The Charnawati Experience - 10 years on

Emergency works, carried out in the Charnawati area in 1988 - 89, provided temporary stabilisation of the landslides caused by the disastrous monsoon flood of 1987 and allowed the road to be reopened. This gave a breathing space to those involved in the project. It became clear during the assessment of possible alternative remedial works that the problems encountered could be solved only by using innovative new technologies. Construction techniques such as passive earth anchors, drilled horizontal drains and concrete armour blocks helped to stabilise the road corridor and significantly reduce the risk of further damage. The introduction of new construction techniques into Nepal, with the associated transfer of technology to the staff of both the Department of Road and the local contractors involved, resulted in a project with long lasting benefits to the community and minimised the maintenance burden to the country. The knowledge passed on to local contractors provided an opportunity for these techniques to be implemented in other locations.

Charnawati – events of 1987

The Lamosangu-Jiri Road (LJR) connects the Arniko Highway (linking Kathmandu and Tibet) with the communities of Dhandabahar, Charikot and Jiri. The road lies to the east of Kathmandu in the hills of Nepal and was constructed be-tween 1975 and 1985 with assistance from the Swiss Government. It opened up access for the people of the area to markets in Kathmandu and provided opportunities for education.

The Charnawati river crosses the LJR through a steep sided valley about 45 from Lamosangu in the so-called mid-hills of the lesser Himalayas. These hills are incised by many rivers and are being continuously reshaped by active erosion processes.

Monsoon floods in 1987 destroyed the concrete bridge of the LJR crossing the Charnawati River. The flooding also caused extensive localised damage to the steep river banks and the process of erosion and sedimentation created a new riverbed. Large landslides were triggered and a section of the road itself was destroyed (see Photograph 1 and Figure 1 on next page).

In September 1987 the Swiss Agency for Development and Cooperation (SDC) requested ITECO to evaluate the damage along the LJR particularly in the area of the Charnawati catchment. The report of this mission presented proposals to reinstate the river crossing and reopen the road to traffic as soon as possible. Based on these findings ITECO prepared designs, implemented a semi-permanent rehab-ilitation of the Charnawati crossing and developed a first vision for a permanent solution for the road alignment within the Charnawati valley.

At the end of the 1987 monsoon season ITECO started the first phase of the CHRP*, and mobilised resources to start reconstruction of the bridge and road approaches in 1988.

* The project was initially known as the Charnawati Rehabilitation Project but was renamed after 1990 to the Charnawati Rehabilitation Programme – here the abbrev-iation CHRP will be used for both.

Rapid response needed

Rehabilitation works were designed to rebuild the road and bridge, stabilise the Charnawati river bed and stabilise the main slide areas crossing the road, all of which needed to fit into any long term rehabilitation concept for the area.

A 21 metre long modular steel truss Bailey bridge replaced the original 10 metre span conc-rete bridge. The bridge type was chosen for its rapid erection to allow the road to be reopened before the start of the 1988 monsoon season and for the ease with which it could be removed if a new structure was constructed some time in the future.

The four metre wide approaches to the bridge were reconstructed according to the standard of HMG Nepal Department of Roads (DoR) for Class II Feeder Roads.

Check dams in the river together with associated drainage and vegetation structures stabilised the river bed and adjacent slopes and provided short term protection to the bridge and the road app-roaches prior to the next monsoon (see Photograph 2).

Design for long term stability

The 1987 flood event and subsequent soil hazard mapping [1] clearly showed that the damaged section of the Charnawati crossing is in a zone with a high risk of instability. Alternative road alignments, with a river crossing further upstream, would have been longer and would have crossed similar risk areas. The engineers therefore decided to retain the original road alignment with two main areas of intervention – in the river bed to protect the slopes from undercutting by flood waters and on the slopes to stabilise the slides.

The river bed ..

The Charnawati River has, in the project area, an average width of 25 metres. The average grad-ient is 11 percent at and just below the bridge and a steeper gradient of 18 percent further downstream.

