11. Case studies

From PRAGMO
Jump to navigation Jump to search

11.1 Mayes Brook

The project

Figure 11.1: Mayes Brook location map

Mayesbrook Park, located in Barking, North-East London (Figure 1), is an area of 45 hectares of parkland which prior to restoration was degraded and under appreciated by the community. The Mayes Brook runs from north to south alongside the west boundary of Mayesbrook Park where in the past it has been engineered to protect against flooding; it is culverted along much of its length and no water quality monitoring is currently undertaken. It was publicaly inaccessible and hidden by high metal fencing. There are historical water quality issues in the Mayes Brook, which have resulted from a series of mis-connections originating from properties and drains that connects to the Mayes Brook. Thames Water is currently dealing this with in two stages.

Restoration will realign the river through Mayesbrook Park, creating more natural bank profiles and introducing river meanders, backwaters and ponds. Through this Mayes Brook will become a feature of the park, public interest and use will increase and it could contribute to local regeneration.

Key themes outlined within the scheme were:

  • Sustainable urban regeneration
  • Recreational amenity (access to nature)
  • Sustainable flood risk management
  • Biodiversity/Conservation
  • Climate change adaptation (demonstration site for this driver in particular)
  • BAP targets for wetland related species and habitats
  • Environment Agency duties to promote the conservation and enhancement of inland waters (Environment Act 1995)
  • Implementation of WFD on the ground in an urban area


Pre-restoration monitoring design

Figure 11.2: Kick sample and Surber sample methods

A baseline survey assessing ecological and hydromorphological aspects was carried out prior to restoration to help develop an evidence base for the Mayes Brook whereby the success of the project may be assessed in future. Environment Agency (2008) outlines this in detail. This baseline survey included:

  • River Habitat Survey (RHS) focusing on natural features recorded; channel and riparian vegetation types; habitat quality assessment (HQA) and habitat modification scores (HMS).
  • River Corridor Survey (RCS) investigating plant communities, species and geomorphological features of the channel, banks and corridor to within 50 meters of bankfull to create a 500 meter visual representation of the section. A biotope map overlay aimed to highlight the different in-stream functional habitat and link in with natural processes.
  • Invertebrate survey 3 minutes kick samples taken at upstream, midpoint and downstream sites following standard Environment Agency methodology. This was analyzed using BMWP (Biological Monitoring Working Party), ASPT (average score per taxon) and Environmental Quality Indices (EQI). Using a 100m representative section, 5 Surber-samples were collected from each of the functional habitats present, determined using the habitat map and personal observation. Unit-area samples were collected to enable densities and individuals per m² to be calculated. Physical parameters including substrate, flow and plants were also recorded for each sub-sample. Photographs of these methods are shown in Figure 2.
  • Fisheries Survey Undertaken using point abundance sampling techniques. Battery powered backpack electric fishing equipment was used and habitat variables were recorded at each point. A small anode ring (20cm in diameter) was used to reduce field size and increase survey efficiency for small species. On a representative section of the channel, 43 point samples were taken at approximately 5 metres intervals. Starting at the downstream end, each point was exposed for 5 seconds and stunned fish were removed by dip net, identified, measured (fork length) and returned.


Pre-restoration outcomes

Pre-monitoring results led to the following basic conclusions:

