Introduction
Literature Review
Rivers are complex ecosystems dependent on intricate relationships between physical geography, geology, ecology, the atmosphere, and climate. Rivers can be assessed at the basin level, which encompasses the entire riverine system, or locally at the catchment level1. In the Denver Metropolitan area, local catchments are the headwaters that feed into the Missouri-Mississippi drainage basins, which empty into the Gulf of Mexico2.
Maintaining the intricate relationships of the catchment provides resources to people and other organisms with food, shelter, scenery, and clean water. A healthy watershed provides for healthy stream morphology that can respond to changes in precipitation events and other natural hazards. A stream channel and the associated floodplain will respond to changes in flow by changing its shape—the depth, width, sinuosity, and bedform—to control the discharge over time.
Hydraulic forces and prominent theories
In addition to precipitation events, channel morphology responds to physical processes, such as erosion and deposition. Both processes occur in relation to the discharge; increased discharge in a stream has the capability of moving more sediment, thus eroding the landscape. Reduced discharge suggests a slower flow, where sediment can be deposited.
Streamflow can be considered two different ways:
1) Hortonian overland flow and saturation flow.
2) Hortonian overland occurs when there is high precipitation and little storage capacity for the precipitation to drain to; therefore the water flows swiftly to and through the channel. This type of flow is prominent in impervious, less vegetated, or semiarid areas. When water reaches a saturation point in the soil runoff toward a river channel occurs (saturation overland flow). This type of flow occurs near slope bases, where vegetation and soil entrainment accumulate and can trap the water before it runs off. The total streamflow between stream gages is measured as volume, known as discharge. Increased discharge can move larger sediment loads, which erodes the surface of the landscape.
Notably, the type of soil and its physical characteristics, such as pore space, texture, and molecular properties influence how much water is retained, and how much is released as overland flow. The amount of vegetation along the river and floodplain aids in determining the rate of water infiltration and overland flow, and limits erosion3.
Why analyze river depth in watersheds?
River depth is an important variable4 to consider when assessing stream health. River depth depends on several factors, including slope, soil, and geologic forces such as erosion, deposition and suspension of soil particles. Each of these play a part in creating a dynamic equilibrium, where all forces are balanced and inputs to the watershed equal the outputs. The South Platte Headwaters, Upper South Platte, and Middle South Platte—Cherry Creek watersheds5 all feed into the Platte River in and upstream of the Denver Metropolitan area. Given the drought conditions Colorado faces today, it is important to consider the watershed health as the metropolitan area expands; variables such as ecological health, water quality and quantity, and stream health will be critical to monitor to maintain the delicate ecosystem services provided.
Questions Being Addressed
This paper seeks to answer the following question: How does channel morphology differ between the Piedmont mountainous region and the High Plains region? This question is further broken down by comparing the differences of three separate variables: river depth, soil water content, and water discharge.
Methodology
Data
30-meter digital Elevation Models (DEM) raster files, the National Hydrography Dataset containing all river centerlines, channels, floodplain, and dam information, and geology data were downloaded from the United States Geological Survey (USGS). Soil data were downloaded from the United States Department of Agriculture, Soil Conservation Service (USDA), developed in tandem with the Colorado Division of Water Resources. Point data were generated in ArcGIS Pro using the Create Transects tool in order to for points to be plotted on lines every 5 km along the Platte River.
The data were streamlined into the Study Points attribute table. Local soil and geology information were recorded, as well as latitude and longitude of the point, and the elevation was recorded.
Data Preparation
Once all data were configured in ArcGIS Pro, transects were generated along the Platte River every 5 km using the Create Transect tool. Points were added to the intersection between the transect and the Platte River centerline. Additional points were added along the left and right floodplains elevation for all points was recorded in the point attribute table. The mean depth was recorded by subtracting the elevation of the floodplain from the river center point. The geology and soil type that the study points resided in were also recorded in the attribute table. The soil dataset contained information on available soil water content, which was joined to the table.
To prepare for statistical testing, the Piedmont and High Plains groups were determined by using data on soil, geology, and elevation. Piedmont areas were defined as areas with steeper slopes, granite bedrock, and Sphinx-Legault rock outcrop soil. These qualities were present along the Platte River upstream of Chatfield Dam. High Plains areas were defined as areas with flatter slopes (elevation difference was less than 1000 feet along the stream segment), Fluvaquents-Alda-Bankard soil, and alluvium (geology). These specifications were met in river segments downstream of Chatfield Dam.
Statistical Testing
Once the data were cleaned, a one-tailed two-sample t-test was used to compare the average depth of the Platte River in Piedmont and High Plains environments. A one-sample t-test was used to analyze if there were any relationship in one group over another. Additional variables of soil water content and discharge were added for analyzation. A linear regression was completed for discharge to evaluate water volume as a function of distance.
Complications
The DEM files showed the elevation of the total study area, but no attribute table was available for reference in the dataset. To derive elevation for the Platte River study points, additional points were plotted on the edge of the floodplain and subtracted from the center point height to yield the depth of the river.
