Dating Torrential Processes on Fans and Cones

Debris cones and alluvial fans involve a range of landform sizes from individual debris-flow lobes, through debris cones and “classic” alluvial fans, to enormous fluvial “megafans” (Harvey 2011). Within the context of this book the focus is on the intermediate scale, debris cones to “classic alluvial fans”. Such landforms are found in three main settings (Harvey 2010): mountain-front, intra-montane, and tributary-junction settings. They occur in all climatic environments, but again, within the context of this book the focus is on dry-region, and temperate upland and mountain environments.

Debris flows generally form when unconsolidated material becomes saturated and unstable, either on a hillslope or in a stream channel. The process is defined as a moving mass of loose mud, sand, soil, rock, water and air that travels down a slope under the influence of gravity. Flows can carry material ranging in size from clay to boulders, and may contain a large amount of woody debris. Volumes of material delivered by single events vary from less than 100 to more than 100,000m³. Generally three factors are necessary for a debris flow to develop: water, sufficient inclination, and abundant sediment supply.

Sediment transport processes have recently gained importance in river engineering, torrent control and reservoir management due to an increasing discrepancy between a surplus of sediments in upstream and a deficit in downstream river sections (Habersack et al. 2010b). This development leads to problems in flood protection (channel change), river engineering (e.g. riverbed degradation), hydropower generation (e.g. reservoir sedimentation) and the ecological status of running waters (e.g. loss of instream structures).

Debris flows are at the interface of several research directions dealing with natural hazards processes. It is therefore not surprising that methods for the prediction of flow and runout of debris flows have similarities to approaches originally developed for snow or debris avalanches and streamflow hydraulics (Körner 1980; Lied and Bakkehoi 1980; Perla et al. 1980; Iverson 1997). However, debris-flow volume and bulk flow behaviour may change during travel through the channel, e.g. by entrainment of loose sediment and/or incorporation of water from a tributary. At present, no generally applicable model is able to cover the range of all possible material mixtures and event scenarios. This complexity results in different torrential processes and results in a large variety of approaches to predict debris-flow mobility.

Debris flows, their volcanic counterparts lahars, and debris floods that may occur as part of a debris-flow event, are the mass movement events that result in most losses worldwide. Hundreds of deaths occur every year, while horrific multi-thousand death tolls are announced every decade or so. Debris-flow related economic and life losses are not limited to emerging and developing countries but regularly occur in the developed world as well. In the few month leading up to the writing of this chapter, debris flows in Madeira (Portugal), Italy and Brazil lead to well over 100 fatalities and created losses well over 100 Million USD. Consequently, debris flows have been the focus of intensive scientific investigation with hundreds of papers appearing annually on various aspects of debris-flow research. Several dedicated conferences have been held whose sole focus is debris flows. In 2005 a book on debris flows and related processes was published (Jakob and Hungr 2005) to offer a more systematic review of the state-of-the-art. The book was published in 2005 and thus reflects mostly knowledge up to 2003 or 2004. With that it is outdated in some fields. Furthermore, it only touched on various topics in chapter form while it is clear to the practitioner that some of the chapters in the Jakob and Hungr book would have deserved to be written up as individual volumes themselves.

Torrential processes like debris flows, debris floods, and intensive bedload transport represent a serious hazard for settlements and infrastructure located on Alpine fans. Protection works and mitigation measures are mostly based on the magnitude of a design event with a defined recurrence interval. However, engineering hazard assessment of torrential processes is often limited to rough estimation of possible future events since the data base is small and statistically reliable relationships between event frequency and magnitude are rare. Some information can be found in historical records of local communities. For this reason some effort has been done recently to collect this information from public administration offices and chronics in Austria (Huebl et al. 2010). Although the data is heterogeneous and its applicability for standard engineering extreme-value analysis is limited, information on historical events represents a helpful tool for hazard assessment in Alpine watershed and should be gathered if available.

Debris fan systems are common features of mountainous environments, particularly in formerly glaciated valleys in the Arctic, Antarctic, and high mountains of the lower latitudes (Rapp 1960a, b; Albjar et al. 1979; Church et al. 1979; White 1981; Caine 1983; Perez 1993). The prevalence of these landforms in these physiographic settings is attributed to rapid weathering associated with the presence of oversteepened valley sides, stress unloading of rock walls following deglaciation, and the severity of the climate (Matsuoka and Sakai 1999; Ballantyne 2002; Curry and Morris 2004). These factors are conducive to the development of fan deposits by rockfall, debris-flow and avalanche activity, each of which represent a significant hazards to fan areas.

