CEE / IES 656 Engineering Applications of Land & Geographic Information Systems FALL 2005

CEE / IES 656

Lab 11: Functional Analysis of Surface Models
Due to beginning of next lab (December 2nd, 2005)

PART I. OVERVIEW
This lab gives an overview of some general ways that surface models can be analyzed: viewing,
measurement (profiles, volume calculation), and hydrologic (flow) analysis. This lab uses digital
elevation models (DEMs), not Triangulated Irregular Networks (TINs), since most ArcInfo surface
functionality uses the DEM data structure. You will use both ArcInfo and ArcMap during this lab. For
descriptions for hydrology modeling in ArcGIS follow the help document for hydrology modeling
sample extension from ESRI. For the more detailed description, please read the full help document
included in the sample extension zip file.

1. Data Download
Work on the C:\TEMP drive or your CAE UNIX account, because this lab can require a large amount
of disk space. At a CAE computer, in an internet browser go to the course home page and download
the compressed data for this lab: DEMORTHO.EXE. By default the files will extract to
C:\TEMP\656LAB11. Extract the following data sets, and prepare them in ArcInfo:
 PVDEM.E00 (DEM in ArcInfo grid format)
 BMNDEM2.E00. (DEM in ArcInfo grid format)
 ORTHO.BIL (Orthophoto in “band interweaved by line” image format)

Arc: workspace c:\temp\656lab11 (set workspace to your NT or UNIX personal drive)

Arc: del demortho.exe (delete the compressed file to save space)
Arc: import grid pvdem.e00 pvdem (import grid – name it PVDEM to save typing)
Arc: import grid bmndem2.e00 bmndem2 (import grid)
Arc: del pvdem.e00 (delete the export file to save space)
Arc: del bmndem2.e00 (delete the export file to save space)
Arc: describe pvdem (view DEM parameters)
Arc: describe bmndem2

2. Import Map Projection Information

When you add a new dataset in vector format into ArcMap, the software is normally looking for *.prj
file to get map projection information. However, it’s unclear how the map projection information is
saved in raster dataset even in the on-line help document.

Fortunately, the DEM datasets contain map projection information. The orthophoto has been
oriented using the same map projection, even though you can’t see any coordinate system
information in the Properties dialog. Therefore, you need to assign those information using given
datasets with ArcCatalog. You may use the same technique for the vector format datasets.
Especially when you create a new shapefile, you need to import the map projection information
from existing datasets for further spatial analysis.

 Start ArcCatalog and go to the folder containing all datasets for this lab.
 Right mouse click on the orthophoto file name (ortho.bil) and choose Properties….
 Under the “Spatial Reference” tab in the Raster Dataset Properties dialog, click the Edit
button.
o In the Spatial Reference Properties dialog, click the Import button.
 Choose a DEM containing the map projection information and click Add button. Note
that all DEMs have the map projection information fortunately with this lab dataset.
 You will see the detailed map projection information in the Details column.
 Click the Apply and OK buttons in order.
o Now you will see the map projection information in the Projection column.
o Click the Apply and OK buttons in order again.
 Start ArcMap and add all datasets into ArcMap.
 Right mouse click on the orthophoto name and choose Properties….
1 of 11
CEE / IES 656 Engineering Applications of Land & Geographic Information Systems FALL 2005
 Under the Source tab in the Layer Properties dialog, check if the orthophoto has the map
projection information or not.

PART II. SURFACE ANALYSIS

1. Perspective Views of Surface & Image Draping

A perspective view of a surface uses perspective geometry from one point of view, rather than the
conventional overhead planimetric view. It can often give a more intuitive understanding of a
surface. The perspective views of surface can be obtained by both ArcInfo and ArcScene.

In ArcInfo, the process of drawing a surface view can be divided into four basic steps: 1)
establishing the surface, 2) orienting the view, 3) specifying where and how the view will de
drawn on the graphics page, and 4) draping and scene rendering. The procedure is described
below.

