Higher Peak Discharge Research Articles

Debris flow dam formation in Southeast Tibet

Glaciers with their deposits abound in the alpine areas of Southeast Tibet. Large debris flows occur frequently from these deposits and form dams that block streams. In this paper, 3 events of large debris flows reported in Peilong Valley located in Southeast Tibet, and which resulted 2 blocking dams resulted, are discussed in details, focusing on the major factors controlling dam formation. The results shows that the first surge group caused by snow and ice avalanches, ice-lake breaks, and large-scale landslides, with a high peak discharge and high velocity, and an abundance of boulders, are most likely to form blocking dams.

Tropical rivers

This paper presents an overview of tropical river systems around the world and identifies major knowledge gaps. We focus particularly on the rivers draining the wet and wet–dry tropics with annual rainfall of more than 700 mm/year. The size of the analyzed river basins varies from 10 4 to 6 × 10 6 km 2. The tropical rivers across the globe drain a variety of geologic–geomorphologic settings: (a) orogenic mountains belts, (b) sedimentary and basaltic plateau/platforms, (c) cratonic areas, (d) lowland plains in sedimentary basins and (e) mixed terrain. All of them show clearly high but variable peak discharges during the rainy season and a period of low flow when rainfall decreases. Some tropical rivers show two flood peaks, a principal and a secondary one, during the year. We computed the intensity of floods and discharge variability in tropical rivers. The relationship between sediment yield and average water discharge for orogenic continental rivers of South America and Asia was also plotted. Insular Asian rivers show lower values of sediment yield related to mean annual discharge than continental orogenic rivers of Asia and South America. Rivers draining platforms or cratonic areas in savanna and wet tropical climates are characterized by low sediment yields. Tropical rivers exhibit a large variety of channel form. In most cases, and particularly in large basins, rivers exhibit a transition from one form to another so that traditional definitions of straight, meandering and braided may be difficult to apply. In general, it is more useful to apply the terminology of single and multi-channel systems or complex anabranching systems at least for selected regional segments. Present-day knowledge of tropical systems and its potential application to improve interpretation of older alluvial sequences and facies models are briefly discussed. Human impact and river management issues including land use changes, mining, dams, interbasin water transference as well as flood hazards are some of the daunting problems in tropical river basins today.

Performance of grass and rainforest riparian buffers in the wet tropics, Far North Queensland. 1. Riparian hydrology

The long and intense storms of the wet tropics present extreme conditions for testing the effectiveness of riparian buffers. This study presents results of a hydrometric investigation of 4 riparian buffers on 2 commercial banana plantations in Far North Queensland, Australia. It investigates runoff generation and riparian hydrology on hillslopes with differing slopes, contributing areas, and topographic convergence. Both grass and rainforest buffers were examined. Surface and subsurface hydrology were measured for 4 wet seasons (December–April) using paired flumes, piezometers, and tensiometers. All buffers experienced large volumes of surface runoff, with peak discharges ranging from 30 L/s on planar hillslopes to 350 L/s on a highly convergent site. Event runoff : rainfall ratios ranged between 0.01 and 0.65. Grass buffers with smaller contributing areas (<0.3 ha) were able to dissipate the energy of surface runoff under all conditions. In a larger (5 ha), highly convergent hillslope, surface runoff became channelised upslope of the buffer and the vetiver hedges and grass were not able to prevent scouring of a channel through the buffer, reducing its performance. Infiltration occurred in all buffers during small events, and at the convergent buffer during large events, most likely due to the presence of deep soil fill. In contrast, exfiltration occurred in the grass buffers on planar and moderately converging slopes during large events. There, the riparian soil approached saturation and return flow and seepage were measured. Under exfiltration, soil strength may be decreased and riparian buffers are needed to decrease erosion hazard. Localised saturation was observed in the rainforest buffer beneath a planar hillslope during large events, where soils were deeper and dried out more quickly than in the adjacent grass buffer. This study documents the high runoff volumes and peak discharges on cropped slopes in the wet tropics, and evaluates riparian hydrological processes. Infiltration is unlikely to be an important buffer function in this environment, but an additional role of buffers is to reduce the erosion hazard presented by exfiltration.