For a realistic hazard assessment, knowledge of past events is of crucial importance. As archival data is generally fragmentary, additional information sources are needed for an appraisal of past and contemporary as well as for the prediction of potential future events. Tree rings represent a very valuable natural archive on past torrential activity as they may record the impact of events in their tree-ring series. In the past few years, dendrogeomorphology has evolved from a pure dating tool to a broad range of applications. Besides the reconstruction of frequencies, tree rings allow - if coupled with spatial positioning methods - the assessment of spread, runout distance, breakout locations or preferred flow path. Similarly, the wide field of applications includes the identification of magnitudes and triggers of debris-flow events if meteorological data is included.

Tree rings were first used for the pure dating of wood, for instance in archaeology. However, in the 1970s, researcher started drawing environmental information from tree rings as their growth reflects the influences to which a tree is exposed during its live. The term “dendrogeomorphology” was first introduced in the 1970s by Alestalo (1971) and refers to the study of geomorphic processes with tree rings. The method was then further developed by Shroder (1978, 1980) and Braam et~al. (1987a, b).

The Ritigraben torrent is located on the west-facing slope of the Mattertal valley (Valais, 46°11′N., 7°49′E.) and spans a vertical range of 2,000 m from its confluence with the Vispa River at 1,080 m asl to the summit of the Seetalhorn (3,100 m asl). A rock glacier occupies most of the headwater basin (1.4 km²) between 2,500 and 2,800 m asl (Fig. 1a), constituting the main starting zone of debris flows. On its downward course to the Mattervispa river, the Ritigraben torrent passes a forested cone (32 ha; 4.3 × 10⁶ m³; Fig. 1b) on a structural terrace (1,500–1,800 m asl), where debris-flow material affects trees composed of European larch (Larix decidua Mill.), Norway spruce (Picea abies (L.) Karst.) and Cembran pine (Pinus cembra L.). Virtually all trees on the cone show growth disturbances (GD) related to past debris-flow activity. The study of 2246 tree-ring sequences sampled from 1102 L. decidua, P. abies and P. cembra trees allowed reconstruction of 124 events since AD 1566. Based on tree-ring records of disturbed trees growing in or next to the deposits, 86% of the 291 lobes identified on the present-day surface could be dated. A majority of the dated material was deposited over the past century. Signs of pre-20th century events are often recognizable in the tree-ring record of survivor trees, but the material that caused the growth anomaly in trees has been completely overridden or eroded by more recent debris-flow activity. Tree-ring records suggest that cool summers with frequent snowfalls at higher elevations regularly prevented the release of debris flows between the 1570s and 1860s; the warming trend combined with greater precipitation totals in summer and autumn between 1864 and 1895 provided conditions that were increasingly favorable for releasing events from the source zone. Enhanced debris-flow activity continued well into the 20th century and reconstructions show a clustering of events in the period 1916–1935 when warm-wet conditions prevailed during summer in the Swiss Alps. In contrast, very low activity is observed for the last 10-yr period (1996–2005) with only one debris-flow event recorded on August 27, 2002. Since sediment availability is not a limiting factor, this temporal absence of debris-flow activity is due to an absence of triggering events, which not only shifted from June and July to August and September over the 20th century, but also seemed to be initiated primarily by persistent precipitation rather than summer thunderstorms.

Vegetation cover has been widely recognized for indicating the status and dynamics of physical components of the environment (Ellenberg 1988; Ozenda 2002; Pignatti et al. 2005; Huc 2008). Phytogeomorphology, the study of interdependence of vegetation and landforms (Howard and Mitchell 1985), has become a central topic in environmental studies (Hession et al. 2010; Marston 2010). According to a geo-ecologic approach (Troll 1971), this subject combines aspects of geomorphology and vegetation ecology with climate, soil science, natural hazards, and environmental management at different spatial scales (Kozlowska and Raczkowska 2002; Walker et al. 2004). Interactions between landforms and vegetation cover are particularly evident in mountain areas where ecological factors change along altitudinal gradients within a limited geographical area (Ellenberg 1988). In particular, landform heterogeneity in extreme plant habitats gives rise to environmental diversity (Kruckeberg 2002; Burga et al. 2004; Baroni et al. 2007) and biodiversity gradients (Gentili et al. 2010).