The design objective for the upper section (upstream and immediately downstream of the bridge) was to stabilise the river banks and to secure and raise the river bed level. This object-ive was achieved by constructing nine gabion check dams (see Photograph 3) and gab-ion retaining walls on both banks over a length of 400 metres. A 10 cm thick layer of concrete on all exposed surfaces prevented abrasion of the gabion wires.

There are no flow records available for the Char-nawati river prior to 1989. The hydraulic comp-utations, based on a design flood of 120 m3/s (estimated as the 20 year flood for the 15 km2 catchment area), showed that average dam heights of 5 metres and a spacing between dams of 40 metres would provide adequate protection. The design took into account that the maximum depth of river bed excavation, using manual labour, was only 1.5 metres.

The design objective for the lower 500 metre long section was to stabilise the river bed at its post-1987 level and to prevent further erosion and triggering of new slides on the left bank.

Design studies showed that the limit of con-ventional gabion structures for a river with such large bed gradients had been reached. The en-ergy dissipation capacity of closely spaced check dams is drastically decreased and would nec-essitate a very large stilling basin at the end of the series of dams. In addition the quantity of gabions needed would have led to very high costs and long construction times. Rocks to fill the gabions would probably have been produced by blasting and crushing boulders from the river bed, a situation that would have been counter productive to an increase in river bed stability.

Conventional rock protection was not possible as there were insufficient large boulders (weight greater than 10 tonnes) in the river and no other source of large quarry stone in the vicinity – even if the means to transport and place such large boulders had been available at the time.

This led to the novel approach of using so called flexible river protection (see Box 1) which was found to be technically more beneficial, feasible to construct using simple techniques and portable equipment and of similar cost to conventional works.

The principle of flexible river protection works is to increase the flow resistance of the existing river bed by placing macro roughness elements – at Charnawati cast in situ concrete elements – in the bed and by moving the large boulders which were present a short distance to locations where they could contribute to the flow resistance [2].

Results from 1:40 scale model tests conducted at the Federal Laboratory of Hydraulics, Hydrology and Glaciology (VAW) in Zurich, Switzerland indicated that optimum performance would be achieved with 2.5 metre high Concrete Armour Block (CAB) elements cast in a three dimens-ionally transposed H-shape for optimum inter-locking and with a minimum weight of 10.5 tonnes (standard CAB element) [3].

The initial model tests indicated that for best results the elements should be placed in the form of discrete, semi circular "carpets" over the width of the river bed (see Photographs 4, 5 and 6). Reference testing showed that severe erosion of the unprotected river bed would start at flows greater than 60 m3/s. With the bed protected by CAB elements local bank erosion would only start at discharges higher than 120 m3/s and would then gradually increase. The limit of control of the discharge was found to be around 160 m3/s (estimated as the 100 year flood). Flows of this (or greater) magnitude would almost certainly modify the whole morphology of the river valley. Any attempt to protect the bed and slopes in such a scenario would be both ineffective and prohibitively expensive. The tests showed also that CAB elements continue to contribute to erosion protection even when moved by the flow from the river bank into a scour hole or when a number of CABs are moved and form inter-locking ramps.

.. and the slopes

The slopes of the slides were between 28° and 35°. The bed rock of the area is predominantly gneissic and is overlain by a 20 to 50 metre thick soil layer. The upper part of this soil layer generally consists of silty-sandy material with boulders. The material origin is residual (weathered gneiss), colluvial (old slide material), alluvial or glacial (former moraine deposits). Soil test results indicated that the internal angle of friction of the soil material was approximately 33° and value of cohesion virtually zero. After toe erosion by the river the slopes thus became

-----------------------------------------------------------------------------------------

Taking ideas from the sea into the mountains

The approach of using concrete elements as flexible river control works was developed, using methods commonly implemented for coastal protection, at the Swiss Federal Institute of Technology in the 1980s. It was applied in Switzerland mainly in the Canton of Uri [5] where floods in 1987 with peak discharges of 600 to 1000 m3/s caused severe damage. The design of these structures was developed partly in parallel to the Charnawati works and the projects mutually benefited from each other in this regard. Element shape adopted in Switzerland, a prism diagonally split, is more suited for industrial production but has a lower interlocking capacity than the shape used at Charnawati.