  1. RHS and RCS highlighted that the channel was over deep, had been greatly modified and lacked any natural features or habitat structures. The banks were very high, steep and reinforced in places with no connection to the floodplain.
  2. HQA scores revealed in terms of habitat quality, that it was ranked in the bottom 40% of the 150 most similar rivers while HMS scored the Mayes Brook in the class of severely modified rivers. The calculated EQI scores for the number of taxa and ASPT put the Mayes Brook into the general quality grade of D (fair quality); considerably different from the grade expected for an unpolluted river of this size, type and location.
  3. The biotope map highlighted only three functional habitats: emergent plant, riffle and run habitats; however emergents were found throughout the channel, and were not just confined to the margins where they would be expected to occur. Surber-samples showed that riffle and run habitats had similar invertebrate compositions (Figure 3).
  4. Pollution tolerant taxa ranging from BMWP score 1 to 3 dominate (snails, leech, hoglice, midges and worms) typical of a modified channel. A low number of pollution sensitive caseless caddis (BMWP 8) indicates that the habitat quality and possibly water quality issues limit invertebrate diversity.
  5. BWMP and ASPT scores indicated that the upstream site had a more varied in- stream habitat suffering least from any organic pollution entering the watercourse.
  6. Roach were dominant in terms of fish abundance and their biomass accounted for 63% of the total catch. Other species included 3-Spined Stickleback (22%), Chub, Dace, European Eel and Perch, so although there is a diversity of species present, the fish community is limited.
  7. Age samples for Roach revealed that the majority were 5 years old and these had above average growth rates compared to the average for Roach in southern rivers; this was linked to low levels of competition and an abundance of pollution tolerant invertebrates providing good food resources.
  8. No juvenile Roach were present suggesting that suitable spawning habitat is limited.
Figure 11.3: Taxa composition and number of individuals for each functional habitat


Current stance and future intentions

The July 2010 launch of the Mayesbrook Park Landscape Master plan: The UK‟s first Climate Change Adaptation Public Park signaled the progress that is being made on the ground – and now presents an important time to identify whether short-term improvements have been made, simply as a result of restoration activities. As the monitoring guidance suggests however, a long-term perspective is just as key to assessing whether the project has been a success for the ecology and the people of the area. A formal monitoring strategy is in the process of being delivered to monitor the restoration project from a physico-chemical, morphological, ecological and social perspective throughout the duration of the project.

Input into the strategy was the first example of using the PRAGMO guidance in regards setting detailed project objectives, and S.M.A.R.T. project monitoring objectives. Four working groups were organized by theme with specialists in each sector (aquatic environment; terrestrial environment; people; climate change) contributing to the overall document. Once all of the desired objectives had been devised and prioritized by the working groups, a decision was taken by other members of the Steering Group to prioritize all four inputs in the context of the wider project, where possible looking to combine or amalgamate objectives with a similar focus or aspiring goals. This has been very successful so far in giving the Mayes Brook Park project a clear vision in regards to;

  • The objectives it would like to achieve;
  • The monitoring strategy to use in this project;
  • The costs and benefits of the monitoring strategy;
  • Who will co-ordinate the monitoring strategy on behalf of the project partners;
  • When, and in what format, data will be available to convey messages to the partners, funders and the public about the success of the project.


Project Partners

The Environment Agency, Natural England, Greater London Authority, London Wildlife Trust, Design for London and the Thames River Restoration Trust (TRRT). The project lead is LB Barking & Dagenham.


References

Environment Agency (2008) Mayes Brook Restoration Scheme: baseline ecological survey. 27pp


Further Information

For more information on the Mayes Brook Project see the RiverWiki case study.

11.2 River Cole

The project

The River Cole at Colehill restoration scheme was one of three chosen for an EU LIFE funded project, “River restoration: benefits for integrated catchment management” as a demonstration site of innovative techniques and best practice. Prior to restoration, the river course was almost entirely artificial having been straightened and deepened over a period of 400 years. In 1997, restoration works involve re-profiling the bed and banks of two sections of the river creating approximately 1.3km of channel in addition to the restoration of a further 1.2km. Further works in 2008 involved the input of gravel and anchoring of large woody debris across the river.


Monitoring design

Monitored aspects of the River Cole restoration at Coleshill. Note that more detailed studies were conducted for water quality, invertebrates, plants and cost-benefit analysis

A monitoring rationale was designed to assess the benefits of the project and evaluate whether the project delivered a low-cost solution. Extensive monitoring addressed physical, chemical, biological and social aspects. Possible benefits identified at the start of the project included:

  • flood storage and flood alleviation
  • nutrient reduction or storage
  • river maintenance costs
  • conservation
  • recreation and amenity.

