may search the entire DEM before realising that no pixel meets the terminating
criteria and moving onto the next flat or pit pixel. This can slow the algorithm down to an
impractical extent. To ovoid this slow-down and enable the algorithm to terminate
prematurely, the PFS break size parameter may be used. This parameter should be set large
enough to ensure genuine solutions paths are found but small enough to restrict unwanted
algorithm slow-down.
The PFS algorithm has several important advantages over other common methods. Firstly, it
is robust and will always find a solution provided a pixel satisfying the terminating conditions
exists. Secondly, it does not distinguish between flat and pit pixels resulting in a consistent
approach to both types of drainage anomalies. Thirdly, it tends to create channel networks
and flow distributions that are more representative of reality than competing algorithms such
as the J&D algorithm utilised by ArcGIS and many other software, which tends to create
many parallel flow paths (Martz and Garbrecht 1998; Tribe 1992).
3.7
3.8
Flow Routing
Following the interpolation and hydrologic conditioning of the DEM, a flow routing algorithm
(Runoff Routing >>DEM Wide Flow Processing page 91) can be applied to delineate
subcatchment boundaries, determine the subcatchment network relationship and calculate
geophysical subcatchment properties.
3.8.1
Theory
The flow routing algorithm forms the basis of CatchmentSIM and is of vital importance to the
quality of any GIS based hydrologic investigation. Most terrain analysis packages adopt the
D8 method for flow routing. However, CatchmentSIM utilises a modified version of Leas
(1992) method which has a number of advantages over the D8 method.
The modified version of Leas (1992) algorithm calculates a downslope flow angle for each
pixel that can be anywhere in the range of 0-360 degrees. The flow direction angle is
determined as the resultant flow vector from the combination of the steepest non-diagonal
pixel flow vector and the next steepest adjacent non-diagonal vector as outlined in Figure
3-15.
Flow from each pixel is then routed through all downslope pixels until a subcatchment outlet
(or DEM boundary) is reached. The algorithm treats the flow path as a line and records the
entry and exit points of the flow path through all pixels as a vector quantity such as illustrated
in Figure 3-16.
35
As shown above, the described algorithm considers flow as a vector quantity flowing through
a raster DEM. This technique has distinct advantages over more common approaches which
consider flow as a raster quantity. In particular, it allows a greater representation of flow
direction over hill-slopes and has greater sensitivity to flow divergence or convergence. For
example, Figure 3-17 depicts parts of the CatchmentSIM flow paths mapped for 9
neighbouring pixels.
The distribution of flow paths can be seen in the lower right pixel in Figure 3-17 where flow
paths from upstream pixels are distributed between both of this pixel's downslope pixels,
based on where the flow paths entered the pixel. This cannot be accommodated in raster
based flow routing techniques and allows for a more accurate representation of flow
distribution, and calculated drainage-path length / slope statistics. CatchmentSIMs flow
routing algorithm is analysed in more detail the following section.
3.8.2
36
This type of error can also be observed in real catchments. For example, the flow paths
generated by the D8 algorithm and the CatchmentSIM algorithm for 6 sample points in a
catchment are compared in Figure 3-19, superimposed over the contours used to interpolate
the DEM.
It can be seen in Figure 3-19 that significant deviations in the calculated flow paths exist for
several of the sample points. The tendency of the D8 method to snap to cardinal or diagonal
directions due to its limitation of eight potential directions can be clearly seen in the lower left
sample points. In these cases the D8 flow paths are snapping to 45 degree lines since this is
the closest approximation to local slope that the D8 algorithm can generate. The
CatchmentSIM flow paths originating from these sample points can be seen to flow more
perpendicularly to the contours. Consequently, they are more hydrologically accurate.
The effect of the errors associated with the D8 method can be seen to follow through into
subcatchment delineation. The subcatchment boundary delineated from the same outlet
point using the D8 and CatchmentSIM flow routing algorithm can be seen in Figure 3-20.
37
The tendency of the D8 method to snap to diagonal and cardinal directions can again be
seen in Figure 3-20, where the D8 generated boundary is biased towards the 45 angle and
does not correctly identify the ridge line between the stream confluences. As shown in Figure
3-20, the two algorithms converge when they reach more defined ridge lines with stronger
contour curvature. As such, the error introduced by the D8 method will be more pronounced
in the outlet areas of subcatchments. The significant problem with quantifying this error is that
it will be a function of the size of the subcatchment being delineated. This is because the
length of the subcatchment boundary segment that does not follow major ridge lines will
become a larger proportion of the total subcatchment boundary as subcatchment area
decreases. As a result, the proportional error associated with the D8 method will become
more pronounced the higher the discretisation of the catchment. This is an undesirable
attribute of the method because higher catchment discretisation is usually undertaken to
facilitate more accurate hydrologic modelling. Thus, if the D8 method is being applied then
a user may inadvertently be introducing greater errors while attempting to gain greater
accuracy.
3.9
3.9.1
The first phase of generating a stream network is identifying channel heads, the points at
which accumulated runoff can be considered a stream. Channel heads can be identified in
CatchmentSIM utilising a number of options. These include Stream Area Threshold (SAT) or
combination of SAT with Minimum Source Channel Length (MSCL). CatchmentSIM also
allows for quantitative assessment of the minimum SAT value that can be adopted whilst
preserving the geomorphologic properties of dendritic stream networks (refer page 115).
38
Mapping of vector streams results in a set of connected vector polylines with Horton /
Strahler orders calculated for each line segment. This can be displayed in CatchmentSIM with
differing colours and line styles for each stream order and analysed with a variety of charts
and derived hydrologic parameters (refer page 68).
3.10
39
Direct input of outlet pixel(s) row and column numbers into a dialog box;
3.10.1
When using the flow accumulation jump algorithm (Subcatchments >> Breakup
Subcatchment page 101), the user simply designates a target number of subcatchments
and the algorithm will automatically break up the catchment into the correct number of
subcatchments. The algorithm works based on finding the largest jumps in the flow
accumulation grid values between pixels in streams and their downstream neighbour which
indicates lateral inflow of a significant tributary. Figure 3-22 illustrates the subcatchment
break-up achieved by the automated subcatchment break-up algorithm using 3 different
target subcatchment values.
40
3.10.2
The Horton / Strahler based catchment break-up algorithm (Subcatchments >> Breakup
Subcatchment page 101) requires the user to first generate a vector stream network by
selection of an appropriate SAT value. This can be based on quantitative analysis of SAT
derived networks as outlined in Section 3.9.1 (page 38) or by generation of a stream network
that closely matches an observed stream layer. At this point, a user can select what order
subcatchments to delineate. For example, if a user selects 3rd order subcatchments then
subcatchment outlets will be placed upstream and downstream of all intersections in the
stream network where two or more of the tributaries have an order greater than or equal to
3. As a result, the total number of subcatchments as well as the precise location of
subcatchment outlets will be a function of the hydrologic properties of the vector stream
network. An example of the subcatchment break-up achieved on a sample catchment when
delineating subcatchments of differing Horton / Strahler orders in shown in Figure 3-23.
42
43
3.11
The labelling methodology is designed to identify the significance of the upstream tributary to
the left of the decimal point, and the subcatchment position in the tributary network to the
right of the decimal place. A number of options are available to customise the labelling
approach (File >> Project Options | Subcatchment Labelling page 59). These include
determining the tributary significance based on decreasing maximum stream length rather
than in downstream order of lateral influx and using alphabetic letters instead of integers to
the right of the decimal point.
44
3.12
Urbanisation Tools
CatchmentSIM includes a range of tools designed to accommodate representation of urban
areas in a project. Firstly, CatchmentSIM allows calculation of impervious area proportions
for subcatchments (Vector GIS Data >> Impervious Areas page 75) which is a parameter
required for almost all hydrologic models. Furthermore, CatchmentSIM offers a more
comprehensive method of urban analysis. Realistic modelling of runoff in urban areas
requires consideration of the numerous processes that may be in play during a rainfall event
in an urbanised catchment. These processes may include roof, downpipe, fencing, roads,
footpaths, gutters and piped drainage systems (Goyen and OLoughlin 1999).
CatchmentSIM accommodates individual representation of several of these processes by
modelling individual flow path deviations occurring as a result of urban hydraulic controls
(Runoff Routing >> Hydraulic Controls page 93) such as roads, gutters and channels
networks.
3.12.1
45
3.12.2
CatchmentSIM includes a number of tools to help model flow paths in urban environments.
Urban structures have a significant effect on flow paths in urban areas and they are usually
not represented in source GIS data such as DEMs or contour and stream alignments. As
such, they need to be added into a CatchmentSIM project as an addition to the source GIS
data. This can be achieved by one of two approaches. Firstly, the urban structures can be
hard-coded into the DEM by changing the elevations of relevant DEM pixels to cause flow
paths to act in a realistic manner in the vicinity of urban structures. Alternatively, urban
structures can be modelled in CatchmentSIM as supplementary objects that control flow
paths when they intersect the alignment of an urban structure. Modelling urban structures in
this way does not require the DEM pixel elevations to be altered. Furthermore, each
individual urban control can be turned on or off, and flow paths and subcatchment layouts
may be regenerated easily. This is valuable when analysing drainage studies for hydrologic
events of differing magnitudes where particular urban structures may only be relevant for
certain storm magnitudes, or during flood mitigation scenario analysis.
CatchmentSIM accommodates both of these modelling approaches. An example of hardcoding urban structures into the DEM is presented in Figure 3-27, where road crown
alignments were hard-coded into DEM by raising all pixels along the road crowns by 0.5
metres using CatchmentSIM's vector data set operations (Digital Elevation Model >> Vector
Dataset Operations page 90). The PFS algorithm was then applied to remove resultant flat
and pit pixels and breach the road crowns at their points of lowest elevation.
46
As shown in Figure 3-27, the calculated stream network has been strongly affected by the
hard-coding of urban structures. For more detail on how hard-coding of urban structures can
aid hydrologic modelling, see the Holland Park Local Stormwater Management Plan case
study shown in Section 6.1 (page 143).