Dating of many types of young (<500 year), dynamic, geomorphic landforms (e.g. mass-movement erosional tracks and deposits, alluvial terraces, flood plains, etc.) for purposes of hazard assessment and mitigation commonly requires greater dating precision than is available through radiocarbon dating or other methods. Ages of trees growing on landform surfaces have been used in a number of studies to estimate the time of landform creation or surface clearing, but the time lag between surface formation or disturbance and the reestablishment of trees can vary from 1 to more than 200 years (Desloges and Ryder 1990; Frenzen et al. 1988, 2005; Larsen and Bliss 1998; McCarthy and Luckman 1993; Sigafoos and Hendricks 1969; Winter et al. 2002). Appropriate lag times for selected tree species and for particular climatic and altitudinal ranges must be determined for the method to be useful.

Debris flow is a dominant mass-movement process in mountain areas all over the world and represents a significant natural hazard. Consequently, knowledge of the frequency and magnitude of debris flows is important for hazard and risk assessments (Rickenmann 1999).

Palaeoflood hydrology is an expanding field as the damage potential of flood and flood-related processes are increasing with the population density and the value of the infrastructure. Assessing the risk of these hazards in mountainous terrain requires knowledge about the frequency and severness of such events in the past. A wide range of methods is employed using diverse biologic, geomorphic or geologic evidences to track past flood events. Impact of floods are studied and dated on alluvial fans and cones using for example the growth disturbance of trees (Stoffel and Bollschweiler 2008; Schneuwly-Bollschweiler and Stoffel 2012: this volume) or stratigraphic layers deposited by debris flows, allowing to reconstruct past flood frequencies (Bardou et~al. 2003). Further downstream, the classical approach of palaeoflood hydrology (Kochel and Baker 1982) utilizes geomorphic indicators such as overbank sediments, silt lines and erosion features of floods along a river (e.g. Benito and Thorndycraft 2005). Fine-grained sediment settles out of the river suspension in eddies or backwater areas, where the flow velocity of the river is reduced. Records of these deposits at different elevations across a river’s profile can be used to assess the discharge of the past floods. This approach of palaeoflood hydrology studies was successfully applied in several river catchments (e.g. Ely et al. 1993; Macklin and Lewin 2003; O’Connor et al. 1994; Sheffer et al. 2003; Thorndycraft et al. 2005; Thorndycraft and Benito 2006). All these different reconstruction methods have their own advantages and disadvantages, but often these studies have a limited time coverage and the records are potentially incomplete due to lateral limits of depositional areas and due to the erosional power of fluvial processes that remove previously deposited flood witnesses. Here, we present a method that follows the sediment particle transported by a flood event to its final sink: the lacustrine basin.

Cosmogenic nuclides (³He, ¹⁰Be, ¹⁴C, ²¹Ne, ²⁶Al, ³⁶Cl) are produced in minerals due to nuclear reactions induced by cosmic rays. Measured cosmogenic nuclide concentrations are used to determine how long rocks and sediment have been exposed at or near the surface of the earth. The timing of abandonment of alluvial fan lobes can be determined directly with cosmogenic nuclide methods. For fans with abundant boulders, the top surfaces of the largest boulders are sampled for exposure dating which is based on the build-up of cosmogenic nuclides. ¹⁰Be, ²⁶Al, and ³⁶Cl have been most commonly applied. On fans with surfaces that were abandoned more than about 100-200,000 years ago, the boulders are often weathered, collapsed and crumbled. Such fans, as well as those that never had boulders at the surface, are dated with individual or amalgamated clast samples. Both post-depositional fan surface modification (cryo- or bioturbation), as well as the presence of inherited nuclide concentrations, may hinder obtainment of accurate ages. The former leads to exposure ages that are younger and the latter to ages that are older than the actual age of fan abandonment. Cosmogenic nuclide depth-profile dating is called upon. Samples are taken every 10-20 cm up to a depth of about 3 m. Alluvial fan sediment at greater depths can be dated with burial dating or burial isochron dating, which are based on the differential decay of two nuclides (for example ¹⁰Be and ²⁶Al). Optimally samples are taken from depths of more than 30 m. Although in the majority of published studies cosmogenic nuclide dating of fans has been used to determine tectonic slip rates, much has been learned about fan construction and, in turn, catchment processes.

