Scottish Government

3 DEBRIS FLOW INFORMATION SOURCES

by P McMillan, D J Brown, A Forster and M G Winter

There is a wide range of information and data available that is relevant to landslide activity in general and to debris flows in particular in Scotland. In this section key findings from the literature are presented along with those from the Project Workshop. The available geological, climatological, topographical and other relevant data is also examined.

3.1 KEY FINDINGS FROM THE LITERATURE

3.1.1 General

To allow the level of understanding of debris flows to be determined and applied to the situation in Scotland, literature was identified from international and more local sources to provide a broad view of the subject. Numerous workers have studied debris flows and a total of more than 100 papers, articles, publications, reports and books were identified and subjected to an initial appraisal, before selecting the most relevant information for full review.

The ability of debris flows to transport and erode large amounts of surface material at high velocities represents a potential hazard to structures, communications, farmland and people in downslope locations. Therefore, the following sections outline key characteristics of debris flows (identified in the literature), particularly with reference to Scottish occurrences.

In the longer term two other sources of information may be worthy of study in order to obtain information on slope stability on the metamorphic rocks of the Highlands and on mass movements of slopes in Galloway, respectively (Watters, 1972; Kirkby, 1963).

3.1.2 Debris Flow Mechanics - Initiation, Transport and Deposition

Debris flows may occur when hillslope sediment cover ( e.g. soil, loose rock and landslipped materials) becomes rapidly saturated with water and flows into a channel, or when excess water on slopes causes extensive hillside erosion and channel scour (Innes, 1983b). Intense rainfall, rapid snowmelt, lake/dam collapse, or high levels of ground water flowing through fractured bedrock provide the water required to trigger movement (Innes, 1983b; Pierson and Costa, 1987; Smith and Lowe, 1991). Three major types of debris flow can be identified: 'valley confined', 'open hillside' and 'slide initiated'. However, valley confined and open hillside flows are geometric descriptions of distinct flow types, slide initiated flows may be viewed as describing a mechanism which is potentially equally applicable to either of the other two types. An additional category of 'peat flows/spreads' can also be included in certain environments.

Flow typically requires relatively steep slopes and high topographic variance. The minimum slope angle for hillslope activity is approximately 30°, although it has been reported as low as 20°. In Scotland, hillslope flows occur on slopes up to 46°, the upper limit governed by the angle at which debris accumulates (Innes, 1983b), with the majority between 32° and 42°. Valley confined flows may occur on angles as high as 75° to 80° as debris emerges from the side walls of gullies, but slumps onto saturated materials in the gully floor can result in flow initiation on angles as low as 15° to 20° (Innes, 1983b).

Takahashi (1978) was able to quantify the upper and lower threshold angles in valley-confined debris flows based on thickness of debris, depth of surface water flow, degree of packing of the sediment, density of the sediment, density of the fluid, angle of internal friction of debris, cohesive strength and gravity.

The role of water is critical to debris flow as pore water pressures facilitate the motion of the granular material and water may initiate colloidal interactions between clay particles (Pierson and Costa, 1987): however, at what point does flow occur? Debris flows contain approximately 50% to 75% sediment by volume (Pierson, 1985) and such mixtures are 104 to 105 times more viscous than water (Johnson and Rodine, 1984). These mixtures possess finite yield (or shear) strength and this must be overcome by applied stress before deformation ( i.e. flow) is possible (Pierson, 1995). This stress is applied by the addition of water to the sediment mass and when the yield strength is overcome, the mass flows as a single viscous, plastic material.

Coarse material, including large boulders, is typically pushed to the head, flanks and upper surfaces of debris flows (Takahashi, 1981; Innes, 1983b; Coussot and Meunier, 1996) and thus inverse grading is observed. The means by which this inverse grading develops, and hence the nature of the flow, are not well understood. Therefore, much has been written in relation to flow mechanics of debris flows, with two main schools of thought. The two schools of thought are broadly as follows:

• Takahashi (1978; 1980; 1981) uses the principles of dispersive (or dilatant) forces (Bagnold, 1954) to explain debris flow mechanics. Dispersive forces transfer momentum from grain to grain and larger particles drift towards the zone with least shear (the upper part of a flow), hence inverse grading is produced.
• Johnson (1970) proposed that granular solids (boulders, gravel, etc.) are supported within the flow mainly by the strength of a fluid matrix comprising clay minerals and water (Bingham and Green plastic-fluid model of 1919) and that grain-to-grain interactions are trivial.