Design of flexible protection works can be based on theory alone (estimating the flow resistance of rough bed layers) [6] for very simple cases but model testing to refine the theoretical results is almost compulsory for more difficult environments (rivers with curves or non homogeneous bed materials).

With theory, model testing and research and using the experience gained from past projects (such as Charnawati) this development process has produced models for use on other projects.

There was thus considerable synergy between CHRP and projects in Switzerland.

---------------------------------------------------------------------------------------

inherently unstable and the slightest increase in pore water pressure triggered the slides.

By increasing the toe load, reshaping the crest and (where necessary) installing a drainage system of gabions (up to two metres deep) with an aggregate filter the slides in the upper river section (near the bridge) were stabilised. The LJR project had developed a variety of bio-engineering techniques - planting of utis (local type of alder), brush and hedge layers and grass seeding for slope protection. Adopting these measures on CHRP greatly improved the stability of the slopes.

The slides in the lower section of the river extended upwards from the river bed to the road, a distance of 100 to 150 metres. Site observations and geophysical exploration rev-ealed that a denser, less permeable layer was located parallel to the slope at depths of two to seven metres. The interface of this and the over-lying material appeared to be the slip plane of the slides.

The ground water table in the area is generally situated at depths of three to nine metres during the dry season and locally rises to the surface during the monsoon season. Stability comp-utations showed that lowering the water table by one metre would increase the slope stability by as much as 10 to 15 percent.

Conventional drainage construction in Nepal at the time was labour based and limited in depth to less than two metres below existing ground. The slide depth and the impossibility of simply reshaping the slide crest (since the road was loc-ated on top of it) demanded the use of more sophisticated techniques combined with the trad-itional design approach.

Drainage works included surface drains and drilled horizontal subsurface drains above the slides together with complementary gabion drains in the slides. Anchored retaining walls were designed to locally stabilise the slide crest (see Figure 2 and Photographs 7, 8 and 9).

The decision to introduce such drilling techniques was based on the fact that stabil-isation of the slides would not be possible without such anchors and drains and that similar problem sites (where this method could be utilised) are found all over Nepal.

The non-prestressed earth anchors, 15 to 25 metre long steel rods each with a design load of 250 kN, extend a minimum of seven metres into stable ground below the slide.

The drilled horizontal subsurface drains, consisting of 20 to 40 metre long polyethylene pipes, continue beyond the slide planes, some six metres below the surface, and drain the material both in the slide area itself and at least a further five metres below the top of the water table. This type of solution contrasted with methods commonly used in Nepal at the time which consisted of a 1.8 metre deep hand dug main drain and a tributary drainage system.

Feedback from site refines the design

One of the successes of CHRP was the way in which feedback from site to the design office fine tuned and optimised the performance and cost of the finished product. (see Box 2 on next page). As a result of an unexpected early flood in May 1990, works at one particular location were significantly affected. The design engineers, together with experts from VAW [4], reassessed the situation and modified the layout of the CAB carpet at a number of critical locations. Tilting of the CABs from the as-cast position further improved their performance (see Figure 3 on next page). In addition the designers suggested that a number of different sizes of CABs could be used to optimise cost and performance. Accordingly the site team modified the existing steel forms (at minimal cost). This enabled the production of a range of sizes of concrete blocks from less than 3 tonnes to more than 24 tonnes. This feedback and optimising process continued over a period of nearly two years and was of direct benefit in achieving the goal of stabilising the slides in the project area.

By combining the results from:

monitoring of slides (to define the best locations for drainage and retaining structures);

in-situ anchor pullout tests (to check the number of passive anchors required for structures); and

monitoring of the ground water table (to decide on the options for the drilled horizontal subsurface drains

the designers provided the site with improved and more cost effective solutions.