Surveys to assess geomorphological change were carried out throughout the EU LIFE+ project‟s duration. In terms of pre-project data, a morphological survey, a topographic survey and a fluvial audit were undertaken and post-construction; a geomorphic survey

was completed in the second and third year; while a follow-up topographic survey was also carried out in the final year of the project. The final report (RRP, 1999) illustrates the effects. An MSc study (Molloy, 2009) assessed hydromorphological change twelve years on. This also reviewed the 2008 works, which aimed to increase the coarseness of sediment to encourage the formation of permanent depositional features such as bars, which increase flow diversity.


Outcomes

Geomorphology

Figure 11.4: Pre, post and following restoration

Two main aspects of morphology were considered:

  • Large-scale channel morphology (i.e. channel planform and cross-section)
  • The frequency of small-scale morphological features (e.g. riffles and pools, point- bars, eroding cliffs).

Fluvial auditing was undertaken to describe the development of channel features following restoration. The River Cole demonstration site is split into two sections, upstream and downstream of Coleshill Bridge. These showed different morphological characteristics, and as such are described separately below.

A control site (upstream of the restoration demonstration site area) is channelized and ponded, and it had no riffles, runs, point bars or mid-channel bars. Over the course of the study, the control section showed little change over the course of the project, while in comparison the newly restored channel upstream of Coleshill Bridge changed dramatically. There was an observed increase in the number of riffles, pools, point-bars, actively eroding banks and overbank deposits.

In the restored section downstream of Coleshill Bridge, erosional channel features all increased and features which increased in abundance were riffles, pools, point-bars, mid-channel bars and overbank deposits. Actively eroding banks were numerous immediately after restoration but these had returned to pre-restoration levels by March 1998. There was also evidence on the River Cole of extensive sedimentation downstream of the restoration site which led to the local raising of bed levels, and a small increase in the number of berms.

Overall, the monitoring indicates that the 1997 restoration works increased channel morphology diversity improving river habitat conditions. This has increased channel length by 10%, reduced the channel width on average about 60% and bed levels have been raised by up to a meter. Molloy (2009) found that the input of coarse gravels also had a positive impact on redistributing sediment throughout the stretch and the talweg had become more sinuous. The large woody debris has also increased flow diversity, and altered the morphology causing the formation of new habitats such as a shallow glide over a deposited gravel bar.


Hydrology

Catchment modelling techniques revealed that should restoration be undertaken at the catchment scale, it would have a significant effect on peak discharge. The most effective theoretical alteration to channel structure was to reduce cross-sectional area to 20% of the existing value. This was predicted to reduce the 1 in a 100-year flood peak by about 10% at the bottom of the catchment, and the more frequent return period floods were reduced by proportionately greater amounts, the 1 in 2-year flood peak for example being reduced by 35% (RRP, 1999).


Biological

  • Macrophytes

There was no evidence that released sediment or other potentially damaging impacts of the restoration work had any adverse impact on the downstream plant communities. The number of wetland species (aquatic and emergent plants) quickly reached pre- restoration levels in the newly created channels, stimulated following the rapid colonization of new muddy banks by marginal wetland ruderals such as pink water- speedwell and celery-leaved buttercup. Interestingly however in the existing channel sections, there was a difference between emergent and aquatic response rates. Emergent plant species demonstrated a rapid recovery while the number of aquatic plant species downstream of Coleshill Bridge appeared to have been little affected by restoration (RRP, 1999) Molloy (2009) similarly reaffirms a lack of marginal bank-side vegetation years later.

  • Invertebrates

In the restored channels, the process of re-meandering eliminated most of the original river channel (which was backfilled with spoil from the newly created sinuous channels). Colonisation therefore began following the completion of the works with no pre-existing invertebrate assemblages.

Aquatic macro-invertebrates recolonisation was rapid and one year following restoration, species richness was only slightly below pre-restoration values. However, the average species rarity of macro-invertebrates recolonising the restored channel was significantly lower than the pre-existing channel data. Prior to restoration, the channel had thirteen local or Nationally Scarce species recorded. Statistically overall, there was a significant interaction between time (before vs. after) and location (control vs. restored), while the two upstream control sites showed similar species richness values to each other throughout the project.

