GEO 465/565
Introduction to GIS

Lab 5 Substitute
Terrain Analysis With Raster Data in ArcView

The objectives of this lab are to explore various raster functions in ArcView and to practice importing data into Arcview.

1.0 Importing DEMs - Preparing Elevation Data for Arcview

Raster data are available in many different formats. You may often need to convert your raster data into a format that can be read by your GIS. This section discusses the converting of digital elevation models into a format suitable for Arcview projects.

There is a three step process to importing raster data into an Arcview project:

  1. Download data onto your computer system.
  2. Convert (if necessary) the data into your GIS's format.
  3. Open the converted data into a view.
In class, we have already met the DEM or Digital Elevation Model topographic data, supplied by the USGS and other agencies. For this assignment, the first thing you must do is download a 7.5' digital elevation model (DEM) from the USGS web site or other location. For a little more information about DEMs see: http://edcwww.cr.usgs.gov/glis/hyper/guide/7_min_dem
 

Question 1.0

There are three important things you need to know about a DEM (or any raster map):

Less "important" is extent: extent is easy to see when you open the DEM in your GIS. Assuming the location of the DEM is appropriate, both resolution and accuracy will help determine whether the DEM will be suitable for a particular application.  Accuracy is expressed in terms of RMSE or root mean squared error.  For a set of points, the elevation is measured both on the DEM and on an independent source of greater fidelity. As some errors are positive and some negative, the values are squared, and averaged across all tested points. This value is in the same units as the value being compared. For the elevation, just elevation is tested, and given in meters. For positional error, the RMSE is normalized over both x and y, much like the standard distance discussed in class. Note that although vertical and horizontal error are measured separately, they are actually closely related. It we don't match the right location, we usually make an error in elevation too.

The next two figures show the accuracy standards of 7.5' DEMs available for the continental U.S.

Figure 1, 7.5' DEMs with an RMSE of 0-7 meters (from USGS web page).

Figure 2, 7.5' DEMs with an RMSE of 7-15 meters (figure swiped from USGS web page).

Now that we've reviewed what DEMs are all about, you are ready to download the DEM that you will use for this lab.

Download, using "Save Link As", the Oregon DEM onto your machine. The file has already been formatted so that it will be readable by ArcView.

For future reference, you can find DEMs filed by quad name (quad name usually includes city name and state) at:http://edcftp.cr.usgs.gov/pub/data/DEM/7.5min/ . There are several other web sites that contain DEM files for regions of the United States. More DEMs for the state of Oregon may be found at www.sscgis.state.or.us/data/dems.html.

1.2 Loading the DEM file into Grid format using Arcview

  1. Start up Arcview.
  2. With the Project Window active, load the Spatial Analyst Extension  by pulling down File/Extensions..., clicking on Spatial Analyst, and clicking on OK. Note: when you save a project with an extension open, the extension will open whenever the project is opened.
  3. Make a view active. Notice that there are now three new choices when you pull down File: Manage Data Sources..., Import Data Source..., Export Data Source... (when a view is active).
  4. Pull down File/Import Data Source.... When you get a dialog box, scroll down to select USGS DEM, then click OK.
  5. Browse to your home directory, and click on the file, then click OK.
  6. You will be prompted for an output Grid Name for your dem, enter a suitable grid name (and ensure that the new grid will be written to your hard disk) then click OK. Please be patient, this conversion takes a few minutes.
  7. You are then prompted to "Add grid as theme to View", click YES. Your grid should now be added to the view.
  8. Save the results to your home directory (you wouldn't want Arcview to "blow up" on you now!).
The following screenshot shows the Oregon DEM (from a seven and a half minute quadrangle near Bend) loaded into ArcView and displayed as an active theme in a view. 

The View/Add Theme... pulldown menu will retain the same option for future queries. You will need to reset the menu to "Feature Data Source" before reading vector themes again.

Once you have the Grid formated raster files, you can copy or move them using File/Manage Data Sources...

1.3 Importing ASCII Grids into Grid format using Arcview

In addition to downloading a USGS DEM and getting it into your view, create a new view and use it to display a portion of a U.S. Defense Mapping Agency (DMA) One Degree DEM (Resolution is 3 arc seconds).

As you probably noticed, one of the options when importing files into Arcview is ASCII Grids. ASCII Grid files are the easist way to currently transfer files between other GIS and Arc/Info or Arcview (however, these files will utilize the most disk space).

Download, once again, using "Save Link As", the ASCII Grid file 90as30.asc into your home directory (by selecting File/ Save As.. ). The 90as30.asc file is a small, modified section of a DMA One Degree DEM. Once you are through importing this data into Arcview's grid format, it is a good idea to delete the file as it takes up a lot of disk space. The DMA DEM has been resampled and interpolated to 30 meters horizontal resolution.

Resolution, in the raster world, indicates how large individual cells in the grids are and how precise the data within each cell is. Horizontal resolution refers to the geographic distance individual grid cells cover. Vertical resolution refers to variation of grid cell values. For example, if a particular DEM only had elevation values 0m, 5m, 10m, 15m, 20m, ..., the vertical resolution would be 5 meters.

