High-Resolution Analysis of Debris Flow–Induced Channel Changes in a Headwater Stream,Ashio Mountains,Japan∗

Coupled hillslope and channel processes in headwater streams (HWS) lead to rapid changes in channel dimensions. Changes in channel size and shape caused by a debris flow event along the length of a headwater stream in the Ashio Mountains, Japan, were captured with the aid of repeat high-definition surveys using terrestrial laser scanning (TLS) techniques. The HWS was classified into three distinct reaches below the debris flow initiation zone. A large knickpoint separated an upper bedrock reach from a colluvial reach along the midsection of the drainage. The colluvial reach transitioned to a lower bedrock reach that terminated at the master stream. Cross-sectional and morphometric analyses revealed no statistically significant changes in channel size or shape along the upper bedrock reach. Debris flow erosion generated significant differences in channel size and shape along a colluvial reach. Sediment bulking associated with erosion along the colluvial reach led to increases in channel size along the lower bedrock reach, but no statistical differences in channel shape. Morphometric analyses from the TLS point cloud revealed that debris flow erosion produced a distinct nonlinear change in channel dimensions in the downstream direction within the HWS. Variations in channel substrate along the length of HWS contributed directly to this nonlinear response. The episodic nature and nonlinearity of erosion associated with the current debris flow event highlights the importance of debris flows in general in understanding the transport of sediment, coarse to fine particulate organic material, and large woody debris, which are critical to the long-term management of riverine environments. TLS sampling methods show promise as one component of a multianalytical approach needed to continuously monitor and manage the dynamics of HWS.     

Debris flow initiation and sediment recharge in gullies

Landslides that enter gullied low-order drainages can either initiate debris flow or stop, depositing sediment in the channel. This process is one of the most common ways that debris flows initiate, but little attention to date has been paid to evaluating the factors that affect whether or not the initial landslide will become a debris flow or deposit sediment in the channel. Statistically significant parameters that determine whether slope failures become debris flows or act to recharge in-channel sediment are channel gradient, angle of entry of failure into the channel, initial failure volume, and the amount of in-channel stored sediment. Steeper channels, low angles of entry, lower volumes of in-channel sediment, and larger initial failures were more likely to result in debris flows. This study found that as the volume of in-channel stored sediment increased, the volume of initial failure required to initiate a debris flow also increased. This result calls into question the simple supply-limited model of cyclical debris recharge and debris flow in low-order gullied drainages and suggests a negative feedback mechanism between debris accumulation and debris flow susceptibility.     

Mapping the spatial and temporal distributions of woody debris in streams of the Greater Yellowstone Ecosystem, USA

The objectives of this study were: (1) to document spatial and temporal distributions of large woody debris (LWD) at watershed scales and investigate some of the controlling processes; and (2) to judge the potential for mapping LWD accumulations with airborne multispectral imagery. Field surveys were conducted on the Snake River, Soda Butte Creek, and Cache Creek in the Greater Yellowstone Ecosystem, USA. The amount of woody debris per kilometer is highest in 2nd order streams, widely variable in 3rd and 4th order streams, and relatively low in the 6th order system. Floods led to increases in woody debris in 2nd order streams. Floods redistributed the wood in 3rd and 4th order streams, removing it from the channel and stranding it on bars, but appeared to generate little change in the total amount of wood throughout the channel system. The movement of woody debris suggests a system that is the reverse of most sediment transport systems in mountains. In 1st and 2nd order tributaries, the wood is too large to be moved and the system is transport-limited, with floods introducing new material through undercutting, but not removing wood through downstream transport. In the intermediate 3rd and 4th order channels, the system displays characteristics of dynamic equilibrium, where the channel is able remove the debris at approximately the same rate that it is introduced. The spatial distribution and quantity of wood in 3rd and 4th order reaches varies widely, however, as wood is alternatively stranded on gravel bars or moved downstream during periods of bar mobilization. In the 6th order and larger channels, the system becomes supply-limited, where almost all material in the main stream can be transported out of the central channel by normal stream flows and deposition occurs primarily on banks or in eddy pool environments. Attempts to map woody debris with 1-m resolution digital four-band imagery were generally unsuccessful, primarily because the imagery could not distinguish the narrow logs within a pixel from the surrounding sand and gravel background and due to problems in precisely coregistering imagery and field maps.     

