Topographic Change Detection

Figure 1 - Photograph of the river Feshie in Scotland, courtesy of Joe Wheaton.

Project Objectives

Learn how to create a Geomorphic Change Detection (GCD) project using this GCD software available from the Riverscapes Consortium
Examine topographic changes on the Feshie River in Scotland (Figure 1) using DEM surveys taken in 2003 and 2006
Identify physical mechanisms of channel change (bank erosion, bar development, channel aggradation, and channel degradation) from maps of topographic change over time
Create a budget segregation to estimate total in-channel change associated with each physical mechanism (method adapted from Wheaton et al., 2013)

Mapping Short-Term Topographic Change in the Feshie

2003-2006

Figure 2 - Hillshades of 2003 and 2006 DEMs of the Feshie River.

As a braided system with high sediment supply, the Feshie River in Scotland regularly experiences topographic channel changes over short periods of time. This is clearly visible in Figure 2, which shows topographic change observed in the Feshie River between 2003 and 2006.

To quantify how much channel change occurred during this three-year period, I input DEMs taken of the Feshie in 2003 and 2006 as well as error surfaces generated from each respective DEM into the GCD software.

Figure 3 - Volume of different changes in elevation in the Feshie River between 2003 and 2006 generated by the GCD software at 80% confidence. Red bars indicate erosion; blue bars indicate deposition.

In total, the Feshie lowered 7160 ± 2202 m3 and raised 4420 ± 1391 m3 during this period. This result indicates that the Feshie eroded more than aggraded between 2003 and 2006. Elevation gains between 2003 and 2006 were generally small-scale and were generally between +0.2 and +0.9 m, whereas elevation losses had greater spread and generally ranged from -0.3 to -1.5 m (Figure 3). The increased spread in volume losses suggests that the Feshie experienced deeper erosion events than aggradation events over the course of this period, supporting the result that the Feshie lost more than gained volume between 2003 and 2006.

Figure 4 - GCD output raster over hillshade of 2006 DEM.

Figure 5 - Hand-annotated areas of bar development, channel aggradation, channel degradation, and bank erosion over hillshade of 2006 DEM.

The GCD software also output a raster data set representing the spatial distribution of detectable changes between the two DEMs (Figure 4). According to the GCD output, 26% of the channel area showed detectable change in elevation between 2003 and 2006. To investigate the physical causes of changes in channel topography, the GCD output raster and hillshades were used to manually determine areas of bar development, channel aggredation, channel degradation, and bank erosion (Figure 5).

Budget Segregation

The GCD software was also used to create a budget segregation based on the four geomorphic mechanisms above. To achieve this, a manually-generated mechanism shapefile (Figure 5) was input into the GCD as a mask. The mask input segregated the total budget (ie topographic changes) into groups based on metrics derived from the shapefile. This provided an estimate of how much each geomorphic mechanism contributed to topographic changes in the reach from 2003-2006.

Figure 6 - Total changes in volume due to bank erosion, channel degradation, channel aggredation, and bar development in the Feshie between 2003 and 2006.

In total, bank erosion contributed the most to channel change by volume and bar formation contributed the least. Additionally, channel degradation contributed more by volume than channel aggradation (Figure 6).

This result for bank erosion makes sense considering how I delineated bank erosion in my shapefile (Figure 5). The changes I observed using the GCD occurred over a three year period. This led to dramatic changes in bank locations at various parts of the reach due to the time interval between the inputted DEMs. I decided that "bank erosion" should include where the bank originally existed in 2003 up to where the bank was recorded in 2006 (Figure 5). The polygons outlining this change likely ended up covering a larger area than bank erosion would cover if the comparative time period was shorter than three years. I believe that if I had outlined the polygons to be as close to the new bank as possible and only considered one year difference between DEMs, the largest mechanism contributing to channel change would be channel degradation.

Regardless of the whether bank erosion or channel degradation was the largest contributing mechanism to topographic changes in the Feshie, it is clear that the Feshie lowering more than it is raising. This could be a result of base level changes, consistent upstream contributions of fine sediment, or another cause of intense, short-term erosion. The GCD analysis gives a strong starting point to investigate topographic change in this reach, but more analysis of the Feshie system as a whole is needed to determine the factors contributing to channel lowering.

More detailed results regarding relative contributions of each physical mechanism to total topographic change by volume in the Feshie can be found below.