Question 1.1

1.3.1 ASCII Grid file format

Open 90as30.asc in the text editor of your choice. Here are the first six rows of the file:
ncols 792
nrows 836
xllcorner 224115
yllcorner 3810285
cellsize 30
nodata_value 0

The first six rows should be in this format.  The ncols and nrows indicate the number of columns and rows that are in the grid. The nodata_value indicates which value refers to cells without any data (for example, a hydrography map might have four categories: no water, intermittent stream, perennial stream, and lake; as well as a no data value indicating places where the category is unknown). The cellsize is the resolution of individual cell -- both in its north-south direction, as well as, its east-west direction. xllcorner and yllcorner represent the lower left-hand corner (not the center of the lower left-hand corner cell as some GIS specify) of the grid as shown in the following figure:

Figure 3, Small grid

The rest of 90as30.asc contains the ncols * nrows of cell values. In the file, each row contains one row's cells, although it is not a requirement.

1.3.2 Importing ASCII Grid files (at last)

It's finally time to download the file. The procedure is nearly identical to importing a DEM file and will be repeated for your convienence:
  1. Start up Arcview (if necessary).
  2. With the Project Window active, load the Spatial Analyst Extension with the by pulling down File/Extensions..., clicking on Spatial Analyst, and clicking on OK (if necessary).
  3. Make a view active. Notice that there are now three new choices when you pull down File: Manage Data Sources..., Import Data Source..., Export Data Source... (when a view is active).
  4. Pull down File/Import Data Source.... When you get a dialog box, scroll down to select ASCII Raster.
  5. Browse to your home directory, and click on the file 90as30.asc, then click OK.
  6. You will be prompted for an output Grid Name for your dem, enter a suitable grid name (and ensure that the new grid will be written to your hard disk), then click OK.
  7. You are then asked if to import "Cell values as integers?", click NO.  Please be patient, this conversion takes a few minutes.
  8. You will now be prompted to "Add grid as theme to View", click YES. Your grid should now be added to the view.
  9. Save the results to your home directory.

1.4 Managing Grids in Arcview

The Grid format in Arc/Info requires software in order to copy, move or delete the Grid themes. Arcview includes this functionality in File/Manage Data Sources...

Grid files in a project are stored in the working directory (File/Set Working Directory...). Make sure your working directory is in an appropriate location.  Your files will only be saved under two circumstances:

Since you will be making many temporary grid themes, remember to save your project or Theme/Save Data Set whenever you generate a useful grid theme.

1.5 Projecting DEM themes in Arcview views

Grids need to be projected just the same as vector maps. USGS DEMs and many US maps use the universal transverse mercator (UTM) coordinate system to locate the raster cells. (See Chapter 2, section 2.3.2 of your textbook, "Getting Started With GIS", for a description of the UTM coordinate system). After you have added a DEM grid theme to a particular  View, you should set the view projection to UTM (in View/Properties...). Your only complication is determining which UTM zone your data is in.The 90as30.asc and other ASCII files available for this lab are located in UTM Zone 11.

As discussed in the text, UTM zones are numbered from 1 to 60, which each zone 6 degrees of longitude wide. The zones begin at 180 degrees west longitude and extend eastwards. Thus, zone 1 is for data from 180 degrees W to 174 degrees W, zone 2 is from 174 degrees W to 168 degrees W... Set the view projection to the appropriate zone for your grid data.

2.0 Looking at continuous maps

In this section, you will explore your elevation theme. The first thing you'll want to do with any grid theme is to look at its properties. With the elevation theme active, pull down Theme/Properties....

Question 2.0.a

Question 2.0.b

Next, you'll want create a pleasing image of the elevation theme. Arcview provides several ways to create good images for continuous maps: contours and hillshading.

2.1 Contour Lines

Contours, traditionally used by Geographers, can be made with Surface/Create Contours....  Go ahead and make a contour map of your data.

Choose a contour interval that provides you with as much
information about the terrain as possible.  Be careful not to
obscure the map with too many contour lines.

Note: there may be a button on the toolbar named "contour". This button is NOT the Surface/Create Contours... function. The contour button will create a shapefile of a single isoline with the value of a spot clicked on the map.

2.2 Hillshade Images

Creating a hillshade image of a continuous scaled grid theme is a two part procedure. First, generate a hillshade map for the active grid theme by using: Surface/Compute Hillshade.../OK. Second, double click on the active grid theme and click on Advanced... in the Legend Editor. Scroll down to the generated hillshade map to define the Brightness Theme and hit Apply on the legend Editor window. As of Arcview 3.1, this is the closest Arcview gets to a three dimensional surface, -- but it makes a decent output map.

Question 2.2.a

A nice feature of hillshading in Arcview is that since hillshading is part of the legend, the DEM will retain the hillshading effect whenever you want to view the DEM.