Neither mechanism has been proven satisfactorily, however Coussot and Meunier (1996), after Middleton and Hampton (1976), suggest a compromise, whereby cohesive or muddy debris flows are supported by the strength of the clay-fluid matrix, whereas cohesionless, or granular, debris flows are supported by grain-to-grain transport and dispersive pressure.

Debris flows are initiated when the applied shear stress exceeds the yield strength of the material involved, thus movement ceases when the shear stress falls below this limit. Deposition occurs en masse as a large plug of material and the flow essentially 'freezes' (Johnson, 1970; Smith, 1986). The deposits are a chaotic mixture of clasts, which are matrix-supported and commonly show a preferred alignment of their long axes parallel to the direction of flow. Flows are derived from heterogeneous debris and can mix with surface materials and flows from other sources, producing mixed populations of rounded and angular clasts of various size, with the exception of the coarsest clasts that dominate the frontal part of the flow.

Flow transformations are defined as changes in flow behaviour between laminar and turbulent states (Fisher, 1983). Surface transformations from the addition of fluid (dilution) or sediment (bulking) are common in debris flows. Landslides or rockfalls may be diluted to form debris flows, by the addition of water from snowmelt or heavy rainfall (Smith and Lowe, 1991), whereas stream or overland flows can bulk up with loose sediment and transform to debris flows (Pierson and Scott, 1985).

3.1.3 Factors Influencing Occurrence

The most important factor in debris flow occurrence is water. Heavy rainfall and/or snowmelt trigger the majority of flows, as the water mobilises the loose sediment. Furthermore, infiltration of this water into the soil is an important contributory factor. Caine (1980) and Innes (1983b) attempted to empirically quantify the amount of rainfall required to initiate debris flow events. Caine (1980) suggested a threshold for debris flow initiation, based upon data from North America, could be expressed in terms of a limiting curve, below which debris flow activity is unlikely to occur:

I = 14.8 D^_0.39

where I is the rainfall intensity (in mm/hour) and D is the duration of rainfall (in hours).

Innes (1983b) developed a similar curve illustrating the rainfall amount-duration relationship that has been reported as triggering a debris flow:

T = 4.9355 D^0.5041

where T is the total rainfall in the period (in mm) and D is the rainfall duration (in hours).

Debris flows in Scotland indicate that anything between 10mm to 75mm of rainfall per hour may be required to initiate these flows, significantly in excess of that predicted by the equation developed by Caine (1980). Current annual rainfall in Britain ranges from 1,000mm to 5,000mm (Meteorological Office) and, therefore, these figures represent significant amounts of rain falling in a short time. An early warning system in California suggests that for a rainfall of approximately 15mm per hour, the threshold time for the onset of mud/debris flows varies from 8 to 14 hours depending on slope angles and available material (Bryant, 1991).

Empirical evidence indicates that many Scottish debris flows are triggered by short intense rainfall events preceded by periods of heavy antecedent rainfall. In this context the two equations presented above will not provide a complete solution to the identification of likely periods of debris flow activity.

Soil type is an important factor in debris flow activity. Ballantyne (1981; 1986) and Innes (1982; 1983b) observed that debris flows are more abundant on slopes mantled by soils with a relatively coarse-grained matrix, including the ablation tills ⁶ common on the side slopes of many of Scotland's glaciated valleys, than on slopes with soils dominated by a fine-grained cohesive matrix. That granular materials are more susceptible to flow probably reflects the high infiltration rates associated with such soils. High infiltration permits a rapid rise in the water table during periods of intense rainfall, leading to an increase in pore water pressures and consequent failure and flow (Ballantyne, 1986). Clearly glaciofluvial ⁷ materials similar to those affected by the A9 flows north of Dunkeld in August 2004 are also vulnerable. Clayey soils, such as the lodgement ⁸ tills common in Scotland, are less susceptible to debris flow as bonds between particles provide cohesion and impede flow (Ballantyne, 1986). This can also be explained in terms of lithology. Where rocks yield sand-rich soils on weathering, such as the Torridonian sandstone of the NW Highlands and the granites of the Cairngorms, debris flow activity is more common (Strachan, 1976; Ballantyne, 1981). Tivy (1962) and Ballantyne (1984) suggest that areas underlain by schist, shale or greywacke, such as the Southern Uplands, yield clay-and silt-rich soils and are subject to debris flows only rarely. However, on-the-ground experience indicates that there is a comprehensive history of instability, including in the form of debris flows, in many areas underlain by schist. Good examples of such instability are the A83 in the vicinity of the Rest and be Thankful, A83 Loch Shira, A890 Stromeferry and the A87 at Invermoriston.