Legacy sediment stored in sinkholes: a case study of three urban watersheds in Tennessee,United States of America

As closed topographic depressions, sinkholes effectively divide karst regions into a series of subwatersheds, where swallets within each sinkhole define the downstream end of each subwatershed. Water and sediment are temporarily stored in sinkholes before continuing to underground streams. While pathways and travel times of water in karst terrain are commonly found using tracing techniques, such techniques do not provide information about the movement and travel time of sediment through these systems. This paper reviews the literature regarding sediment storage within sinkholes and presents a case study, the objective of which is to estimate the magnitude and timing of sediment deposition within three sinkholes in an urbanized area in Tennessee, United States of America. Sedimentation rates in sinkholes were estimated based on age determination of cultural artifacts (buried trash and coal fragments) and by ¹⁴C dating. For the three study sinkholes, minimum deposition rates (maximum residence times) were 15?cm?yr ^?1(6 yrs), 1.5?cm?yr ^?1(81?yrs), and 0.2?cm?yr^?1(640 yrs). The analysis revealed that sinkholes undergo fill and flush cycles at different timescales caused by changes in upland sediment supply and morphologic changes at swallet openings. Sinkholes that lack swallet openings, or with swallets that have become blocked by debris, are effectively disconnected from master streams with regard to sediment. Sediment stored in sinkholes may be released downstream when swallet conditions change.     

Geomorphic Effects of the Extreme Rainfall Event of 20–21 July, 2004 in the Latnjavagge Catchment, Northern Swedish Lapland

Mass transfers triggered by a rare rainfall event on 20–21 July, 2004, with 58.4 mm of rain within 24 h and 71.7 mm of rain within 48 h in the Latnjavagge catchment (9 km² , 950–1440 m a.s.l.; 68°20'N, 18°30'E) in the higher Abisko mountain region (Swedish Lapland), are quantified and analysed in direct comparison with mean annual mass transfers in this drainage basin. In years without rare rainfall events the Latnjavagge catchment is characterized by restricted sediment availability resulting in low mechanical denudation and mass transfers. During the rare rainfall event of 20–21 July, 2004, major stability thresholds on the slope systems (triggering debris flows and slides) and in the channel systems (break‐up of channel debris pavements and step–pool systems) in the Latnjavagge catchment were passed and mass transfers by debris flows, slides and fluvial debris transport in creeks and channels were several times higher than the mean annual mass transfers in Latnjavagge. In the calculation of longer‐term mass transfers and sediment budgets, rare events like the 20–21 July, 2004 rainfall event have to be considered as essential components. A reliable estimation of the recurrence intervals of such rare events is especially problematic. The general problem of defining an adequate length of process monitoring programmes is pointed out.     

Riparian reforestation and channel change: A case study of two small tributaries to Sleepers River, northeastern Vermont, USA

Measurements of two small streams in northeastern Vermont, collected in 1966 and 2004–2005, document considerable change in channel width following a period of passive reforestation. Channel widths of several tributaries to Sleepers River in Danville, VT, USA, were previously measured in 1966 when the area had a diverse patchwork of forested and nonforested riparian vegetation. Nearly 40 years later, we remeasured bed widths and surveyed large woody debris (LWD) in two of these tributaries, along 500 m of upper Pope Brook and along nearly the entire length (3 km) of an unnamed tributary (W12). Following the longitudinal survey, we collected detailed channel and riparian information for nine reaches along the same two streams. Four reaches had reforested since 1966; two reaches remained nonforested. The other three reaches have been forested since at least the 1940s. Results show that reforested reaches were significantly wider than as measured in 1966, and they are more incised than all other forested and nonforested reaches. Visual observations, cross-sectional surveys, and LWD characteristics indicate that reforested reaches continue to change in response to riparian reforestation. The three reaches with the oldest forest were widest for a given drainage area, and the nonforested reaches were substantially narrower. Our observations culminated in a conceptual model that describes a multiphase process of incision, widening, and recovery following riparian reforestation of nonforested areas. Results from this case study may help inform stream restoration efforts by providing insight into potentially unanticipated changes in channel size associated with the replanting of forested riparian buffers adjacent to small streams.     