Innovation demands site management changes

Prior to 1988 small, local contractors carried out most construction work in Nepal on a set unit rate basis. Contract documentation was almost never used and management contracts for con-sultants were unknown.

Contracts

The early works in 1988 and 1989 were of an emergency nature and, for expediency, more than 70 local contractors, mainly acting as suppliers of labour carried out the initial work. ITECO staff, together with counterparts from DoR, directly managed the works on site. Where feasible, in terms of quality and progress, construction works were carried out using labour based methods well established in Nepal. After 1989 the nature of the project changed and required the project organisation to adapt to the use of innovative technology. The profile of the contractors involved changed from local, piece work contractors to regional and national class contractors.

Three contracts for:

river control works at Charnawati;

landslide stabilisation works at Charn-awati; and

stabilisation works at Kalimati

were awarded. This latter contract related to works at Kalimati, a large slide area upstream of the LJR crossing of the Charnawati. The project continued to execute directly some of the specialised work such as anchoring and drilling of drains.

Management

Management ideas new to Nepal introduced at that time included:

a `Steering Committee' was set up to inv-olve DoR and SDC, together with ITECO as consultant, in transparent decision making processes and to integrate all involved stake-holders;

FIDIC contract documents were intro-duced; and

consultancy services were provided under a management contract.

DoR and other donor agencies adopted and refined these approaches and management tools in subsequent projects.

Workforce and plant

Over the entire project period, the labour input amounted to more than 380,000 person-days (see Box 3 on next page). The project procured and imported large quantities of equipment. Items for the anchor and CAB works included crawler drill, portable drilling equipment with comp-ressors, cable crane and concrete batching plant. Standard construction equipment (dumper, concrete mixers, vibrators, hand tools etc.) was also imported into Nepal at the start of the start of the

-----------------------------------------------------------

Nature lends a helping hand

Boulder Shifting: Soon after casting the first CABs the idea of including natural boulders in the flexible river protection scheme was raised. Trial runs with available hand tools - crowbars, winches and jacks proved successful and larger hydraulic jacks were procured. Boulders with a weight of up to 15 tonnes were regularly moved over several metres (see photograph 10 above). Under favourable conditions even larger boulders could be shifted.

Boulder Anchoring: CHRP intro-duced the use of anchored retaining structures in Nepal and in June 1989 drilling of passive earth anchors for retaining walls and anchor points to hold drainage structures was comm-enced. The method proved effective and ways were investigated to red-uce the cost. This led to what was called ‘boulder anchoring’, a new and economic anchoring technique appropriate for use in steep colluvial slides. The basic idea of boulder an-choring is to use a strategically imp-ortant boulder as a retaining struct-ure. Three or four 15 m long passive anchors were drilled (see photo-graph 11 at left) and installed thr-ough the boulder, thus saving the cost of a concrete structure. Alth-ough this technique proved very su-ccessful these anchors would now be prestressed to prevent initial movements required to stress anchors.

------------------------------------------------------------------

project due to the absence of sufficient numbers of equipment available locally. Introducing the WYSSEN cable crane and the LUMESA drilling machine to Nepal’s contractors allowed new construction methods which previously would have been totally impractical.

Training

The early work of CHRP, from 1988 to 1989, used techniques already well established in Nepal. With the application of the new tech-niques in design and construction after 1989 training became an important part of the project (see Box 4).

Outputs and Costs

The main quantities of work carried out on CHRP are shown in Box 5 (see next page). The overall project cost, in 1991 value, amounted to USD 8.5 million (12.4 million Swiss Francs or NRs 300 million), resulting in a present day value of about USD 10 million.

Ten years on

At the end of the project in late 1991, the newly rehabilitated area was still fragile. Consequently, DoR and SDC agreed to finance a maintenance and monitoring project which continued until 1993. During this period, a local team consisting of one headman and 6 labourers implemented further bio-engineering structures. ITECO and DoR engineers continued to observe the entire area and to record data from all the installed extensometers, piezometers, anchor gauges and the more than fifty monitoring points set up in the former slides.

