The effectiveness of a soil bioengineering solution for river ... - Fpv umb

bs_bs_banner

Area (2012) 44.4, 479–488

doi: 10.1111/j.1475-4762.2012.01132.x

The effectiveness of a soil bioengineering solution for river bank stabilisation during flood and drought conditions: two case studies from East Anglia Lenka Anstead*, Rosalind R Boar** and N Keith Tovey** *Department of Geography, Geology and Landscape Ecology, Faculty of Natural Science, Matej Bel University, Tajovskeho 40, Banska Bystrica, 974 01 Slovakia E-mail: [email protected] **School of Environmental Sciences, Faculty of Science, University of East Anglia, Norwich NR4 7TJ Revised manuscript received 24 August 2012 The increasing frequency of extreme flow events as a consequence of climate change could potentially impact the stability of lowland clay streams in East Anglia. Some of these rivers act as conveyors of additional water that is pumped downstream to satisfy the growing demand for water in this dry region. To accommodate the additional flows, the river channels have been deepened and straightened and the combined effect is that riverbanks are becoming more unstable. Willow spiling, a bioengineering stabilisation measure, could be an effective and sustainable mitigation option for managing riverbank erosion. There is a growing interest in understanding the performance of such interventions, particularly under changing climatic conditions. Two willow spiling projects were implemented in March 2009 and studied for their biological and geomorphological function during the first year after installation. This study shows that droughts and floods can impact the effectiveness of these measures, but through careful design, implementation and post-project monitoring this method can be effective in withstanding such stresses and protecting riverbanks. Key words: soil bioengineering, streambank stabilisation, vegetation revetment, environmental management, hydrological extremes, East Anglia

Hydrological and geomorphological challenges in East Anglia In East Anglia, engineering interventions mainly during the 1960s and 1970s (such as dredging, straightening of river channels, installation of weirs or sluices and removal of riparian vegetation) caused the banks to become higher and therefore more prone to instability. Apart from situations where banks are highly erodible, most instances of fast progressing erosion rates are a result of human activity (Thorne et al. 1996). In addition, climate change has an impact on river bank erosion, acting on both sides of hydrological extremes: floods and droughts. East Anglia is the driest region in the UK, with an effective annual rainfall of

only 147 mm and long dry summers where evaporation rates greater than rainfall are typical for the region. East Anglia has less water available per person than many hotter and drier countries (EA 2009). The UKCIP02 scenarios (Hulme et al. 2002) predict that by 2050, the annual winter rainfall in East Anglia will increase between 15 and 20 per cent (for low to high emissions scenarios) and summer rainfall will decrease by between -20 and -40 per cent. Heavy winter precipitation will become more frequent and so will summer droughts. Any further decrease in spring and summer flows due to climate change will intensify drought conditions (EERA SDR 2004) and increase pressures on the already stressed water supply in the region (EA 2009). Increases in the amount of transferred water during winter and less

Area Vol. 44 No. 4, pp. 479–488, 2012 ISSN 0004-0894 © 2012 The Authors. Area © 2012 Royal Geographical Society (with the Institute of British Geographers)

480

Anstead et al.

natural water in rivers during dry summer months would thus put more strain on the ecology and hydromorphology of aquatic ecosystems. Historically, both an increasing need for flood protection and a growing demand for water led to engineering interventions. These involved dredging and straightening of river channels and installing channel structures or the removal of bank vegetation. Although such works helped to achieve higher channel capacity and conveyance, they often initiated erosion and sedimentation problems. Silted, uniform and over-managed channels are usually of poor ecological quality. At present, all major rivers and many smaller tributaries in East Anglia are classified as heavily modified, modified or artificial (EA 2011). The UK has adopted the EU Water Framework Directive (WFD), which sets targets to prevent deterioration of the status of all surface water and groundwater bodies and to protect, enhance and restore all bodies of surface water and groundwater with an aim of achieving a good status for surface water and groundwater by 2015 (OJC 2000) (page 12 of the WFD)

with a possible extension of this target by up to 12 years (UKTAG 2008). Currently in East Anglia only 12.3 per cent of rivers have a good ecological status, 72.7 per cent fall under moderate and 12.3 per cent are of poor or bad ecological status. No rivers have been classified as having a high ecological status (EA 2011). Willow spiling could be one of the options to improve the ecological status of the engineered rivers in East Anglia whilst maintaining riverbank stability.