Radiocarbon dating has contributed considerably to the understanding of alluvial fan and debris cone evolution, as well as their hazard and risk potential. This development mirrors the contributions made in investigations of other depositional environments since invention of the technique in the 1950s (Cook and van der Plicht 2007; Jull 2007). Radiocarbon dating involves measuring the amount of the radioisotope ¹⁴C preserved in fossil organic materials and using the rate of radioactive decay to calculate the age for a given sample. The application of radiocarbon dating to secure the development and evolution of alluvial fans requires the deposition and preservation of organic materials within the sedimentary stratigraphy. The dating technique has been used widely to reconstruct the development of alluvial fans in temperate to sub-arctic regions where environmental conditions are conducive to the presence and preservation of organic materials. The age limit to the application of radiocarbon dating ranges from approximately 45,000 years to typically several hundred years. Thus the technique is relevant for constraining the age of alluvial fans since the middle of the last cold stage, Marine Isotope Stage (MIS) 2 (45–11.5 ka) through the current interglacial (11.5–0 ka), the Holocene.

Alluvial fans and debris cones are important buffers in the sediment cascade from mountains to the sea. Interactions between tectonic, climate, and base-level controls often render fans and cones dynamic geomorphic features – even on timescales of millennia and beyond (Harvey 2010). Unraveling such dynamics has proven difficult mainly because establishing chronologies for fans and cones is challenging: Buried organic remains are usually scare (Chiverrell and Jakob 2012, this volume), dendrochronology and lichenometry (Schneuwly-Bollschweiler and Stoffel 2012, this volume; Jomelli 2012, this volume) will often be limited to cone and fan surfaces and help accessing the recent past only (Lang et al. 1999), and rarely are tephra layers preserved or synsedimentary carbonates developed that would permit radiometric dating. Luminescence dating can offer an alternative as it allows determining the time when sediments were laid down. But limitations exist that are related to (i) limited exposure of sediment grains to daylight during the short and rapid transport of many fan deposits and (ii) low luminescence sensitivity of quartz from many high mountain environments. After a brief introduction on luminescence dating techniques, an overview of relevant dating applications is given and the possibilities and limitations of applying luminescence dating to alluvial fan and debris cones are discussed.

In this chapter, we present a successful application of optically-stimulated luminescence (OSL) dating for reconstructing Quaternary alluvial and colluvial terrace sequences in the Pisco valley of western Peru. Here, climate-driven changes in debris-flow activity resulted in sediment aggradation and subsequent dissection forming distinct terrace levels. The terraces are made up of rather well-sorted, polymict, clast-supported conglomerates at the base and monomict, matrix-supported breccias towards the top. Both the abundance and thicknesses of breccias increase upsection and form the top of each sequence, leading to fan-shaped geometries. The well-sorted and clast-supported fabric of the conglomerates is interpreted as deposition by the perennial braided Pisco trunk stream. In contrast, the matrix-supported breccias were deposited by debris-flows sourced locally in the adjacent tributary valleys. The superposition of breccias on conglomerates implies that tributary debris-flow fans prograded from the valley margin towards the centre. OSL dating reveals that initiation of sediment accumulation was concurrent with a shift to more humid conditions. OSL ages also show that sediment aggradation started at the valley outlet, with the locus of sediment deposition then propagating farther upstream. Subsequent erosion and dissection of the valley fill commenced later during the same humid interval and continued throughout the following drier climate. Interestingly, the destructive phases also started at the valley outlet and propagated upstream, presumably as the hillslope sediment reservoirs became depleted. It is concluded that the sedimentary fill of the Pisco valley, and presumably that in other mountain belts, records a transient, non-steady response of debris-flows and fluvial processes to climate change, where non-steady sediment flux by hillslope erosion is buffered for a limited time span.

Monitoring of debris flows, aimed at assessing their physical parameters, is important both for research purposes, such as the determination of the rheological behaviour and the calibration of mathematical models, as well as for planning countermeasures and designing warning systems.

Accurate and comprehensive hazard assessment, as one part of integrated risk management, demands the application of a wide-range of methods (Fig. 1). Such methods include:

In mountain regions worldwide, rainfall-induced landslides and associated debris flows erode slopes, scour channels, and contribute to the formation of alluvial fans that may harm humans and destroy buildings. Rainfall-induced slope failures are frequent and widespread in Italy, where individual rainfall events can result in single or multiple slope failures in small areas or in very large regions. Most of the harmful failures were rainfall-induced, and several were shallow slides or debris flows. In the 60-year period 1950–2009, casualties due to landslides were at least 6,349, an average of 16 harmful events per annum. The large number of harmful events indicates the considerable risk posed by rainfall-induced shallow landslides and debris flows to the population of Italy (Guzzetti et al. 2005a; Salvati et al. 2010).