Question 2.2.b

2.3 Descriptive Functions for a single theme

Arcview has several ways of quantitatively describing information in a raster map. One is the histogram function (in the toolbar). Another is the Statistics button in the Legend Editor. Both do the obvious.

3.0 Raster Analysis

This section will provide several examples of raster analysis in ArcView, as well as discuss general issues important to raster analysis.

3.1 Analysis Properties

As mentioned earlier, cell resolution should be considered before doing any analysis. With an active view, click on Analysis/Properties.... You can then choose the resolution, and extents of analysis using various techniques. For example, if you zoom in to the extent of the area you want to analyze, pull down Analysis Extent and click on Same as Display (NOT Same as View which sets the extent to the extent of the themes in the view). It is a good idea to keep track of how long a particular application takes for each type of analysis based on the number of cells (ncols * nrows) in the analysis extent. That way, you don't start running an application that will take 14 hours when you only have time for 18 minutes. Arcview is SUPPOSED to have a stop button in the lower right corner of the Arcview window. However, it might be minutes before the button appears for an analysis.

3.2 Slope for Elevation (and all continuous maps)

The first derivative of a surface is often more important than the values of the surface itself. For example, when hiking in the Northeast, the absolute elevation doesn't matter at all (we're under the same relative stress at 350 meters as we are at 150 meters). However, hiking up a 30% slope is far more difficult than a 3% slope.

3.2.1 Making a Slope map

To make a slope map in Arcview, make your surface map (in this case, your USGS DEM) active. Pull down Surface/Derive Slope. And that's it!

3.2.2 Understanding Slope

Do the following commands:
  1. Rename the existing slope map (from 3.2.1) including the extent resolution value in the name.
  2. Double the resolution using Analysis/Properties....
  3. Make another slope map with this greater resolution.
  4. Rename the slope map from step 3 with the extent resolution value in the name.
  5. Make histograms from both slope maps (histogram button).
Compare the two histograms.

Question 3.2.2.a

3.3 Distance and Proximity

You are  playing 18 rounds with Tiger Woods, Michael Jordon, and the Governor of Illinois. After several spirited minutes of discussing geography with the governor, he asks if its possible to estimate how much of Illinois is relying on businesses in other states when purchasing computer supplies and equipment. You whip out your laptop computer, which contains Arcview, and proceed to put together the following analysis, while Michael struggles in a sandtrap.

3.3.1 Distance analysis

Your first analysis is to roughly estimate the likelihood that computer users in various parts of Illinois use mail order to buy computer equipment. Use the following assumptions: Prepare your analysis by doing the following:
  1. Create a new view for this section.
  2. Open a USA cities shapefile and a USA states shapefile in this new view (Try the directory c:\esri\av_gis30\avtutor\arcview\qstart).
  3. Join (c:\esri\av_gis30\avtutor\arcview\qstart) cidemog.dbf to the cities shapefile.
  4. Select all cities with a population greater than 20,000.
  5. Load the Spatial Analyst if its not already loaded.
  6. Zoom into an area just larger than Illinois including any selected cities that computer users in Illinois might travel to.
  7. Set the Analysis Extent and Resolution using Analysis/Properties.... Set the distance to kilometers.
  8. Set the resolution to one kilometer. Make sure the number of rows and columns are reasonable numbers (1,000,000 cells can be quickly done on the current computers).
And now your ready for some analysis! Your analysis is in two parts: generating a distance map and calculating the likelihood that people within a distance will purchase from a computer store.
  1. With selected cities active, pull down Analysis/Find Distance
  2. Pull down Analysis/Map Calculator to run the calculation:
(([Distance to Cities.shp] < 160.AsGrid) * (0.8.AsGrid - (0.005.AsGrid * [Distance to Cities.shp])))

This represents the equation:

likelihood purchasing supplies = 0.8 - 0.005 * (km from city).

Read the help section on "Map Calculator" to see what the map calculation does. The Map Calculator is probably the most important Grid function that Arcview offers. It is likely that 90% of Grid analyses can be done in the Map Calculator (there should be a program such as the Map Calculator in any robust raster GIS - e.g., GRASS has r.mapcalc).

Question 3.3.1.a

Question 3.3.1.b

3.3.2 Proximity analysis

Michael is out of the sandtrap, but the governor is having a horrible slicing problem involving a nearby lake. Nevertheless, having the world's highest paid geographer with you helps!

After accounting for mail order purchases, you realize you can also estimate which cities computer users will shop at. You go back to your laptop. Since your goal is to determine whether a computer user purchases their supplies in Illinois, when you do a proximity analysis, you request the computer use the state name field of Cities.shp for the output map from Analysis/Assign Proximity. This way, your output map will contain areas named after the states the cities are from, completing your analysis.

Question 3.3.2.a

Last updated 6 February 1999
Originally developed by Chuck Ehlschlaeger,, CUNY-Hunter College
http://dusk.geo.orst.edu/gis/demlab/

Return to GEO 465/565 Labs