A. Establishing the surface

Arc: Arcplot
Arcplot: disp 9999 3 (open a wide graphics window)
Arcplot: mape pvdem (set the map extent)
Arcplot: gridpaint pvdem # linear # gray (draw grid using elevation values: lighter =
higher)
Arcplot: image ortho (display orthophoto image)
Arcplot: surface lattice pvdem (define which surface will be viewed)

B. Orienting the view

Arcplot: surfacedefaults (use default view orientation settings)
Arcplot: surfacetarget * (select a target position in the middle of the DEM)
Arcplot: surfaceobserver relative <azimuth> <elevation_angle> <distance> (first, try
180, 30, 2000 – see below)

C. Specifying where and how the view will be drawn on the graphics page
(we won’t set anything special here)

D. Drawing the view by overlaying objects over the surface through draping or scene
rendering.
Arcplot: clear
Arcplot: surfacedrape mesh fishnet (display surface by draping a square
grid over it)
Arcplot: surfacescene image ortho (display surface by draping an
orthophoto over it)

Change the point of observation and generate a new perspective scene. Try this for different
azimuths and elevation angles:
Arcplot: sobs relative <azimuth> <elevation_angle> <distance> (now try 120 20
2000)

2 of 11
CEE / IES 656 Engineering Applications of Land & Geographic Information Systems FALL 2005
Arcplot: ssc image ortho (display surface by draping an orthophoto
over it)

Change the elevation scale to make it more exaggerated, and view the surface again:
Arcplot: sz 3
Arcplot: sobs relative 120 10 2000
Arcplot: ssc image ortho
ArcScene has the similar functionalities for perspective views of surface.
 Start ArcScene by choosing Start > CAE Applications > Engineering > Arc Desktop 9.1

> ArcScene or clicking the ArcScene ( ) button in the 3D Analysis toolbar of ArcMap.
 Add an orthophoto (ortho.bil) and a DEM (bmndem2) into the ArcScene. Initially, all raster
datasets (grids and images) are drawn as though they were resting on a flat surface in
ArcScene.
 Practice changing perspectives using the Navigate ( ), Fly ( ), and Set Observer ( )
buttons.
 For image draping, you need to convert the DEM into TIN. To build TIN from the DEM, choose
3D Analyst > Convert > Raster To TIN… from the 3D Analyst toolbar.
 In the Convert Raster to TIN dialog, choose “bmndem2” as the input raster, set “5” as the Z-
tolerance, enter the output TIN filename, and click OK button. Note that the smaller the Z-
tolerance, the more the computing time required.
 Right mouse click on the orthophoto filename (ortho.bil) and choose Properties….
 Under the Base Heights tab, choose “Obtain heights for layer from surface” option and select
the TIN created above. Click Apply and OK buttons.
 Now you will see there are elevation variations over the terrain.

2. Viewshed
The SURFACEVIEWSHED command in ARCPLOT analyses the temporary viewing lattice created from
the current tin or lattice surface. It creates a planimetric display highlighting those areas that can
be seen from a single observation point. SURFACEVIEWSHED uses the surface viewing environment
commands to establish the surface, the location of the point of observation, and the extent of the
surface that will be considered during the analysis.

A. Establish the surface in Arcplot:

Arc: Arcplot
Arcplot: disp 9999 3 (open a graphics window – 3 makes it wide)
Arcplot: mape bmndem2 (set the map extent)
Arcplot: gridpaint bmndem2 # linear # gray (draw grid using elevation values: lighter =
higher)
Arcplot: surface lattice bmndem2 (define which surface will be viewed)
Arcplot: surfaceshade (displays the DEM using analytical hillshading)

B. Orient the view:

Arcplot: surfacedefaults
Arcplot: surfaceresolution 15 (set resolution for analyses to 15 x 15 meters
(same as DEM))
Arcplot: surfaceprojection panoramic (instead of a perspective view)
Arcplot: surfaceobserver * # 2 (in graphics window, select point of view for viewshed –
“2” is a vertical offset, so point of view is 2 meters above
DEM)
Arcplot: surfaceviewfield 180 180 (do viewshed analysis 180 degrees to left & right
of target)

Draw the view:

Arcplot: surfaceviewshed (compute & display viewshed – areas visible from
observer position & elevation will be shaded dark)

3 of 11
CEE / IES 656 Engineering Applications of Land & Geographic Information Systems FALL 2005
Try a few other observation points (valleys, hills) for viewshed calculation. Also, modify the
observer’s elevation offset for the same observation point to compare results.
Arcplot: clear
Arcplot: ssh (displays the DEM using analytical hillshading)
Arcplot: sobs * # 2 (select new observer position, and/or change elevation
offset of 2 meters)
Arcplot: svs (compute & display viewshed – areas visible from observer
position & elevation will be shaded dark)

In ArcMap, you can perform the viewshed analysis.