Managing river bank instability Hard engineering intervention has traditionally been used to ‘repair’ a failing riverbank, often with little consideration for the geomorphological processes that are causing the instability (Thorne 1978). Although ‘hard’ methods such as sheet piling, concrete slabs or gabion (stone) baskets can be designed to a high degree of precision, increasing failures of such solutions, noted mainly during the 1980s and 1990s, started to raise questions about their relevance in every situation (Li and Eddleman 2002; Hoag and Fripp 2005; Hey 2006). A range of alternatives exist that use living natural materials and collectively these are known as ‘soft engineering’ or ‘soil bioengineering’ (Schiechtl and Stern 1997). These approaches have the potential to contribute to the improvement of the ecological status of rivers and could also be used to build up the resilience of a river bank against climate change effects. The main benefits lie in the strength of root systems of the living revetments reducing erosion and reducing the near bank velocity and shear stress peaks (Thorne et al. 1998).

However, quantitative evaluations of soil bioengineering projects are scarce. Some authors have monitored biotechnical projects after one growing season (Shields et al. 1995; Akridge et al. 1999; Simon and Steinemann 2000), including observations of willow stake sprouting (Watson et al. 1997; Li et al. 2006; Petrone and Preti 2008 2010). However, despite the wide application of willow spiling in the UK, with the exception of some unpublished reports and personal observations (Murphy and Vivash 1998; Goodson 2002; Laing 2003), quantitative information regarding the biological and geomorphological performance of this method has not yet been published.

Willow spiling Willow spiling is a mixture of craft and engineering that involves weaving long willow stems around vertically driven willow poles (Plate 1). The use of willow as a material for riverbank protection dates as far back as 28 BC when willow bundles were used to stabilise the banks of the Yellow River in China (Hoag and Fripp 2005). A wealth of experience in using willows lies in civil engineering projects across Europe, notably in Austria (Rey 2009), Switzerland (Evette et al. 2009) and Germany (Simon and Steinemann 2000). Today, willow spiling is the most widely used willow-based system in the UK. A review of the method alongside the experience from nearly 140 projects has been summarised by Anstead and Boar (2010). Willows occur naturally along many rivers in the UK. making the use of willows for bank protection both practical and ecologically sustainable. They reproduce easily from cuttings and grow rapidly into long pliable stems suitable for weaving. Willows tolerate flooding, relatively infertile soil and polluted substrates (Schiechtl and Stern 1997). Their benefits to river ecosystem are manifold and alongside their environmental function, willow revetments provide economic and social benefits such as the

Plate 1

Completed willow spiling wall before backfilling, The River Stour, East Anglia


Soil bioengineering solution for river bank stabilisation involvement of the local community through volunteering projects (A. Walters personal communication). In the absence of other woody plants, coppiced willow spiling could last around 100 years (Schiechtl and Stern 1997).

The catchment of the River Stour The River Stour rises in eastern Cambridgeshire, flows through Suffolk and forms a boundary between Suffolk to the north and Essex to the south for most of its course before joining the North Sea (Mills 2003; Figure 1). The river is 108.5 km in length and drains an area of 1044 km2. The drainage catchment is underlain by Cretaceous chalk covered with boulder clay (70–100%), fluvial sands and gravels. Average annual rainfall in the whole catchment is 586 mm, ranging from 550 to 600 mm. The mean long-term summer rainfall (April to September) is 282 mm and winter (October to March) is 292 mm – during the study period from April 2009 to March 2010, the summer rainfall was 221 mm and winter 329 mm. Maximum discharges usually occur between November and February. During dry months, flows can be enhanced via the Ely Ouse to Essex Water Transfer Scheme that provides up to 3.8 m3/s additional water into the river (ESW 2009). Water transfers may account for most of the river flow in the upper reaches. The upper reaches of the River Stour drain arable land, while the land along the lower reaches is primarily used for grazing by cattle and sheep. The catchment is mainly agricultural (75.5%) with grassland (13.9%), woodland (6.1%) and settlement (1.7%). Agriculture contributes diffuse pollutants and the river floodplain has been designated as a nitrate vulnerable zone. The River Stour was one of the first channelised rivers in England and has been engineered variously for water mills, navigation and more recently, for the Ely Ouse to Essex Transfer Scheme.