高位滑坡的运动转化形式

高位滑坡剪出口高于坡脚,它一旦滑离滑坡发生区,运动可能转化成四种形式:1.崩塌:由滑体经分级解体滑过剪出口处依次向前倾倒而成;2.碎屑流动或3.碎屑滑动:由滑动块体经碎屑化而成;4.泥石流:在适当的细粒物质量和水体条件下生成的碎屑流动。     

CHANNEL CHANGES IN ARID BADLANDS,BORREGO SPRINGS,CALIFORNIA

Geometric, hydraulic, and sediment characteristics in arid badlands near Borrego Springs, California, are examined in relation to precipitation events of varying magnitude and frequency. The longitudinal and cross profiles of five ephemeral channels occupying a 2.5 km² catchment were surveyed under pre-and post-storm conditions during the February 1976-December 1978 period. Such arid region channels offer the opportunity to observe and explain rates and methods of profile change under different flow types in a short period of time. Catchment responses to light winter events include substantial lags between initial precipitation and channel runoff, the limited downstream movement of small slugs of sediment, high losses of discharge into channel alluvium, and prolonged mass movement of debris from adjacent hillslopes into the channels following the storm events thus promoting aggradation along certain channel reaches. Responses to intense summer storms include explosive channel and hillslope runoff and localized scour and fill, both during and following such events, thereby promoting substantial aggradation and erosion along portions of the channels. Although ephemeral flow conditions may produce channel profiles which are distinct from those in perennial streams, the evaluation of the methods of sediment transport and the storage of debris in arid catchments offer useful explanation for other environments.     

Transport and deposition of large woody debris in streams: a flume experiment

Large woody debris (LWD) is an integral component of forested streams of the Pacific Northwest and elsewhere, yet little is known about how far wood is transported and where it is deposited in streams. In this paper, we report the results of flume experiments that examine interactions among hydraulics, channel geometry, transport distance and deposition of floating wood. These experiments were carried out in a 1.22-m-wide×9.14-m-long gravel bed flume using wooden dowels of various sizes as surrogate logs. Channel planforms were either self-formed or created by hand, and ranged from meanders to alternate bars. Floating pieces tended to orient with long axes parallel to flow in the center of the channel. Pieces were deposited where channel depth was less than buoyant depth, typically at the head of mid-channel bars, in shallow zones where flow expanded, and on the outside of bends. We hypothesize that the distance logs travel may be a function of the channel's debris roughness, a dimensionless index incorporating ratios of piece length and diameter to channel width, depth and sinuosity. Travel distance decreased as the ratio of piece length to both channel width and radius of curvature increased, but the relative importance of these variables changed with channel planform. Large pieces can move further than our debris roughness models predict if greater than 50% of the active channel area is deeper than the buoyant depth of the piece, or if momentum is high enough to carry pieces across shallows. Our debris roughness model allows first-order prediction of the amount of wood transport under various channel geometries.     

Channel and woody debris characteristics in adjacent burned and unburned watersheds a decade after wildfire, Park County, Wyoming

Large variability in responses of stream sediment and large woody debris (LWD) to severe fire has limited the accurate prediction of the magnitude and duration of fire effects on streams. Conditions in one Absaroka Range stream that was severely burned in 1988 were compared to those in an adjacent, undisturbed stream to improve understanding of fire effects on channel and LWD characteristics beyond the first few years. Ten reaches of each stream were sampled during summer 1999.Average bankfull channel width was greater in burned Jones Creek than in unburned Crow Creek. LWD frequency and overall frequency of LWD accumulations were greater in Crow Creek than Jones Creek. Debris-jam frequency was greater in Jones Creek after accounting for differences in the frequency of pieces with length greater than channel width. Larger piece size and better anchoring contributed to more frequent, small accumulations of LWD in Crow Creek. Differences between streams in LWD frequency are consistent with greater mobility of debris in burned Jones Creek. LWD-associated fine-sediment deposits were thicker but less frequent along Jones Creek than Crow Creek.     

The geomorphic function and characteristics of large woody debris in low gradient rivers, coastal Maine, USA

The role, function, and importance of large woody debris (LWD) in rivers depend strongly on environmental context and land use history. The coastal watersheds of central and northern Maine, northeastern U.S., are characterized by low gradients, moderate topography, and minimal influence of mass wasting processes, along with a history of intensive commercial timber harvest. In spite of the ecological importance of these rivers, which contain the last wild populations of Atlantic salmon (Salmo salar) in the U.S., we know little about LWD distribution, dynamics, and function in these systems. We conducted a cross-basin analysis in seven coastal Maine watersheds, documenting the size, frequency, volume, position, and orientation of LWD, as well as the association between LWD, pool formation, and sediment storage. In conjunction with these LWD surveys, we conducted extensive riparian vegetation surveys. We observed very low LWD frequencies and volumes across the 60 km of rivers surveyed. Frequency of LWD ≥ 20 cm diameter ranged from 15–50 pieces km^− 1 and wood volumes were commonly < 10–20 m³ km^− 1. Moreover, most of this wood was located in the immediate low-flow channel zone, was oriented parallel to flow, and failed to span the stream channel. As a result, pool formation associated with LWD is generally lacking and < 20% of the wood was associated with sediment storage. Low LWD volumes are consistent with the relatively young riparian stands we observed, with the large majority of trees < 20 cm DBH. These results strongly reflect the legacy of intensive timber harvest and land clearing and suggest that the frequency and distribution of LWD may be considerably less than presettlement and/or future desired conditions.     