  • Fish

Following restoration, fish biomass and density quickly returned to pre-restoration levels, with the highest values found in areas of gravelly eroding substrate. Surprisingly however, these values were recorded in the downstream impact reach below the restoration scheme itself. It is suggested that the improvement is however more than likely a reflection of the improved habitat in the restored section, as fish may be expected to rest in the downstream areas before moving into the faster flowing shallower water of the restored area to spawn.

Fish species richness generally remained unchanged both in the restoration site and the control and impact reaches. Twelve years on, it was determined that it may take much longer to see fish populations change (Molloy, 2009) and that factors other than geomorphic change may be limiting the establishment of fish at varying scales of the life cycle.

  • Social appraisal
    • 42% of Coleshill residents thought that the restoration scheme was of „good‟ or „quite good‟ value, with many at the time wising to determine the cost-benefit of the project at a much later date following re-establishment of the river environment, and in-particular ecology.
    • Overall the view of the scheme was consistently favourable. 53% broadly approved the restoration work; perhaps a reflection that the River Cole in part did not appear too degraded at the beginning of the project.


Lessons Learnt

While it was hoped that the restoration features would act as in-stream nutrient concentrations, the absence of any clear reduction (RRP, 1999) should not be unexpected. In the medium to longer term, effective buffering and nutrient removal processes may lead to an improvement in water quality, but it is likely that should releases of nutrients still occur elsewhere in the catchment, the water quality may remain at a similar level to pre-project.


References

  • Molloy, H. (2009). Hydromorphological changes to the River Cole over a twelve year period following restoration. Submitted in partial fulfilment of the requirements for the degree of MSc Water Management (Environmental option). 65pp.
  • RRP (1999). The effects of river restoration on the R. Cole and R. Skerne demonstration sites. Final report. 60pp plus appendices. River Restoration Project, Huntingdon.


11.3 River Quaggy

The project

For years the River Quaggy at Sutcliffe Park was lost underground in a culvert. Local residents only became aware that a river was there when their homes flooded more frequently as development increased. Rather than further deepening and widening the hidden channel, a decision was made to combine flood risk management with a strategy for river restoration that would benefit the local community.

The Sutcliffe Park restoration scheme was part of a series of flood alleviation schemes along the River Quaggy, and provided a floodwater storage area upstream of Lewisham town centre, which has suffered from severe flooding in the past. Since restoration, Sutcliffe Park has won two awards, the Living Wetlands Award 2007 and the Natural Environment Category of the 2007 Waterways Renaissance Awards.

A new 'low-flow' meandering channel was cut through the park, following its original alignment. The previous culvert was retained, enabling it to take excess water in times of extreme flood events. Flow is now regulated between the two watercourses by a sluice. To provide further flood water storage, the park itself was lowered and re-shaped to create a floodplain capable of storing a maximum of 85,000 cubic metres of flood water. A network of boardwalks, pathways and viewing points were designed to encourage access to the river and ponds, all of which were an integral part of the scheme for community and wildlife enhancements.


Monitoring design

The monitoring objectives were;

  • To determine the post restoration adjustments in the reach‟s geomorphology;
  • To determine the physical habitat diversity of the restored reach;
  • To assess the riverbed and floodplain sediment quality and;
  • To assess the water quality of the restored reach
  • To assess the ecology of the restores reach


Methods Used

Geomorphological appraisal

  • Surveys of river level and bankfull dimensions (i.e. width and depth);
  • Calculation of Manning‟s „n‟, hydraulic radius, wetted perimeter;
  • Measurement of flow velocity and discharge;
  • Completion of a River Habitat Survey.

Sediment quality appraisal

  • Particle size analysis of riverbed and floodplain sediment;
  • Measurement of organic matter content of riverbed and floodplain sediments;
  • Measurement of trace heavy metals within riverbed and floodplain sediments.

Water quality appraisal;

  • Measurement of trace heavy metals within river water;
  • Measurement of dissolved oxygen and pH;
  • Measurement of Nitrate, Phosphate, Nitrite and Chloride.