However, clay content is an important constituent in the mobilisation of flows. Although debris flows are rarely initiated in these soils, a cohesive debris flow has the potential for longer run-out distances. Clay impedes soil water movement and hence increases the possibility of soil saturation (Innes, 1983b). This fluid matrix is highly mobile and capable of travelling long distances, and Innes (1983b) found that debris flows in deep tills "may be two or three orders of magnitude larger" than in areas of thin cover. However, in the grain-tograin interactions of cohesionless (or granular) debris flows (Bagnold, 1954; Takahashi, 1978; 1980; 1981) energy is dissipated more rapidly and therefore, run-out is shorter.

Channel/slope geometry is an important control on the nature of debris flows. While confined flows will often travel further, relatively unconfined flows (floodplains/large U-shaped valleys) will frequently spread out to a greater degree forming a large lobate geometry. Where flow run-out is confined to tight valleys it will usually terminate close to the source, but the flow itself may incise deep channels (up to 5m) (Yarnold, 1993; Berti et al., 1999).

Plant roots play a critical role in stabilising colluvium ⁹ against failure on hillsides. Furthermore, vegetation cover provides interception of rainfall and encourages evapotranspiration, thus reducing both direct and indirect infiltration into the soil which can de-stabilise colluvium. Removal of vegetation by deforestation and heather burning increases the possibility of debris flow (Bovis, 1993; Benda and Dunne, 1997) by increasing water ingress into the soil. The effects of deforestation are known to endure for up to 10 years, with an associated elevated likelihood of instability during that time.

3.1.4 Hazard Identification, Assessment and Management

The body of literature on hazard identification, risk assessment and management of debris flows grows as our understanding of the phenomenon increases. Knowledge of debris flows may not allow us to prevent debris flows. However, with sensible hazard identification, assessment and management some degree of control is possible. The following paragraphs identify some approaches to the identification, assessment and management of debris flows. These issues are discussed further in later sections of this report.

Detailed hazard identification measures have been adopted which allow hazard mapping and zoning to be carried out. Hungr et al. (1987) selected various parameters (such as slope angle and channel geometry) to identify the potential impact of a debris flow in an upland area. Using an assumed mean deposit thickness and empirical run-out formulae (Takahashi, 1981) they calculated the extent of the debris flow and delineated three hazard zones (direct impact, indirect impact and flood zone). A year later, a debris flow occurred in the study area and its outline closely followed that of the predicted flow. Wilford et al. (2004) used 'watershed morphometrics' to recognise debris flow hazards. This method considers key attributes of debris flow generation including watershed area, length and shape, drainage density, relief, forest cover and extent of terrain greater than 30°. These data were statistically analysed to identify boundaries, used to determine whether debris flow would occur.

Larsen and Parks (1997) evaluated the correlation between roads and landslide distribution in Puerto Rico, as a measure of landslide risk. Where a landslide hazard had been identified as impacting a stretch of road, information on road type and traffic volume were used to provide an assessment of the risk posed by the landslides. Similar methods can be adopted for debris flows (Wieczorek et al., 2004).

In North America and Japan, warning systems are in place to manage debris flow activity. Advanced warning systems use rainfall data to predict debris flow occurrence ( e.g. Caine, 1980; Innes, 1983b) and provide an alert approximately 12 hours before the anticipated event. However, these systems are often unreliable and rainfall or its intensity may not be the sole cause of debris flow. In British Columbia, current contingency plans include monitoring by highway patrols. Certain river crossings and areas identified in debris flow hazard mapping are under full-time surveillance during periods of extreme weather. Patrols observe water discharge and flow discolouration and if significant changes are observed roads and/or bridges can be closed. Post-event warning systems include slide-warning fences. These consist of lengths of wire connected to control stations which, if impacted by debris flows/landslides, send a signal back to the control station. Appropriate stretches of road or railway can then be closed and emergency services dispatched (Hungr et al., 1987).