鼎湖山木荷(Schima Superba)粗死木质残体的分解研究

以鼎湖山国家自然保护区季风常绿阔叶林优势种之一木荷的粗死木质残体为研究对象,采用目前国际较认可的腐解等级划分方法,对木荷粗死木质残体3个腐解等级(Ⅰ、Ⅱ、Ⅲ)的密度及养分含量进行测定分析.结果表明:①木荷粗死木质残体的密度随腐解等级的增加从0.58 g/cm3下降到0.16 g/cm3,分解速率常数K值为0.144 7/a,在自然状态下,木荷粗死木质残体分解95%所需时间约为21 a;②粗死木质残体C/N比值随分解的进行先降后升而N/P则相反;③K、Ca、Na、Mg离子浓度随腐解等级的增加出现不同程度的增加,但增幅不明显.研究结果旨在为系统评估粗死木质残体的分解行为在生态系统养分循环中的作用提供基础数据.     

The effects of logging on frequency and distribution of landslides in three watersheds on Vancouver Island, British Columbia

Three hundred and sixty three landslides in three watersheds that totaled 382 km² were identified from air photographs, beginning at a date that preceded logging to the present. The three watersheds all lie on Vancouver Island; however, they have different precipitation regimes, topography, and amounts logged. Landslide areas in the watersheds varied in size from 200 m² to more than 1 ha. Nearly 80% of the landslides were debris slides; 15% were debris flows, and the remainder primarily rock falls. Following logging, the number of landslides increased substantially in all watersheds although the amount of increase was variable: approximately 11, 3, and 16 times in Macktush Creek, Artlish River, and Nahwitti River, respectively. Other analyses of changes in landslide density also produced highly variable results, with the number of landslides increasing between 2.4 and 24 times. Further, 2–12 times more landslides reached streams following logging activities. Densities for landslides impacting streams increased for the period of record from 1.5 to 10 times following logging activities. The densities were substantially greater where only landslides that reached streams since development began in a watershed were considered. Roads had the greatest spatial impact in the watersheds (compared to their total area), with frequencies determined to have increased by 27, 12, and 94 times for Macktush, Artlish, and Nahwitti, respectively. The results highlight the relative impact of roads and their role in slope stability.     

FLOW DISTURBANCE CAUSED BY CROSS-STREAM COARSE WOODY DEBRIS

Coarse woody debris affects many streams in forested regions throughout the world. These effects include lateral channel migration, bank slumping, and aggradational or erosional features related to flow redirection. The extent of effect is dependent on the ability of the system to resist the new forces derived from flow redirection. This study on the Pine River, Ontario, looks at how obstructions that are perpendicular to downstream flow modify fluid behavior. Results show that fluctuations in speed and approach azimuth vary considerably depending on the position of the sample relative to the obstruction. The use of time-averaged (1 sec., 30 sec.) recordings of fluid speed and azimuth at selected channel locations shows how flow adjusts to external controls as it moves away from the obstruction zone, giving an indication of the spatial extent of the obstruction influence. These data are represented as a function of the diameter of the obstruction relative to the surrounding flow depth (obstruction ratio), and then are compared to results found in other debris obstructions on the Pine River and Wilmot Creek. Flow obstruction dimensions in the study site equal 26.5 trunk diameters (the average diameter of the tree trunk measured five times along its length), and range between 16.9 and 56.7 trunk diameters on the Pine River (n = 48) and between 7.4 and 63.5 trunk diameters on Wilmot Creek (n = 1066). Knowledge of these spatial relationships may allow for better management of woody debris in streams, primarily from the perspective of aquatic habitat. Multiquadric interpolation formed the basis for plotting fluid vector fields, showing the behavior of flow as it approached and moved through the obstruction zone. This is compared to studies of flow in unobstructed meanders in an attempt to quantify obstruction influence, and is used to provide a depiction of flow under these circumstances. [Key words: flow patterns, woody debris, influence zones, stream management.]     