 Start ArcMap and ArcCatalog, and create a new point-type shapefile in the lab folder using
ArcCatalog. Note that you need to import the map projection information from any existing
datasets to this new shapefile.
 Add a DEM (or a TIN) and the new point shapefile into ArcMap.
 Activate the Spatial Analyst (or 3D Analyst) toolbar.

 Activate the Editor toolbar by clicking the Editor ( ) button.

 From the Editor toolbar, choose Editor > Start Editing. Make sure that the Task is “Create
a New Feature” and the Target is the new point shapefile.
 Click the Sketch tool ( ) button and locate a view point over the terrain as an observer’s
location.
 Choose Editor > Stop Editing and click Yes button to save your edits.
 From the Spatial Analyst (or 3D Analyst), choose Spatial Analyst (3D Analyst) > Surface
Analysis > Viewshed….
 In the Viewshed dialog, choose the DEM (or TIN) as the input surface, the point shapefile as
the observer’s point, set “1” for the z-factor and “1” for the output cell size, and enter the
output grid name. click OK button.

QUESTION 1: Make a map plot showing the viewshed analysis from the highest location within the
terrain from bmndem2. Use “5” for the output cell size. The map plot should include an
orthophoto, TIN or DEM, viewpoint shapefile, and viewshed result grid. Use appropriate transparent
values for the best display. The map plot should also include legend, title, north arrow, scale bar,
and your personal information.

QUESTION 2: Why might the viewshed results derived here not match the actual viewshed in the
real world?

3. Cross-Section Profiles
SURFACEPROFILE is an AML that automatically computes and draws a profile graph for one or more
section lines. The section lines can be entered as arcs contained in a line coverage, by interactively
pointing at the planimetric display, or by entering coordinate pairs as the command line. To
establish surface in Arcplot:

Arc: Arcplot
Arcplot: disp 9999 3 (open a graphics window – 3 makes it wide)
Arcplot: mape bmndem2 (set the map extent)
Arcplot: surface lattice bmndem2 (define which surface will be viewed)
Arcplot: ssh (displays the DEM using analytical hillshading)
Arcplot: usage surfaceprofile
Usage: SURFACEPROFILE <* | 'xmin ymin xmax ymax'> <* | line_cover | 'xy...xy'>
{profile_info_table} {sample_distance}
Arcplot: surfaceprofile * *
• Define box for profile plot: (in the graphics window, outline a box containing the top
quarter of the DEM – this is where your profile plot will be displayed)
• Define the profile line (in the graphics window, click on the middle west
(left) edge of the DEM, and then the middle east (right) edge of the DEM.
While the view window is still active, hit 9 on the keyboard to end).

4 of 11
CEE / IES 656 Engineering Applications of Land & Geographic Information Systems FALL 2005
The profile plot should show the profile for the line you selected across the center of the DEM from
west to east. Compare the profile with the hillshaded DEM surface, and verify that valleys & hills
correspond.

Arcplot: quit

In ArcMap, you can perform the cross-section profile analysis similarly.

 Start ArcMap and add a DEM (or a TIN) into the ArcMap.
 Activate the 3D Analyst toolbar.
 From the 3D Analyst toolbar, click the Interpolate Line ( ) button.
 Click the surface and digitize a line. When you are finished, double-click to stop digitizing.
 Click the Create Profile Graph ( ) button.

QUESTION 3. Make a map plot showing the cross-section profile analysis within the terrain from
bmndem2. The map plot should include an orthophoto, TIN or DEM, cross-section lines, and profile
graph. Use appropriate transparent values for the best display. The map plot should also include
legend, title, north arrow, scale bar, and your personal information.