Ecological appraisal;

  • Surveys of instream macrophytes and riparian plant species;
  • Surveys of macroinvertebrate species, calculation of BMWP, ASPT and number of taxa.


Outcomes

Comparison of the channel design specifications from 2004 and the measured channel geometry from 2006 found that the reach had a more diverse form after restoration and that the mean bankfull width and depth had increased. The presence of instream aquatic plants (macrophytes) had a major impact on water velocities, and was responsible for the creation of extensive slackwater areas. The sediment assessments showed that traces heavy metal in both the riverbed and floodplain sediments exceeded a number of sediment quality guidelines, although they were within the „normal‟ range for UK urban rivers based on limited published data. This highlighted the need for a comprehensive set of sediment quality guidelines for assessment of UK urban rivers. Ecological appraisal based the classification of Holmes et al. (1998), found that instream and marginal plant species composition was characteristic of both lowland rivers with minimal gradients and rivers with impoverished ditch floras in lowland England Mean Trophic Rank results suggested that in future the monitoring of nitrate and phosphate

levels would be required. Macroinvertebrate surveys indicated that the watercourse was populated by mainly pollution-tolerant taxa.


Lesson Learnt

Results from the appraisal of the Sutcliffe Park reach of the River Quaggy highlighted a number of implications and recommendations for monitoring river restoration projects in of urban rivers;

  • If on-site geomorphological input is not available at the construction phase, then geomorphological monitoring should take place as soon as the channel is constructed to obtain accurate as-built data which can be confidently used in future geomorphological assessment of the restoration scheme.
  • The results from appraisal post restoration need to be validated with results from annual monitoring. For example, assessment of macroinvertebrates in the Sutcliffe Park reach showed limited abundance and diversity of high BMWP scoring species, but research has suggested that a restricted taxa range after restoration is not uncommon, and changes in macroinvertebrate assemblages may be subtle within the first few years, particularly in urban areas.
  • The requirement for a standard set of guidelines to assess the quality of urban river sediments. The comparative guidelines used to assess riverbed sediments were Canadian (OPSQG) guidelines and were not ideal for assessing the quality of UK river sediments. Furthermore, the guidelines used for floodplain sediment quality assessment were for use in assessing trace heavy metals in soils rather than fluvial sediments.


11.4 Seven Hatches

The project

The STREAM project was a £1 million four-year conservation project centred on the River Avon and the Avon Valley in Wiltshire and Hampshire. The River Avon and its main tributaries are designated as a Special Area of Conservation (SAC), and the Avon Valley is designated as a Special Protection Area (SPA) for birds. The STREAM project has undertaken strategic river restoration activities and linked management of the river and valley to benefit the river habitat including water crowfoot and populations of Atlantic salmon, brook and sea lamprey, bullhead, Desmoulin's whorl snail, gadwall and Bewick's swan.

A Conservation Strategy for the River Avon Special Area on Conservation (2003) identified the main issues affecting the ecological health of the River Avon SAC, and agreed on a range of actions required to address them. It also highlighted the complex relationship between the river and the Avon valley. In December 2002, work began on securing substantial new funding to do the following:

  • Restore, to favourable condition, the River Avon Special Area of Conservation/Special Site of Scientific Interest (SSSI) and the Avon Valley Special Protection Area/SSSI.
  • Tackle wider biodiversity issues outside the European protected sites including additional priority species and associated habitats, and
  • Improve public access, awareness and support for the natural heritage importance of the river and valley.

The project identified six sites where conservation-led restoration of the watercourse habitat may be used to demonstrate techniques and disseminate knowledge and experience of this work


The River Wylye at Seven Hatches

Just upstream of Wilton, „Seven Hatches‟ was one of the six sites within STREAM. It had historically been over widened and over deepened, and sluices prevented fish migration and caused a backwater effect on flows upstream of the structure. The project objectives aimed to;