Alternatively, urban structures can be modelled separately using CatchmentSIM's hydraulic
control tools. These tools allow representation of channels and gutters in a CatchmentSIM
project that act as overriding flow controls. These hydraulic controls are described in the
following sections.
3.12.3
Channel type hydraulic controls (Runoff Routing >> Hydraulic Controls page 93) are
drawn in CatchmentSIM as a solid line with triangles pointing in the direction of the line
towards the channel outlet
or
. These controls have the effect of forcing
Hydraulic Controls Form
flow paths that intersect with these controls to follow the channel until its outlet point
regardless of whether this involves upstream flow, or flow in a direction that does not
represent the steepest downslope direction at that point in the DEM. These hydraulic controls
should be used when a user wishes to completely override the DEM and force flow to follow
a pre-determined path. Channel type hydraulic controls should be used to represent
drainage channels or urban flow paths that are not represented in the DEM or source
contour data. A stream network can also be imported as channel type hydraulic controls to
force flow to follow an existing stream network as shown in Figure 3-28.
47
3.12.4
Gutter type hydraulic controls (Runoff Routing >> Hydraulic Controls page 93) are drawn
in CatchmentSIM as a solid line with triangles pointing at a perpendicular angle to the
direction of the line
Controls form
or
. The triangles point in the direction in which flow is
allowed to pass over the hydraulic control.
When a flow path intersects a gutter type hydraulic control, flow will still follow the DEM
calculated steepest descent within the restrictions imposed on the flow direction by the gutter.
If the flow direction would see the flow path crossing the gutter (against the direction of the
arrows) then it is restricted from doing so. Instead, flow is allowed to travel along the gutter
provided this direction represents a downslope gradient. If neither of the along-gutter
directions are downslope then the flow path is trapped. The gutter processing algorithm will
then search along the gutter in both directions within a specified tolerance for a pixel of
lower elevation. If such a pixel is found within the tolerance then the flow path will be
mapped to this point, and the algorithm is re-applied at the new location. The tolerance may
be in the form of a set number of pixels or a set elevation. The effect of this algorithm is to
simulate ponding at low points behind gutters which would in reality fill the pixel elevation
and allow flow to progress to pixels of higher elevation provided they are lower than the
height of the gutter. The gutter height is simulated in CatchmentSIM by utilising an elevation
tolerance. If a pixel of lower elevation is not found within the specified tolerance then a
hydraulic breach is formed and the flow path is permitted to breach the gutter. This can be
seen in Figure 3-29 where flow paths travel along the gutter until the searching algorithm
fails to find a suitable downslope pixel along the structure and a hydraulic breach occurs.
48
3.13
Cross-section charting;
More information and an in-depth description of how these charts can be used to examine
subcatchment hydrologic characteristics is given on page 106.
CatchmentSIM also has the ability to generate animations of parameter variation over a
catchment (Analysis >> Create Animation page 117).
49
3.14
RORB
URBS
DRAINS
CSTalk scripts can also be written by the user to allow integration with in-house software
applications or 3rd party software applications that do not yet have available CSTalk scripts. A
number of users have developed their own CSTalk scripts and some of these are also
available on the website. These include:
A modified version of the RAFTS-XP scripts developed by Patterson Britton & Partners
in Sydney, Australia. This script was developed to accommodate an alternative type of
subcatchment routing of impervious areas.
The coupling methodology provided by the CSTalk language is illustrated in Figure 3-30. A
manual and reference guide for writing CSTalk scripts is also available.
50
Development of GIS Techniques for Automated Topographic and Hydrologic AnalysisCase Studies and Applications
Since CSTalk scripts can produce files in any format, they are also commonly used to
develop standardised report formats for professional organisations. For example, Patterson
Britton and Partners developed a CSTalk script that automatically generates a Rich Text File
(*.rtf) with the standardised report template shown in Figure 3-31. The data columns that
cannot be read directly from a CatchmentSIM project, such as rainfall loss coefficients are
obtain through a series of script-generated dialog boxes.
3.15
Summary
As outlined in the preceding sections, CatchmentSIM incorporates a range of algorithms for
digital terrain analysis and hydrologic analysis. These include algorithms for the interpolation
of a raster DEM from contour and watercourse alignment data. These include advances on
current implementations of profile based DEM interpolation algorithms by utilising variable
53
4 Program Reference
This Chapter documents every menu item and dialog box within CatchmentSIM. They are
described in the order that they appear in the menus.
4.1
54
File Menu
Program Reference
New Project (Ctrl+N)
The new project form is used to setup a new CatchmentSIM project. This form requires the
project projection to be set as well as the design plane for the project. The project must also
be assigned a filename and location. Optional text information can also be added at this
stage.
Project information
(optional)
Set project
Filename
Projection
Properties
Project Information
The project information is optional and these fields may be filled to provide information
regarding the project, when it was created and by whom. This information can later be
accessed by CST macro scripts to transfer the information into downstream models.
55
Like most GIS systems, CatchmentSIM supports projections of the earth into Cartesian based
coordinates. CatchmentSIM projects can be undertaken in any projection and custom
projections can be created to suit a particular project minimising distortion for the project.
While CatchmentSIM can be used in any projections, it only supports conversion of data
from one projection to another for a limited number of transformation types, these include:
Mercator;
Transverse Mercator;
Polyconic
56
Program Reference
Note:
If you wish to create a custom projection for the project, select the Custom Projection button.
This will bring up the Custom Projection form.
The custom projection form allows the selection of a wide range of ellipsoids, projection type
and designation of the required parameters for the selected projection. A custom ellipsoid
can also be defined.
Note:
Design Plane
The design plane needs to be assigned in the New Project form. The design plane is the
easting and northing extents of the project. The design plane can be as large as desired and
may be larger (but not smaller) than the intended DEM. All input data will be clipped to the
extents of the design plane minimising file sizes and processing times. The design plane
extents must be in the same units as the projection before the OK button is pressed.
However, they can be converted from other coordinates using the Convert Projection button.
Simply enter the coordinates of the design plane extents in any coordinates, then click
57
Scaling Factors
Scaling factor allow for the use of data that is not in the units of metres. If data is in feet or
other units then enter appropriate scaling factors for the x and y dimensions.
Note:
58
Program Reference
Project Options
The Project Options form is composed of 6 tabs containing a range of parameters required
for the CatchmentSIM algorithms. Each tab will be described in succession.
The General tab of the Project Options form includes a number of parameters including:
Stream Burning Depth: This decimal parameter governs how deep stream will be
burnt when the stream burning function is implemented (Digital Elevation Model >>
Develop DEM >> Stream Burning page 81). For more information see
Section 3.5.2, page 28.
Time Out Error Threshold: When using hydraulic controls such as gutters and
channels, circular flow paths can be created which will crash CatchmentSIM if flow
processing is initiated. To identify any problem areas, the Runoff Routing >> Draw
Pixel Flow Path (page 91) tool can be used. This algorithm will inform the user of any
invalid flow circles in the downslope flow path. This parameter is multiplied by
59
The Flat and Pit Removal tab of the Project Options form includes parameters for
CatchmentSIMs two methods of flat and pit removal, namely the Filling Algorithm and PFS
algorithm. These parameter include:
60
Pixel Filling Increment: This parameter indicates the elevation added to flat pixels in
each iteration of the filling algorithm. For more information see Section 3.6.1
(page 29) or Digital Elevation Model >> Remove Flat and Pits Using Filling
Algorithm (page 86).
PFS Minimum Gradient: This parameter governs the minimum gradient that must be
observed over the PFS solution channel before the algorithm will determine that a
solution has been found. This parameter can be adjusted to ensure legitimate PFS
solutions are found and that rounding errors do not cause PFS solution channels to
become flat. For example, if the PFS minimum gradient is too low then a very shallow
slope may be calculated over the PFS channel. After linear interpolation is applied to
Program Reference
determine pixel elevation over the channel, these elevation will be rounded to the
precision of the DEM (single or double). This can cause these pixels to become flat.
More information regarding the PFS algorithm and adjustment of its parameters is
shown in Section 3.6.2 (page 30) and Section 5.3.1 (page 129).
Treatment of Boundaries and Non-Defined Pixels: This parameter instructs the PFS
algorithm what to do if a potential solution channel encounters a DEM boundary or
blank pixel. Either the algorithm will terminate (the pit or flat will remain) or the
algorithm will continue to search ignoring the boundary or blank pixel. The default
option is to terminate the algorithm as ignoring the pixel can slow down the algorithm
considerable and may also find solution channels that are not realistic (ie., in reality
the flow path from the flat or pit pixel may actually flow across the DEM boundary). In
practise, the flat and pit pixel near the boundaries should not be in the catchment of
interest anyway. More information regarding the PFS algorithm and adjustment of its
parameters is shown in Section 3.6.2 (page 30) and Section 5.3.1 (page 129).
PFS Break Size: This parameter has been introduced for speed reasons, once the PFS
algorithm has grown very large looking for a potential solution channel over a wide
area it can slow down considerable. Once the PFS algorithm has grown to the PFS
Break Size the algorithm will terminate. If time is not a constraint this parameter can
be adjusted upward to ensure the PFS algorithm finds even the most complex
solutions.
The Stream Networks tab of the Project Options form contains the following parameters:
Stream Area Threshold: This parameter determines the number of pixels that
must drain to a single pixel before that pixel is considered a channel head
during generation of raster streams (Runoff Routing >> Draw Raster Stream
61
The Subcatchment labelling tab of the Project Options form governs how CatchmentSIM
labels subcatchment in the Nodal Link Arrangement and consists of the following
parameters:
The basic CatchmentSIM labelling system starts by identifying the main stream by flow length
and will number subcatchments along this stream as 1.01, 1.02 etc downwards until the
catchment outlet. Subcatchment 2.01 will be at the head of either the next downstream
lateral inflow tributary or the head of the second longest stream depending on the option
selected in the Downstream Subcatchment Names option.
62
Link Drawing: This options wether links will be draw between subcatchment centroids
and their outlet and then to the downstream subcatchment centroid or simply between
subcatchment centroids and the next downstream subcatchment centroid.
