Linking debris-flow hazard assessments with geomorphology
Introduction
Debris flows occur in various environments. In particular, in arctic and alpine regions, steep slope gradients and the availability of loose debris precondition these areas for landslides. Triggers are commonly prolonged or heavy rainfall often accompanied by snow melt. These natural processes are an important factor for landscape evolution. However, if humans are exposed either deliberately or unintentionally to these processes, natural events turn to natural hazards with the potential to cause disasters. Therefore, solutions which offer a reduction of hazard and in particular risk to the exposed population are demanded. These preventive solutions can be either direct countermeasures mostly based on geotechnical engineering such as deflecting dams, stop bars, and reinforcement of endangered structures or indirect procedures such as raising awareness of potentially affected people and land use planning.
To avoid the confusion resulting from the wide range of definitions, the term debris flow used within this study refers to the internationally accepted definitions given by Cruden and Varnes (1996) and Dikau et al. (1996). Debris-flow hazard analysis is generally carried out either by applying numerical and physically based models or using empirical approaches. While advanced numerical models are able to calculate debris-flow characteristics in detail, they have their limitations in application due to detailed requirements of input data. Rheological and physical–mathematical based modeling of debris flows needs detailed information on rheologic, hydrologic, and hydraulic properties. Numerous authors are working with such physical models (e.g., Costa and Wieczorek, 1987, Iverson, 1997a, Iverson, 1997b, Major and Iverson, 1999). A recent review on different approaches is given by Hutter et al. (1996) and Jan and Shen (1997). Most recent research on debris-flow modeling is summarized in Chen (1997), Wieczorek and Naeser (2000), and within the proceedings of the International Symposium Interpraevent, 2000a, Interpraevent, 2000b, Interpraevent, 2000c. In contrast, empirical models are based on few parameters and consequently allow generalized conclusions only. However, they generally offer an easily applicable and verifiable approach. Various authors have developed empirical relationships between debris-flow characteristics and conditioning parameters. A sound review of worldwide study has been published by Rickenmann (1999). Numerous other authors have developed site-specific relationships (e.g., Evans and Hungr, 1993, Hungr, 1995, Corominas, 1996, Bathurst et al., 1997, Wieczorek et al., 2000).
The need to carry out debris-flow hazard assessments in Iceland became evident as a result of a series of catastrophic snow avalanche and debris-flow events. Snow avalanches in Iceland have been studied for several decades. Monitoring of snow avalanches was established after an accident occurred in Neskaupstaður in 1974 with 12 casualties. Snow observers were hired in the most endangered villages as a local contact for Civil Defence Authorities. Responsibilities include the registration and analysis of snow conditions as well as specific snow avalanche events. Despite these efforts, two snow avalanches in Súðavík and Flateyri in 1995 caused 34 fatalities. As a consequence, the snow avalanche department of the Icelandic Meteorological Office (IMO) was extended, and the laws and regulations concerning hazard mapping for snow avalanches and landslides (including debris flows) in Iceland were revised. Landslides were included because of numerous failures also posing a serious threat to communities. Along with these revisions, older hazard maps became invalid. According to this new regulation (The Ministry of the Environment, 2000), risk zones for snow avalanches, debris flows, and rock fall have to be prepared for and applied to the endangered communities on request. The Icelandic Meteorological Office (IMO) is responsible for carrying out the avalanche and landslide hazard assessments.
Landslides, and in particular debris flows, occur regularly and cause considerable damage in Iceland (Jóhannesson, 2001). A historical review of landslide events was first undertaken by Ólafur Jónsson in 1957. This review is based on magazines, newspapers, old annals, etc. and demonstrates a nationwide landslide occurrence throughout Iceland. Often, only the largest events or those causing server damage were noticed and/or recorded, which is indeed a problem common to historical reports on landslides (Glade, 1998). Consequently, a direct comparison of triggering events is rather difficult because the consequences of a ‘nontriggering event’ might just have not been recorded. Thus, the derived frequency is a minimum information on landslide occurrence only, real frequency might be higher. The current landslide database is still in paper format only, but a digital inventory and a GIS database are being developed by IMO in cooperation with the Icelandic Institute of Natural History (IINA). If such a database becomes available, further analysis of landslide-triggering conditions might be possible despite the limitations given by the recording procedure (e.g., Glade, 2000).
