Spatial Geomorphic
Heterogeneity
The design of metrics to describe riverine forms and processes is limited only by the imagination, and this is especially true in the case of heterogeneity metrics, particularly when considering the array of geomorphic unit schemas that could be used to parameterize such metrics. Heterogeneity metrics have unique descriptive power in certain cases. An example is the ability of a river corridor to trap wood. Investigations of wood load variability and wood deposition patterns indicate that river corridors with a mix of landforms that transport wood and those that trap wood tend to accumulate the highest wood loads (Okitsu et al., 2021; Ruiz-Villanueva et al., 2016; Scott & Wohl, 2018). How would one measure the degree to which geomorphic units are mixed together, as well as their spatial arrangement? Traditional metrics like channel width, slope, or even surface roughness are insufficient, but heterogeneity metrics provide descriptors aptly suited to this task, specifically the evenness and subdivision of landform patches.
Here, we discuss a few classes of metrics that we have found useful in investigating geomorphic processes. We refer readers to summaries of landscape pattern metrics from ecology for more detailed information (Magurran, 2021; McGarigal, 2012).
Measuring Geomorphic Unit
Diversity
Diversity measures both richness (total number of classes) and evenness (relative abundance of classes). If richness varies between sites of interest, diversity can be normalized by maximum diversity to describe evenness alone. Diversity is commonly measured using either the Shannon or Simpson diversity index. Shannon’s diversity index is more sensitive to rare classes and more heavily weights richness, whereas Simpson’s diversity index is less sensitive to rare classes, more heavily weights evenness, and can be readily interpreted as the probability of two randomly placed points landing in different geomorphic units (Nagendra, 2002; Somerfield et al., 2008).
Diversity metrics are useful when the evenness or number of geomorphic unit classes reflects processes of interest. For example, a high evenness of canopy heights (i.e., even area covered by low, medium, and high canopy) on young floodplains might indicate the prevalence of ongoing vegetation succession. Comparing canopy height evenness over time to flow or sediment flux could indicate whether either of those factors are limiting vegetation succession (e.g., a lack of cottonwood recruitment due to flow regulation; Braatne et al., 2007)). Alternatively, the richness of in-channel geomorphic units (e.g., pools and bars of varying origin, further subdivided by surface grain size) could indicate the effectiveness of in-channel structures like wood and patches of vegetation at producing more variable sediment erosion and deposition patterns, assuming that greater hydraulic variability caused by wood and vegetation would result in a greater variety of geomorphic units
(Fryirs & Brierley, 2021). Finally, the diversity (i.e., both richness and evenness) of floodplain geomorphic units (e.g., scroll bars, oxbow lakes, relict channels, all further classified by vegetation community) could reflect the variability and relative dominance of processes like avulsion and channel migration in reshaping the river corridor (Slingerland & Smith, 2004).
Measuring the Spatial Configuration of Geomorphic
Units
Diversity (i.e., richness and evenness) does not account for the spatial configuration, or arrangement, of geomorphic units. As such, diversity may not indicate overall complexity. This means that in some cases, diversity metrics will need to be supplemented with spatial configuration metrics (Figure 2).