Hungr et al. (1987) suggest defensive measures against debris flows in source, transportation and deposition areas. In source areas these include reforestation and 'controlled harvest' schemes to reduce debris production resulting from deforestation or natural loss of vegetation. Road construction and management involves the avoidance and elimination of unstable cuts and fills, which could provide debris sources or initiation points. Channel beds and side slopes should be cleared of debris, and channels lined or controlled with check dams. In transportation zones flows may be trained by chutes, tunnels and deflecting walls or the channel can be diverted. In the deposition zone, measures such as stilling basins or retention walls can be utilised.

A Project Workshop was convened by the Scottish Executive as an integral part of this project and details are given in the Appendix. The information presented at the Project Workshop and the results from the discussion sessions form the framework of this report. Subsequently, work packages for the preparation of this report were allocated to targeted individuals.

3.2.1 Hazard Factors

Hazard factors are those conditions from the past ( e.g. geology), present ( e.g. slope angle) and future ( e.g. forecast rainfall) which determine either individually or in combination with other factors the potential for a debris flow event to occur, and thus the existence of that type of hazard.

Many hazard factors were identified at the Project Workshop and these are divided into a number of categories. These are developed further in Sections 5 and 6.3, but for the moment are listed with no attempt to relate factors to each other, to eliminate repetition, omission or, indeed, to ensure that each factor resides in the correct category.

3.2.2 Hazard Exposure Factors

Hazard exposure factors are those conditions, usually from the present, which determine,either individually or in combination with other factors, the potential for a debris flow hazardto interact with the trunk road network and road users.

A number of categories of hazard exposure factor were determined at the Project Workshop.In common with the hazard factors these are developed further in Section 6.3, but for themoment are listed as previously with no attempt to relate factors to each other or to eliminaterepetition or omission.

a) Road usage - traffic flows.
b) Road usage - traffic type.
c) Strategic importance.
d) Road geometry.
e) Sightlines.
f) Client expectations.
g) Environmental implications..
h) Road class.
i) Road gradient. s) Remoteness.
j) Serviceability.
k) Traffic management.
l) Availability of alternative routes.
m) Services/utilities.
n) Structures.
o) Proximity to hazards.
p) Pathway.
q) Emergency service access
r) Communications.

The identification of areas of potential slope instability in the form of debris flows will require two main data types.

First, information on the factors that cause slope instability is required. Such data include the following:

• Geometric data ( e.g. slope angle) which are best obtained from data sets such as digital terrain models as these can be interrogated to determine which slopes lie within a range of slope angles for example.
• Information on slope materials is also required from sources such as geological maps and geotechnical databases. In addition, data on land-use may also be required.
• Data to define the water condition of the slope may include rainfall data, storm track data, wind (drying by evaporation), plant cover (drying by transpiration), hydrology (surface water maps), hydrogeology (subsurface water maps), ground permeability maps and artificial drainage plans.

Second, information on past landslide locations that have affected the road network, their type of movement, their date of occurrence and, if relevant, reactivation dates may help to identify sites of current landslide activity and the factors that control their occurrence under present climatic conditions. Such data are contained within geological maps, landslide databases, ground investigation reports, PhD theses, and papers in technical/scientific journals.

3.3.1 Geological/Geotechnical Information

The British Geological Survey ( BGS) holds a large amount of geological, engineering geological and geotechnical data. These data are increasingly being held in digital form and are being accessed, viewed, analysed and presented using sophisticated computer systems (relational databases and Geographical Information Systems, or GIS) that enable them to be combined in different ways. Thus BGS offers not only large relevant data holdings but also the ability to manipulate the data to user needs incorporating new types of data into the system as the need arises.

6"/1:10,000 scale geological maps: The area covered by modern 1:10,000 scale and older 1:10,560 scale maps is shown on Figures 3.1 and 3.2. However, this gives no indication of the geological content of each sheet. Every geological map is to some extent a personal product, with a content reflecting the experience and professional interests of the geologist. The age of the mapping is not necessarily a guide to content or quality. Primary survey field slips (1:10,560 scale) often contain a wealth of data compared to those produced during more modern mapping when a more focused and time constrained mapping style was the normal procedure.

Figure 3.1 - Availability of 1:10,000 geological maps.