Based on this history and legislative demand, landslide hazard assessments have been developed for Seyðisfjörður and Eskifjörður, east Iceland and for Bíldudalur, Bolungarvík, and Patreksfjörður in the Westfjords. In eastern Iceland, an Austrian method was applied, which is based on semiphysical models. The description of the method and the calculated landslide hazard zones are given in Jensen and Sönser, 2002a, Jensen and Sönser, 2002b. For the Westfjord communities Bolungarvík, Bíldudalur, and Patreksfjörður, Glade and Jensen (2004) developed a landslide hazard assessment scheme based on empirical models. Both approaches are applicable to other Icelandic villages.
An assumption inherent to most debris-flow assessments is unlimited sediment supply. Although already conceptually addressed by various authors (e.g., Zimmermann and Haeberli, 1992, Haeberli, 1996, Zimmermann et al., 1997, Bovis and Jakob, 1999), this underlying assumption will be investigated in more detail for the Bíldudalur study area. The availability of sediments in different types of sediment storages (Moore et al., 2002, Schrott et al., 2002) has important consequences for any hazard analysis. Clearly, rheology and mechanics of debris flows are an important considerations in many situations, however, if the sediment has been removed in the source areas, further hazard in the near future can be ranked as low. A different hazard ranking might be derived from physical or empirical models based on past events because they assume unlimited sediment availability. Therefore, not only the debris-flow process itself has to be determined, but also sediment sources, sinks within catchments, and rates of refill of sediment storages after removal (i.e., after a debris flow has occurred) have to be evaluated (Bovis and Jakob, 1999). This refill of different sediment storages is termed sediment reproduction in the current study.
This study aims to evaluate the Icelandic Westfjord debris-flow hazard assessment with respect to the assumption of unlimited sediment availability. The overall aim is thus to assess debris-flow hazard by taking into account geomorphic preconditions, processes, and sediment supply. This includes approximation of sediment reproduction rates and delivery of sediment from other sources which is finally available for debris flows. To determine the potential of these processes correctly, it is necessary to assess the availability of sediment to be mobilized by the debris flow (e.g., Keaton and Lowe, 1997, Bovis and Jakob, 1999). Only the high content of sediment, and in particular the content of large clasts, changes a large flood into a disastrous event. Investigations of sediment reproduction and sediment movement are common procedures in geomorphic analysis and are highly applicable to debris-flow hazard assessments. Specific goals of this study include
- •
identification of geomorphic processes within the study area;
- •
determination of spatial distribution and level of activity of geomorphic processes;
- •
detailed characterisation and categorisation of debris-flow incidences;
- •
definitions of areas most susceptible to debris flows based on field evidence;
- •
assessment of sediment available for debris-flow occurrence;
- •
reconstruction and scenario modeling of debris-flow movement to define potential run-out zones and depositional areas;
- •
comparison of modeled sediment requirements with current estimated sediment reproduction rates; and
- •
approximation of over- or underestimation of debris-flow hazard schemes.
While the first five goals require a detailed field investigation, the last three goals are most crucial and important. For example, if a debris-flow has occurred and removed all the material from the stores, and if these stores will not be refilled, or if the rate of storage refill is very low (e.g., 500 years), there will be no apparent danger for the near future from that particular site. Therefore, under these conditions, even rainstorm events with a 10-year return period do not initiate debris flows. They trigger only floods with reduced damaging potential. But if the store is filled within the next 2 years, the return period of the debris flow is equivalent to the return period of the triggering climatic event. Therefore, it is critical to investigate different sediment sources, mobilization, and reproduction in more detail in a specific study area.