Program Reference
The Hydraulic Structures tab of the Project Options form includes parameters designed to
control the operation of Hydraulic Controls within CatchmentSIM, specifically:
HC Breach Sensitivity: This option governs how hydraulic breaches will be determined.
They can either be derived by a distance measure (pixel number threshold) or an
elevation tolerance. In the case of a distance measure, if a pixel of lower elevation
can not be found along the gutter within the pixel tolerance then a hydraulic breach is
formed and the flow path overtops the gutter. In the case of an elevation tolerance, if
a pixel of lower elevation can not be found along the gutter without gaining more
than the elevation tolerance (over the flow path along the gutter to get to the lower
elevation pixel) then a hydraulic breach is formed and the flow path will overtop the
gutter. This is designed to simulate ponding behind a gutter and eventual
overtopping. More information of Hydraulic Controls and breaches can be found in
Section 3.12.2 (page 46) and Section 5.6 (page 139).
63
The Digital Elevation Model tab of the Project Options form contains parameters related to
the DEM created by CatchmentSIM, specifically:
64
DEM Precision: The DEM precision parameter governs wether the DEM will be stored
as single or double precision floating point values. Single precision stores 7-8 digits
while double precision stores 15-16 digits. Selecting double precision will increase the
size of the DEM file size by 100% and total project files by around 20% and runtimes
by 20% as well as RAM requirements. Double precision may be required when using
elevation data with several digital on the left side of the decimal point such as
10450.056 ft or very flat DEMs or if you wish to use a low PFS minimum gradient
parameter.
Vector to Raster Conversion: This parameter governs wether all pixel underlying an
imported contour will be assigned the contour elevation or only those identified by the
selective vector to raster conversion algorithm. This algorithm is described in
Section 3.5.2 (page 21). It is generally accepted that selective assignment of
underlying pixels creates more realistic DEMs and thus, this is the default option.
Multiple Contours Option: This option governs what elevation will be assigned to
pixels which underlie multiple contours with differing elevations. The options are
average value (default), max, min or a max/min threshold. The max/min threshold
option is designed to ensure that the extremities in the DEM are represented by
picking the maximum contour elevation if the average value is greater than a
threshold value and the minimum contour elevation in the average value is less than
the threshold.
Program Reference
Projection Properties
The Projection Properties form lists the attributes of the projection defined for the
CatchmentSIM project. The project projection is established based on selection of a specified
projection class and subclass in the New Project form or by using the Custom Projection
form.
Save Project
The save project menu item will save the project to the previously selected filename. This will
replace the previously saved version. If you wish to retain the previously saved version then
you should use the Save As option to save to a different file name.
Save Project As
The save project as menu item will prompt the user to save the project to a new filename.
65
The print preview form allows you to visually select between portrait and landscape model
and access the print manager form in order to change printers, paper sizes etc. This form can
be accessed from the File >> Print menu or the Print Chart button on the Chart Wizard
form.
66
Program Reference
4.1.2
View Menu
Vector Layers
Change colour of
selected vector
layer
Add vector layer
(eg., ESRI ShpFile,
MID/MIF)
Project information
(optional)
Update display
order (raster always
on bottom)
DEM DEM
Display Dialog
Add raster image
(eg., air photo)
Remove raster
image (eg., air
photo)
67
This form allows stream segments of each Horton order to be displayed in a different colour,
width and style (see Section 3.9, page 38).
68
Program Reference
There are two interactive methods for geo-referencing a raster image as well as two
numerical standards for storing the geo-referencing data. At this stage, CatchmentSIM does
not support rotation of images, thus, raster images must be aligned with the north/south,
east/west axis of the adopted projection.
If the image boundary coordinates or pixel size is known then the image may be manually
geo-referenced by typing the information in after selecting the appropriate referencing
standard. If the coordinates of the image boundary are known then the 3 Point Image Corner
Coordinates standard can be used. However, if the bottom left coordinates and pixel size (in
map units) are known then the Bottom Left Coordinates and Pixel Size standard can be used.
If coordinate information is not known, the image may be geo-referenced visually. The Drag
and Drop On Screen button will cause the form to disappear and then the user is requested
to select the diagonal corners of the images location by reference to the displayed map
information. A rectangle will be drawn as the cursor is moved (after input of the first point) to
help visualise the image location. Once the second point has been added, the form will
reappear and the relevant coordinate information is added. At this stage the Move Image By
2 Point Dragging option may be used to fine tune the image location. This option allows the
image to be re-scaled by associating two known points on the image with their real locations.
After clicking the button, the form will disappear. First click a point on the image and then
click its proper location, after repeating this twice, the image will be automatically rescaled,
the form will reappear and coordinate information will have been updated.
Note:
The Drag and Drop On Screen option must be used before the Move
Image By 2 Point Dragging option can be used.
Display of DEM
To customise the display of the DEM, select the Digital Elevation
Model layer and select the Properties button.
There are a number of options regarding how the DEM is displayed.
It may be displayed using a variety of colour gradients and the
maximum/minimum display elevations may be adjusted. The DEM
may also be displayed as a Shaded Relief Diagram.
69
Max/min elevation
for DEM Display
Set max/min to
project max/min
Shaded relief
parameters
Zoom In
The zoom in menu option will change the cursor to a magnifying glass and allow you to
zoom into a selected portion of the screen. Click with the left mouse button and that location
will become the centre of the next view.
Zoom Out
The zoom in menu option will change the cursor to a magnifying glass and allow you to
zoom out of the current view. Click with the left mouse button and that location will become
the centre of the next view.
Pan View
The pan view menu option allows the screen to be panned without changing the zoom level.
It requires the user to click and hold the left mouse button. The location of the click will move
to the location of the click release in the new view.
70
Program Reference
Fit View
The fit view menu option will zoom the current view to display the full extent of the design
plane.
Refresh (F1)
The refresh menu option redraws the current view.
71
4.1.3
Import 3D Contours
The import 3D contours menu option allows for the import of a contours GIS files (eg.,
ShapeFile, mid/mif) into the CatchmentSIM contours database. This option should be used if
a DEM is to be interpolated from contours.
Since contour elevation needs to be included, you will be asked to designate the data
column in the GIS file that holds the elevation data relevant to each polyline or line segment.
This data column should be a decimal or integer value.
72
Program Reference
Since stream alignment GIS files do not usually have elevation tags, you will not be
requested to identify a GIS data column for elevation values as for the import contours
option.
73
74
Program Reference
Impervious Areas Database
The Impervious Areas form allows the user to import and manage impervious area polygon
databases from GIS files. These polygons are used to derive impervious area proportion
statistics for the mapped subcatchments.
Turn on / off
individual polygons
Polygon description
Polygon impervious
proportion
Import polygons
from GIS file
Delete currently
selected polygon(s)
Impervious
proportion of areas
not covered by a
polygon.
Multi-select
polygons based on
specific criteria
Edit specific
property of
currently selected
polygon(s)
Filter polygon table
(show all, on, off)
75
If the GIS file does contain appropriate data columns these should be selected in the
corresponding drop down box. The Polygon Description data columns should be a text
columns whereas the impervious proportion data column should be a decimal or integer
column. If the GIS file does not contain such data columns, the user should select None for
each drop down box. In this case, all polygons will be imported with a blank description and
100% impervious proportion. This can later be altered using the tools described bellow.
76
Program Reference
Deleting Impervious Area Polygons
Impervious area polygons can be deleted by individual or group selection and then clicking
the Delete Polygon(s) button. Deleting polygons should only be used when polygons are to
be permanently deleted, otherwise turning them off may be preferred.
77
4.1.4
Square or
rectangular DEM
pixels
Define DEM by
number of
rows/columns or
specific pixel size
78
Program Reference
DEM Extents
The DEM extents are the easting and northing boundaries of the DEM. They can be assigned
to be equal to the design plane (set in the New Project form) or as a new rectangle that is
fully contained within the design plane using the Design Plane / New Extents radio button. If
new extents are to be used then they can be typed in manually or a rectangle can be
dragged over the map by selecting the Drag Over DEM button.
Note:
If the square grid pixels option is selected the New Extents will
automatically be selected as the max easting will need to be need to be
slightly adjusted to maintain square pixels. The reduction in the DEM
extents of the max easting will be shown under the row edit box.
Similarly, if the Set Cell Size option is selected then adjustments will
need to be made to the max easting and max northing. This will be
displayed under the cell size edit box.
DEM Resolution
The DEM resolution refers to the pixel size of the DEM as well as the number of rows and
columns. The resolution can be defined in 3 ways:
Use 2a) Setting the number of rows and columns (rectangular DEM);
Use 2a) Check the square grid pixels option and set the number or rows (square DEM);
or,
Note:
If you setup a DEM after one has already been created then the original
DEM and all derived data layers will be removed.
79
Interpolate Watercourses
If watercourse alignments have been imported (Vector GIS Data >> Import Stream / Flow
lines page 72) then this data is likely to be 2D. Thus in order to rasterise this information
into the DEM linear interpolation is undertaken between intersected contour lines along each
stream alignment. This algorithm will piece together complete stream alignments from a
network of stream segments and process each in turn from the main stream to last tributary.
The watercourse interpolation algorithm is described in more detail in Section 3.5.2
(page 20).
Interpolate ITL
Interpolation Training Lines (ITL) allow ridge lines and valleys to be better defined in the DEM
prior to DEM interpolation. ITL placement tips and algorithm details are given in
Section 3.5.2 (page 25).
80
Program Reference
Interpolate DEM
The DEM interpolation algorithm is a vector ray based approach and is described in Section
3.5.2 (page 26). The algorithm requires the corner DEM pixel elevation to be defined as well
as the resolution of the interpolation technique (number of search rays). CatchmentSIM will
approximate the DEM corner elevation based on nearby contour, watercourse and ITL detail
but these can be over-ridden if required.
DEM corner
elevations
Number of search
rays (8,16,32 or
user defined)
Note:
Increasing the number of search rays will increase the accuracy of the
resulting interpolation but will also increase the algorithm run-time. See
Section 3.5.2 (page 26) for more details.
Note:
The DEM interpolation algorithm will only derive elevations for blank
DEM pixels. Therefore it can be used to fill in missing data in imported
/ sampled DEMs.