Figure 1

481

Case study sites Two sites on the River Stour (gravel and clay site) were chosen to test willow spiling as an alternative to proposed hard-engineering approaches to manage river bank erosion. Both sites are located in an Area of Outstanding Natural Beauty (AONB) and river banks are grazed by livestock and used as public footpaths. The proposed willow spiling presented an aesthetically and ecologically sustainable option to address local erosion. The riverbank at the gravel site (G) consists of noncohesive sands and gravels that are easily entrained by flow, while the bank at the clay site (C) is composed of cohesive clays and silts where interparticle forces are making the banks more resistant to erosion. The river bank at both field sites was subjected to significant erosion; the bank at the clay site has previously eroded at a rate of up to 0.3 m per year and at the gravel site at a rate of up to 1.3 m per year. Erosion rates were calculated from recent aerial photographs and from field topographic surveys carried out between 2007 and 2009 (Anstead, unpublished data). From this historical and field analysis, it appeared that the instability was triggered by human intervention (installation of a weir upstream of the clay site and gabion deflectors upstream of the gravel site). Both sites are positioned on the outside of meander bends (concave banks). The water surface slope at the clay site, located 125 m downstream from a weir, was 0.0060 (calculated as drop in elevation in m/horizontal distance in m), comparable to some UK upland rivers (see Ferguson 1981), with the average water surface slope for this reach over several kilometres being considerably lower, 0.0009. The gravel site, located 250 m downstream of a major confluence, had a water surface slope 0.0023, while the average water surface slope for the reach was much lower, only 0.0005.

Study catchment with the location of the two project sites (blank circles): clay site (C) and gravel site (G) and the location of gauging stations on the river and main tributaries (shown by grey squares)


482

Anstead et al.

The gross stream power, which is a channel characteristic dependent on the water surface slope and the bankfull discharge (Simons et al. 1965), was 2.01 kW/m for the clay site (where bankfull discharge was 34.10 m3/s) and 0.88 kW/m at the gravel site (with bankfull discharge 38.77 m3/s). The specific stream power expressed as the ratio of gross stream power to the channel width was found for the clay site to be 0.08 kW/m2 and for the gravel site 0.04 kW/m2, values comparable with some upland streams in the UK (Ferguson 1981). The site sinuosity was 1.15 at the clay site and 1.71 at the gravel site. At both sites, the bank angles (35° at the clay site and 27° at the gravel site) and heights (1.8 m at both sites) were similar. Width/depth ratio at the clay site was 13.68 and at the gravel site 9.57.

Revetment design Different lengths of spiling were installed at the sites, but design features were similar in other respects (Table 1). Revetments were made from local and recycled materials that would blend with the natural environment. Two-tiered spiling was installed at each site: (1) a lower tier (LT) was placed at the mean summer water level as recommended by Schiechtl and Stern (1997) and (2) an upper tier (UT) was installed 1 metre above the LT to account for the maximum retaining height of a willow revetment (Polster 2002). The lengths of spiling were the minimal lengths to support the section of unstable bank at both sites.

Table 1

Willow stakes, about 2.0–2.5 m in length, were inserted into the river bank roughly 0.5 m apart and then tightly interwoven with long pliable willow canes, recommended to be at least 2.5-m long (Agate and Brooks 2001). Key steps for successful design were to have at least two-thirds of the stake’s length firmly embedded in soil (Schiechtl and Stern 1996), deep enough to hold the woven wall, but also allowing roots to reach the water table during dry periods (Crowder and Pullman 1995). The depth of insertion was similar to cantilever sheet piling in loose soil (United States Steel 1984). After weaving, the structures were backfilled with soil. The original riverbanks and the revetments shortly after installation are shown in Plate 2. Projects were installed in March, which is the optimum planting time for the river bank in this region; after the winter high flow season and before the vegetation growth season starts.