Section snippets
Bíldudalur, northwest Iceland
This study is carried out for Bíldudalur, one of the three previously mentioned communities in the Westfjords. Bíldudalur is located in Bíldudalsvogur in the Arnarfjörður fjord in the southern part of the Westfjords in Iceland (Fig. 1). Legally, it is part of the Patreksfjörður community. This community is exposed to various natural processes such as snow avalanches, slush flows, rock falls, debris flows, and flooding.
The top of Bíldudalsfjall mountain above the village is 460 m a.s.l. Two
Field mapping and aerial photography
To determine overall dynamics of the general geosystem, the geomorphic situation has been mapped. This includes all major past and recent processes and forms in the landscape. Focus was given to the specific features of the particularly important landslide processes. The applied geomorphic mapping legend is based on the Geomorphological Legend Key, developed within a large German scientific project on standardising geomorphologic mapping (Leser and Stäblein, 1975). This legend key has been
Geomorphic map
The geomorphic map shows that, in particular, periglacial, gravitational, and fluvial processes and forms dominate the glacial-shaped landscape (Fig. 9). Periglacial processes on the plateau of Bíldudalsfjall include stone sorting, solifluction, and gelifluction, as well as a high activity of bedrock weathering. Stone sorting features on plateaus change with increased slope angle towards the slope from rings (flat area) to stripes (creeping area), indicating a slow but continuous material
Sediment supply considerations
It has to be pointed out that the absolute values of both debris-flow calculations and sediment reproduction rates should be considered as general trend. Despite the limitation of accuracy, however, this analysis demonstrates that there is an inherent danger of overestimating the hazard posed by debris flows when considering rainfall events only. As the conceptual model in Fig. 11 shows, rainstorms with similar totals (circled numbers) do not necessarily always produce similar debris-flow
Discussion and conclusions
It is evident from local observations that debris flows constitute a potential threat to the communities. The geomorphic map displays the dominant natural processes operating in the study area and gives some indication of state of activity. It is a first approximation of temporal occurrence and current sediment distribution. The information gained provides the basis for further scenario modeling.
The more detailed debris-flow map indicates the distribution of debris flows and their forms and
Perspectives
This analysis is based on a number of assumptions and gives approximations only. Numerous issues have been addressed, which should be investigated in more detail in the future. Work might include
Acknowledgements
Rainer Bell helped during field work and supported this study continuously. Esther Jensen and Kristján Ágústsson discussed related issues during field work, and Tomas Jonassen and Trausti Jonasson, all from the Icelandic Meteorological Office, supported this study. Thomas Sönser and Siegfried Sauermoser discussed field observations. Ole Humlum, Colin Ballantyne, and Albert Pissart gave helpful general comments on weathering rates and sediment mobilization in Arctic and in particular for
References (121)
Do periglacial landscapes evolve under periglacial conditions?
Geomorphology
(2003)The geomorphic significance of rock glaciers: estimates of rock glacier debris volumes and headwall recession rates in west Greenland
Geomorphology
(2000)- et al.
The paleomagnetism and geology of the Patreksfjördur–Arnarfjördur region of northwest Iceland
Tectonophysics
(1975) Solifluction rates, processes and landforms: a global review
Earth-Science Reviews
(2001)- et al.
Rockfall activity from an alpine cliff during thawing periods
Geomorphology
(1999) Slow mass movements and climatic relationships, 1972–1994, Kapp Linné, West Spitzbergen
Dating slope deposits and estimating rock wall retreat in Northwest Spitsbergen by lichenometry
Geografiska Annaler
(1986)Holocene rockfall retreat in Svalbard: A triple-rate evolution
Earth Surface Processes and Landforms
(1997)- Arnalds, O., Gretrasson, E., 2001. Soil Map of Iceland. Soil Conservation Service, Agricultural Research Institute....
Eine Abschätzung von Schuttproduktion und Schutttransport im Bereich aktiver Blockgletscher der Schweizer Alpen
Zeitschrift für Geomorphologie, Neue Folge
(1977)