Stream Burning
Stream burning will lower the pixel elevation of all pixels underlying an imported stream line
by the increment set in the Project Options form (File >> Project Options |General) as
outlined on page 59. Stream burning (assuming sufficient depth) will force derived flow paths
to follow the imported stream network. For more information, please see Section 3.5.2
(page 28).
81
Import DEM
The Import DEM form should be used if an external DEM is to be used as the base DEM for
CatchmentSIMs analysis. This option does not require a DEM to be setup previously as
apposed to populating a new DEM by sampling an external DEM. After selecting this option
you will be prompted to select the DEM file, be sure to select the appropriate DEM format in
the Filter (ASCII-GRID, SRTM 1/3).
Source DEM
parameters
Note:
82
Boundaries of design
plane vs import region
Program Reference
populate a new DEM that has previously been setup (Digital Elevation
Model >> Setup Digital Elevation Model).
Once a DEM has been setup (Digital Elevation Model >> Setup Digital Elevation Model
page 78) elevation can be assigned by sampling an external raster DEM or Triangular
Irregular Network.
Note:
Only DEM pixels that have no data will be sampled from the external
DEM. This allows for several DEMs to be sampled to complete the
CatchmentSIM DEM. For example, a highly detailed DEM with some
missing data could first be sampled followed by a lower resolution DEM
to fill in the blanks. Alternatively, the blanks can be filled in by using the
interpolate DEM function (Digital Elevation Model >> Develop DEM
>> Interpolate DEM page 81)
Raster DEM
Sampling a raster DEM can be useful because it allows the DEM information to be preserved
while changing the DEM resolution and projection. When this function is selected you may
be asked to provide the projection of the source DEM (unless it is inherent eg., SRTM DEMs
are always in lat/long). If you decline to select the DEM projection when prompted, the DEM
will be assumed to be in the same projection as the project.
Note:
TIN Dem
A Triangular Irregular Network (TIN) may be sampled to assign elevations to a
CatchmentSIM DEM. CatchmentSIM supports ASCII-TIN format. You will be asked to define
the projection of the TIN, otherwise it will be assumed to be in the same projections as the
project.
83
Select SRTM 1 or 3
(arc second resolution)
Select region
(affects URL)
The URL is displayed in case the source files have been moved by NASA. If this has occurred,
the URL can be manually updated.
84
Program Reference
After selecting OK, the appropriate DEMs will be determined and downloaded from the web
server, unzipped and automatically sampled.
Note:
Subdivide DEM
Subdividing the DEM is a unique option to increase the resolution of the DEM in order to
smooth subcatchment boundaries. Subdividing the DEM will increase the number of pixels by
400%. The number DEM pixels will be placed as per the diagram and interpolated from their
surrounding neighbours.
Expand or reduce
DEM boundaries
(see diagram)
Diagram illustrating
new DEM
positioning.
85
Note:
This is not a permanent mapping layers and will be removed when the
screen in refreshed.
86
Start successive
iterations.
Filling
increment
End successive
iterations.
Program Reference
Apply PFS Flat and Pit Removal Algorithm Submenu
The apply PFS Flat and Pit Removal Algorithm submenu contains procedures to apply the PFS
algorithm to the entire DEM or an individual pixel. The PFS algorithm is an advanced leastcost breaching algorithm designed to resolve flat and pit pixels by forming a channel to their
least cost outlet where cost is defined by a hierarchy of least elevation gain and shortest flow
path. The PFS algorithm theory is described in detail in Section 3.6.2 (page 30).
The PFS algorithm will be applied using the parameters defined in the Flat and Pit Removal
pane of the Project Options form (see page 59).
The PFS algorithm can be used in conjunction with the filling algorithm to resolve the flat and
pit pixels within a DEM. This process in described in Section 5.3 (page 128).
Entire DEM
The entire DEM menu option of the PFS algorithm submenu will apply the PFS algorithm to
all flat and pit pixel in the DEM in order from lowest to highest elevation. Pixels are processed
in this order to reduce the length of average solution paths and increase the hydrologic
fitness of the resulting DEM. The solutions paths for those pixels that are successfully treated
by the PFS algorithm are temporally drawn in yellow on the screen. If a pixel is not resolved
by the PFS algorithm then this may be remedied by adjusting the PFS parameters in the
Project Options form (such as., reducing minimum gradient, changing boundary / blank
pixel treatment or increasing DEM precision), see Section 5.3 (page 128) for more
information on treating troublesome DEMs.
Individual Pixel
The individual pixel option of the PFS algorithm submenu will apply the PFS algorithm to an
individual DEM pixel. Simply left mouse click on the pixel to apply the algorithm to a pixel. A
yellow solution path will be drawn if the PFS algorithm was able to find a successful outlet for
the pixel.
3D Visualisation Submenu
The 3D Visualisation submenu includes functions to visualise part or all of the DEM and
generate subcatchment boundaries in 3D.
87
3D Image
Zoom in / out
Directional controls
(pan or rotate)
Axis extents
Subcatchment
boundaries
Show subcatchments
/ Exaggerate Z Axis
Resolution of 3D
mesh
The best way to manipulate the 3D visualisation image is to click and drag on the image.
Further options include the ability to display the image as a wire frame drawing, filled
drawing or combined wire frame / filled drawing (as shown is filled) in the dropdown box.
The Options button brings up a small dialog offering further options including a transparency
mode, display of axes and background colour. The operation of the directional controls is
dependent on wether pan mode or rotate model is selected.
The mesh resolution governs the details shown in the image, higher mesh resolutions will use
significantly more memory as the Open GL graphics are memory consumptive. The mesh
resolution is set at a default of the greater of 150x150 or the number of rows and columns
of the DEM segment being visualised. The colouring on the visualisation is the same as the
DEM gradient selected in the DEM View Attributes form and the Shaded Relief colouring can
also be used. The subcatchments checkbox governs wether subcatchments are shown in the
images (this box is only available if you are visualising more than subcatchment) and the
edge height governs how visible thee boundaries are in the image. The Exaggerate Z check
box governs whether the z axis will be stretched to exaggerative the topography of the DEM
since vertical range is often small compared to x and y range over a DEM.
88
Program Reference
Subcatchment
The subcatchment menu option will launch the visualisation window and display only those
DEM pixels within the selected subcatchment. After selecting this menu option,
subcatchments should be cross-hatched as you move the cursor over them. Clicking on one
will launch the visualisation window.
DEM Segment
The DEM Segment menu option will enable you to draw a rectangle on the screen in a
similar fashion to the Set Zoom Window option in order to identify a rectangular section of
the DEM to visualise.
Utilities Submenu
89
Action to perform on
DEM pixels
For example, if certain features such as road crests are not picked up in a course DEM then
they can be manually input into the DEM using the Vector Data Set operations by using the
Pixels on Polyline checkbox and Raise by 2m (for example). The PFS algorithm can then be
run to breach the road alignments at their lowest point to determine road overtopping points
during a flood. For an example of this, please see the case study in Section 6.1 (page 143).
Delete DEM
The delete DEM menu option will remove the DEM from the project. All DEM derived
datasets such as subcatchment boundaries and vector streams will also be removed.
90
Program Reference
4.1.5
Reprocess Subcatchment
The reprocess subcatchment menu option undertakes flow processing for only those DEM
pixels within a certain subcatchment. After selecting this menu option, moving the cursor over
a specific subcatchment should cause it to become hatched in yellow. Clicking with the left
mouse button will initiate flow processing for the subcatchment.
91
This form prompts the user for the Stream Area Threshold (SAT) and Minimum Source
Channel Length (MSCL) parameters with witch to generate the vector stream network. The
SAT value refers to the number of pixels flow paths that must accumulate in a single pixel
before it will be considered a potential channel head. The MSCL parameter (if used see
checkbox) defines how long a stream segment must be to be considered valid. Thus if many
potential channel heads exist in an area only the highest elevation pixel will be an actual
channel head. No other channel heads will be accepted unless their source channels are
greater than the MSCL before intersecting the vector stream network.
The effect of using the MSCL parameter can be seen bellow. The left figure does not use the
MSCL parameter whereas the right figure has used a MSCL of 100 metres. Both vector
stream networks were generated with the same SAT value. It can be seen that using the
MSCL parameter can reduce the duplication of streams from nearby channel heads. It will
also effect the calculated bifurcation of the stream network.
92
Program Reference
Duplicate streams
removed
The stream network includes derivation of Horton orders for each segment of the stream
network. The network can be coloured based on the derived Horton orders in the View
Attributes form by selecting Properties from the Vector Layers panel when Synthetic Stream
Network is selected (see page 68).
Vector stream networks are also useful for analysis of subcatchment drainage properties as
well as many of CatchmentSIMs charts and calculated subcatchment parameters (refer
page 106).
93
Import Hydraulic
Controls from GIS
Delete selected
Hydraulic Controls
Group select
Edit property of
selected HC(s)
Filter table
Update background
image with changes
Note:
94
Program Reference
Importing Hydraulic Controls
The Import Hydraulic Control button allows hydraulic controls to be imported from a GIS file
containing line or polyline elements. A data column can be selected to provide the Hydraulic
Control description or if this is not available, it can be set to none and the description will be
left black (it can be manually edited later).
Note:
95
Draw Channel
The draw channel menu option allows you to draw channel type hydraulic controls directly
on the screen. Use the left mouse button to place line vertices and the right mouse button to
finish. If you find that the arrows are pointing the wrong way after the first left mouse click
either right click and start again or finish the line and later toggle the A/B type in the
Hydraulic Controls form.
Draw Gutter
The draw gutter menu option allows you to draw gutter type hydraulic controls directly on the
screen. Use the left mouse button to place line vertices and the right mouse button to finish. If
you find that the arrows are pointing the wrong way after the first left mouse click either right
click and start again or finish the line and later toggle the A/B type in the Hydraulic Controls
form.
Note:
96
Program Reference
View Hydraulic Breaches
Hydraulic breaches occur in certain circumstances when flow paths intersect Gutter type
Hydraulic Controls. Basically, a hydraulic breach occurs when a flow path can not find a
pixel of lower elevation in which to flow to within a set elevation or number of pixels
traversed threshold. This is designed to simulate ponding behind the gutter and eventual
breach due to overtopping. Hydraulic breaches are further explained in Section 5.6
(page 139) and are highly dependent on the parameters set in the Project Options |
Hydraulic Structures form.