Post-project evaluation one year after installation In this study, the biological performance of spiling on the River Stour was measured by stake survival, shoot extension and increase in the number of shoots. The geomorphological performance was assessed in terms of structure undercutting and the erosion of backfill. Biological sampling was carried out six times at monthly intervals between May and October 2009 and the geomorphological surveys were performed in November 2009 and after

Design features for the two spiling projects: clay (C) and gravel site (G) Clay site (C)

Gravel site (G)

Length of revetment Number of tiers Number of stakes Length of stakes Salix species Average stake diameter Way of planting stakes Age of most withies Soaking of material Date of installation

6 m (upper tier), 7 m (lower tier) 2 22 2.0 m Salix alba L. 6.3 cm (⫾0.92 cm) Upside down 3–4 years One week before installation 10 March 2009

Person-hours*

33 person-hours (2.75 per linear m) One week after installation

45.5 (upper tier), 13.5 m (lower tier) 2 141 2.0–2.5 m Salix alba L. 6.5 cm (⫾2.3 cm) Upside down 3–4 years Freshly coppiced material 12 March, 14 March, 18 March (upper tier), 25 March, 31 March (lower tier) 108 person-hours (1.66 per linear m)

Timing of backfill Amount of soil used for backfill Additional materials

0.5 tonnes

One week after installation (for top tier); one day after installation (bottom tier) 3 tonnes

None

Recovered jute geo-textile, grass seed (lower tier)

* as total person-hours for the project and per linear m of spiling Area Vol. 44 No. 4, pp. 479–488, 2012 ISSN 0004-0894 © 2012 The Authors. Area © 2012 Royal Geographical Society (with the Institute of British Geographers)

Soil bioengineering solution for river bank stabilisation

Plate 2

Table 2

483

View of the riverbanks, looking upstream: clay site before (A) and after (B) treatment and gravel site before (C) and after (D) treatment Percentage of stakes with at least one living shoot, May to October 2009

Clay site (C) Gravel site (G)

May

June

July

August

September

October

96.5 97.8

100.0 100.0

100.0 100.0

100.0 98.6

100.0 82.3

90.91 17.02

the flood season in March 2010. The component tiers of the two installations were assessed separately and referred to as: C-UT (clay site upper tier); C-LT (clay site lower tier); G-UT (gravel site upper tier) and G-LT (gravel site lower tier).

Biological performance Considering that willow spiling survival and growth in the initial stages after installation is regarded as one of the key criteria for the long-term effectiveness of the live revetment (Jarvis and Richards 2008; Anstead and Boar 2010), much attention was given towards monitoring the biological growth and survival of the willow stakes. For measuring trends in the survival of willow spiling, the presence of living buds and shoots was recorded for all vertical stakes. The increase in shoot length (referred to as shoot extension) and number of shoots were sampled on stakes chosen at random and these data were collected from 12 stakes at the clay site and 31

stakes at the gravel site. The length of all visible shoots on the sampled stakes was measured to an accuracy ⫾ 0.1 cm for shoots up to 1 cm long and to within ⫾0.5 cm for shoots longer than 1 cm. The whole population of stakes in the upper tier at the gravel site was sampled to find the true mean and to establish the minimum sample size. The number of samples was determined as the point at which the cumulative mean of the sample began to approximate to the true mean of the population and values fell within an accuracy interval of the true mean of ⫾1.00 cm (value based on the maximum error of sampling). Most stakes sprouted within four to eight weeks after installation, and all stakes were still alive one month later. June, July, August and, in the case of the clay site also September and October 2009, were very successful months. The first decline in stake survival occurred at the gravel site in September. Here, the end of season field visit showed a large decline in stake survival and only about 17 per cent of the stakes were living (Table 2). Most of the dead stakes were located on the upper tier.


484

Anstead et al.

At the clay site, the shoot extension rate was greater in the lower than in the upper tier, but this reversed in August 2009. Shoot length increased in the lower tier by almost 35 per cent in October 2009. Overall however, both tiers performed similarly to each other (p > 0.05). At the gravel site, shoots were initially longer in the upper tier, probably due to the later installation of the lower tier (by four weeks) and access to light early in the season. The situation reversed in June. There was a considerable decline in mean shoot length in both tiers in July, which continued in the upper tier until the end of the season. In October, the mean shoot length in the lower tier was 9.5-times the mean shoot length in the upper tier, with shoots up to 98 cm long. Overall, there was a significant difference (p < 0.0001) in the mean shoot length between the two tiers at this site. The upper tier had the lowest mean rank, while the lower tier had the highest. When the same tiers at both sites were compared, the upper tier at the gravel site had shorter (p = 0.0007) shoots than the same tier at the clay site and the lower tier at the gravel site also had longer (p = 0.04) shoots than the corresponding tier at the clay site (Figure 2). Comparing the two sites, the clay site had longer mean shoot length (p = 0.03). With regards to the number of shoots, initial sprouting on most of the stakes sampled was vigorous with abundant new shoots. Later in the season, the number of living shoots decreased and shoots located mostly around the top of stakes died or were removed by grazing animals. At the clay site, the highest number of shoots occurred during May 2009, with a mean SD of 7 ⫾ 7.1 shoots per stake in the upper tier and 11.7 ⫾ 6.7 shoots in the lower tier. July had the greatest decrease on all stakes (down by 43.75 per cent). There was some regrowth and further decreases towards the end of the season, but recovery was limited. At the gravel site, the fluctuations were more marked, with the largest increase in living shoots occurring during May and June 2009. The mean number of shoots was then