The applicability of a parsimonious model for local and regional prediction of runoff

The Hydrological Recursive Model (HRM), a conceptual rainfall-runoff model, was applied for local and regional simulation of hourly discharges in the transnational Alzette River basin (Luxembourg-France-Belgium). The model was calibrated for a range of various sub-basins with a view to analysing its ability to reproduce the variability of basin responses during flood generation. The regionalization of the model parameters was obtained by fitting simultaneously the runoff series of calibration sub-basins after their spatial discretization in lithological contrasting isochronal zones. The runoff simulations of the model agreed well with the recorded runoff series. Significant correlations with some basin characteristics and, noticeably, the permeability of geological formations, could be found for two of the four free model parameters. The goodness of fit for runoff predictions using the derived regional parameter set was generally satisfactory, particularly for the statistical characteristics of streamflow. A more physically-based modelling approach, or at least an explicit treatment of quick surface runoff, is expected to give better results for high peak discharge.

ワイヤ放電加工のワイヤ断線防止制御の研究 （第３報） 大電流放電の場合の放電位置と放電電圧

Recently used wire EDM machines generate a triangular high peak discharge current of 500 ampere or more. It is very difficult to detect the true discharge voltage, because the discharge voltage is transformed by the influence of the high discharge current and impedance of the wire. In this study, we investigate the variation of discharge location and the discharge voltage immediately before the wire breakage in the case of high discharge current on wire EDM, in orderto predict the wire breakage. Three methods that compensate the detected discharge voltage are proposed and the investigated results using these methods are described It was found that in the case of low discharge current, phenomena immediately before the wire breakage such as local concentration of discharges and increase of short circuits in the gap, are clearly observed, but it is difficult to observe them in the case of high discharge current.

A review of catastrophic drainage of moraine-dammed lakes in British Columbia

Moraine-dammed lakes are common in the high mountains of British Columbia. Most of these lakes formed when valley and cirque glaciers retreated from advanced positions achieved during the Little Ice Age. Many moraine dams in British Columbia are susceptible to failure because they are steep-sided, have relatively low width-to-height ratios, comprise loose, poorly sorted sediment, and may contain ice cores or interstitial ice. In addition, the lakes commonly are bordered by steep slopes that are prone to snow and ice avalanches and rockfalls. Moraine dams generally fail by overtopping and incision. The triggering event may be a heavy rainstorm, or an avalanche or rockfall that generates waves that overtop the dam. The dam can also be overtopped by an influx of water caused by sudden drainage of an upstream ice-dammed lake (jökulhlaup). Melting of moraine ice cores and piping are other possible failure mechanisms. Failures of moraine dams in British Columbia produce destructive floods orders of magnitude larger than normal streamflows. Most outburst floods are characterized by an exponential increase in discharge, followed by an abrupt drop to background levels when the water supply is exhausted. Peak discharges are controlled by dam characteristics, the volume of water in the reservoir, failure mechanisms, and downstream topography and sediment availability. For the same potential energy at the dam site, floods from moraine-dammed lakes have higher peak discharges than floods from glacier-dammed lakes. The floodwaters may mobilize large amounts of sediment as they travel down steep valleys, producing highly mobile debris flows. Such flows have larger discharges and greater destructive impact than the floods from which they form. Moraine dam failures in British Columbia and elsewhere are most frequent following extended periods of cool climate when large lateral and end moraines are built. A period of protracted warming is required to trap lakes behind moraines and create conditions that lead to dam failure. This sequence of events occurred only a few times during the Holocene Epoch, most notably during the last several centuries. Glaciers built large moraines during the Little Ice Age, mainly during the 1700s and 1800s, and lakes formed behind these moraines when climate warmed in the 1900s. Twentieth-century climate warming is also responsible for recent moraine dam failures in mountains throughout the world. Warming from the late 1800s until about 1940 and again from 1965 to today destabilized moraine dams with interstitial or core ice. The warming also forced glaciers to retreat, prompting ice avalanches, landslides, and jökulhlaups that have destroyed some moraine dams.