  • Modify the operation of Seven Hatches sluices, reducing height by an average of 0.15 metres, therefore increasing ecological connectivity between reaches and improving upstream habitat quality;
  • Restore the historic bed level and increase the heterogeneity of bed morphology in previously dredged reaches, by the reclamation and re- introduction of excavated gravel/stone bed material;
  • Narrow over wide channels where necessary in order to re-establish a sinuous channel of appropriate cross-sectional area with respect to present day hydrograph data;
  • Increase the amount of large woody debris in the channel in order to increase both the availability of this habitat type and morphological diversity of the channel;
  • Break out and remove the tractor bridge footings and replace with a single span bridge.
  • Remove the impounding effect of the structure;
  • Enhance the availability and quality of habitat for SAC species (and habitats);
    • Bullhead (increase the number of hard bed pools, insert large flints in new riffles/fast glides and increase shading/ large woody debris for juveniles);
    • Salmon (improve migration routes, source viable spawning sites, and more appropriate habitat for fry and parr);
    • Brook lamprey (increase the availability of well sorted, fine sediment in shaded, marginal areas with large woody debris for ammocoetes and gravel/sand dominated shallows <40cm deep for spawning adults);
    • Desmoulin‟s whorl snail (marginal zone enhancement of all channels); Ranunculus (increase heterogeneity in velocity and bed morphology).


Monitoring design

Pre, post and following restoration

Detailed monitoring was carried out at Seven Hatches, with a control site with comparable physical characteristics. Field mapping was converted into a suitable digital GIS format to allow calculation of the area of habitats within the two sites to monitor change following repeat surveys. The GIS recorded physical and ecological features, sample and cross-section locations and any other spatial data collected in the field.

Pre-restoration surveys intended to establish a record of biological and physical conditions at the site prior to restoration. The post-restoration surveys recorded modifications to the channel after restoration. It should be recognised that there is a limitation to the comparisons that could be made over this still relatively short duration. The relationship between physical and biological conditions was assessed, taking into account other factors and processes that might have influenced these.

The following datasets were collected:

  • Included different aspects including visual and social elements; physical characteristics; vegetation; fish and aquatic invertebrates and; mammals, terrestrial invertebrates and birds.

Note that this was probably more comprehensive than is necessary for a typical individual UK restoration project. Please note that further data has been collected post 2008 by Royal Haskoning.


Outcomes

Hydromorphology

The introduction of gravels and the creation of riffles were largely successful as heterogeneity in flow types was achieved. Additionally, large woody debris pinned into the substrate increased the flow variability locally and there was evidence of scour on the downstream side of the structures. In the long term it is predicted that the turbulence at moderate to high flows generated by the woody debris will help to ensure that the riffles remain free of excessive siltation.

Appraisal of overall effectiveness is limited by the apparent lack of direct response to restoration measures; as while there has been an observed increase in variation of channel width, depth and flow velocities, there has also been natural variation in an increase of water level and processes within the catchment. Significant geomorphic change is likely to occur over much a longer timescale as the river naturally readjusts (Royal Haskoning, 2010).


Biological

  • Macrophytes

Channel narrowing techniques (berms) were successful in terms of providing marginal vegetation features. The system of brashings and log deflectors upstream of the hatches trapped silt and sediment. Deflectors have improved heterogeneity of the habitat, providing shallow well vegetated margins close to the existing deeper water. Larger structures would have had a more significant narrowing effect; however, this would likely have had an adverse impact on flood flow conveyance.

The narrowing above the hatches could have been bolder than was actually carried out. The log deflectors could have protruded much further into the channel and the brushwood infill and log staked wet ledge could have then been wider. However, what was installed is developing well. The result of the planting scheme was not as varied as was originally planned because many of the plants did not survive as a result of water levels being higher than expected due to wet winters and wet summers. It could be a number of years before the ledge vegetation reaches its full potential.

Post monitoring statistics revealed that macrophyte cover increased, the proportion of macrophyte species preferring swift flows increased and the proportion of macrophyte species preferring slow flows decreased. However, in comparison to the control reach, these were insignificant to demonstrate an attributable direct response to restoration measures. Macrophyte composition remained the same.

  • Invertebrates

Assemblages at the restoration site does not appear to have changed significantly more than that observed in the control site however with high taxonomic richness observed at both sites prior to restoration; there was unlikely to have been an significant increase.