12.1 ⫾ 7 per stake in the upper tier and 13.6 ⫾ 5.7 per stake in the lower tier. Some new shoots appeared at the end of summer, including the first upstream stake in the upper tier, which sprouted with a record number of 39 shoots. However, a large decline occurred at the end of the growing season, when the total number of living shoots fell by 92.4 per cent. Statistically, there was no difference in the number of shoots between the upper and lower tiers at the same sites (p > 0.05). However, when testing the same tiers at different sites, both upper and lower tiers at the clay site had fewer shoots than the same tiers at the gravel site (p = 0.03 and 0.02 respectively). To account for both the shoot extension and the number of shoots, summed shoot length was calculated. This showed that initially the gravel site performed better than the clay site. However, due to rapid shoot removal and higher mortality later in the season, the final summed shoot length at the gravel site was around 45 times less in the upper tier and two times lower in the lower tier than for the corresponding tiers at the clay site (Figure 3).

Geomorphological assessment To evaluate the geomorphological function of the spiling and to make a prognosis of its potential for the future, conditions on the river bed and on the backfill were recorded during the first winter after installation (November 2009 and March 2010). The basic point elevation data were initially collected in the field and gridded using the Radial Basis Function (RBF) with the Multi-Quadric option in Surfer computer software (Golden Software 2010). This gridding is a function that enables extrapolation of randomly spaced data to evenly spaced grid nodes that can be shown as surface plots. The RBF technique is effective for reconstructing smooth, manifold surfaces from point data (Carr et al. 2001) and the data can be subsequently ‘sliced’ and plotted in any desired cross-sectional profile (Figure 4).

Figure 2

Mean and median values of each cohort for mean shoot length and number of shoots per stake between May and October 2009 Note: Data are for the entire monitoring period. Vertical bars represent one standard error around the mean Area Vol. 44 No. 4, pp. 479–488, 2012 ISSN 0004-0894 © 2012 The Authors. Area © 2012 Royal Geographical Society (with the Institute of British Geographers)

Soil bioengineering solution for river bank stabilisation

Figure 3

485

Summed shoot lengths (cm) per number of sampled stakes in 2009

Table 3 Eight cross-sectional areas (CS1–8) at clay and gravel sites in m2 and the percentage of difference (D) between November 2009 and March 2010 Clay site river bed area (m2)

CS1 CS2 CS3 CS4 CS5 CS6 CS7 CS8

Gravel site river bed area (m2)

Nov 09

Mar 10

D (%)

Nov 09

Mar 10

D (%)

0.2005 0.3681 0.3737 0.3754 0.3618 0.3134 0.2979 0.1095

0.1671 0.3180 0.3084 0.3376 0.3363 0.3052 0.3226 0.4791

-16.66 -13.60 -17.49 -10.06 -7.05 -2.62 8.27 -3.36

0.7743 0.8503 0.7555 0.8344 0.6713 0.8255 1.1072 1.0547

0.4524 0.5359 0.4222 0.4838 0.5029 0.9434 1.3774 1.3530

-41.57 -36.98 -44.12 -42.02 -25.09 14.28 24.40 28.28

Note: Erosion is shown in negative values

Figure 4 Examples of cross-sectional plots at the gravel site. CS1–4 are the numbered codes for the cross sections in a downstream direction Note: Curves with black symbols are based on data collected in November 2009, while curves with blank symbols are based on data obtained in March 2010

