ABSTRACT We used field-based case studies to examine how spatial heterogeneity influences the response of river corridors to changes in water and sediment fluxes after wildfire. Spatial heterogeneity describes the degree to which a river corridor differs from a spatially uniform feature. Biogeomorphic processes, especially those that involve beaver ( Castor canadensis ) dams and large wood in the channel and floodplain, both influence and respond to spatial heterogeneity. These feedbacks can in turn strongly influence the degree to which a river corridor is resistant, resilient, or sensitive (RRS) to wildfires. A resistant system experiences little change in process or form following a disturbance. A resilient system absorbs disturbances without diminishing or changing process or form. A sensitive system undergoes persistent change following disturbance. Just as spatial heterogeneity can be characterized with respect to different spatial scales in a river network or river corridor, so different components of a river corridor can vary in their response to disturbance. Consequently, spatial scale and the specific process or landform under consideration will strongly influence designation of RRS. We drew two inferences from the case studies: (1) Reach-scale details of spatial heterogeneity and biogeomorphic feedbacks can be important in determining both reach-scale and network-scale responses to major disturbances such as wildfire, and (2) because of the potential for biogeomorphic feedbacks that either attenuate or exacerbate postfire inputs to river corridors, protecting and fostering reach-scale spatial heterogeneity and the biota involved in biogeomorphic feedbacks can increase river network resilience to wildfire.

Abstract This one-day field trip highlights recent research into the late Holocene geomorphic evolution and land use history of Piedmont stream valleys near Raleigh, North Carolina. European settlers began building water-powered milldams in the eastern United States in the 1600s, and dam construction continued until the early twentieth century. At the same time, regional-scale land clearing associated with agriculture and development increased upland erosion rates 50–400 times above long-term geologic rates. Much of the eroded sediment was subsequently aggraded on floodplains and impounded behind milldams. This trapped "legacy" sediment, commonly mistaken for natural floodplain deposition, has gone largely unrecognized until recently. This study focuses upon 1st to 4th order streams in W.B. Umstead State Park that drain into the Neuse River basin. There are seven water-powered milldam locations within the park and adjacent areas. Geomorphic mapping demonstrates that upland soil erosion and valley bottom sediment aggradation was substantial following European-American land acquisition and their conversion of large amounts of forest land for agricultural purposes. We observe three distinct sedimentary units in stream bank exposures that are corroborated by 14 C dating. Pre-European sediments range from ca. 4400–250 yr B.P. and consist of quartz-rich axial stream gravels and off-channel organic rich clays. Two legacy sediment units are differentiable; pre and post-dam, and range in age from ca. 300–100 yr B.P. The pre-dam sediments consist primarily of fluvial sands, and are interpreted as channel aggradation in response to soil erosion from upland land clearing prior to dam construction. Post-dam sediments are distinguished by finer grain size and sedimentology consistent with slackwater deposition, including sandy "event" layers, interpreted to be the result of large floods into the former mill ponds. Stream bank magnetic susceptibility (MS) measurements exhibit large and consistent increases at and above the pre-European-legacy sediment contact, suggesting that MS is a suitable proxy for legacy sediment identification along North Carolina Piedmont streams. Estimates of aggraded legacy sediment from two stream reaches in Umstead State Park indicate that the volume of eroded upland soils is approximately balanced by valley bottom sediment aggradation, and that area-averaged depth of upland soil loss was equivalent to 3–15 cm across this part of the Piedmont. We evaluate the current impact of legacy sediment erosion on stream water quality by capturing the total suspended sediment load (TSS) during discharge events using ISCO samplers at 5 sites on Reedy and Richland Creek. We document a TSS increase as water passes through reaches containing milldam deposits. This suggests that modern stream water impairment in the Piedmont may result where milldams were constructed and legacy sediments impounded. The field trip concludes by examining an active beaver (Castor canadensis) pond–wetland meadow complex above the historic Yates Mill pond. Beavers may prove to be valuable assets in the restoration of Piedmont stream systems still suffering from centuries of poor land and soil management.