  • Fish

Initially measures appear to have improved salmon and trout populations in the first year following restoration – species that prefer swift flows – and this was not replicated at the control site. Grayling populations may also have been positively impacted however further catchment scale analysis is required. However, salmon numbers decreased the following year, and while restoration has been successful in altering the age composition of bullheads with an increase in juveniles; overall bullhead and lamprey populations appear to have declined following restoration.


Lessons Learnt

  • Prior to implementing a restoration scheme, SMART objectives should be set following the approach outlined in RRC (2009). The subsequent monitoring protocol should then aim to assess if these SMART objectives have been met ensuring that it is linked directly to the objectives of the restoration scheme and the Water Framework Directive.
  • It is recommended that a 10 years monitoring programme is undertaken to include sufficient replicates to enable detailed statistical analysis. For example, the majority of the river remains over-deep and over-wide and it is recognised that it may take some time for the channel to gradually adjust through year-on-year sediment deposition and vegetation growth.
  • Monitoring should be undertaken throughout the entire reach rather than specific sections. This will enable holistic conclusions to be drawn on the effectiveness of the scheme.
  • Cross sections, macrophytes, invertebrates and fish were monitored in detail however it is apparent that measurements over a longer timeframe are required to enable statistically robust analysis to be undertaken. Macrophyte and invertebrate sampling should continue for at least 5 years to determine what direct effect the restoration work is having. Though macrophytes and invertebrates are not the designated interest, they often provide a more reliable indication of river health than more mobile fish populations.
  • Monitoring of velocities and substrate was not effective in producing data of sufficient quality and resolution. Velocity measurements should be undertaken in a variety of flow conditions and repeated when water levels are similar. Substrate measurements should be taken using a sediment sieve to collect grain sizes and enable a detailed analysis of sediment distribution.
  • Monitoring approaches are going to require more ‘vision’ in terms of immediate works versus long term results.
  • Change was limited by sub-optimal operation of hatches with restoration potential constrained by their existence as structures. Planned changes to the hatch operation were not carried out because of the concerns about reduced flows and the potential effect on salmon in Butchers Stream and flooding downstream in Wilton. The project demonstrated the impact that in-stream structures can have, and a hatch operating protocol developed through this four year project.


References

Royal Haskoning (2010). Seven Hatches Case Study Draft Report: Appraisal of River Restoration Effectiveness.

RRC (2009). Post works assessment of the STREAM restoration project sites at Seven Hatches (R. Wylye).


11.5 Kissimmee River Restoration Project, Florida

Case study written before project delivery

The project

Tom Palmer, The Ledger

The Kissimmee River is located in south Florida, arising from the Kissimmee Lake headwater streams just south of Orlando before flowing in a southerly direction into Lake Okeechobee, which is the second largest lake in the USA. The river once meandered for 103 miles through Central Florida and inundation of its floodplain as a result of for long periods by heavy seasonal rains, meant that wetland plants, wading birds and fish thrived. This was up to two miles in width. However, prolonged flooding was seen to cause severe impacts to humans, and the U.S Army Corps of Engineers cut and dredged a large 30-foot deep straightway canal cut between 1962 and 1971. While it achieved flood reduction benefits, it detrimentally impacted upon the river-floodplain ecosystem.

When restoration is complete in 2015, more than 40 square miles of river-floodplain ecosystem will be restored, including almost 20,000 acres of wetlands and 44 miles of historic river channel. Phase 1 in the lower Kissimmee basin began in 1999 and was completed in 2001, while Phase 2 was completed in 2009 respectively, together restoring continuous flow to 19 miles of the Kissimmee River. The third phase involves backfilling to the canal cut and restoring flow to a further eight miles of river. About 98% of the land required to complete the River Restoration has now been acquired – a total of 102,061 acres; and the only sections that will remain untouched are those that are still required to quickly drain floodwaters. The total project cost is anticipated to be in the region of $620 million.