More significant erosion and sedimentation processes occurred at the gravel site compared with the clay site. At both sites, erosion prevailed in the upper sections of the spiling and the rate decreased downstream. At the gravel site, sedimentation was noticeable on the river bed as well as on the backfill, but overall more material was deposited than was eroded within the surveyed section. The downstream end of the spiling was stable, while the upstream end has been subjected to presumably higher shear stresses. Examination of the river-bed profile immediately revealed substantial undercutting of the spiling (up to 29.14 cm at the gravel site). This probably represents a rate of bed material loss that the spiling won’t be able to compensate for with its roots, and without a further human intervention this part of the structure may fail. The cross-sectional areas expressed as percentages of change are shown in Table 3. Elevation of the backfill in relation to the spiling was recorded at each stake. At the clay site, the mean material loss from the upper tier ⫾SD was -3 ⫾ 4.6 cm,


486

Anstead et al.

Figure 5 Erosion of backfill that occurred around stakes in the upper and lower tiers at the clay and gravel sites, in a downstream direction

while at the lower tier the corresponding value was -6.2 ⫾ 2.9 cm. Erosion of up to -14.5 cm (15% of depth) and sedimentation of up to 1 cm occurred in the backfilled material. In the upper tier, the backfill behind 27 per cent of the stakes was eroded by at least -5 cm, while in the lower tier, this occurred behind 63 per cent of stakes. At the gravel site, the mean erosion behind stakes in the upper tier was -0.7 ⫾ 2.6 cm and in the lower tier, -5.9 ⫾ 9.6 cm. Erosion of up to -35 cm (35% of backfill depth) and sedimentation of up to 9.5 cm occurred at the gravel site. Figure 5 shows erosion trends with distance along the spiling.

Discussion of the key factors Both installations were tested by a range of stress conditions such as low flows during the establishment of cuttings, extreme floods during winter and extensive grazing. Loss of soil from around the spiling caused by high flows is one of the most common reasons for failure in the UK (Anstead and Boar 2010). The two spiling installations

were subjected to four high-flow events during winter 2009/2010. Rainfall in February 2010 was double the long-term monthly average (Met Office 2011). Between October 2009 and March 2010, the rainfall was 120 per cent of the normal. This increased the magnitude of bed scouring and was more significant at the gravel site. In contrast, the period between March 2009 and October 2009 was extremely dry and soil moisture deficits were well above average (Met Office 2011). Dryness during the first weeks after installation greatly disadvantages root formation and establishment. The trend in biological growth showed that strong initial shooting without rooting did not secure long-term survival and shoot mortality occurred when the energy stored within stake tissues was depleted. Without an effective root system, individual stakes could not survive independently from their parent tree. Stake mortality primarily affected the upper tier at the gravel site because gravel differs from cohesive soil in its ability to store and chemically bind water (Thorne 1978). The soil moisture content was tested in the field and showed 10–17 per cent deficit in the available water content at the gravel site. Shoots in the upper tier were removed by grazing livestock and did not contribute to soil shading (which would have reduced evaporation) and as a result of capillary rise and greater shade from sun and wind, the backfill behind the lower tiers had approximately double the moisture content at both sites. Both installations were fenced to prevent disturbance by cattle. However, a small group of goats at the gravel site got through the fence in July 2009 and grazed and trampled the spiling. Stakes responded to initial removal and damage to primary shoots by producing secondary shoots. However, constant grazing and irreversible removal of the bark from the stakes resulted in the increase in stake and shoot mortality during July and August 2009.

Management recommendations Alongside project monitoring, a number of recommendations for the design stage arise from the authors’ observations and these are summarised below: 1 Spiling should not be used on river banks where bed scour occurs. 2 Coppicing, if required, should be scheduled after a high flood period because emergent shoots greatly reduce flow erosive forces. 3 Stakes should be inserted in soil as deeply as possible to minimise drying and splitting. 4 Upper spiling in less cohesive soils should be placed further down the bank so the stakes maintain access to water during dry periods.


Soil bioengineering solution for river bank stabilisation 5 Any dead stakes/sections of spiling should be replaced by new willow material. 6 Backfilling should not be delayed. Willows must be kept in constant contact with soil and water. Spiling should also be placed to allow enough space for easy backfilling to avoid creating air cavities behind the structure. 7 The soil behind the lower tier should be protected before the winter season and if there is no natural growth, it can be seeded or covered with a geotextile. 8 Shade early in the season should be prevented because light is important for growth and establishment of new shoots. 9 Grazing and trampling by livestock should be prevented if possible.