The location of hydraulic breaches can be viewed using the Runoff Routing >> View
Hydraulic Breaches (page 97) menu option.
As shown above, by selecting the rows in the Hydraulic Breach form the location of the
hydraulic breaches is shown as a hatched yellow circle in the background. It can be seen that
these hydraulic breaches occur in areas where flow would pond behind the gutter.
97
Navigation, use
arrows or type in
row and column
98
Program Reference
4.1.6
Subcatchments Menu
Add Subcatchment
The add subcatchment menu option allows the user to visually designate the outlet for a
catchment or subcatchment. Outlets are plotted as a polyline and can be as long or short as
desired. After selecting this menu option use the left mouse button for each vertex of the
outlet line and right click to finish.
Other methods of adding subcatchment(s) include importing the subcatchment outlets from a
GIS file (Subcatchments >> Import Subcatchment Outlets page 101) or by using the auto
catchment breakup algorithms (Subcatchments >> Breakup Subcatchment page 101).
After a subcatchment has been added, it will be rasterised into a set of outlet pixels. These
can be manually edited in tabular or visual form using Subcatchments >> Edit
Subcatchment Outlet Pixels page 100) or the Outlet Details button on the Subcatchment
Manager form.
Delete Subcatchment
Subcatchments can be deleted in two main ways. Firstly, using the delete subcatchment menu
option, the cursor will go into subcatchment selection mode, where subcatchments will be
highlighted when the cursor is moved over them. After left clicking a subcatchment in this
mode you will be prompted for confirmation to delete. After selecting Yes, you will be asked
wether to reprocess flow paths for pixels within this subcatchment. It is recommended that you
do this to ovoid the potential for errors. However, if you are deleting a large number of
subcatchments, it can be more time effective to select no, delete further subcatchments and
then run Runoff Routing >> DEM Wide Flow Processing (page 91).
If the subcatchment you wish to delete has no area or can not be selected in this way, it may
be deleted from the Subcatchment Manager form by selecting the appropriate row and
clicking Delete.
99
This form presents a tabular display of the outlet pixels that form the outlet for the selected
subcatchment. A new outlet pixel can be added by clicking the New Outlet Pixel and then
typing in the row and column numbers. Outlet pixels can be deleted by selecting a row and
clicking Delete Outlet Pixel.
100
Program Reference
Breakup Subcatchment
The breakup subcatchment form allow discretisation of a previously mapped catchment or
subcatchment into a number of smaller subcatchments based on one of two catchment
breakup approaches, Flow Accumulation Analysis and Horton Order Analysis.
The catchment or subcatchment to be discretised can be selected by either typing in its
integer code (the number in the first row of the Subcatchment Manager form) or clicking
From DEM which allows the subcatchment to be manually selected by left mouse click when
the subcatchment is highlighted.
Note:
101
102
Program Reference
Once satisfied with the suggested dual outlets, click Accept New Subcatchment and the new
outlets will be introduced. You will be prompted to reprocess flow paths for all pixels within
the subcatchment being broken. This is synonymous with the Runoff Routing >> Reprocess
Subcatchment (page 91) menu option and is recommended to update the subcatchment
mapping and derived parameters.
Subcatchment Manager
The Subcatchment Manager form holds parameters for all subcatchments in the project as
well as the catchment as a whole. These parameters are dependent on the projection,
scaling factors and DEM quality. It also has buttons enabling exporting the table to a
Comma Separated Variable file (CSV) for import into Excel (for example), deleting
subcatchments (select row in table and click Delete), editing outlet pixels and printing.
103
km 2
the perimeter length squared and is dimensionless (
(km )2
104
).
Program Reference
Advanced Submenu
105
4.1.7
Analysis Menu
Graph Wizard
General
The graph wizard form provides access to a range of charts which are described individually
in the following sections. However in general, first select the type of chart required from the
tabs, either flowline profile, cross-section profile or subcatchment analysis. The subcatchment
analysis chart includes a dropdown box further specifying the type of chart required. At this
point, add the pixels (flowline profile), cross-sections (cross-section profile) or subcatchments
(subcatchment chart) for analysis. This can be done visually or manually. In the case of the
subcatchment analysis charts, select the Single Sub(s) From DEM button and manually click
on as many subcatchment as desired (they will highlight when the cursor is over them) and
then click Finished. You will find that the integer code associated with the subcatchments has
been added to the list box. Alternatively, there are buttons to manually add single
subcatchments by typing them in, adding the catchment as a whole, adding all
subcatchments or adding subcatchments with/without upstream inputs.
When the pixels, cross-section or subcatchment to analyse have been selected, press
Generate Chart to process the graphing. The resultant graph can be edited using the Edit
Chart Attributes button and zoomed by clicking and dragging on the graph window. The
charts can be printed using the Print Chart button or copied to the clipboard for pasting into
your preferred application. The charts can also be exported to a Comma Separated Variable
(CSV) file for further analysis in spreadsheet applications.
The subcatchment analysis charts are designed to be used in combination to interrogate
subcatchment parameters and gain a better understanding of their hydrologic characteristics.
106
Program Reference
Specific information on each chart follows:
Note:
The option to plot the vectored slope line will be disables once 6 or
more pixels are selected to reduce clutter in the chart.
107
108
Program Reference
Pixel (%) vs Downslope Distance
This chart plots distance vs the proportion of pixels within a subcatchment that will have
reached a raster stream pixel by the time they have travel such a distance. Thus, as distance
increases, pixels (%) will approach 1 (100%). This chart illustrates how rapidly a
subcatchment can be drained. For example, the chart below indicates that after 120m of
flow path, flow from 78% of pixels from subcatchment 1.07 will be in-stream whereas 82%
of pixels from subcatchment 3.01 will be in stream. This (considered in conjunction with
slope and land use can help assign relevant runoff routing parameter in subsequent rainfall
runoff models.
109
110
Program Reference
Bifurcation (Stream Order vs # Streams)
The bifurcation chart is based on the Horton ordered vector stream network outlined in
Section 3.9 (page 38). For instructions on generating a stream network, see page 92. The
bifurcation charting adds two series to the chart for each subcatchment. The circular dots are
the stream order vs log number of stream of that order and the line is the best fit
(bifurcation). The bifurcation (slope of the line) and an indication of the fit (R) can be seen in
the legend.
Note:
The steeper the line, the higher the bifurcation and this generally means that the generated
vector stream network is more fractal (higher proportion of 1st order streams) and may be
better drained. It is also wise to assess wether the vector stream network is
geomorphologically valid with reference to the law of stream drops (Broscoe 1959) by using
the Stream Order vs Stream Drop and Stream Area Threshold Comparison charts.
111
112
Program Reference
Hypsometric Curve (Area vs Height)
The hypsometric curve chart provides a measure of the geomorphologic state of the
subcatchment. Horton observed that subcatchments in different state of their
geomorphological evolution had different relative height vs relative area relationships. This
chart includes a background picture illustrating Hortons hypsometric profiles.
113
It can be see that for this example, the vector stream network generated with an SAT of 100
results in a t of -17 indicating that there is a relationship in the stream drop distribution
hence the network is not geomorphological valid (ie., the SAT is too low and there are too
many first order streams resulting in two low a 1st order mean stream drop). To determine an
appropriate SAT to use to generate a vector stream network the Threshold Area Comparison
chart can be used.
114
Program Reference
Threshold Area Comparison
The Threshold Area Comparison chart can be used to determine the highest level of vector
stream network discretisation that fulfils the constant law of stream drops (Broscoe 1959). For
more information on how this law is applied to analyse vector stream networks, please see
the Stream Order vs Stream Drop chart (page 114).
This chart will derive a set of vector stream networks, determine their Students t-test values
and bifurcation and plot them for the range of SAT values. When the Generate Chart button
is clicked the following form will be generated.
Add CatchmentSIM
generated series of
SAT values
SAT values to chart
Add a new SAT
value, remove
selected value, clear
entire list
Do you wish to use a
MSCL for generating
vector stream
networks.
This chart governs the list of SAT values that will be analysed. To start by imputing a
CatchmentSIM derived list of SAT values, click Add Default Sequence. This button will add a
selection of SAT values which should include an SAT appropriate for ensuring consistency
with the constant stream drop law. Individual SAT values can be added or removed using the
Custom Sequence controls. A Minimum Source Channel Length (MSCL) option can also be
used (for more information see page 92). After clicking Ok, CatchmentSIM will generate a
vector stream network for all the SAT values selected and determine associated parameters
including bifurcation and Students t-test (this algorithm can be time consuming if many SAT
values have been added). The values are then plotted such as in the chart bellow.
115
The Student t-test values are shown on the bottom. As described under the Stream Order vs
Stream Drop chart (page 114), the minimum SAT that yields a stream network where the ttest indicates that the means are not statistically different with a 95% confidence interval (ie., t
2) should be adopted. It can be seen that for this example, this occurs around an SAT of
1500 which is the finest vector stream network consistent with the law of constant stream
drops.
116
Program Reference
Create Animation
The create animation menu option will generate the Create Animation form which holds
controls for developing custom animations within CatchmentSIM.
Type of animation
Note:
117
118
Play, pause, go to
start, go to end and
frame rate buttons
Program Reference
Save Animation As AVI
In the Play Animation form menu there is an option to save the animation as an AVI file. This
function will stitch together all the images associated with each frame of the animation using
a third party program that is distributed with CatchmentSIM. After this menu option has been
selected a prompt will be presented asking you to click OK when the third party application
has finished. At this stage a DOS style window should be visible. Once the DOS window says
the animation is complete, click Ok.
Note:
119
4.1.8
Export Menu
Check the checkbox associated with the data layers that you wish to export and press Ok.
120
Program Reference
Exporting Individual Data Series
For each export data series that is selected in the Export GIS data form, a form will be
generated to customise the specific setting of the exported data series. The form bellow
corresponds to the vector stream network export form (in this case synthetic stream network).