1:50,000 scale geological maps: The 1:50,000 scale maps offer complete coverage of Scotland. However, the quality and content of the mapping is variable depending on the age of the map and the mapping requirements for the sheet. In many cases the primary 'one-inch' maps have been revised to modern standards but some are still to be revised and the available maps are rescaled versions (to 1:50,000) of the earlier maps at the one-inch scale. In some cases the revised sheets in the Highlands have reused the primary survey superficial geology line-work. The extent of modern revision mapping (newly completed and ongoing) and the areas in need of revision are shown in Figure 3.3. Detailed mapping of Quaternary deposits (also called superficial deposits or drift) is largely a recent and ongoing commitment, and approved map-work is so far limited to areas around Aberdeen, Caithness, the Cairngorms and the Solway Firth (Figure 3.4).

Figure 3.2 - Availability of 1:10,560 (six inch) geological maps.

The National Landslide Hazard Assessment: The BGS national assessment of the potential for landslide hazard is based on geology, slope angle, and inferred material properties such as strength, plasticity, grain size, and discontinuity spacing. It is based on the UK 1:50,000 scale digital geological map and the assessment indicates how near conditions at a place might be expected to be to the onset of slope instability. As such it is an ideal land management tool with regard to maintaining slopes in a stable condition. If the component causative factors of the assessment are carefully examined appropriate stabilising actions, such as drainage or the reduction of slope angle may be identified.

It is a generalised assessment of the potential for a variety of types of landslide movement. As such its accuracy in the identification of the likelihood for any one type of movement is limited. However, the methodology is such that it is possible to recalculate the assessment using values that model more precisely the conditions for a single type of movement or reflect a particular geological environment. Refinement of the methodology has been achieved successfully in, for example, North London for failures in London Clay by talking into account its geotechnical properties and also in Builth Wells, Wales by including factors for the weathering behaviour of the local rock types. Field visits to sites of high landslide potential in the North London exercise have confirmed the presence of past and currently active landslides.

Figure 3.3 - Age of available 1:50,000 geological maps.

Figure 3.4 - Age of available 1:50,000 Quaternary (superficial deposits or drift) geological maps.

The National Landslide Database: The national landslide database is believed to be the most advanced of any landslide database in Britain and comparable to the best internationally. It stores up to 70 different types of spatial, temporal, physical and environmental data as well as socio-economic impacts. The reference for the original source of the information such as BGS maps, journal references, PhD theses (and so on) is easily retrievable through the user interface. This enables more detail to be obtained by reference to the source material. Information is stored in 30 fully relational data tables, complemented by a series of history and trigger tables that provide a secure audit trail for data entry and update. Four interfaces are available according to needs and there is a choice of hardcopy, computer or graphical information system front end. It currently contains nearly 10,000 entries for Great Britain over 1,200 of which relate to Scotland. The dataset has been compiled from a wide range of sources apart from BGS maps and it contains landslides that do not appear on the BGS maps. Similarly there are landslides on the more recently revised geological maps that have yet to be entered into the database. Revision of the database to include recent events or recently mapped landslides is an ongoing task but is constrained by available resources and other commitments.

The National Geotechnical Database: The national geotechnical database is used primarily to hold and analyse geotechnical data collected for the geological formation studies of the 'Engineering behaviour of British rocks and soils project' and the British geological hazards project. Thus, while it may not contain many data relevant to the study of debris flows and the Scottish trunk road network, it does offer an advanced, highly developed and proven geotechnical database that could be used with immediate effect to contain and analyse those data contained within the ground investigation reports relevant to the assessment of the potential for debris flow hazards.

Borehole records and site investigation records: The BGS borehole records collection for Scotland comprises borehole logs from site investigations, well bores, mineral bores, research bores which are held as paper copies and digital scans. They are available to BGS staff through a graphical data interface ( GDI) that enables borehole locations to be viewed in conjunction with a wide range of other data layers. The digital scan may be retrieved through the GDI (Figure 3.5). The areal distribution is irregular but mainly concentrated in urban areas and along linear routes.

The BGS site investigation report collection for Scotland contains mainly factual reports that describe the location, purpose and test data relating to the borehole database but may also contain interpretative reports and reports from other investigations such as penetrometer data and geophysical survey data. The geographical location of the investigation, as outline boxes, may be viewed using the BGS graphical data interface ( GDI) (Figure 3.6). The reports are held as a microfiche archive and have also been scanned (from the microfiche archive) but they have not yet been made available (readily) to BGS staff this is due to the difficulties of indexing them to a borehole level. It is intended to make an index to SI report level available in November 2004 within BGS.