Hillslope hydrology in tropical rainforest steeplands in Brunei

Many remaining areas of tropical rainforest in south-east Asia are located on landscapes dominated by deep valleys and very steep slopes. Now that logging activities are extending into these steeplands, it is essential to understand how the natural rainforest system behaves if any kind of realistic assessment of the effects of such disturbance is to be made. This paper examines the hydrological behaviour of an undisturbed rainforest system on steep topography in the Temburong District of Brunei, north-west Borneo. The physical and hydrological properties of the regolith material are generally typical of tropical residual soils. The regolith has a clay texture and a low dry bulk density beneath a superficial litter/organic horizon. The infiltration capacity of the surface soil was several hundred mm h−1. That of the exposed mineral subsoil was an order of magnitude less, similar to the saturated hydraulic conductivity (Ksat) of around 180 mm h−1 at a depth of 150 cm. There was no indication that Ksat reduced with depth except very near the bedrock interface. Soil tensions were measured using a two-dimensional array of tensiometers on a 30° slope. During dry season conditions, infiltrating rain-water contributes to soil moisture, and drying of the soil is dominated by transpiration losses. During wet season conditions, perched water tables quickly develop during heavy rainfall, giving rise to the rapid production of return flow in ephemeral channels. No infiltration excess or saturation overland flow was observed on hillslopes away from channel margins. Subsurface storm flow combined with return flow produce stream flow hydrographs with high peak discharges and very short lag times. Storm event runoff coefficients are estimated to be as high as 40%. It is concluded that the most distinctive feature of the hydrology of this ‘steepland rainforest’ is the extremely ‘flashy’ nature of the catchment runoff regime produced by the combination of thin but very permeable regolith on steep slopes. Copyright © 2000 John Wiley & Sons, Ltd.

Sensitivity Analysis of Three-Dimensional Steady-State and Transient Spray Irrigation Models

We conducted a sensitivity analysis of ground water flow on a spray irrigation site to evaluate key parameters affecting use of spray irrigation for sewage disposal. Spray irrigation is considered successful when water recharges the aquifer (low runoff) and mounding of the water table is below set limits. It can be difficult to predict success and allowable irrigation rates given uncertainty in parameters. The U.S. Geological Survey (USGS) code MODFLOW was used to compare two- and three-dimensional steady-state simulations, and three-dimensional transient simulations. We compared Latin hypercube sampling and systematic adjustment of parameters as methods of sensitivity analysis, transient versus steady-state sensitivity, and variations in layer configurations. These models were not calibrated because they were created during planning stages before field data became available. The Latin hypercube sampling effectively tested a large number of parameters simultaneously, and can be used as a screening tool for systematic adjustment of parameters. Transient and steady-state simulations predicted similar sensitivity to hydraulic conductivity, but steady-state modeling produced higher peak heads and discharges. Again, we recommend that the steady-state simulations could be used as a screening tool to evaluate trends. Transient modeling is likely to provide more realistic predictions of heads and discharges. When different layer configurations were compared, we found that two-dimensional modeling overpredicted mounding and drain discharge because flow in a permeable soil layer is neglected. Two-dimensional modeling would not be effective for predicting allowable irrigation rates. These results provide modelers an indication of the effects of model simplification, which can be used to plan modeling strategies.

A view on some hydrological triggering systems in landslides

In this paper different types of hydrological triggering systems for debris flows, shallow and deeper landslides are described. The generation of surface run-off and high peak discharges in first order alpine catchments is an important triggering mechanism for debris flows. Failure conditions in shallow landslides can occur when at a critical depth, which is determined by the cohesion of the soil and the slope angle, the moisture content in the soil becomes close to saturation, resulting in a considerable reduction of soil strength. Deeper landslides (5–20 m depth) are in most cases triggered by positive pore pressures on the slip plane induced by a rising ground water level. The assessment of meteorological threshold conditions for shallow landslides (1–2 m) needs more detailed meteorological information than for deeper landslides. In the analyses of the hydrological triggering systems of deeper landslides the presence of a permeable top layer and fissures has to be taken into consideration.