Results
Statistical Results
Piedmont stream segments had depths averaging 0.8 meters with more variance between the averages, depending on local environmental conditions. River depth in the High Plains averaged 0.45 meters, with less variance across sample points. Percent available water content (AWC) in the Piedmont section had a mean of 0.74 with a large variance around the mean whereas the High Plains mean AWC was 0.11 with very little variance around the mean. Neither t-test could be confidently determined as statistically significant as both tests had p-values greater than 0.05.
See Maps and Graphics Below
Limitations
The results above are limited to a relatively small section of the Platte River, which invokes spatial autocorrelation. This principle exemplifies the idea that areas closer in space tend to be more similar than areas further apart. In this context, the 50-km river segment is continuous and traverses through the Rocky Mountains and High Plains area in the Denver Metropolitan area. In future work, further analysis should consider the length of the river segment, whether different segment groups should be compared, or whether human impacts has significantly altered the landscape and rendered a significant difference that impacts the analysis. Additionally, due to the limited area of interest, it is important to consider the results only as they pertain to this area. Expanding the boundaries could change the results.
| t-Test Results | ||||||
|---|---|---|---|---|---|---|
| Piedmont Mean | Plains Mean | Piedmont Variance | Plains Variance | t-Value | p-Value | |
| Mean Depth (m) | 0.803218223 | 0.457565842 | 2.391007027 | 0.522248436 | 0.805249567 | 0.214646659 |
| AWC (%) | 0.74871875 | 0.107604 | 6.447781532 | 0.00091364 | 0.976742396 | 0.168387363 |
Figure 4: Discharge in the Platte River as a function of length from beginning of study area. The discharge decreases due to
the flow being restricted in Chatfield Dam.
Figure 1: Map of the Platte River study area, including all study points and watershed boundaries.
Figure 2: Geology type was used also to determine the divide between the Piedmont and High Plains groups. Beginning at Chatfield Dam, south of Littleton, the soil changes from granite to alluvium. At that point, Piedmont is considered upstream of Chatfield Dam with granite underlying geology and High Plains is downstream of Chatfield Dam with underlying alluvium geology.
Discussion and Conclusion
Data Interpretation
The t-test results indicate that there is no statistically significant difference between Piedmont and High Plains river depths. This was not altogether surprising because the area of interest followed approximately 150 km of the Platte River, which may not be enough distance to result in statistically different mean depths. The mean depth in the Piedmont section was 0.803 meters ±2.39 meters. The wider variance around the mean could be explained by the bedform, which is made up of larger rocks or boulders, rather than fine sand or sediment. The deeper distance to the bottom of the stream suggests that there is more sediment transport (erosion) due to the increased discharge7 volume in the Piedmont segment. The mean stream depth was 0.457 meters ± 0.009 meters in the High Plains. This segment of the stream is slightly wider than the Piedmont headwaters8, meaning that the river morphology changed the way it responds to the discharge. The wider and shallower form allows the water to slow down because of friction force from the streambed. The mean available soil water content was 74.8% in the Piedmont and 10.7% in the High Plains. The soil water content depends on pore size, topography and vegetation9. The available soil water content was added as an additional variable to compare to grasp where water is contained overall within the watersheds. Given that 74.8% of water is available in the Piedmont, this suggests more water is retainable in the area. Sphinx-Legault rock outcrop, however, is sandy10, offering the idea that there is more vegetation that can retain the water in the Piedmont area. 10.7% soil water content in the High Plains submits the notion that less water can be retained in the fluvaquents Bankard-Alda soil, which is a sandy type of soil11. The stark difference could be indicative of less vegetation along the river channel. The amount of vegetation within the floodplains can affect how channel morphology reacts to streamflow, as increased vegetation along channel banks reduce the amount of stream discharge. Impervious surfaces reduce the amount of available water content and result in increased, “flashier” streamflow in a channel that can cut banks, if the channel is not cement, flood nearby areas, or decrease sediment particle size via the erosion process. The linear regression of stream discharge shows that discharge decreased by 0.005 cubic feet per second for every one-foot distance downstream. This decrease in discharge results from Chatfield Dam retaining water to prevent flooding in the metropolitan area downstream. Theoretically, according to the River Continuum Concept , it is expected for discharge to increase as distance increases. For the results to present findings potentially similar to the River Continuum Concept, a meta-analysis of rivers in the Missouri-Mississippi Basin could be a better scale compared to this study.
Future Directions
To get a deeper insight to the Platte River channel morphology, field studies should be undertaken to provide accurate ground measurements. Cross-referencing the data with aerial imagery will provide stronger data for statistical testing. Width, depth, soil cores, and discharge measurements can be taken at each of the study points to analyze if urbanization impacts urban streams. Similarly, sediment type and geology may also prove interesting to study and understand if there are statistically significant changes near the floodplain within urban stream reaches. Furthermore, this study should be compared to similar streams to see if urbanization causes similar general geomorphic alterations or if it is better to consider geomorphic changes case-by-case.
Brooke Safely is a Master of Science candidate in Environmental Science.