In-Class Exercise 5 [SRTM DEM]
A. Data Acquisition
B. Data Processing
C. Generate Contours
Acquiring 30m resolution topographic data.
Shuttle Radar Topography Mission (SRTM) Links to an external site. data has been available for most of the globe at 3 arc seconds (~90m) cell resolution for over a decade. In 2014 the USGS decided to release the 1 arc second data which works out to about 30m cell resolution. SRTM is useful in it's coverage and consistency, although there have been some issues with data voids Links to an external site.. Certain countries may have higher resolution products available. For example the United States has the National Elevation Dataset (NED) Links to an external site. for the contiguous lower 48 states at 1/3th arc sec (~10m) and some portions at 1/9th arc sec (~3m) cell resolution.
|
US lower 48 states DEM available from NED at ~3m resolution in 2017 |
Select regions have DEM products at resolutions even higher than 3m typically these are LiDAR derived. Check Open Topography Links to an external site., local and national agencies, and commercial vendors for these products.
The steps outlined below are an introduction to acquiring elevation data that will work for most of the planet. If you are doing further work with elevation data in Arcmap using other DEM sources (including LiDAR) it is recommended you read ArcGIS Desktop Help: Managing Elevation Data, parts 1,2,3. Links to an external site.
 SRTM DEM (~30m resolution) extends north as far as 60° latitude |
A. Data Acquisition
1. Login to Earth Explorer
Next, we will acquire topographic data for your study area from http://earthexplorer.usgs.gov Links to an external site. (SRTM 1 arc sec or ~30m, higher res within US)
Login (top-right) using the EarthExplorer account created earlier.
2. Define your study area bounding box.
When in doubt it is better to get slightly more data than you think you'll need with terrain data (lest your analysis falls of the edge of the world!)
Zoom map into study area. Make sure the outline of your map window reflects your desired study area.
On left side where it reads "Coordinates" in blue choose Decimal if your coordinates are in Decimal Degrees.
Click "Use Map"
Zoom back one level to see the selection box you created. If you're satisfied with the study area.
Note other options for selection such as "Date Range". These can be important for datasets like satellite imagery.
At the top of the window click "Save Criteria" and name your study area.
3. Determine datasets
Click "Data Sets > " at the bottom.
Open "Digital Elevation" and inside that "SRTM"
Check "SRTM 1 Arc-Second Global"
Note the various datasets. The other 30m resolution data of potential interest here is ASTER Global DEM with it's limitations (2009 evaluation Links to an external site.) but we will use the newly available SRTM 1 Arc Second.
4. Download data
Click "Results >"
Depending on the extent of your study area you'll have a number of tiles available.
Click the Footprint icon to see the extent of each tile
Other options to the right of the Footprint are Preview, Compare Browse, Metadata Browse, Download, Bulk Download (add to queue) and Delete.
Click Download. Bulk Download requires installing extra software but it makes sense with 10+ tiles.
Click Download on each tile and choose BIL 1 Arc-Second
Save each to your Project folder in a subfolder called "Sources" that you create. X closed the File Format window.
Once complete return to the File Explorer and browse to your Documents/Project/Sources folder and Right Click each and choose "7 Zip > Extract to \FILENAME\".
You should ideally create a Master "Project" folder, within that folder create subfolders: DEM, TEMP, and SOURCES
Look inside one of the folders. Check file sizes. Can you tell which one is a Worldfile?
B. Data Processing
1. Convert BIL to GRID
New Blank Map in Arcmap
Add all BIL files to a study area. Create Pyramids: Yes
Look at the coordinates in the lower right -- do they make sense?
Recall EarthExplorer metadata said WGS1984 Decimal Degrees. The project should have inherited the system of the first layer added. This is known as "Unprojected" in Arcmap because it uses angular (degree) units.
Examine the layers and the seams between your tiles. Are they properly tiled or are there gaps? With raster files it's important to do preliminary steps like joining tiles in the original projection so that the tile alignment is as close as possible.
Are there white gaps in the data? SRTM had problems with "data voids". These have been filled in many cases but it's possible we'll have to do some processing in your study area.
Right click each layer and choose Export Data > Export...
Leave the Extent, Spatial Reference and Cell Size alone.
Why do you think Cell Size cx, cy is "0.00027777" -- what kind of units are these?
Note the Spatial Reference in the box is Geographical Coordinate System (GCS) and No Compression.
Location: Click the folder on the right. Browse into your Project folder. Click New Folder tool at the top and call the folder "DEM". Select DEM and Click "Add" so the output is saved in that folder. Confirm the path ends in \Project\DEM
Name: S1_s16_w072 (or whatever your coordinates are)
File: GRID
Click Save. Add to Map? Yes
Note that if you have a large area to convert it is possible to automate this in the Toolbox with the Batch command.
Save your project MXD as DEM_GCS.mxd at this point.
2. Mosaic Tiles
Combine tiles with tool Mosaic to New Raster tool found in
ArcToolbox > Data Management Tools > Raster > Raster Dataset > Mosaic to New Raster Tool
- Add all layers
- Output Location: Choose DEM folder and click "Add"
- [ProjectName]SRTM1 (use no extension to get a GRID output)
- Leave Spatial Reference blank so it stays with GCS WGS 1984
- Pixel Type: 16_BIT_SIGNED (the format of the source)
- Number of Bands: 1
Why is that? What else do you think the Mosaic to Raster could be used for that would involve more bands?