4. Volumetric Analysis (Presented for Reference Only)

The VOLUME command calculates the area and volumetric space between a tin surface and a
specified datum value. Partial volumes can be calculated by setting the datum to any value greater
than the minimum z value for the tin. Calculating partial volumes is similar to calculating partial
areas and can be used for comparative analysis of volumes between ranges of z values.
Arc: usage volume Usage: VOLUME <tin> {base_value} {out_info_file} {z_factor}

The CUTFILL command summarizes the areas and volumes of change during a cut-and-fill operation
on an area represented by two lattices, one before and one after the cut-and-fill operation. CUTFILL
creates a lattice, a polygon coverage, and an INFO report file describing the difference between the
two lattices.
Arc: usage cutfill Usage: CUTFILL <before_lattice> <after_lattice> <out_lattice>
<out_cover> {z_factor}

In ArcMap, the Spatial Analyst toolbar has a functionality to calculate Cut/Fill only whereas the 3D
Analyst toolbar has both functionalities to calculate Cut/Fill and Area/Volume. In this lab, you are
going to practice to calculate the area and volume of terrains using the 3D Analyst toolbar due to
unavailability of comparison surface.
 Start ArcMap and add a DEM (or a TIN) into the ArcMap.
 Activate the 3D Analyst toolbar.
 From the 3D Analyst toolbar, choose 3D Analyst > Surface Analysis > Area and
Volume….
 In the Area and Volume Statistics dialog, select an input surface and choose “Calculate
statistics above place” in the Reference parameter column.
 In the Output Statistics column, click the Calculate Statistics button. Then you will see
three values, 2D area, surface area, and volume. Note that the 2D area is different from the
surface area due to terrain variation.
 If you want to save the statistics as a text file, mark “Save/append statistics to text files” and
enter a text filename, and click the Done button.

PART III. HYDROLOGIC ANALYSIS USING ARCINFO

The slope and aspect are two characteristics from which subsequent hydrologic analyses are
derived:
 Slope = steepness, in percent or angular units
 Aspect = direction in which slope face is pointing, measured in degrees clockwise from
north (0 to 360)

5 of 11
CEE / IES 656 Engineering Applications of Land & Geographic Information Systems FALL 2005
From slope & aspect, the flow direction of each cell can be determined. Flow direction is then used
to determine other hydrological parametres. However, the analytical computation of flow direction
can be thrown off by the presence of artifical “sinks” (local minimum elevations) in the surface
which don’t really exist (or do exist, but wouldn’t serve as the “bottomless pits” that the flow
direction algorithm would consider them to be). For most hydrological surface analyses, sinks are
removed before computing flow direction.
 Sink Filling = removing local depressions in a surface which don’t really exist.
 Flow Direction = which direction a cell drains downhill (1=east, 2=SE, 4=S, 8=SW, 16=W,
32=NW, 64=N, 128=NE)

Flow direction is then used to determine other hydrological parameters for a surface. We will look at
 Accumulated Flow (or Upslope Contributing Area) = sum of uphill cell area flowing into a
given cell. A cell in a major river will have a large accumulated flow. A cell on top of a hill will
have no accumulated flow.
 Stream Networks The linear features along which water flows – also derived from the flow
direction. From the stream network and flow direction, other stream grids can be produced,
showing stream order (which streams are farther downstream) and stream length (the
upstream or downstream distance at each cell in a stream)
 Watershed An area that drains to a common outlet point. Watersheds are also called
basins, drainage basins, or catchments, or contributing areas. The flow direction information
for a surface model can be used to delinate watershed boundaries.

1. Sink Filling & Flow Direction

First display the DEM we will examine:

Arc: grid (initiate grid module)

Grid: disp 9999 3
Grid: mape pvdem

6 of 11
CEE / IES 656 Engineering Applications of Land & Geographic Information Systems FALL 2005
Grid: gridpaint pvdem # linear # gray (show grayscale image of DEM: lighter = higher
elevation)
Grid: surface lattice pvdem (define surface for other viewing capabilities)
Grid: ssh (display DEM using hill shading)

To compute our hydrological products we will use the Grid module of ArcInfo. Most (not all) Grid
commands are structured as mathematical functions. Commands can be nested. For example, the
following command first computes the flow direction of our DEM, and from that it finds the sinks.
The resulting grid is the name on the left side of the equation.