Relationships between land-use and forced-pool characteristics in the Colorado Front Range

Land use practices in Colorado during the last two centuries altered the supply of sediment and water to many channels in the upper South Platte Basin. As a result of increased supply of sediment and mobility and reduced peak flows, the characteristics of pools associated with channel constrictions, referred to as forced pools, may have been altered. Increased supply of sediment and reduced transport capacity of high flows could lead to aggradation in forced pools. Channel confined by road corridors could lead to high velocities at normal flows, increased energy dissipation from riprap, or even increased pool frequency resulting from failed riprap. To assess potential alterations, four hypotheses were tested: (1) impacted streams will show significantly different mean volume of pools than a control stream; (2) mean volume of pools on streams where land-use activities increased the supply of sediment will be significantly different from streams solely affected by flow regulation; (3) the strongest change in characteristics of pools of impacted streams will be a reduced volume of pools; (4) streams affected by road corridors will show statistically lower spacing of pools than streams unaffected by roads. The downstream spacing and residual volume of twenty consecutive forced pools were surveyed on five streams in the Colorado Front Range that varied from no contemporary impact to multiple historical and contemporary impacts. ANCOVA with stepwise model selection indicated that degree of land-use (categorical), bankfull spacing of pools, upstream riffle slope and expansion ratio were all significant (α = 0.1) predictors of volume of pools (R² = 0.73). Simple linear regression of mean volume of pools and stream specific variables (gradient, drainage area and discharge) and least square means comparison of mean volume of pools indicated a need to standardize volumes of pools by slope and discharge so that the volumes of pools could be compared among different levels of land-use. Significant correlations between drainage area and volume of pools allowed volume of pools to be standardized by drainage area and thereby discharge. This dimensionless variable was also significantly correlated with channel slope, which permitted the construction of a new variable, PV_QS (volume of pools standardized by discharge and slope). Least square means comparison of mean PV_QS revealed that the control reach was significantly different from road-impacted reaches. Mean volume of pools was significantly larger in the control reach compared to all but one road-impacted stream. This was likely a function of higher wood loading in the control reach and the competence of high flows in the road-impacted reach. Streams affected by road corridors did not have significantly different bankfull spacing of pools from streams not impacted by roads. The multiple interactions among control and response variables explored in this study indicate the need to identify the most constrained and sensitive response variables when attempting to assess channel response to land use.     

Physical and Human Factors Influencing Potential Fish Habitat Distribution along a Mountain River, France

Potential fish habitat along the Drôme River, France, is a function of the distribution of large woody debris, boulders, undercut banks, gravel substrate, and pools. The distribution of these features is, in turn, a function of channel geomorphology, watershed and riparian forest characteristics. We conducted field work and analysed aerial photographs for 190 elementary segments of 500 m length along the Drôme River's 95 km course from the Alps westward to its confluence with the Rhône River near Loriol. The Drôme River does not follow the classic pattern of a monotone downstream decrease in gradient and change in channel characteristics. Although channel gradient, braided index and channel incision all decrease downstream, stream power is independent of longitudinal distance. These variables are largely controlled by geomorphic, human or hydrologic factors at the reach scale. Potential fish habitat richness decreases downstream, but individual habitat variables affecting habitat richness do not necessarily decrease downstream, many being controlled by local factors rather than by position along the continuum. Large woody debris is more abundant in braided reaches located directly downstream of confluences with main tributaries or downstream input sites. Boulders are most abundant downstream of failed bank protection works or in gorges. To improve fish habitat in the Drôme River, we recommend taking a long-term and large-scale perspective. Because structures placed in this unstable channel are likely to be washed downstream, we propose to emulate natural river dynamics and to permit large woody debris to enter the channel in unstable reaches via bank erosion, and that this debris not be removed (as is routinely done now) but permitted to migrate downstream through the system, creating fish habitat en route.     

