212 • Z.RUBIN ETAL.
<br /> bottom and to alter the manner in which sediment is trans- respectively. The 100MHz antennas were selected for the
<br /> ported or retained by reducing flow volumes, removing comprehensive survey of the wetland, providing a balance of
<br /> beaver,and altering valley-bottom vegetation. penetration and vertical resolution.The comprehensive reflec-
<br /> tion-type survey in step mode was conducted during July
<br /> 2009 using a Sensors and Software PulseEkko 100 GPR system.
<br /> Methods Seven cross-sectional (perpendicular to the valley axis) trans-
<br /> ects (identified as XSO, XS1, ..., XS6) and two longitudinal
<br /> Multiple methods were used to map sediment deposits,quantify (parallel to the valley axis)transects(identified as Long W and
<br /> aggradation rates, and identify processes (in-channel and over- Long E)were surveyed with the 100MHz antennas(Figure 2).
<br /> bank fluvial deposition,direct hillslope input,beaver pond filling, Three cross-sections(XSO,XS1, XS2), and the portion of Long
<br /> peat accumulation) creating alluvial fill within the Lulu City W between XSO and XS2 were further surveyed with the
<br /> wetland.Analyses included historic aerial photograph interpreta- 200 MHz antennas to provide higher resolution of the near-
<br /> tion, ground penetrating radar (GPR) surveys, and trenching, surface deposits.In sum,—3 km of GPR surveys were conducted.
<br /> coring,and radiocarbon dating of valley-bottom sediment. Antennas were oriented parallel to the direction of motion
<br /> along transects and spaced 1 m apart, with a 25 cm spacing
<br /> between radar traces along the profile. Common midpoint
<br /> Aerial photo analysis (CMP) surveys were conducted on cross-sections XS2 and
<br /> XS4 (Figure 2)to measure radar velocities. Both CMP surveys
<br /> Aerial photographs of the Lulu City wetland were analyzed for yielded average radar velocities of —0.05 Wns. This velocity
<br /> evidence of debris flows, channel planform changes,channel is consistent with established rates for wet sand (Baker et al.,
<br /> migration, and beaver ponds. The wetland area was defined 2007),a widespread sediment in the Lulu City wetland.A topo-
<br /> as non-forested valley bottom, the extent of which has not graphic survey was conducted along the GPR transects using a
<br /> changed substantially since 1937. Imagery from 1937 (1: 23 TOP-CON Total Station. Points were surveyed at a spacing of
<br /> 000), 1953 (1: 63 360), 1969 (1: 15 840), 1987 (1: 15 840), —5 m across the low gradient wetland, with additional points
<br /> 2001 (1: 4000), 2004 (digital image, resolution=15 cm), and surveyed at breaks in slope. A topographic correction was
<br /> 2009 (resolution =1 m) cover the entire study area. The 1937 applied to GPR transects so that reflections throughout the
<br /> and 1953 images are black and white while later images are wetland were referenced to a fixed datum.
<br /> in color. Photos were scanned, imported into ArcGIS v9.3, GPR penetration depths averaged —4m, with a range from
<br /> and rectified. Debris-flow deposits, channel locations, beaver 2-6m. Penetration was shallowest in coarse substrate (e.g.
<br /> ponds,and vegetation changes in the wetland were interpreted cross-sections XSO and XS6) and deepest in fine deposits (e.g.
<br /> by comparison with contemporary examples;i.e.modern bea- XS3) (Figure 2). In general, GPR penetration depth is similar
<br /> ver ponds and debris-flow deposits were identified in the field to the maximum depths reached with hand augers and the
<br /> and in recent imagery, and similar features were recognized depths of the oldest radiocarbon samples. Thus, GPR surveys
<br /> in historical imagery. In the wetland, unvegetated alluvium is generally represent the last 4000 years; GPR interpretation was
<br /> identified as debris-flow deposits because in the low-gradient verified laterally and vertically by auger cores and excavator pits.
<br /> wetland,distinguished from fluvial deposits which are typically GPR reflection patterns were initially interpreted using a
<br /> thin deposits of fine sediment and unlikely to bury or remove radar facies analysis(Beres and Haeni, 1991)in which packages
<br /> vegetation.Channel migration was assessed by tracing channel of reflections were identified in the GPR data based on the spatial
<br /> centerlines in each image wherever dominant channels were continuity, configuration, amplitude, and frequency of the re-
<br /> present. Beaver ponds were identified from ponded water or flections. Those packages were traced and compared with soil
<br /> broad, unvegetated sections of channel upstream of a cross- descriptions from the cores and pits. The interpretation of radar
<br /> channel obstruction(beaver dam).Temporal and spatial variation images was then iteratively adjusted as needed so that the catego-
<br /> in vegetation was not distinguished because of the resolution of rization of radar-facies was in agreement with soil observations.In
<br /> the aerial photos, although establishment of trees is discernable most cases, the resolution of the GPR did not allow individual
<br /> and noted. The short growing season at high elevation and the beds to be distinguished. An approach was therefore taken to
<br /> inherent obstacles to aerial photo interpretation suggest that trees broadly categorize depositional regimes. Depositional environ-
<br /> would probably have to be a decade or more in age before being ments were grouped into higher energy (debris flow, overbank,
<br /> recognized in photos. and in-channel fluvial)and lower energy(peat, beaver pond fill,
<br /> and overbank)process regimes.Overbank deposits are included
<br /> in both low-energy and high-energy divisions.This was necessary
<br /> Ground penetrating radar (GPR) because overbank deposits are interbedded throughout both
<br /> regimes. Thus, low-energy deposits are regimes that fluctuated
<br /> GPR has proven to be an efficient, non-invasive,and effective between peat,pond accumulation,and received overbank depos-
<br /> method of imaging sedimentary deposits in a number of envir- its,whereas the high-energy division received coarse(gravel and
<br /> onments including internal structures in sediment(Hugenholtz larger)fluvial and colluvial deposits,but also received finer over-
<br /> et al., 2007), peatlands (Jol, 2009), alluvial fill (Hickin et al., bank deposition.This division proved useful for the broad charac-
<br /> 2007), and wetlands (Bristow and Jol, 2003;Jol, 2009). GPR terization of aggradational processes over thousands of years.
<br /> is a method in which a high-frequency electromagnetic pulse
<br /> is transmitted into the ground and partially reflected back to
<br /> the surface from interfaces where there are changes in the Sediment descriptions and inferred aggradational
<br /> electromagnetic properties (dielectric permittivity) of the sub- processes
<br /> surface materials. These reflections may identify contacts
<br /> between deposits of different grain size, mineralogy, water Sediment descriptions from 15 hand-augered and 19 excavator-
<br /> content,organic content,and bedrock(Davis and Annan, 1989). dug soil pits were used to verify GPR interpretation and quantify
<br /> A trial survey was conducted at the study site to test different depositional processes(Figure 2).Hand-augered holes were dug
<br /> frequencies.Frequencies of 50,100,and 200 MHz were tested, with an 8.3 cm diameter, bucket-type auger. The maximum
<br /> with depths of penetration of approximately 8, 4, and 2 m, depth was 3.5 m, and depths typically ranged from 1-3 m.The
<br /> Copyright©2011 John Wiley&Sons,Ltd. Earth Surf.Process.Landforms,Vol.37,209-222(2012)
<br />
|