Examine the results and especially the locations that were seams between the BIL tiles by zooming in to > 1:100000 to inspect.
3. Convert to Metric units in UTM
Geographical units are useful for global datasets but to take measurements and assess patterns in metric space using these data we need to project them to linear units (meters). Universal Transverse Mercator (UTM) is the most common system used for local studies by field scientists.
Choose
ArcToolbox > Data Management > Projections and Transformations > Raster > Project Raster
- Input raster: SRTM1
- Output Location: Project Folder / DEM /
- Output Raster Dataset:
- [ProjectName]_S1_UTM
- Output Coordinate System:
- Projected Coordinate Systems > UTM > WGS 1984 or NAD 1983 > [choose appropriate zone]
- If you're not sure of the UTM Zone of the study area see this interactive map Links to an external site.. A Shapefile showing UTM Zones Links to an external site.can be added to your project.
- Right-click "Add to Favorites" when you've found your appropriate UTM Projection. Then choose it and click OK.
Leave the rest of the settings at their defaults. Note the X and Y Cell Sizes aren't exactly 30 m.
Save your GCS MXD.
C. Generate Contours
Contours or isolines are an effective way of showing relief. Contours are typically more of a presentation feature than an analytical layer in GIS and as such there the priority is on aesthetics and communication. If it means moving and smoothing segments to communicate better then so be it.
1. Create contour vectors
Open a new blank MXD Map file so that the map is in the appropriate UTM projection.
Add the Mosaic DEM in UTM as your first layer. Inspect the lower right corner of the map window to ensure the coordinates are Metric and that they're reasonable for your area.
ArcToolbox > Spatial Analyst > Surface > Contour
- Input Raster: [Project]_S1_UTM
- Output Polyline Feature: Accept default location
- Contour Interval: 50 [25 to 100 depending on terrain. This can take a while]
- Base contour: blank
- Z factor: 1
If this takes too long you may have selected too many contours for your computer to handle. Either be patient or cancel.
When the procedure completes open the attribute table and note the different fields. CONTOUR contains the metric value of each vector contour line.
2. Symbolize
Index contours get numerical labels and help the eye distinguish major contour intervals. These typically occur on a round number every 4 or 5 contour lines. If you made 25m contours consider putting an index contour every 250m interval.
Create a Index copy
- Rename Contour layer to Contour_25 [whatever your interval was]
- Create a copy of the Contour output:
- Right click Contour in table of contents
- Edit > Paste
- Rename Pasted layer "Contour_250 Index" [interval * 5]
Note that this creates a second reference to the same original vector file it does not take double the space on your hard drive.
Definition Query to show only some of the vectors
Right click the Index layer and choose Properties.
Choose the Definition Query tab
Paste the following into your query box [for a 250m index interval]
Floor(Contour/250) = Contour/250
Symbolize appropriately
Right click layer > Properties > Symbolize
- Make the Index layer 1.5 px wide and dark brown
- Make the Contour[interval] layer .6 wide and tan color.
Or choose other values going for aesthetics and communication.
3. Label Index Contours
Right click the Index layer and choose Properties > Label...
Check "Label Features in this layer"
Label Field: Contour
Consider changing the color to dark brown and bold face.
Click Placement Properties... > Position: check only "On the Line". OK.
Since these are lying directly on the contour lines we'll need a buffer mask behind the text. Click Symbol... and Edit Symbol... then Mask tab. Select Halo and give it a Size: 1.5. OK, Ok. Ok.
Note that there is a "Scale Range..." option. Perhaps you don't want the contours to have labels out past 1:500,000 scale?
File > Export a map as a PDF at 300 dpi and turn this in as your assignment.
4. Further Refinements
These are optional and not due for this In-Class Exercise. Things to consider for better representing terrain in your study area:
- Look into choosing better symbology for your UTM GRID raster layer. I often use a dark brown to light brown as you saw in the Callalli terrain. This is something that can take some experimentation. Raven Maps Links to an external site.makes beautiful relief maps and they a palette sequence that uses reddish shades in the mid-upper elevations.
- A hillshade effect can be added to the raster symbology or generated separately. The Callalli_DEM_hs was a separate Hillshade layer with transparency.
- The index contour labels are often repetitive. You can explore the Properties > Labels > Placement Properties... options for showing fewer labels per object. For fine control convert to annotations or "dumb" graphics and place your labels carefully.
- you can eliminate the clutter of many tiny contours by looking at the their length in the attribute table "Shape_Length" (Shapefiles can have these but they must be added). One method for removing these is sorting on the Shape_Length and removing the tiny contours under, say, 100m in length.
There are tools for cleaning up contour lines such as "Simplify Line" (remove vertices) and "Smooth Line" which has a threshold.
Give credit for these data
These US Government data are in the public domain, but in publications and websites using these data you should include a line thanking the source. It might read something like this
"We acknowledge NASA for providing SRTM elevation data"