Available tabular
data export columns
(for database file)
Decimal places
(decimal data
only)
Remove data
field
Reproject
export data
(mandatory for
custom
projections)
Convert polygons to
polylines or polylines to lines
The data fields available to export to the GIS data file are listed in the Available Data Fields
list box. These are a function of the CatchmentSIM data layer to be exported. They can be
added to the set of data fields to be exported by double clicking on them or clicking the
>>> button, they may be removed using the Remove button. If the data layer is a decimal
data type then the number of decimal places can be customised using the Decimal Places
spin edit control. The order of the exported data series can be edited using the up and down
arrows.
The checkboxes under the Available Data Fields list box govern the shape types for regions
and polylines in the exported data series. Unchecked, these shapes will remain in the native
shape type, however, if checked, regions may be dropped to polylines or all the way to
simple lines.
121
After clicking Ok, the user will be prompted to enter the file location.
122
Program Reference
Setup Display Window
The setup display window menu option allows the user to precisely setup the CatchmentSIM
window size and location by specifying its height, width, anchor points as well as the zoom
characteristics by setting the extents of the visual part of the project.
This form is designed to be used to help ensure consistent views for export of screenshots or
video capture, mostly for education and training purposes.
123
The CST scripts are stored in the Installation Directory \ CST Files and those that are
available are shown in the file tree view. By selecting a directory in this window, the available
CST scripts in that directory are shown in the top right window. By selecting a script in this
window, the description of this CST script (stored in the CST file header) will be shown in the
Macro Description window.
After selecting the Run Macro button, CatchmentSIM will behave in accordance with the
macro instructions.
124
Note:
Note:
Program Reference
4.1.9
Help Menu
Help
The help menu option will bring up the CatchmentSIM offline help.
Tutorials Submenu
The tutorials submenu gives access to CatchmentSIMs tutorials.
On-line Description
The on-line description menu option links to the Overview section of the CatchmentSIM
website which provides information on the CatchmentSIM algorithms (internet access
dependent).
Technical Support
The technical support menu option provides a link to a web form where requests for technical
support can be made (internet access dependent).
125
About
The about menu option will display the About screen.
126
Key Challenges
5 Key Challenges
5.1
Introduction
This Chapter outlines some of the key challenges associated with undertaking a terrain
analysis project with CatchmentSIM. These include re-projection of raster DEMs, flat and pit
removal from DEMs, automated catchment breakup, using CatchmentSIMs hydraulic control
features and geomorphologic analysis.
5.2
127
To reproject a DEM, setup a project with the desired projection (File >> New Project
page 56), setup the DEM with the desired boundaries and resolution (Digital Elevation Model
>> Setup DEM page 78) and then populate the DEM pixel elevations with the external
DEM (Digital Elevation Model >> Assign DEM Elevations by Sampling External DEM
page 83). You will be prompted to select the projection of the external DEM or lat/long if
appropriate. CatchmentSIM will then reproject the DEM accordingly. When a DEM type is
selected where the projection is known (such as SRTM data which is lat/long) then
CatchmentSIM will automatically recognise the correct projection.
5.3
128
Key Challenges
5.3.1
PFS Algorithm
Has the lowest elevation gain over the course of the path;
If two paths have the same elevation gain, the shorter path is selected.
As discussed previously, there a 4 parameters that may be edited in the Project Options form
(refer page 59) that affect how the PFS algorithm operates. These are repeated here for
convenience.
PFS Minimum Gradient: This parameter governs the minimum gradient that
must be observed over the PFS solution channel before the algorithms will
determine that a solution has been found. This parameter can be adjusted to
ensure legitimate PFS solutions are found and that rounding errors do not cause
PFS solution channels to become flat. For example, if the PFS minimum gradient
is too low then a very shallow slope may be calculated over the PFS channel.
After linear interpolation is applied to determine pixel elevation over the
channel, these elevations will be rounded to the precision of the DEM (single or
double). This can cause these pixels to become flat.
PFS Break Size: This parameter has been introduced for speed reasons, once
the PFS algorithm has grown very large looking for a potential solution channel
over a wide area it can slow down considerable. Once the PFS algorithm has
grown to the PFS Break Size the algorithm will terminate. If time is not a
constraint this parameter can be adjusted upward to ensure the PFS algorithm
finds even the most complex solutions.
DEM Precision: The DEM precision parameter governs wether the DEM will be
stored as single or double precision floating point values. Single precision stores
7-8 digits while double precision stores 15-16 digits. Selecting double precision
will increase the size of the DEM file size by 100% and total project files by
around 20% and runtimes by 20% as well as RAM requirements. Double
precision may be required when using elevation data with several digital on the
left side of the decimal point such as 10450.0560 ft or very flat DEMs or if you
wish to use a low PFS minimum gradient parameter.
129
Problem:
The pixel elevation is so low that an outlet can not be found
within the DEM that is lower by a sufficient amount to satisfy the minimum
downslope gradient criteria.
Solution:
Problem:
A potential solution pixel is too far away (PFS algorithm exceeds
minimum break size).
Solution:
Problem:
A potential solution pixel is to far away (PFS algorithm
encounters a boundary or blank pixel prior to finding the solution pixel.
Solution:
Assign elevations for blank pixels, either by interpolation or
sampling. Alternatively, change blank pixel treatment to Ignore (note: that this
may increase the time of the algorithm and cause unrealistic solution paths for
pixels near the boundary)
Problem:
Solution:
Manually raise the pixel and its neighbours (using the Inspect
Pixel form page 98), or manually lower a pixel near the desired catchment
outlet to serve as a outlet for the PFS algorithm.
Problem:
You may find that if the minimum downslope gradient is too low
then the PFS algorithm will create a path of flat pixels along its solution path. This
is caused by a very shallow gradient being converted to a series of stepped flat
pixels when the elevations are rounded to the DEM precision (single or double).
Solution:
Increase the minimum downslope gradient or change the DEM
precision to double.
130
Key Challenges
Advanced Use of PFS Algorithm
In extremely flat terrain or poor quality DEMs you can apply some special techniques to
ensure flat and pit pixels are removed. These may include:
Manually lower the elevation of a pixel near the catchment outlet (using the Inspect
Pixel form page 98) so as to serve as a solution for the PFS algorithm. You will
need to ensure this pixel is not too close to the boundary and that this pixel is low
enough to provide adequate slope to satisfy the minimum downslope criteria. You
will also need to ensure the minimum break size is large enough to provide a
solution for all flat and pit pixels.
Burn a vector stream network into the topography prior to running the PFS
algorithm. This could be an external stream network (imported) or a calculated
vector stream network. To use a calculated vector stream network, first export it as a
GIS file (Export >> Export GIS Data page 120) and then import it as an imported
stream network (Vector GIS Data >> Import Stream / Flow lines page 72), prior
to applying stream burning (Digital Elevation Model >> Develop DEM >> Stream
Burning page 81). This will lower pixels along the stream network and increase
the gradient along potential PFS solutions paths, hence increasing the effectiveness
of the algorithm.
Using the Digital Elevation Model >> Utilities >> Interpolate Streams Over
Existing DEM (page 89) algorithm (see Section 5.3.2, bellow).
5.3.2
An imported vector stream network can be useful during the interpolation of a DEM to ensure
it is hydrologically accurate. However, an imported stream network can also be valuable
when using an imported / sampled DEM. This network can be burnt into the DEM (Digital
Elevation Model >> Develop DEM >> Stream Burning page 81) or the Digital Elevation
Model >> Utilities >> Interpolate Streams Over Existing DEM (page 89) can be used to
ensure pixels underlying the vector network decrease in a downstream direction. This
algorithm ensures that pixels underlying the stream network decrease in a downstream
direction by applying linear interpolation to remove any rising part of the long-section along
the stream as outlined in Figure 5-2.
131
Pixel elevations
under stream line
Interpolation of DEM
pixels resulting from
algorithm
In Figure 5-2, the blue line represents the interpolated elevations assigned to the underlying
pixels after application of the algorithm. The algorithm can only interpolate DEM pixel
underlying a stream segment if two criteria are met.
Note:
132
The algorithm can only interpolate new pixel elevation if it can find a pixel of
lower elevation along the line segment at some point. For example, in Figure
5-2, the channel illustrated by the blue line will only be interpolated by the
algorithm in the case of the green DEM stream line profile as apposed to the
red which never dips bellow the minimum elevation of the black profile. Thus
the algorithm may be affected by very low pit pixel along the stream profile.
This can be avoided by running one iteration of the filling algorithm prior to
running the algorithm.
Key Challenges
A combination of the filling algorithm and the PFS algorithm can form a hydrologically
realistic DEM very quickly in virtually all case. In certain circumstances (very flat terrain or
poor quality DEMs) it may be necessary to adjust the parameters of the PFS algorithm to
achieve the desired results. In even rarer cases, where the desired catchment outlet may not
be the lowest part of the DEM, some manual editing of the outlet pixel elevations, stream
burning or interpolating an observed stream network over the DEM can fix the problem. This
ensures that these algorithms can work where more simple approaches (such as the J&D
algorithm) fail.
5.4
5.4.1
The lateral inflow analysis algorithm and the Horton Ordering based automatic breakup may
produce different catchment breakups even when the same number of subcatchments are
generated. In general, the Horton Ordering algorithm appears to produce better results. In
the example presented in Figure 5-3 a Threshold Area Comparison (refer Section 5.5.1,
page 135) was undertaken to determine the appropriate SAT value for generation of the
vector stream network (800 pixels in this case) and vector stream network was generate. This
can be seen (coloured by Horton Order refer page 68) in Figure 5-3. A Horton Order
based catchment breakup was then undertaken based on delineating all 2nd order and higher
subcatchments. For comparison, the Flow Accumulation Analysis algorithm was also run with
the target subcatchments set to the same number that were delineated in the Horton
Ordering based breakup.
133
Figure 5-3: Comparison of Horton (top) and Flow Accumulation Analysis (bottom) Breakup Algorithms
It can be seen in Figure 5-3 that the two automated catchment breakup algorithm produce
different result. The Horton Ordering breakup seems to deliver a better distribution of
subcatchments with a similar size.