Uplift and landscape stability at Taroko, eastern Taiwan

Abstract The Taroko area of Taiwan represents an extremely dynamic environment, of rapid geological and geomorphological change. The location of the area on the collision zone between the continental Eurasian plate and the oceanic Philippine Sea plate leads to a high frequency of seismic activity, intense internal deformation and high rates of tectonic uplift. The rate of uplift, which is believed to be one of the highest measured, is approximately 5.5mm per annum, with higher rates inferred for the last million years. In consequence, the area, which consists predominantly of relatively weak, intensely deformed marble and schist, has a high potential energy providing ideal conditions for rapid landform evolution. In addition to high rates of uplift, high rates of denudation are also favoured by the extreme climatic conditions. In particular, the area is seasonally affected by tropical cyclones (typhoons); an average of 16 of these occur in the northwest Pacific per annum, about three of which affect Taiwan. The passage of a typhoon is marked by heavy rainfall which may reach intensities of 100mm per hour and 1500mm in 24 hours. The high rainfall intensities generate intense fluvial activity, including extensive overland flow, high peak discharges, the mobilization of debris channels, landslides and rockfalls. Intense tectonic and geomorphological activity provide the ideal environment for studying the rapid evolution of landscapes. In this paper, uplift data are combined with daily hydrological records from a major river for the last fifteen years. The hydrological records, which include discharge, mean flow rates and sediment volumes, are used to estimate the sediment budget for the river and hence to estimate denudation rates. It is demonstrated that average aerial denudation rates approximately equal uplift rates. These data are then used to generate a simple computer model for the evolution of Taroko Gorge, which demonstrates that current landforms may be created without substantial changes in the rate of uplift.

Outburst and rainfall-induced peak runoff events in highly glacierized Alpine basins

The effect of rainfall on runoff from the highly glacierized basins containing Findelengletscher and Gornergletscher, Pennine Alps, Switzerland, is assessed for the periods May-September 1987 and 1993. Several major rainstorms produced high peak discharges in both seasons. Runoff response to prolonged heavy rainfall varied during the ablation season in both basins, but also differed between the two. Annual outbursts from the ice-dammed Gornersee also produced peak discharges from Gornergletscher. The transient snow-line partitions a glacierized basin into snow-covered and snow-free portions which, respectively, retain and return rapidly to runoff both meltwater and rain. Elevation of the 0°C isotherm in the atmosphere interacts with basin hypsometry to determine the proportion of the catchment area that receives rain in a given event. Progressive rise of the transient snow-line in summer increases the size of the snow-free area, and hence the area of basin that rapidly responds to rainfall. Heavy rainfall appears to have produced an outburst from Findelengletscher in September 1993. The largest rainfall-induced floods are most likely to occur in summer between mid-July and late-August, terminating periods of elevated air temperature, when thermally induced runoff is at a maximum and subglacial drainage pathways are well developed, or when rainfall induces outbursts in spring and late summer at times of declined internal drainage capacity.

Outburst and rainfall‐induced peak runoff events in highly glacierized Alpine basins

The effect of rainfall on runoff from the highly glacierized basins containing Findelengletscher and Gornergletscher, Pennine Alps, Switzerland, is assessed for the periods May–September 1987 and 1993. Several major rainstorms produced high peak discharges in both seasons. Runoff response to prolonged heavy rainfall varied during the ablation season in both basins, but also differed between the two. Annual outbursts from the ice-dammed Gornersee also produced peak discharges from Gornergletscher. The transient snow-line partitions a glacierized basin into snow-covered and snow-free portions which, respectively, retain and return rapidly to runoff both meltwater and rain. Elevation of the 0 °C isotherm in the atmosphere interacts with basin hypsometry to determine the proportion of the catchment area that receives rain in a given event. Progressive rise of the transient snow-line in summer increases the size of the snow-free area, and hence the area of basin that rapidly responds to rainfall. Heavy rainfall appears to have produced an outburst from Findelengletscher in September 1993. The largest rainfall-induced floods are most likely to occur in summer between mid-July and late-August, terminating periods of elevated air temperature, when thermally induced runoff is at a maximum and subglacial drainage pathways are well developed, or when rainfall induces outbursts in spring and late summer at times of declined internal drainage capacity. Copyright © 1998 John Wiley & Sons, Ltd.

1995 Ruapehu lahars in relation to the late Holocene lahars of Whangaehu River, New Zealand

Observations of active flows, their deposits, and the effects of lahars generated during the 1995 eruptions of Ruapehu are compared to those of the last 2000 years. The 1995 lahars were generated by similar mechanisms, and had similar volumes and flow rheologies, to those of the last 135 years. However, the large number of lahars in the 1995 sequence and the eventual emptying of the entire Crater Lake on Mt Ruapehu by lahars is distinctive in the context of the historic record. The 1953 and 1975 lahars, although of smaller total volume than the largest 1995 lahar, had higher peak discharges and consequently greater effects on life and property within 57 km from the source. This implies that failure of a lake dam is the most efficient mechanism for generating a fast lahar with high peak discharge, and that eruptions through Crater Lake are more efficient at generating a lahar with high peak discharge when the lake is full. The scant 1995 lahar deposits preserved indicate that the geological record of lahars is likely to be incomplete and biased to events of greatest magnitude. Given this limitation, lahars which formed deposits of the prehistoric (>135 years old) Onetapu Formation were of larger magnitude in all respects, and often more concentrated with sediment, than those of the last 140 years (including 1995). In order to generate such events, catastrophic total emptying of Crater Lake is required. Alternatively, some of the events may have been the result of small flank collapses on the eastern slopes of Ruapehu. These larger events are less frequent than small eruption‐triggered lahars on Ruapehu, but they may be preceded by little or no warning. Future lahar hazard mitigation strategies in the Whangaehu River should begin with improving the detection and early warning of such flows.

