GIS Lab Projects‎ > ‎

Lab 08: Morphometric Analyses


Overview


The objective of Lab 8 is to gain proficiency in using elevation models to analyze surfaces or landscapes to produce useful and interesting information about an area.  In this case, we use elevation data from the Salt Lake County area to delineate the Big Cottonwood Canyon Watershed and produce a number of analyses to help characterize the watershed.  The data utilized, a 10 meter resolution DEM and streams from the National Hydrography Dataset, are both publicly available through the Utah Automated Geographic Reference Center (AGRC).  

Learning objectives

1.  Illustrate advanced skills beyond making pretty maps.
2.  Delineate a watershed boundary from elevation data.
3.  Derive a drainage network including the delineation of stream orders.
4.  Apply common morphometric analyses to characterize the watershed.

Context

For this lab project, we are once again working for the fictitious employer "Big Cheese Watershed Management."  Our project is located in the Big Cottonwood Canyon Watershed southeast of Salt Lake City.  The maps on the page below show the project location in Salt Lake County and the State of Utah to provide some context for the area that will be further analyzed through the course of the following steps.

To view a larger version, click on the image below.


Task 1: Delineate the Big Cottonwood Watershed.

Step 1:  Fill the pits in the DEM.

Digital Elevation Models often have "sinks" or "peaks" as anomalies that can create discontinuity in the surface.  If not corrected, these anomalies will affect the accuracy of surface flow models and many other analyses that rely on the DEM.  

Image from Wheaton, 2014


Step 2:  Calculate Flow Direction.

A watershed, also known as a catchment area or drainage basin, is an area of land where all  water that falls upon and flows across the surface drains toward into the same common body of water such as a river, stream, lake or aquifer.  Since water reliably flows downhill, we can use digital elevation models to calculate the direction that water should flow.  The flow direction tool, included in the ArcGIS spatial analyst extension, calculates the steepest downward descent (and thus the direction of flow) for every cell in the surface grid.

Surface Flow Direction

The image below shows the output of the flow direction raster each color represents a different direction of flow.  Click on the image to view a larger version.



Step 3:  Delineate the Catchment Area (ie watershed).

Using the flow direction raster that was created in Step 2, we can then calculate flow accumulation.  The Accumulation tool in ArcMap will calculate the number of other cells within the surface grid that ultimately drain into any particular cell.  By defining a threshold value and reclassifying this information to show only the cells with the highest accumulation values, we can begin to see the general drainage pattern of the watershed.  

Surface Flow Accumulation

The image below shows the output of the flow accumulation tool near the mouth of the canyon.  Click on the image to view a larger version.

Establishing a Pour Point

By using the flow accumulation data (above) in conjunction with the topography, we can then determine potential locations to establish a "pour point" that will be used to delineate the watershed boundary. These are typically points at the edge of the grid or just downstream of major confluences.  By looking at the topography (in this case represented by a hillshade), we can visually estimate where the most appropriate pour points are likely to be located based.  In this case different pour points were established where the line formed by the accumulation data exits through the mouth of the canyon.  

Delineating the Watershed

The watershed tool was then used to delineate the watershed area using the flow direction raster from Step 2 in combination with the different pour points.  The output rasters were then inspected in relation to the topography to select the one that best represented the watershed. 

The map below represents the final delineated watershed developed through the steps listed above.  To view a larger version, click on the image below.


Verifying the accuracy of the watershed delineation

To verify the accuracy of the delineation, the area (square kilometers) of the watershed was calculated and compared to the Big Cottonwood Canyon Watershed available from the AGRC. The delineated watershed has an area of 129.308 square kilometers while the file available from the AGRC covers 129.24 square kilometers.  These two numbers are really very close and represent that our delineation was fairly accurate.  The most likely difference between the two is the resolution of the data that was used to delineate the watershed. 

The figure below maps the delineated watershed boundary produced in this task on top of the watershed boundary provided in the AGRC dataset.  The AGRC boundary had to be made quite thick to show up underneath the delineated boundary - which is a good thing. The most significant difference is at the mouth of the canyon on the south side, but as shown in the map, the difference is barely discernible unless you're specifically looking for it.

To view a larger version, click on the image below.



Task 2: Perform a Stahler Stream Order Classification identifying the drainage network in the watershed.


Step 1: Adapt the flow accumulation data from Task 1 to better represent drainage pattern

This step utilizes the same surface flow accumulation data that was generated to establish the pour point in Task 1 above.  The primary objective for this task is to represent the drainage network within the Big Cottonwood Watershed.  In Task 1, a threshold value of 500,000 was applied to identify the primary drainage area near the mouth of the canyon.  In this case, we are interested in a more complete perspective of the drainage pattern.  Therefore, a new threshold value had to be identified.  This was done by overlaying the accumulation raster layer onto streams data from the National Hydrography Dataset.  The symbology of the accumulation raster was adjusted to represent different threshold values.  Higher values show fewer drainage features, while lower threshold values capture more (and longer) branches of the drainage network.  Through a process of trial and error, a threshold value of 3000 was selected as an appropriate balance.  