Other data: The BGS data holding also contains numerous reports, sheet descriptions, sheet memoirs, geologist notebooks and photographs that date from the formation of the survey to the present which contain information relevant to hazard assessment but are too diverse to be listed or described in detail in this report. These data are most likely to be useful at the local or site specific level of hazard assessment.

Figure 3.5 - Illustration of the BGS borehole record holdings for Scotland.

Figure 3.6 - Illustration of the BGS site investigation record holdings for Scotland.

3.3.2 Scottish Executive Data

Generally Scottish Executive ( SE) data is linked to Arcview/Arcinfo, except as specified below. It should be noted that SE data holdings are not targeted towards the problem in hand nor do they include information of a specifically geological or geotechnical nature. They do however include topographical data which forms an important and integral part of any assessment and other data that may be integral to the process.

The Browser: This gives a forward, bi-directional video view of the entire trunk road network as viewed from the front seat passenger eye-view of a saloon car (the view is actually slightly downwards and to the left). This is useful for locating points of the network and getting a general feel for the landscape. Some detail can be picked out in the near-field and general shapes, for example whether a given area is or is not mountainous, from the far-field.

Topographical Data: Topographical data is available at 1:50,000 from Ordnance Survey ( OS) and at 1:10,000 to 1:1,250 from Landline. These are about to be replaced by Mastermap which will cover 1:50,000 downwards and better link with the GIS (Geographical Information System). Mastermap has been in production for a while now. It is understood that it is intended to run in parallel with Landline until Landline is phased out.

The Scottish Executive has also recently completed the purchase of full coverage NEXTMap coverage for Scotland, including the full digital terrain model. However, the use of the data is restricted to flood prevention studies, but it is expected that the use of the data could be expanded for a reasonable sum.

OS have a DTM (Digital Terrain Model) product called Landform PROFILE but it is not clear whether this forms part of the SE contract package and therefore whether it is/could be made available ¹⁰.

SERIS (Scottish Executive Road Information Service): This essentially comprises high-speed survey ( HRM) data from SE's Pavement Management System ( PMS), developed from the proprietary WDMPMS. The data includes bendiness, gradient, bend radius, bend start/finish points, bend length, crossfall (and so on) all of which might be useful in assessing the exposure of vehicles to debris flow hazard. The SERIS system is not linked to Arcview/ Arcinfo. Data can, however, be extracted and exported as shape files.

Traffic Flow Data: Traffic flow data is a fundamental requirement for determining the exposure of vehicles to debris flow hazard. The likely requirement is for 24-hour, 2-way AADT (Annual Average Daily Traffic). This is available for the entire network, although the level of confidence in this data can be variable. Traffic data at various levels is available from the SRTDb (Scottish Road Traffic Database) team at SE.

3.3.3 Climate Information

The Meteorological Office ( MO) has large quantities of weather (short term) and climate (long term) data over the UK and the ability to process such data with its powerful super-computing facility. Thus it is experienced at producing products to meet a wide variety of customer needs. BGS and the MO are already in discussions regarding the use of MO climate data in the assessment of geohazards and it has been apparent that such an application requires both sides to work together closely to unite their complementary skills and datasets in the most effective and appropriate outcome. It is unlikely that an 'off the shelf' data set for rainfall or climate could be applied successfully to the problem and a MO participation in the team using a customised dataset offers the best prospect of a successful assessment of the potential for debris flow hazard.

Rainfall Data: The UK observing network is made up of various categories of station, including synoptic stations that provide comprehensive hourly data; climatological stations providing daily (0900-0900 GMT) means, extremes and totals; and rainfall stations providing daily (occasionally hourly or sub-hourly) rainfalls. Synoptic sites ¹¹ provide data in 'real time', as do some climate stations, but the majority of climate and rainfall sites send in data at the month-end. There are also land stations of various types, including climatological stations ( SAMOS, auxiliary, SAWS/ SIESAWS/ MAWS are sub-sets of the synoptic type whereas CDLs and Health Resorts are mainly climatological). Rainfall stations are far more numerous and are shown on Figure 3.7. For rainfall there are also 5/15 minute areal data on 1km, 2km or 5km grids from the weather-radar network. In Scotland the radar sites are Hill of Dudwick (near Aberdeen), Stornoway and Corse Hill (near Glasgow). The data collected from these stations can be expressed over longer terms, commonly 30-year averages, and shown as yearly or monthly averages (Figure 2.7) ¹².