134
Key Challenges
5.5
5.5.1
CatchmentSIM offers a range of analysis tools to help quantitatively assess the appropriate
SAT for stream network generation. This is based on the law of constant mean stream drops
as first observed by Broscoe (1959). This law states that the mean drop of streams of different
Strahler orders should be statistically similar. Tarboton (1997) suggests comparing the 1st
order mean stream drop to the mean stream drop of all other streams using Students t-test
with a 95% confidence interval (ie., T 2). The lowest SAT that yields a stream network that
has statistically similar means within this confidence interval should be applied. Alternately,
the MSCL can be adjusted with a highly detailed stream network until Students T value is less
than 2. CatchmentSIM can chart the mean stream drop scatter for a generated stream
network (Analysis >> Graph Wizard | Steam Order vs Stream Drop page 114) as shown
in Figure 5-4.
135
In Figure 5-4, the small circles represent the individual stream segment elevation drops while
the larger circles represent the mean elevation drop for that order. It can be seen in the
legend that the calculated Students T value for this stream network is -1.57 which is well
within the 2.00 95% confidence interval. Hence, the stream network generated at this SAT
(500 pixels) does obey the law of constant mean stream drops.
In order to help the user identify the appropriate SAT for use in an analysis, CatchmentSIM
also enables the Students T value to be calculated for a range of SAT derived stream
networks and displayed as a chart (Analysis >> Graph Wizard | Threshold Area
Comparison page 115). The same project used in Figure 5-4 was used to generate Figure
5-5. It can be seen that the general trend is for the Students T value (left axis) to reduce as
SAT increases. This means that SAT values smaller than 500 produce stream networks that
do not obey the law of constant mean stream drops within the 95% confidence interval. This
chart can be used to evaluate the minimum SAT that can be applied with confidence. The
chart also indicates the catchment bifurcation ratio for these stream networks which stays
quite constant for the range of stream networks analysed.
136
Key Challenges
For the example presented in Figure 5-5, it can be seen that the Students T-value trend line
crosses the threshold value of 2.0 at approximately 490 pixels. Thus, 490 pixels is the
minimum SAT value that should be used in order to generate a stream network that conforms
to the law of constant mean stream drops.
5.5.2
Horton / Strahler ordering provides the functionality to derive some key geomorphologic and
fractal stream network coefficients, that can be used to examine the hydrologic properties of
subcatchments. Once streams have been defined as a specific order, a number of useful
parameters and geospatial statistics can be calculated. The most important of these is the
bifurcation ratio which is a measure of the relationship between the numbers of streams of
different orders. Strahler found that a strong log-normal relationship exists between the
logarithms of the number of streams of each order versus stream order. The gradient of this
relationship is deemed the bifurcation of the subcatchment. Figure 5-6 shows the calculation
of the bifurcation for the sample stream network illustrated in Figure 3-21 (page 39).
137
CatchmentSIM can automatically calculate the bifurcation ratio for the catchment and all
subcatchments, as well as a number of related parameters such as Horton drainage density.
Interestingly, Strahler's work on topographic maps found strong bifurcation relationships, with
bifurcation ratios that were consistently within the range of 3-5. It has been found that
CatchmentSIM generated vector stream networks also exhibit this strong relationship, with
bifurcation ratios also around this range. This lends weight to the argument that vector
stream networks generated over DEMs can closely resemble the fractal nature of natural
systems.
One of the unfortunate attributes of all techniques for analysis of stream networks is that
most derived parameters are dependent on the SAT value adopted for generation of the
stream network. However, it has been found that the bifurcation ratio is not highly dependent
on the SAT value used to generate the network. This is illustrated in Figure 5-7, which shows
the bifurcation ratios calculated for two different subcatchments for a range of vector stream
networks calculated at different SAT values.
138
Key Challenges
It can be seen in Figure 5-7 that there does not appear to be a strong trend in each of the
Bifurcation vs SAT plots. This fact is useful because it implies that deviations in bifurcation
ratios across subcatchments within a model may convey important information about the
hydrologic properties of the subcatchments and not simply be a function of the adopted SAT
value.
For example, in Figure 5-7, it can be seen that regardless of the SAT value used to generate
the vector stream network the bifurcation ratio for subcatchment 2.01 is significantly higher
than subcatchment 1.01. The implications of this relationship means that subcatchment 2.01
has a greater proportion of 1st order streams than subcatchment 1.01 (bifurcation plot will
be steeper) and the drainage network is more fractal. This information has important
hydrologic implications and may mean that subcatchment 2.01 will respond quicker to
rainfall and may need to be allocated a smaller lag time or related lag coefficient in any
downstream hydrologic or hydraulic model.
This example is presented simply to illustrate some methods of analysis using CatchmentSIM
that can provide the basis for assignment of runoff routing parameters in hydrologic models
that may be applied. Aside from vector stream network analysis, CatchmentSIM includes a
number of other hydrologic analysis tools to help quantify relationships between various
subcatchments and even individual pixels which are outlined on page106.
5.6
5.6.1
Channel type hydraulic controls (Runoff Routing >> Hydraulic Controls page 93) are
drawn in CatchmentSIM as a solid line with triangles pointing in the direction of the line
towards the channel outlet
139
140
Key Challenges
5.6.2
Gutter type hydraulic controls (Runoff Routing >> Hydraulic Controls page 93) are drawn
in CatchmentSIM as a solid line with triangles pointing at a perpendicular angle to the
direction of the line
Controls form
or
. The triangles point in the direction in which flow is
allowed to pass over the hydraulic control. When a flow path intersects a gutter it is only
permitted to cross the gutter in the direction of the arrows, otherwise it must flow along the
gutter in a downslope direction. The exception occurs when a downslope direction can not
be found. In this case, the algorithm will search within a given tolerance, either a number of
pixels or a given elevation (File >> Project Options | Hydraulic Structures page 59) for a
pixel of lower elevation. If one can be found within the tolerance, then flow is mapped to this
point, if not then a hydraulic breach occurs as outlined in Section 3.12.4 (page 48). Figure
5-11 and Figure 5-12 illustrate the flow paths and subcatchments mapped from the same
gutter hydraulic control with two different elevations based hydraulic breach tolerances.
141
It can be seen in Figure 5-11 that 3 hydraulic breaches are formed (yellow circles) where the
gutter hydraulic control is breached. This is designed to simulate ponding behind the low
elevation gutter and indicate the points at which it will breach. It can also be seen that this
gutter has a relatively small effect on the subcatchments. However in Figure 5-12 where a
higher gutter is implemented (using a higher elevation tolerance in the Project Options form),
that only one hydraulic breach is formed and a significant change in the mapped flow paths
is evident.
142
Case Studies
6 Case Studies
This chapter profiles two case studies that demonstrate some of CatchmentSIM algorithms.
6.1
143
6.1.1
Project Methodology
The project consisted of importing an external DEM developed from aerial survey into
CatchmentSIM and using internal algorithms to remove all flats and pits from the DEM. The
road crown database in the area was represented in the DEM by a 'road burning' approach.
Inlet gullies were directly imported from a GIS database and subcatchment boundaries and
parameters were automatically calculated prior to export of catchment and subcatchment
characteristics to the DRAINS model. The DEM, imported road network and inlet gullies are
shown in Figure 6-1.
The DEM shown in Figure 6-1 was originally developed by aerial photogrammetry and was a
highly complex TIN of several million points. This was converted to a raster DEM with the
12D geo-processing software, since this project was completed before CatchmentSIM
included the functionality for sampling TIN models. The DEM resolution was 2499 rows by
3249 columns, which forms a DEM with over 8 million pixels.
Flat and pit pixels within the DEM were treated by the PFS algorithm until flow could be
processed from all points within the catchment.
The key challenge in modelling the Holland Park environment lies in the highly urbanised
nature of the catchment. Most of the topographic features that will control flow in the area
are not represented in surveyed GIS data (even highly accurate aerial photogrammetry based
survey). The single most important urban features in the area are the roads. Consequently, it
was crucial that these were adequately represented in the DEM. As outlined in Section 3.12
(page 45), CatchmentSIM provides two different approaches for representation of urban
features in projects, hard-coding of these features into the DEM, or simulation of these
features as overriding external controls. For this project, the road network was hard-coded
into the DEM. This was achieved using CatchmentSIMs vector data set operations to raise all
DEM pixels underlying road crown alignments by 0.5 metres.
144
Case Studies
Following the 'road burning' the PFS algorithm was applied to treat all flat and pit pixels that
were formed during the road burning procedure. This caused the roads to be breached by
major drainage paths at their points of lowest elevation. This may be seen by looking closely
at the CatchmentSIM generated minor drainage network displayed in Figure 6-3 (page 146).
The subcatchment inlets were derived based on the Asset Database of Brisbane City Council
representing inlet gullies. 466 inlets from this database were directly imported into
CatchmentSIM.
Following import of the subcatchment outlets, flow routing was undertaken for the entire
catchment producing the subcatchment boundaries shown in Figure 6-2.
The level of detail that is accommodated by the CatchmentSIM flow routing algorithm and
resultant modelling of urban features can be clearly seen by examining the CatchmentSIM
results superimposed over aerial photography. Figure 6-3 illustrates the subcatchment
boundaries, road crown alignments, inlet gullies and calculated minor drainage paths for the
project (a vector stream network calculated at a low SAT value).
145
- 146 -
Case Studies
It can be seen in Figure 6-3, that the road crown alignments strongly influence the
subcatchment boundaries and minor drainage paths, yet in other areas the flow is controlled
by the DEM. This has allowed for realistic flow paths and inlet gully drainage areas to be
calculated by combination of a sampled DEM with an urban features (road crown) database.
The Brisbane City Council City Plan GIS database was also imported into CatchmentSIM to
help represent the urban environment. This database was imported as a CatchmentSIM
impervious areas database in order to calculate impervious proportions for each inlet gulley
drainage area. This GIS layer may be seen in Figure 6-4. The different colours represent
different land-use types and different impervious proportions were assigned to these polygons
to realistically represent their varying degrees of imperviousness.