The monitoring system confirmed the effective-ness of the remedial measures and revealed that all slide movement had virtually stopped one year after the completion of the construction works.

The records available indicate that the level of the water table decreased in the Charnawati slides (resulting in an increased sliding res-istance). The anchor forces in the retaining stru-ctures were found to have increased.

The monitoring work ceased at the end of 1993. Unfortunately the instrumentation, together with most of the survey points is now totally destroyed. Accordingly the performance of the works from 1994 to 2003 can be assessed only by comparing photos taken in the late 1980s and early 1990s with the situation on site today. The site is now extensively covered with vegetation and most of the slides are stable. The gabion structures have suffered only minor damage, including some human intervention (wire mesh removed) and the river bed is now at more or less its 1991 level. The anchored structures are all intact with only some localised damage due to rockfalls to the anchor heads. Some movement of the CABs is evident but in general the overall interlocking effect of the carpet is now even better than at the end of construction.

The drains and tributaries continue to function and trees and vegetation surround most of the structures away from the river bed itself. The bio-engineering structures themselves are fully established and are performing as intended.

Impressions of Charnawati for visitors today

Photographs 12, 13 (next page) and 14 to 18 (page 8) compare the devastated barren ground left behind by the 1987 flood and subsequent landslides and the situation today. The first impression that a visitor today gets is of course the green scenery followed by surprise at the amount of vegetation growing on the gabion and around the concrete structures in the river bed (Photograph 19 on page 8). A closer inspection reveals the variety of plants and the integration of the `soft' bio-engineering measures with the `hard' structures. The local community now uses the upper parts of the former slide areas either for grazing of cattle and sheep or as source of fodder for the animals. Provided the area is not over-grazed, over-cut or – more importantly that the forest is not reconverted into irrigated fields – the Charnawati valley seems to have reached a state of equilibrium and the environment can continue to be compatible with the low-impact activities currently taking place.

There is a need to carry out some corrective measures to restore the works to their 1991 state. These include the conversion of a closed drain to an open channel and installation of ten additional CABs in two locations which would correct local flow and scour patterns. There are a few small isolated slides which could be stabilised with vegetation structures. The total estimated cost, to bring the site and structures back to the state existing at the end of 1991, would be about USD 15,000 or about 0.3% of the original construction cost.

Lessons learnt

Development of the area

The rehabilitation works, besides providing short term employment and training opportunities to local people assured uninterrupted access to the area from Charikot to Jiri. This contributed to the economic and social development of the area. The importance of road access is illustrated by the traffic which increased from 35 vehicles per day in 1987 to 150 vehicles per day in 2002.

Sustainability

The assessment more than 10 years after completion of the works has clearly shown that the CHRP works have protected the LJR in the years since 1991. While other unstable areas in Nepal are ‘stabilised’ by nature itself – mainly re-vegetation – the extent of instability at Charn-awati was such that intervention by man was required to lend a helping hand to the healing processes of nature.

The fact that no money for recurring maintenance has been spent at Charnawati in the last ten years indicates that the works carried out have proven long term viability.

The initial design approach which explicitly focused on low maintenance requirements proved to be appropriate.

Management

On the project management level, CHRP introduced the FIDIC contracting system to the Nepalese contractors. The contractors became aware that they had both duties and rights under a contract and made the construction industry more efficient, effective and transparent.

Design

The successful merging of conventional structures, low cost bio-engineering techniques, and advanced technology such as drilling for landslide stabilisation makes this site an ideal showpiece for such works in South Asia (see comments of some of the players from CHRP in Box 6).

While CAB protection is useful in providing and improving stability in terrain such as Charnawati there are limitations to its use. A CAB system, similar in concept to that used at Charnawati, was constructed along the bed of the Rapti Khola river adjacent to the Tribhuvan Highway between Bhainse and Hetauda (in the south east of Nepal) in 1994. A large flood in this river in June 2002 caused significant damage to the protection works.