The map below shows the drainage network at a threshold level of 3000 overlaid on top of the NHD streams data.  In some areas, the streams extend further up into the watershed than the identified drainage lines.  In other cases, we have captured more of the network than was evident from the NHD data and even identified branches that didn't show up as streams at all.  The reason for these differences is relatively simple.  The NHD data maps streams as water features, while we are essentially mapping potential drainage accumulation.  Ultimately, there are many other factors such as soil texture and permeability, precipitation patterns, etc., that also determine whether water actually accumulates on the surface to form a stream. 

To view a larger image of the drainage pattern in relation to the NHD stream data, click on the image below.


Step 2: Classify stream segments according to the Stahler Stream Order hierarchy.

A common method of classifying stream segments within a watershed is the use of Stahler stream orders, which identifies a hierarchy of tributaries.  Two first order tributaries create a second order tributary, the confluence of two second order tributaries creates a third order tributary and so on.  To quickly generate this hierarchy for Big Cottonwood Canyon, the stream order tool within ArcGIS spatial analyst applied to the flow direction and accumulation raster data.  The following figures represent the final classification of the four orders that were identified for the threshold value of accumulation (3000) described above.  To enhance the display of this data at the scale of the watershed, the raster output of the stream order tool was converted to polyline features and broken into individual layers representing each stream order classification.  This also allowed for the use of the geometry calculator and summary statistics to determine the total length of the tributaries that fell into each of the four classes.  These lengths are shown in the table to the right of the map below.

To view larger versions, click on the appropriate image below.  A PDF document of the map can also be downloaded by clicking here.























Task 3: Perform common morphometric analyses to help characterize the watershed.

In addition to delineating watershed boundaries and illustrating drainage patterns, there are many other analyses that can be done to further characterize a watershed using elevation data. In this task, we will calculate slopes, illustrate the importance of curvature, and calculate some zonal statistics that help summarize important information about the area.

Step 1:  Slope Calculation

Slope is an important characteristic of a watershed that affects the velocity of surface flow, vegetative cover, erosion, and many other aspects that become important to watershed management.  The map below represents different categories of slope percentages within the Big Cottonwood Canyon Watershed.  From the map, it is obvious that slopes are generally steeper in the lower (western) portion of the watershed than in the upper elevations to the east.



Step 2:  Calculate curvature within the watershed.

Curvature is defined as the second derivative of a surface, or the slope of the slope.  Curvature is an important consideration in watershed hydrology because it affects how water moves across the surface - water may converge, diverge, accelerate, or decelerate as it moves across the surface.  For cartographers, curvature can also be used to enhance different features within mapped data.  

The graphics below illustrate the differences between curvature, plan-form curvature, and profile curvature.  For all the graphics, the black arrows represent how the pattern of surface flow is affected by the curvature of the surface.

Plan-form Curvature

Image from Buckley, 2010

Profile Curvature

Image from Buckley, 2010.


Combinations of Curvature

Image from Buckley, 2010

Step 3: Terrain Mapping using Curvature in Conjunction with Hillshading

The images on the page below compare how the visual appearance of terrain is altered by overlaying plan, profile, and combined curvature raster layers.  The plan curvature overlaid onto the hillshade emphasizes the horizontal features of the terrain surface.  An example where this may be beneficial would be in mapping areas with distinct geologic strata.  The profile curvature emphasizes the vertical features of the terrain such as ridges and canyons.  Both plan and profile curvature can be combined and represented simply as the curvature of the landscape or terrain, but their ability to enhance certain land forms will be diminished.  This can be done using the curvature tool in ArcGIS or, for mapping purposes, be done manually by overlaying the raster layers and adjusting transparency until the desired effect is achieved.

To view a larger version of this comparison, click on the image below or download the PDF here.


Step 4:  Zonal Statistics

To facilitate comparison between different watersheds, the average slope and curvature of the watershed can be useful.  In this final step, zonal statistics functions within ArcGIS were used to make some of these calculations.  The following table shows the statistics for slope, plan-form curvature, and profile curvature.




References


Buckley, Aileen (2010).  "Understanding Curvature Rasters." ArcGIS Resources Blog.  Available at:  http://blogs.esri.com/esri/arcgis/2010/10/27/understanding-curvature-rasters/

Wheaton, Joe (2014).  Advanced GIS Course Website.  Utah State University.  Available at:  http://gis.joewheaton.org/home

Utah Automated Geographic Reference Center (AGRC).  Utah Department of Technology Services.  Available at:  http://gis.utah.gov/


Ċ
Matt Coombs,
Mar 3, 2014, 3:17 PM
Ċ
Matt Coombs,
Mar 3, 2014, 11:09 AM
Ċ
Matt Coombs,
Mar 4, 2014, 8:47 AM
Ċ
Matt Coombs,
Mar 3, 2014, 9:32 AM
Comments