Fans and cones are the results of recurring geomorphic processes and they mark preferred areas for settlements and other forms of land use. Due to their position below larger slopes and valleys, processes like landslides, rock falls, debris flows and snow avalanches often endanger people, buildings and infrastructure. In the past centuries and decades, damage causing natural hazard events initiated the construction of structural mitigation measures like snow supporting structures against avalanches or check dams to reduce erosion in torrents. The goal of these efforts was to reduce the probability of damage causing events – decisions were primarily hazard-based. In the last 10 years the damage potential was increasingly taken into account, which led to risk-based decisions. The fundamental basis for both approaches is a solid knowledge of the event history allowing to derive realistic scenarios of potential future events (Jakob 2012, Chap.​ 6). Since the frequency and magnitude of these scenarios have a large impact on the calculated risk, dating of past events are of pivotal interest for risk analyses.

Floods and debris flows have the potential to cause fatalities, displacement of people and damage to the environment, to severely compromise economic development and to undermine economic activities. As floods and debris flows are natural phenomena they cannot be prevented. However, some human activities (such as the increase in human settlements and economic assets in floodplains and the reduction of the natural water retention by land use) and climate change contribute to an increase in the likelihood and adverse impacts of (flash) flood and debris-flow events. Damage caused by these processes may vary across countries and regions.

The term torrent control refers to the implementation of specific measures in the catchment or stream bed of a torrent. The purpose of torrent control is many fold – to stabilize the bed and adjacent slopes, to regulate discharge of floods, to dose runoff and solid transport, to filter large components (blocks, drift wood), to dissipate the energy of debris flow or to deviate (bypass) hazardous flow processes from objects or areas at risk (ONR 24800:2008). Torrential barriers (functional types, see Fig. 1) form part of the structural protection measures in torrent control that aim at reducing the risk to a tolerable level.

In recent years, flood and flood-related disaster management has shifted from protection against flood to managing the risks of floods. In Europe, this shift is reflected in the Flood Risk Directive of October 2007 (2007/60/EC; FRD) (Quevauviller et al. 2009). The FRD requires EU Member States to undertake a preliminary assessment of flood risks and, for areas with a significant flood risk to prepare flood hazard and flood risk maps and flood risk management plans. In this framework, the chain of forecasting, warning and response systems is becoming an integral part of integrated risk management. Forecasting and warning are particularly interesting as risk management options for flash flood events and attendant geomorphic response like debris flows, which occupy a central position in this chapter.

There is no doubt, the world’s climate is changing. But, as soon as we are interested in the impacts at a special location and a certain time in future uncertainties grow tremendously. Concerning meteorological parameters affected by climate change IPCC already uses verbal phrases to describe uncertainties (IPCC 2007). The highest degree of certainty (“virtually certain”), and therefore the most reliable proof of climate change can be found for the temperature increase. The uncertainty for other climate components is much higher. Simple physically based assumptions seem to be able to describe further effects (e.g. warmer temperature – increased water vapour – increased precipitation – more floods). But the climate system is more complex. Concerning the impact on increasing Alpine natural hazards such as flood or debris-flow magnitudes and frequencies there is no clear and direct relation at the first glance. For the Danube River in Austria it was found out that there is an upward trend for floods. But, a deeper insight shows that “small floods” are responsible for this trend. It is more likely that – beside other reasons – river regulation measures for flood protection and hydropower use lead to a channelization of the river and a reduction of retention areas, which result in a reduced flow time (Blöschl and Merz 2008).

Practitioners need data on past events for hazard and risk assessments or for planning of mitigation measures. The dating methods described earlier in this book are appropriate to provide at least part of the data required by practitioners. To give an overview on the type of data that can be obtained through the different methods and to display and compare the advantages and limitations of each method in a structured way, an overview (“checklist for practitioners”) is presented.

Torrential activity on alluvial fans and debris cones represents a major hazard in many regions where human settlements and infrastructure expand into potentially endangered areas. For the assessment of hazard and risk, detailed knowledge about past torrential activity and reliable predictions of possible future evolutions are of crucial importance. Data can be obtained from various data sources as has been shown in the previous chapters of this book. It is obvious that all methods have their advantages and their limitations and that their application always depends on site specific conditions and on the type of information sought. To obtain the best possible data with highest accuracy for a better estimate of hazards and risk on a fan or cone, a multi-method approach often represents the best solution. It is also important to notice that current frequencies and magnitudes will not necessarily remain the same in the future. We should be aware that there is a long chain of dependencies and uncertainties from a changing climate to the impact on natural hazards.