Discharge response to channelization of a coastal plain stream

The Obion River in western Tennessee was channelized by the U.S. Army Corps of Engineers in the 1960s. The primary purpose was to reduce flooding that inhibits agricultural productivity in the lower bottomlands. Channel enlargement and straightening has improved hydraulic efficiency. The increase in water velocity has effectively decreased flooding in the upper sections of the Obion River. However, runoff from the channelized portion of the drainage basin converges at downstream locations faster than the stream channel can accommodate, resulting in higher peak discharges and an increase in flood frequency. During the growing season months (May–October), the number of floods on the lower Obion River increased 140 percent following channelization. The greater flow efficiency allows water to move rapidly out of the Obion River drainage, decreasing the average duration of flood events. Brief periods of inundation, however, can destroy crops; therefore, the change in flood duration caused by channelization is not a desirable alternative to the higher number of floods that also occur. Conditions for bottomland cultivation have improved in the upper channelized sections of the Obion River. However, the higher flood frequency during the growing season on the lower river segments will limit agricultural productivity, which is contrary to the initial justification for channelization.

The Hydrological Effects of a Wildfire in a Eucalypt Afforested Catchment

SYNOPSIS A high-intensity wildfire through a humid, upland catchment which was partially afforested to Eucalyptus fastigata, markedly increased stormflows and caused high soil losses off the afforested slopes. Total streamflow yield was unchanged but baseflow was reduced after the fire. The change in stormflow generation was caused by increased overland flow which resulted in shorter times of concentration and higher peak discharges during storms. Increased overland flows are linked to the widespread presence of water repellency in the plantation soils. The fire also increased the erodibility of the surface soils which were readily entrained by the overland flow, hence the accelerated soil erosion. Measured soil losses on the afforested slopes (55 t/ha) were roughly five times higher than those estimated from sediment in the stream (10 t/ha). The difference is attributed in large part to a healthy riparian buffer zone which was observed to have trapped large amounts of eroded soil and ash. The potential loss of site fertility and the threat of downstream flooding and sedimentation after such wildfires, lead us to advise forest managers to adopt a more positive approach to forest fuel management.

An 85-kJ high performance capacitor bank with double-mode trigatrons

The present article describes the constructional aspects of an 85-kJ capacitor bank at 20 kV for the production of pulsed high magnetic fields. The discharge current magnitude is 56 kA at a ringing frequency of 1.1 kHz. The bank employs 8 double-mode trigatrons in parallel which are triggered into conduction by the application of 16 simultaneous low jitter high voltage pulses generated by a multiple trigger pulse generator. The performance of the bank with double-mode trigatrons is compared with that obtained by conventional single-mode trigatrons. The waveforms of discharge current in time domain and fast fourier frequency spectrum for single- and double-mode trigatrons are presented and analyzed for performance evaluation. The performance with double-mode trigatrons is characterized by distinctly superior synchronization in the closing of trigatron switches over a wide range of operating voltage resulting in high peak discharge current and low waveform distortion.

Characteristics of glacier outburst flood in the Yarkant river, Karakorum mountains

1. The drainage of the Kyagar glacier dammed lake and the Tram Kangri glacier dammed lake at the upper Shaksgam is the main reason for glacier outburst floods in the Yarkant river. The Kyagar glacier dammed lake is characterized by subglacial drainage, while the Tram Kangri glacier dammed lake by mainly lateral drainage and, secondly, by subglacial drainage. 2. The drainage mechanism of the Tram Kangri ice dam determines the main characteristics of flood hydrography of the Kagun station, while the Kyagar glacier dammed lake plays an important role in the formation of floods. 3. Glacier outburst floods in the Yarkant river are characterized mainly by high peak discharge, big rising rate, small total volume and short duration. The floods happen mostly from late summer to early autumn. A period of 6 to 10 years in occurrence of large scale glacier outburst floods exist. The periodicity depends mainly on large scale drainage in the Kyagar ice-dammed lake. 4. Formation and dimensions of glacier dams at the upper Shaksgam were determined by long-term variations of the regional climate, whereas the changes of storage capacity in the lake reflect cold and warm changes of alpine region. Therefore, frequent glacier outburst floods indicate glacier advance and climatic variations.