Figure 3.7 - Illustration of the Meteorological Office rainfall sites (image courtesy of the Meteorological Office).

Snow: Snowfall data is available nationally with 30-year average snowfall expressed on a monthly, seasonally or annual basis available on the MO web site (Figure 3.8). More detailed site-specific data would be available as required.

Storm Track Data: The MO have done some work on storm tracks across the UK and the number of storms passing through each year it has not yet appeared in print and it has probably not been done for Scotland alone. However, a contact at the MO has informally advised that it should be possible to generate such information from the available data.

Figure 3.8 - Example of Meteorological Office 30-year monthly average snowfall data (image courtesy of the Meteorological Office).

3.3.4 River and Stream Data

River and stream gauging data is also available from the Scottish Environment Protection Agency ( SEPA) ( www.sepa.org.uk). Preliminary attempts have been made to relate debris flow activity at Stromeferry to stream gauging data on the River Carron, some 5km to the north east, by (Nettleton et al., In Press). These indicate that a good correlation can be achieved provided that the data available is relevant to the area under consideration. Four events would have been forecast rather than the three experienced over a three year period. This work also shows a similar correlation with rainfall, albeit from a station at Plockton around 10km to the west.

The Flood Estimation Handbook (Anon, 1999) contains river and stream catchment data ( e.g. return period, capacity and flow) that could be useful in relation to determining slopes prone to debris flows.

3.3.5 Land Use Data

The majority of relevant land use data appears to be available from the Centre for Ecology and Hydrology. The Land Cover Map of Great Britain (1990) is a digital dataset, providing classification of land cover types into 25 classes, at a 25m (or greater) resolution .

The map provides:

• The first complete map of the land cover of Great Britain since the 1960s.
• The first comprehensive map of the land cover of Great Britain created from satellite information.
• The first digital map of national land cover.
• Accuracy to the field scale, checked against ground survey.

The Land Cover Map comprises 25 classes as listed in Table 3.1 those classes that are particularly relevant to the assessment of debris flows in upland areas are highlighted ¹³.

Data Availability: Data are available in two ways - within the Countryside Information System (1km resolution only), and as stand-alone datasets, at 25m and 1km resolutions. Stand-alone datasets are provided, to the customer's requirements, for any area of the country. Areas of data are cut out as a box by using Ordnance Survey grid references. Data is available at 25m resolution or 1km resolution in either a percentage or dominant value dataset. Other intermediate resolutions can be created as well.

Charges and licensing: Data charges are in three bands, according to end use, in accordance with NERC's Data Policy. These bands are commercial (highest rate), non-commercial, and research use (lowest). UK academics may be entitled to further reductions, subject to NERC arrangements. Data is supplied under licence. A wide variety of licences can be provided, from single user research licence to a corporate multi-user, multi-site licence.

Application to potential debris flow hazard assessment: This data set will have potential use in inferring the groundwater conditions because the nature of the vegetation will be influenced by the available moisture ( e.g. 'bog' implies permanent saturation). There are also implications for the reinforcement of slopes by plant roots and for the removal of moisture by plant cover, which will depend on the species present, and the maturity of the cover. Conversely, recently felled tree cover may both reduce the strengthening effect and increase the presence of water due to reduced root uptake in addition to the potential for rotted tree roots to aid infiltration during high rainfall events.

Limitations: It is acknowledged that some misclassification of the land use will have been made at the time of survey but this is thought to be relatively minor in nature. Also, the dataset was based on 1990 information and it is possible that the land use has changed since that time

Table 3.1 - The correspondence between the 25 'target' cover-types and the 17 'key' cover types of the Land Cover Map of Great Britain. Those classes denoted thus are considered to be of particular relevance to this study.

3.3.6 Digital Elevation Models

Digital elevation models ( DEMs) are models of the Earth's surface that can be used within a GIS environment to identify and quantify many aspects of topography such as slope angle and slope height, which can be incorporated into an assessment of potential slope instability.