Following assignment of impervious proportions to each inlet gulley drainage area, the
project was exported to the hydrologic / hydraulic DRAINS model. This was achieved by
exporting CatchmentSIM data to a CSV file and then using DRAINS spreadsheet functions to
import subcatchment attributes. CatchmentSIM now includes a CSTalk macro script designed
for simpler coupling with the DRAINS model.
6.1.2
CatchmentSIM Contribution
This project would not have been possible without CatchmentSIMs vector data set
operations tools (for raising road crown alignments) or the PFS algorithm for treatment of
resultant flat and pit pixels within the DEM. Raising DEM pixels underlying the road crown
alignments creates large closed depressions bounded by road crowns. Application of
traditional filling algorithms and the J&D Algorithm would have simply filled the closed
depression forfeiting any benefits associated with the exercise. The PFS algorithm is crucial to
the success of this project because it finds the least cost path at which to breach road crowns
as shown in Figure 6-3.
147
6.2
Introduction
This project was undertaken to verify CatchmentSIMs algorithms against the geo-processing
undertaken by the United States Geological Survey (USGS) during their on-going project
called Elevation Derivatives for National Applications (EDNA). The EDNA project utilises the
existing National Elevation Dataset (NED) (a raster DEM data set), and the National
Hydrography Dataset (NHD) (a vector watercourse alignment data set) in combination with
automated hydrologic DEM analysis algorithms to produce the following derivative data sets:
Aspect
Contours
Filled DEM
Flow Accumulation
Flow Direction
Reach Catchments
Shaded Relief
Sinks
Slope
Synthetic Streamlines
This project is probably the most comprehensive automated terrain analysis work currently
being undertaken in the world. A sample data set is available on the EDNA website
(http://edna.usgs.gov/Edna) for the 835,000 hectare Upper Washita catchment in the
Arkansas-White-Red River Basin. This data has been made available as a sample of a
complete EDNA analysis and to provide a testing ground for new and improved methods
and tools.
As such the source data sets were downloaded and imported into CatchmentSIM to enable
comparison of CatchmentSIM with the various algorithms used to compile the EDNA data
sets.
6.2.2
Project Methodology
The Upper Washita NED DEM in raw format was imported into CatchmentSIM with 4475
rows and 5187 columns. The NHD watercourse alignments were imported into
CatchmentSIM and the Digital Elevation Model >> Utilities >> Interpolate Streams over
Existing DEM (page 89) tool was applied. This algorithm is similar to the watercourse
integration algorithm documented in Section 3.5.2 (page 22) but it is designed for
application to sampled DEMs as apposed to those in the process of being interpolated (refer
Section 5.3.2 page 131). The algorithm processes the watercourse alignment network to
assign flow directions and then ensures that DEM pixel elevations underlying the watercourse
alignments decrease in a downstream direction. Where pixel elevations are flat or increase in
a downstream direction, the algorithm continues to process the watercourse alignment until a
pixel of lower elevation is found. Then all pixels between these points are assigned linearly
interpolated elevations along the watercourse alignment.
148
Case Studies
Following this the filling algorithm was applied to treat some hill-crest areas where the limited
vertical precision of the DEM had flattened the hill-crest. Finally, the PFS algorithm was
applied to remove all remaining flat and pit pixels. Catchment and subcatchment delineation
was then processed based on the outlet points identified in the EDNA database. This yielded
a subcatchment network that is very similar to the ENDA results and is shown in Figure 6-5.
A 3D representation of the Upper Washita catchment boundary and calculated vector stream
network derived by CatchmentSIM can be seen in Figure 6-6.
149
6.2.3
The catchment boundaries generated by CatchmentSIM and the D8 algorithm were closely
matching in most regions. However, some areas of discrepancy were found. These were not
due to the flow routing algorithm, rather, were a consequence of the differing approaches to
treatment of flat and pit pixels. A sample area of catchment boundary delineation differences
is shown in Figure 6-7.
As shown in Figure 6-7, the CatchmentSIM and EDNA derivations of the catchment
boundary differ slightly. However, both of the automated catchment boundaries are more
accurate than the Hydrologic Unit Catalog (HUC) boundary which had previously been done
by hand. It is not entirely clear from the contours, which of the automated catchment
boundaries is more correct. However, it can be argued that the CatchmentSIM boundary is
more correct because the dashed black lines are realistic flow paths derived from the
contours which breach the USGS derived boundary.
Other regions of catchment delineation deviation between the two approaches are shown
draped over the 3D surface of the original NED DEM in Figure 6-8.
150
Case Studies
2.
1.
Similarly to Figure 6-7, it can be seen in Figure 6-8 that both automated approaches are
more accurate than the hand delineated HUC catchment boundary. The observed
differences in the automated catchment delineation in Figure 6-8 are due to the use of the
PFS algorithm as apposed to the J&D algorithm for treatment of flat and pit regions. A
quantitative judgement of which boundary is correct is difficult, however, it could be argued
that the CatchmentSIM algorithms produce a catchment boundary that is more able to
identify ridge lines in areas of low sampling definition (Circle 1) and is better able to bisect
flat hill-crest areas (Circle 2).
The differences between the flow routing algorithms adopted in the USGS approach (D8
method) and CatchmentSIMs flow routing algorithm are more pronounced in areas near
catchment outlets as outlined in Section 3.8.2 (page 36). An example of the subcatchment
boundaries generated near the outlet of one of the Upper Washita subcatchments can be
seen in Figure 6-9.
151
It can be seen in Figure 6-9 that the CatchmentSIM boundary is more successful at
identifying the expected ridge line between the 1140 contour alignment towards the bottom
the figure. The D8 generated boundary is tending toward the 22.5 degree increment error as
outlined in Section 3.8.2 (page 36).
6.2.4
CatchmentSIM also produced a very different stream network than the USGS approach even
when using the same SAT value for channel head identification. The USGS selected a SAT
value of 5,000 pixels (4.5 km2) for deriving a calculated stream network based on matching
the NHD data set used to condition the DEM. Figure 6-10 illustrates the NHD, EDNA and
CatchmentSIM vector stream network generated at the same SAT value.
It can be seen in Figure 6-10 that the CatchmentSIM stream network is better able to match
the NHD stream network and does not exhibit the 8 direction angular bias that is commonly
associated with the D8 method and evident in the centre section of the figure.
To investigate the effect of increasing the detail level of the calculated stream networks, the
SAT was lowered to 200 pixels (0.18 km2) to examine the ability of the software products to
represent minor drainage paths and to test the effectiveness of the flow routing algorithm.
Figure 6-11 illustrates a comparison of these stream networks over the 1:24,000 Digital
Raster Graphic (DRG) mapping supplied by the USGS.
152
Case Studies
Northern Tributary
1.
Southern Tributary
2.
The parallel flow paths associated with the J&D algorithm and the D8 method are clearly
evident in the USGS D8 generated stream network as shown in Figure 6-11 (Circle 1). The
CatchmentSIM network has a more realistic fractal nature due to the PFS flat and pit pixel
treatment algorithm and the more advanced CatchmentSIM flow routing algorithm.
Furthermore, the increased resolution of the vector stream networks has revealed a significant
deviation in the two networks (Circle 2). The USGS model incorporates the drainage area in
the bottom left of the figure (green dashed line) to the northern tributary while the
CatchmentSIM model indicates this area draining to the southern tributary input. Thus, if
subcatchments were generated for these two tributaries, the two models would produce very
different results. It is not entirely clear from the DRG mapping which model is correct since
there are no contours in the vicinity. However, the significance of this deviation is that the
different hydrologic conditioning and flow routing algorithms can have major impacts on
subcatchment delineation and stream network generation.
6.3
Conclusions
These two case studies profiling research that has been completed with CatchmentSIM, aim
to demonstrate the capabilities of CatchmentSIM and the algorithms it employs. The Holland
Park Local Stormwater Management Plan project demonstrates how the hard-coding of
urban structures in conjunction with CatchmentSIMs improved PFS flat and pit removal
algorithm can accommodate accurate modelling of urban features that may have otherwise
prohibited application of automated hydrologic analysis algorithms in such a catchment.
The Upper Washita Catchment EDNA project demonstrates how CatchmentSIM was able to
generate superior catchment boundaries and calculated stream networks than the
approaches adopted by the USGS. These improved results are largely due to
CatchmentSIMs improved flow routing algorithm and the PFS flat and pit pixel resolution
algorithm.
153
References
Broscoe, A. J. 1959. Quantitative analysis of longitudinal stream profiles of small watersheds,
Office of Naval Research, Project NR 389-042, Technical Report No. 18. Department of
Geology, Columbia University, New York.
Flint, J. 1974, Stream gradient as a function of order, magnitude and discharge, Water
Resources Research, vol. 10, no. 1, pp. 969-973.
Goonetilleke, A. & Jenkins, G. A. 1996, GIS Display Tools or Analytical Tools for
Hydrologists? Proceedings of the 23rd Hydrology and Water Resources Symposium, Hobart,
Tasmania, pp 631-637.
Goyen, A. G. & OLoughlin, G. G. 1999, Examining the Basic Building Blocks of Urban
Runoff, Proceedings of the 8th International Conference on Urban Drainage, Sydney, pp.
1382-1390.
Jones, R. 2002, 'Algorithms for using a DEM for mapping catchment areas of stream
sediment samples', Computers & Geosciences, vol. 28, no. 9, pp. 1051-1060.
Martz, L. W. & Garbrecht, J. 1998, The treatment of flat areas and depressions in
automated drainage analysis of raster digital elevation models, Hydrological Processes, vol.
12, no. 1, pp. 843855.
Martz, L.W. & Garbrecht J. 1998 The Treatment of Flat Areas and Depressions in
Automated Drainage Analysis of Raster Digital Elevation Models, Hydrologic Processing, vol.
12, no. 1, pp. 843-855.
Tarboton, D G., Bras, R L. 1991, 'On the Extraction of Channel Networks from Digital
Elevation Data', Hydrological Processes, vol. 5, no. 1, pp. 81-100.
Tribe A. 1992, 'Automated recognition of valley lines and drainage networks from grid digital
elevation models: a review and a new method', Journal of Hydrology, vol. 139, no. 1, pp.
263-293.
154