Conclusions Willow spiling is an old technique, which, if implemented carefully, can be effective in stabilising river banks in East Anglia with relatively high local gradient. Willow spiling has the potential to provide effective, long-lasting protection against river bank instability triggered by human intervention and climate extremes. But it needs careful management in the early years, including the exclusion of all livestock. Guidance on some of the critical factors that may influence the effectiveness of willow spiling has been provided. The bank material is important in terms of water retention and gravel banks require deeper planting of the willow stakes. Long-term monitoring of width/depth ratio would help to detect whether the revetment is effective or if it triggers erosion on the opposite bank, for example in channels that are widening as a response to engineering interventions upstream. Taking into consideration these recommendations and general design principles for the method, willow spiling should, alongside other vegetation-based methods, be given a priority over hard engineering techniques. Acknowledgements Authors would like to thank Sudbury Common Lands Charity, Hector Bunting and associated workers and conservation volunteers for their help with the construction of the two projects. Furthermore, the authors are grateful to the two referees who provided valuable feedback on the manuscript.

References Agate E and Brooks A 2001 Waterways and wetlands: a practical conservation handbook British Trust for Conservation Volunteers, Reading Akridge A, Eigel J D and Athanasakes J G 1999 Stream restoration and soil bioengineering Public Works 130 48–51 Anstead L and Boar R R 2010 Willow spiling: review of streambank stabilisation projects in the UK Freshwater Reviews 3 33–47

487

Carr J C, Beatson R K, Cherrie J B, Mitchell T J, Fright W R, McCallum B C and Evans T R 2001 Reconstruction and representation of 3D objects with radial basis functions SIGGRAPH Computer Graphics 2001 Conference Proceedings ACM SIGGRAPH Springer, Berlin 67–76 Crowder W and Pullman P M C 1995 Collecting willow, poplar and redosier dogwood hardwood cuttings for riparian site plantings: technical note Plant Materials 29 US Department of Agriculture, Natural Resources Conservation Service, Spokane Washington (http://plant-materials.nrcs.usda.gov/pubs/ wapmctn290195.pdf) Accessed November 2011 EERA SDR 2004 Living with climate change in the east of England East of England Regional Assembly and Sustainable Development Roundtable (http://www.hef.org.uk/news/ climate-change.pdf) Accessed November 2011 EA 2009 Water for people and the environment: water resources strategy for England and Wales Environment Agency, Bristol (http://publications.environment-agency.gov.uk/PDF/ GEHO0309BPKX-E-E.pdf) Accessed November 2011 EA 2011 The state of our environment: water – east of England 2010 Environment Agency, Peterborough (http://www. environment-agency.gov.uk/static/documents/Business/SOE__Water.pdf) Accessed November 2011 ESW 2009 PR 09 water resources management plan: strategic environmental assessment report – non-technical summary Essex and Suffolk Water, Chelmsford (http://www.eswater. co.uk/PR09_WRMP_SEA_Non-technical_Summary_for_Web_ Site.pdf) Accessed November 2011 Evette A, Labonne S, Rey F, Liebault F, Jancke O and Girel J 2009 History of bioengineering techniques for erosion control in rivers in Western Europe Environmental Management 431 972–84 Ferguson R I 1981 Channel form and channel changes in Lewin J ed British rivers George Allen and Unwin, London 90– 125 Golden Software 2010 Surfer 9.11.947 Golden Software Inc, Colorado Goodson J M 2002 Environmental controls on the colonisation and establishment of vegetation on river banks under varying grazing pressure Unpublished PhD thesis Oxford Brookes University Hey R D 2006 Fluvial geomorphological methodology for natural stable channel design Journal of the American Water Resources Association 42 357–86 Hoag C J and Fripp J 2005 Streambank soil bioengineering considerations for semi-arid climates Riparian/Wetland Project Information Series No. 18 US Department of Agriculture, Washington DC (http://plant-materials.nrcs.usda.gov/pubs/ idpmcar5981.pdf) Accessed November 2011 Hulme M, Jenkins G J, Lu X, Turnpenny J R, Mitchell T D, Jones R G, Lowe J, Murphy J M, Hassel D, Boorman P, Mc Donald R and Hill S 2002 Climate change scenarios for the United Kingdom: the UKCIP02 scientific report Tyndall Centre for Climate Change Research, School of Environmental Sciences, University of East Anglia, Norwich (http://www.ukcip.org.uk/ wordpress/wp-content/PDFs/UKCIP02_briefing.pdf) Accessed July 2012 Jarvis R and Richards I G 2008 Engineering and the environment – perfect partners? The use of willows in bioengineering


488

Anstead et al.