The case highlights the fact that:

the availability of accurate hydrological information in Nepal is still insufficient to adequately estimate floods and hence the det-ermination of the design flood remains a critical design issue;

the ability of flexible river protection works to cope with massive floods is limited by the max-imum practical size of CABs able to be cast and handled on site (generally around 30 tonnes); and

the design process is still largely empirical. In areas of high flow or unusual geometry physical model testing is recommended.

Technology transfer

The transfer of knowledge and hardware (such as drilling equipment and accessories, formwork for CABs etc.) from CHRP continues to be of sign-ificant benefit to the local construction industry, donors, and planners. They are now much more aware of alternative landslide and riverbed stab-ilisation techniques and, equally importantly, have on hand suitable equipment for the implementation of those solutions.

The market for these new approaches however developed slowly (see Box 7).

Replicability beyond Charnawati

Rehabilitation of roads had traditionally focussed on pavement rehabilitation. CHRP introduced and reinforced the concept that stabilisation of the road corridor was equally important. CHRP concepts are becoming more widely considered and designs produced on CHRP are being modified and adapted for use elsewhere.

The following list includes projects that applied ‘new’ CHRP techniques and indicates the source of funding, the years over which the project was implemented, the project location and the works undertaken which relate to CHRP.

Project: Road Flood Rehabilitation Works, Arniko Highway and Thankot-Naubise Road

Funding: World Bank and HMG/N, 1991-1995

Work: 3,800 anchors and 1,400 CABs installed (note – many CABs were smaller and lighter than those used at CHRP), 6,700 m of french drains constructed and significant numbers of boulders shifted

Project

Project: Arniko Highway Project

Funding: SDC and HMG/N, 1992-2003

Work: 660 anchors and 3,700 CABs installed

Project: Reconstruction of Bhainse Bridge and Bhainse to Hetauda section of the Tribhuvan Highway

Funding: KfW and HMG/N, 1995-1999

Work: 465 CABs installed

Project: Malekhu-Dhading Besi Road Project

Funding: KfW and HMG/N, 1996-2000

Work: 550 anchors and 200 CABs installed, 3,000 m3 of boulders shifted

Project: Dharan-Dhankuta Road (Maintenance)

Funding: ODA

Work: 200 anchors installed

References:

[1]KRAEHENBUEHL Juerg, OSTERWALDER Walter, WAGNER Alexis, 1991, Flood Disaster Rehabilitation, Nepal: A Case Study, Transportation Research Record No. 1291

[2]OSTERWALDER Walter, 1992, River Control with Concrete Elements, International Sym-posium Interpraevent 1992

[3]VAW, 1992, Charnawati Report on the Hydraulic Model Tests concerning the Charnawati River stabilisation project at the Lamosangu-Jiri Road (Nepal), using flexible structures of concrete elements’, Report No 4003/3,Federal Laboratory of Hydraulics, Hydrology and Glaciology (VAW), ETH Zürich (unpublished)

[4]VAW, 1990, Assessment and Evaluation of the Charnawati flood event of May 31/June 1, 1990’, Report No 4021/1,Federal Laboratory of Hydraulics, Hydrology and Glaciology (VAW) , ETH Zürich (unpublished)

[5]SCHLEISS Anton, AEMMER Martin, PHILIPP Ernst, WEBER Heinz, 1998, Erosion protection at mountain rivers with buried concrete blocks, Wasser, Energie, Luft No. 90;3/4

[6]BEZZOLA Gian-Reto, 2002, Einfluss von Makroauigkeiten auf die Stabilität alpiner Gewässer (Influence of River Bed layer com-position on stability of high gradient rivers, Wasser, Energie, Luft, 2002.

Authors:

Kurt Waber

Walter Osterwalder

Kalyan Gyawali

Alexis Wagner

Martin Fox

c\o ITECO Engineering Ltd

PO Box, 8910 Affoltern am Albis

Switzerland

Email: iteco@iteco.ch