There are several sources of DEMs in the UK, as follows:

• CEH Enhanced OS Landform PANORAMA Dataset 10m contours.
• OS Landform Profile Contours 5m contours.
• NEXTMap Britain Digital Surface Model.
• NEXTMap Britain Digital Terrain Model.

The Ordnance Survey's Landform Profile Contours and Intermap Technologies Inc's NEXTMap are the main sources of DEMs for national and regional scale studies. Lidar data are more accurate but expensive to fly and have only been acquired for a relatively small number of areas, largely for flood prediction studies.

Currently the NEXTMap Britain dataset is the most accurate national digital elevation dataset of Great Britain. It is a homogenous dataset that was captured with one sensor over two years. The other current national elevation datasets available for GB offer a much coarser resolution and lower vertical accuracy and were captured with multiple techniques over the course of forty years, thus affecting the level of detail, currency and consistency of the data. Therefore, the NEXTMap Britain dataset can provide a previously unattainable level of detail about the surface of the earth in Great Britain. Although both OS and Intermap are due to launch updated, more accurate versions of their DEMs in the near future it is likely that NEXTMap will continue to contain the more accurate data.

NEXTMap Britain is a modern and accurate digital elevation and image data set covering England, Wales and Scotland. It utilises Intermap's STAR-3i¨ IFSAR (Interferometric Synthetic Aperture Radar) which can quickly collect large areas of high resolution image data irrespective of the conditions. The product includes both elevation data and orthorectified radar imagery ( ORRI). The Digital Elevation Models ( DEM) comprise two further data formats, Digital Surface Model ( DSM) and Digital Terrain Model ( DTM), whereas ORRI provides an enhanced image with a ground resolution of up to 1.25 metres (Table 3.2).

Table 3.2 - Data contained within Intermap NEXTMap.

DEM comparisons: Staff in the BGS Remote Sensing laboratory compared the results of the available datasets for a section of the Trent Valley to determine their suitability for a range of geological projects. Their findings were as follows:

• CEH enhanced OS landform panorama: The Centre for Ecology and Hydrology DTM is based on Ordnance Survey ( OS) Landform PANARAMA dataset with 50m cell spacing. The CEH have enhanced the base OS product by enhancing the hydrological features. Full coverage of Great Britain is available to BGS's Science Budget projects but commercial projects have to enter complex negotiations with OS over derived data licensing. It is understood that OS have stopped supporting the base product and intend to withdraw it.
• OS Landform Profile contours: Ordnance Survey ( OS) Landform PROFILE contours, UK coverage. Useful for studying large areas but lacks detail in low-relief areas. Supplied as DXF format contours and spot heights and gridded in-house. The effective cell spacing is 10m.
• NEXTMap BRITAIN SK64 DSM: NEXTMap Britain Digital Surface Model ( DSM), good detail, some subtle topographic features visible. Sample supplied by NEXTMap Britain,
• NEXTMap Britain SK64 DTM - 5M cell: NEXTMap Britain Digital Terrain Model ( DTM), subtle topographic features are less distinct compared to the NEXTMapDSM. The vegetation and buildings have been removed, resulting in a 'bare-earth' representation of the surface.

3.3.7 Summary

The foregoing represents a fair summary of the available data. Inevitably there are some deficiencies in the data compared to the ideal. These are as follows:

• Slope materials: The information on the superficial geological deposits adjacent to the trunk road network is not of uniform age or detail and in some area it is likely to be less detailed than would be needed to assess the potential for slope instability. Localised mapping, especially regarding three-dimensional superficial material characterisation, may be needed to assess accurately the problem for areas identified as likely to be unstable based on initial assessment using available data.
• Climate: The rainfall and associated climate data are essential to assess debris flow hazards. Although the raw data exists it is likely that it will require Meteorological Office involvement to enable them to be understood, processed and applied in the most appropriate manner.
• Water: The available river and stream flow/catchment information has been relatively little used in the context of landslides and, more specifically, debris flows. This data shows considerable potential, in tandem with rainfall data, to assist in identifying precursor debris flow conditions.
• Slope:NEXTMap appears to be well-suited to the geometric requirements of slope instability assessment. However, access to the data needs to be organized. BGS is experienced in its use and has cover as far north as the Highland Boundary Fault but is currently in the process of obtaining the remaining cover through the NERC. Use in commercial contracts by BGS appears to be possible. The Scottish Executive has complete NEXTMap cover for Scotland, but its use is restricted to flood prevention studies.