3.2 | Residual pool depth analysis
Results from the repeated measures ANOVA for riffles indicated that structure type did not affect the change in RPD, nor did the interaction between structure type and years post construction (Table 6). However, the number of years post construction did have a significant effect on RPD (p < 0.0001). Pairwise comparisons indicated that the change in RPD within preexisting riffles was greatest in year 0, which represented the as-built condition, compared to all subsequent years (p < 0.0001; Figure 4). RPD then decreased in year 1 following the first runoff cycle (p < 0.0001), with all subsequent years (2-4) exhibiting less change in RPD (p< 0.001). The change in RPD indicated that pools that were developed within preexisting riffles were likely over excavated in year 0, filled in to some degree in year 1, and then remained relatively stable in years 2-4 (Figure 4). The estimated marginal means for each year indicated that the net change in RPD was 0.26 m across the study period for pools that were developed within preexisting riffles, suggesting that the treatments effectively increased RPD in these locations. Detailed results from the RPD analysis, including pairwise comparisons for any significant effects, were included in the Supporting Information.
Analysis of RPD at preexisting pool locations indicated that there was a significant interaction between structure type and years after construction (p = 0.003; Table 6). Structure types that performed differently in year 0 included log toe (p = 0.034), log vane/log toe combination (p = 0.071), and no structure (p = 0.011), all of which exhibited a greater change in RPD when compared to log vanes. The boulder vane/boulder toe combination treatment also performed differently in year 1 when compared to other structure types (Figure 4), exhibiting a greater change in RPD when compared to boulder clusters (p = 0.002), boulder vanes (p = 0.017), log toe (p= 0.094), log vanes (p = 0.007), and no structure (p = 0.002). The log vane/log toe combination also demonstrated a greater change in RPD at preexisting pool locations when compared to no structure in the first year after construction (p = 0.048). No differences in RPD were observed between structure types for preexisting pool locations in years 2, 3, or 4. The change in RPD was generally smaller at preexisting pools when compared to riffles (Figure 4), with a net change in RPD of 0.13 m across the study period.