Climatic and Hydrologic Effects on the Regeneration of Populus angustifolia James Along the Animas River, Colorado

I investigated the dates of origin of riparian for- summer Arizona Monsoon. Extant stands originated in ten ests dominated by Populus angustifolia James, and recent to thirteen discrete periods between 1848 and 1976, in years interannual fluctuation in P. angustifolia seedling with both high spring and fall peak discharges. Expected abundance on a relatively undisturbed 6-mile reach of the seedling abundance and stand-origin dates since 1914 were Animas River in southwestern Colorado. The goal was to reconstructed using climate data, and were extended to 1556 develop plausible hypotheses about the roles of floods and using tree-ring chronologies. Model results suggest good interannual climatic fluctuation in structuring these forests. seedling years occurred more frequently (about every 3.4 I determined the year of origin of 242 recently-established years) than stand-origin years (about every 10-15 years). seedlings and fifty-seven forest stands, and then developed Good seedling years were 2-3 times, and stand-origin years empirical models relating seedling abundance and stand- were 5 times more common from 1848 to 1985 than from origin events to climatic and hydrologic fluctuations. Seed- 1556 to 1848. Recent expansion of P. angustifolia may have lings were most abundant in years with cool winters, wet been favoured by more frequent cool, wet years since 1848. springs, and cool, wet falls (R2adj=0.98). Both good seed-

Muscle afferent contribution to control of paw shakes in normal cats.

1. The discharge of various hindlimb muscle afferents was recorded during paw shakes in normal cats with the use of floating dorsal root electrodes. 2. Muscle spindle group Ia-afferents and tendon organ group Ib-afferents fired during muscle lengthening, reaching very high peak discharge rates and then silencing at or shortly after the onset of shortening. The timing of Ia firing was consistent with the predictions of a linear model as well as the responses of Ia endings subjected to identical length variations in separate anesthetized cats. 3. In the latter "reconstruction" experiments, waxing and waning dynamic fusimotor action straddling whole paw-shake sequences gave the most consistent matches with the data from the normal cats. The reproducibility of the inferred fusimotor action justifies the inclusion of paw shakes as a class of movement in which fusimotor set is high. 4. The peak ensemble Ia activity from single hindlimb muscles was estimated to be approximately 20 kiloimpulses/s (Kips). Ankle extensor and hamstrings length variations were nearly in phase in the first cycles of a paw-shake sequence. From published data on spindle populations in these muscles, this indicated that peak Ia input to the spinal cord exceeded 0.2 megaimpulses/s (Mips). 5. The phase relationship between origin-to-insertion muscle length and Ia firing during paw shakes was little affected by doubling or tripling the moment of inertia of the foot. We argue that this refutes the notion that in paw shakes phase reversals occur between muscle fibers and tendons in the muscles studied. 6. Inertial loading of the foot led to small but significant reductions in mean paw-shake frequency. This is consistent with an afferent contribution to the generation of these movements. 7. We conclude that in paw shakes in normal cats, the CNS "chooses" to sensitize Ia-afferents to muscle length variations by increasing dynamic fusimotor action. The resulting ensemble Ia input is very large and is likely to play a significant role in reflexly shaping the alpha-motoneuronal activity responsible for the paw shakes.

On the generation of delayed peaks in stream discharge

Where deep permeable soils overlie impermeable bedrock and where steep hillslopes border a narrow floodplain, subsurface stormflow dominates the hydrograph. In such cases delayed peaks in stream discharge may occur several days after rainfall, providing the major volumetric response and possibly also the highest peak discharge. The general occurrence of such delayed peaks is considered, making reference to field data from two small drainage basins in southwest England, and to the results of computer simulations using a distributed hillslope model. The importance of soil-water convergence into hillslope hollows is stressed. The relative roles of rainfall and antecedent soil moisture in controlling the size and timing of the delayed peak is discussed. The implications of the results for hillslope runoff modelling are briefly considered.