Grid: psink = sink(flowdirection(pvdem))

Grid: gridpaint psink

If this DEM were used for subsequent hydrological analysis, all of the sinks displayed would be
treated as bottomless pits, where flow ended. This is almost certainly not the case, so we will fill
these sinks before computing flow direction.

Grid: fill pvdem psinkfill sink (bsinkfill = filled DEM)

Grid: pflowdir = flowdirection(psinkfill) (bflowdir = flow direction of filled DEM)
Grid: gridpaint pflowdir (display flow direction grid: each color = flow in different
direction)

2. Accumulated Flow
From the flow direction grid, compute the accumulated flow at each cell. Examine the command
usage – notice the weight grid option. Without a weight grid, the result is simply the number of cells
which flow into a given cell. With a weight grid, each cell that flows into a given cell is first
multiplied by a weight value (specific to that cell). Instead of adding cells, the weight-multiplied
cells are added to give accumulated flow.

Grid: usage flowaccumulation (F) = FLOWACCUMULATION(<dir_grid>, {weight_grid})

QUESTION 4: Describe how the weight grid option in the Flow Accumulation command
could be used to estimate total non-point-source pollutant discharge (for example,
sediment runoff) at any given point. However, what is a primary limitation of this
estimate, if only based on the assumptions of the Flow Accumulation model?

Grid: pflowacc = flowaccumulation(pflowdir)

Grid: gridpaint pflowacc # linear # gray (lighter areas = more accumulated flow)

3. Delineating (Theoretical) Stream Network

The flow accumulation grid can be interpreted to analytically derive the (theoretical) stream
network. The level of detail desired corresponds to the minimum value of flow accumulation we
consider to constitute a stream. The lower the flow accumulation value we accept, the more
streams there will be.

A graphical way to select the minimum flow accumulation value is by interpreting the brightness
display of flow accumulation. Find the flow accumulation value at the point you consider to be the
beginning of a stream:

Grid: clear
Grid: gridpaint pflowacc # linear # gray (lighter areas = more accumulated flow)
Grid: surface lattice pflowacc
Grid: surfacevalue * (click in the grid to find the flow accumulation
value – hit 9 to end)

You can also interpret a histogram of the flow accumulation grid to try and determine a reasonable
cutoff value:

7 of 11
CEE / IES 656 Engineering Applications of Land & Geographic Information Systems FALL 2005
Grid: usage histogram
HISTOGRAM <value_grid> {item} {n_levels} {max_count} {zone_grid}
Grid: clear
Grid: histogram pflowacc (can’t interpret, because distribution is concentrated. Magnify vertical
scale)
Grid: clear
Grid: histogram pflowacc # # 10

For this histogram, the horizontal axis is the range of flow accumulation values on the grid. The
vertical axis is the number of cells having a given flow accumulation. The histogram derived from
PVDEM appears stepped. Those steps probably correspond to different classes (sizes) of streams.
On the horizontal axis of the histogram, you could select a flow accumulation value which includes
the number of steps you want (you don’t need to pick a value to finish this lab).

For this lab, we’ll try two minimum flow accumulation cutoffs: 100 cells and 50 cells. Reclassify the
flow accumulation grid into two classes: stream (flowacc >= min) and no stream (flowacc < min)

Grid: strnet100 = con(pflowacc >= 100, 1) (cells with pflowacc > 100 are assigned value of 1;
others = NODATA)
Grid: clear
Grid: gridpaint strnet100

Grid: strnet50 = con(pflowacc >= 50, 1) (cells with pflowacc > 50 are assigned value of
1; others = NODATA)
Grid: clear
Grid: gridpaint strnet50

Compare the 100-cell and 50-cell stream networks. Alternate the GRIDPAINT commands to see one
and then the other (use up and down arrows on keyboard to reproduce commands).

QUESTION 5: Why might the theoretical stream network derived from the flow
accumulation algorithm be different than the same stream network mapped
independently (by visually identifying rivers & streams and determining their positions)?