Meeting report CIWEM Birmingham, UK (http://ciwem. helixsoftware.co.uk/media/51265/04_willow_in_bioeng_ paper.pdf) Accessed July 2012 Laing S 2003 Investigating the application and long-term performance of ‘soft’ riverbank protection techniques: 30 case studies from the Thames region Unpublished internal report Environment Agency Thames Region Li M-H and Eddleman K E 2002 Biotechnical engineering as an alternative to traditional engineering methods: a biotechnical streambank stabilization design approach Landscape and Urban Planning 60 225–42 Li X, Zhang L and Zhang Z 2006 Soil bioengineering and the ecological restoration of riverbanks at the Airport Town, Shanghai, China Ecological Engineering 26 304–14 Met Office 2011 2010 Weather summaries (http://www. metoffice.gov.uk/climate/uk/2011/october/averages.html) Accessed November 2011 Mills A D 2003 Oxford dictionary of British place names Oxford University Press, Oxford Murphy D and Vivash R 1998 Revetment techniques used on the River Skerne restoration project Environment Agency R&D Technical Report W83 Environment Agency, Bristol OJC 2000 Directive 2000/60/EC of the European Parliament and of the council of 23 October 2000 establishing a framework for Community action in the field of water policy Official Journal of the European Communities 60 L327 1–72 Petrone A and Preti F 2008 Suitability of soil bioengineering techniques in Central America: a case study in Nicaragua Hydrology and Earth System Sciences Discussions 5 379–403 Petrone A and Preti F 2010 Soil bioengineering for risk mitigation and environmental restoration in a humid tropical area Hydrology and Earth System Sciences 14 239–50 Polster D F 2002 Soil bioengineering techniques for riparian restoration in Proceedings of the 26th Annual British Columbia Mine Reclamation Symposium The Technical and Research Committee on Reclamation Dawson Creek, British Columbia, 230–39 (http://www.trcr.bc.ca/docs/2002-polster.pdf) Accessed November 2011 Rey F 2009 A strategy for fine sediment retention with bioengineering works in eroded marly catchments in a mountainous

Mediterranean climate – Southern Alps, France Land Degradation and Development 20 210–16 Schiechtl H-M and Stern R 1996 Ground bioengineering techniques for slope protection and erosion control Blackwell, Oxford Schiechtl H-M and Stern R 1997 Water bioengineering techniques: for watercourse, bank and shoreline protection Wiley-Blackwell, Oxford Shields F D Jr, Cooper C M and Knight S S 1995 Experiment in stream restoration Journal of Hydraulic Engineering 121 494– 502 Simon K and Steinemann A 2000 Soil bioengineering challenges for planning and engineering Journal of Urban Planning and Development 126 89–102 Simons D B, Richardson E V and Nordin C F 1965 Sedimentary structures generated by flow in alluvial channels SEPM Special Publication 12 34–52 Thorne C R 1978 ‘Processes of bank erosion in river channels’ 2 vols Unpublished PhD thesis University of East Anglia Thorne C R, Allen R G and Simon A 1996 Geomorphological river channel reconnaissance for river analysis, engineering and management Transactions of the Institute of British Geographers 21 469–83 Thorne C R, Amarasinghe I, Gardiner J, Perala-Gardiner C and Sellin R 1998 Riverbank protection using willows Scoping Study Environment Agency R&D Technical Report W154 Environment Agency, Bristol UKTAG 2008 Environmental standards and conditions (phase 2) UK Technical Advisory Board on the Water Framework Directive (http://www.wfduk.org/tag_guidance/Article%20_11/ POMEnvStds/sw_07) Accessed November 2011 United States Steel (USS) 1984 Steel sheet piling design manual US Department of Transportation with permission from Federal Highways Association (http://www.mcipin.com/publications/ sheetpiles/USSteel_1984-SteelSheetPiles.pdf) Accessed November 2011 Watson C C, Abt S R and Derrick D 1997 Willow posts bank stabilisation Journal of the American Water Resources Association 33 293–300