Stream Channel Enlargement Response to Urban Land Cover in Small Coastal Plain Watersheds,North Carolina

Stream channel response to urban land use has not been well documented for southeastern Coastal Plain streams. In this study, urban channel response was evaluated in small Inner Coastal Plain watersheds (<5 km²) in eastern North Carolina. Reaches were selected across a range of watershed total impervious area (0-67% TIA). Channel dimensions and sediment grain size data were collected along 20 urban (>10% TIA) and 20 rural reaches (<10% TIA), and at 10 stormwater outfall sites (180 cross-sections). Urban cross-sectional area, channel incision ratio, and channel grain size (gravel%, D₅₀, and D₈₄) were greater, relative to rural channels. Bankfull cross-sectional areas were approximately 1.78 times greater for urban watersheds than for rural watersheds. Channels in urban watersheds were incised and had median full-channel capacities approximately 3.4 times greater than channels draining rural watersheds. Watershed TIA explained 65-72% of channel capacity enlargement. Urban expansion in the region began in the 1960s, with major urbanization occurring over the last 25 years. Channels draining urban watersheds are still responding to this land use change by downcutting and widening. Urban channel incision has frequently cut off streams from their floodplains, reducing floodplain sediment retention and water quality functions.     

The influence of relative sediment supply on riverine habitat heterogeneity

The diversity of aquatic habitats in streams is linked to physical processes that act at various spatial and temporal scales. Two components of many that contribute to creating habitat heterogeneity in streams are the interaction between sediment supply and transport capacity and the presence of local in-stream structures, such as large woody debris and boulders. Data from previously published flume and field studies and a new field study on tributaries to the South Yuba River in Nevada County, California, USA, were used to evaluate the relationship between habitat heterogeneity, local in-stream structural features and relative sediment supply. Habitat heterogeneity was quantified using spatial heterogeneity measures from the field of landscape ecology. Relative sediment supply, as expressed by the sediment supply/transport capacity ratio, which controls channel morphology and substrate textures, two key physical habitat characteristics, was quantified using a dimensionless bedload transport ratio, q. Calculated q values were plotted against an ecologically meaningful heterogeneity index, Shannon's Diversity Index, measured for each study reach, as well as the percent area of in-stream structural elements. The results indicate two potential mechanisms for how relative sediment supply may drive geomorphic diversity in natural river systems at the reach scale. When less mobile structural elements form a small proportion of the reach landscape, the supply/capacity ratio dictates the range of sediment textures and geomorphic features observed within the reach. In these settings, channels with a moderate relative sediment supply exhibit the highest textural and geomorphic diversity. In contrast, when less mobile structural elements are abundant, forced local scour and deposition creates high habitat heterogeneity, even in the presence of high relative sediment supply.     

泥石流输沙及其对山区河道的影响

泥石流能在很短时间内将大量大小混杂的固体物质输入主河,影响主河河床演变,形成灾害。在连续观测资料的基础上,对泥石流输沙的强度、级配和时空分布特征进行了分析。通过实际测量,分析了泥石流在沟道内冲淤特征以及影响泥石流冲淤特征的因素,如泥石流活动规模和局部沟道条件等。通过水槽实验,分析了泥石流与主河交汇的机理,将泥石流入汇主河的模式概括为掺混模式、潜入模式、推进模式和堵河模式,并且从能量角度阐释了汇流区的水沙交汇特征,提出了泥石流堵江的判据。最后,分析了泥石流多发区受泥石流入汇影响,主河河床在平面形态、横断面形态、纵断面形态和河型等方面的变化特征。     

Bed load transport in an obstruction-formed pool in a forest, gravelbed stream

This paper examines channel dynamics and bed load transport relations through an obstruction-forced pool in a forest, gravel-bed stream by comparing flow conditions, sediment mobility, and bed morphology among transects at the pool head, centre, and tail. Variable sediment supply from within and outside of the channel led to a complex pattern of scour and fill hysteresis. Despite the large flood magnitude, large portions of the bed did not scour. Scour was observed at three distinct locations: two of these were adjacent to large woody debris (LWD), and the third was along the flow path deflected by a major LWD obstruction. Bed material texture showed little change in size distribution of either surface or subsurface material, suggesting lack of disruption of the pre-flood bed. Fractions larger than the median size of the bed surface material were rarely mobile. Sediment rating relations were similar, although temporal variation within and among stations was relatively high. Relations between bed load size distribution and discharge were complex, showing coarsening with increasing discharge followed by fining as more sand was mobilized at high flow. Lack of local scour in the pool combined with bed load fining and net fill by relatively fine material implied that the dominant sources of mobile sediment were upstream storage sites and local bank collapse. Patterns of flow, channel dynamics, and sediment mobility were strongly affected by a LWD flow obstruction in the pool centre that created turbulent effects, thereby enhancing entrainment and transport in a manner similar to scour at bridge piers.