4. Determine Stream Order

Stream networks are heirarchical. They have a tree-like topology, with smaller streams flowing
together to make larger ones. Given this heirarchical structure, streams are classified by an “order”
which indicates its relative position in the heirarchy.

There are two stream order classification rules available in ArcInfo, named Strahler and Shreve.
They differ in how they determine the order of a stream after two or more streams have joined.
Below is a graphical explanation:

Using the Shreve stream order classification, compute stream order both of the stream networks
delineated above (for the two different flow accumulation cutoff values):

Grid: usage streamorder STREAMORDER(<net_grid>, <dir_grid>, {STRAHLER | SHREVE})

Grid: pord100 = streamorder(strnet100, pflowdir, shreve)
Grid: gridpaint pord100
8 of 11
CEE / IES 656 Engineering Applications of Land & Geographic Information Systems FALL 2005
Grid: surface lattice pord100
Grid: sv * (identify stream order values in graphics window – write down stream order
values at specific locations, especially at points farther downstream in the
network. Compare them to values found from the next grid. Hit 9 key to quit.)

Grid: pord50 = streamorder(strnet50, pflowdir, shreve)

Grid: gridpaint pord50
Grid: surface lattice pord50
Grid: sv * (identify stream order values in graphics window – compare to previous
results)

QUESTION 6: Were the Shreve stream order classifications different for the two different
stream networks? If so, why were they different?

5. Delineate Watershed Boundaries

A watershed is defined as the area flowing to a single point. Thus the extent of a watershed
depends on which point is selected. The flow concentration point is called an “outlet” or “pour
point”.

Watershed boundaries can be determined automatically or interactively. First, we will determine a

watershed boundary by interactively selecting a pour point using the flow accumulation grid:

Grid: disp 9999 3 (make the graphics window larger)

Grid: clear
Grid: gridpaint pflowacc # linear # gray
Grid: pwshed1 = watershed(pflowdir, selectpoint(pflowacc,*)) (select a point along a
stream)
Grid: gridpaint pwshed1 (display the watershed boundary)

9 of 11
CEE / IES 656 Engineering Applications of Land & Geographic Information Systems FALL 2005
The BASIN function delineates watersheds automatically by identifying ridge lines in the surface
from the flow direction grid Ridge lines define watershed boundaries. Pour points are located at the
edge of the DEM boundary (unless there are sinks, which would also be considered pour points).

Grid: pbasin = basin(pflowdir)

Grid: gridpaint pbasin

There is also a function called SNAPPOUR which assists a user in automatically “snapping to”
(locating) the pour point with highest flow accumulation within a search area. This is important if a
set of pour points is slightly mis-registered with respect to the flow accumulation grid. A shift of
only one cell may result in selection of a pour point which is off the stream network, and hence will
not produce the desired watershed.

CEE / IES 656

Lab 11: Functional Analysis of Surface Models
Due to beginning of next lab (December 2nd, 2005)

1: Make a map plot showing the viewshed analysis from the highest location within the terrain from
bmndem2. Use “5” for the output cell size. The map plot should include an orthophoto, TIN or DEM,
viewpoint shapefile, and viewshed result grid. Use appropriate transparent values for the best
display. The map plot should also include legend, title, north arrow, scale bar, and your personal
information.

2: Why might the viewshed results derived here not match the actual viewshed in the real world?

3. Make a map plot showing the cross-section profile analysis within the terrain from bmndem2.
The map plot should include an orthophoto, TIN or DEM, cross-section lines, and profile graph. Use
appropriate transparent values for the best display. The map plot should also include legend, title,
north arrow, scale bar, and your personal information.

4: Describe how the weight grid option in the Flow Accumulation command could be used to
estimate total non-point-source pollutant discharge (for example, sediment runoff) at any given

10 of 11
CEE / IES 656 Engineering Applications of Land & Geographic Information Systems FALL 2005
point. However, what is a primary limitation of this estimate, if only based on the assumptions of
the Flow Accumulation model?

5: Why might the theoretical stream network derived from the flow accumulation algorithm be
different than the same stream network mapped independently (by visually identifying rivers &
streams and determining their positions)?

6: Were the Shreve stream order classifications different for the two different stream networks? If
so, why were they different?

11 of 11