2.2. Sample Design
We conducted a preliminary field study in 2020 where we selected six 300
m sample reaches that were representative of: a) the historically
degraded habitat found in the lower SRR, b) boundary habitat located
along the edges of the valley plug and c) habitat located within the
valley plug. Each of the three habitat types were represented by two of
the six sample reaches. In each reach, we measured mesohabitat
composition by estimating the area of geomorphic units within each reach
(i.e., glide-runs, pools, backwaters, riffles and large wood
accumulation) and calculated percent-area of each type.
In analysis of geomorphic change over time, as well as habitat
persistence in the form of floodplain inundation, we used modified
methods from the Riverscape Inundation Mapper tool
(RIM; Bartelt, 2021). Based on these
methods, we conducted relatively rapid and manual digitization of
landscape features to analyze readily available (e.g., Google Earth) and
easily acquirable (e.g., consumer-grade drones) high-resolution aerial
imagery. The digitization of visible features from high-resolution
orthoimagery is a widely used method in landscape analysis (e.g.,
Carbonneau, Fonstad, Marcus, & Dugdale, 2012; Carbonneau et al., 2020;
Donovan et al., 2019; Green, Hagon, Gómez, & Gregory, 2019), and we
used the approach to examine geomorphic change resulting from the valley
plug and the persistence of this feature over time.
To establish a baseline of geomorphic condition in the lower 90 km of
the SRR, we chose two degraded and laterally-confined or partly-confined
reference reaches, one upstream of Spring Canyon and the other upstream
of Moonshine Wash (Figure 1). Selection of these reference reaches was
access-limited due to a lack of roads and the rugged terrain of the
region. The selected reaches were similar in length to the valley plug
(7.36 km) and were intended to represent simplified and degraded
habitat, offering context for metrics of change within the valley plug.
For each reference reach, we followed a five-step process where we 1)
acquired basemap imagery, 2) digitized features that represent: a)
valley bottom extent, b) inundation, c) geomorphic units and d) riparian
and upland vegetation, 3) split the valley bottom and metrics into 13
equally spaced segments, 4) quantified the metrics from each segment and
5) calculated measurements of variance for each site (see Table 2). We
repeated these data capture methods over two time periods: 2009 and
2015, chosen because these time steps allowed us to examine significant
change over time pre- and post- valley plug formation. We then repeated
these data capture methods for the Cottonwood Wash valley plug over
three time periods: 2009 (pre-valley plug), 2015 and 2021.