Monitoring design

Extensive monitoring is an integral component of the scheme, and since the project aims to restore the entire Kissimmee ecosystem, studies have and continue to be undertaken not only the river itself but also on the wetlands, and the hydrology, hydraulics, water chemistry, algae, plants and macro invertebrates of the environment, looking in terms of the diversity, productivity and functional processes. The driving force behind this is a set of „61 expectations‟ which considers not only the desired end point, but also the natural processes upon which they depend. This is based on a variety of data whereby inferences can be made including historical records, professional expert judgement and empirical/computational models.

A key element of the monitoring design is the Kissimmee River Restoration Evaluation Program (KRREP), a comprehensive monitoring and assessment program designed to evaluate ecosystem-scale responses to the restoration program. This involves the collection of baseline datasets, after construction and following re-establishment and if an expectation is not met; adaptive management strategies may have to be adopted. KRREP will:

  • Assess achievement of the project goal of ecological integrity.
  • Identify linkages between restoration project and observed changes.
  • Support adaptive management as construction proceeds and after project completion.


Outcomes

Phase 1:

This involved the removal of a water control structure, the creation of a new river channel and the infilling of an eight mile flood control canal. The measured improvements have been compared with the condition before restoration began and the results are extremely encouraging (SMWFD, 2008):

  • Continuous flow of water since 2001 has improved biodiversity value of the river its floodplains and surrounding wetlands, and the biological community composition.
  • Organic deposits on the riverbed decreased by 71%, which has helped re- establish sand bars, providing new habitat for invertebrates and shorebirds.
  • Emergent plants native to the historic river, to replace undesirable plants, are developing.
  • Dissolved oxygen concentrations, critical for the survival of fish and other aquatic organisms, have increased to levels similar to those in relatively pristine rivers in South Florida.
  • Aquatic invertebrates are more characteristic of free-flowing water (e.g. caddis and mayflies)

Phase 2:

Works were undertaken upstream of Phase 1 towards Lake Kissimmee and below the Avon Park Bombing Range in a similar vein to those completed in 2001. It is still only recently that these works have been completed but the initial signs remain promising.

  • The restored river‟s water quality is better, which is reflected in the fish populations. Native Largemouth bass and sunfish populations have increased significantly.
  • The length of river channel through the restored section has increased from 13 miles to 25 miles, and Mike Cheek, an environmental scientist with the South Florida Water Management District, has compared the old canalised cut to “an interstate highway”, whilst the new channel, is a “two-lane country road; much more scenic and biologically diverse” (Palmer, 2010).
  • Bird species which were historically lost have returned, and the mixture is now relatively rich, consisting of more than 300 species; largely as a result of the wetlands rebounding. At least six species of shorebirds, which are smaller and harder to spot in aerial surveys used to monitor trends in bird populations, have been documented.


Lessons Learnt

  • While efforts to coordinate and integrate wildlife research projects have been widespread, and largely successful, there is a continuing drive to share and collect even more data to allow scientists a more complete picture of the extent of ecological ecosystem restoration.
  • It is apparent that with in addition to a diverse range of datasets, there are also hydrologic modelling studies and wider regional studies, that this case study in particular of all of those within this monitoring guidance, presents an example of the archetypal “all-singing, all dancing” example of appraisal in river restoration.
  • While this project is perhaps an exceptional case in terms of the scale, cost and ambitious expectations set, much will still be learnt about the ecosystem-scale impact of undertaking not only river, but also floodplain and wetland restoration on their associated biotopes and species. With very little data available at this scale, following the completion of headwater projects to increase water storage capacity in the Upper Chain of Lakes by 2013, restoration evaluation of ecosystem recovery will continue through to 2018, creating a monitoring record of almost two decades, probably the longest river monitoring record worldwide.


References

11.6 Shopham Loop

Shopham Loop

The project

In 2004, on the Western River Rother in West Sussex, an 18th century canal which cut off approximately 850 m of meander loop was blocked with a dam, forcing the flow back round the loop. Previously, remnant flow in the loop had caused excessive sedimentation, and so this sandy material had to be removed. At the same time, parts of the floodplain were lowered and a levee augmented to encourage flooding on the inside of the loop; a scrape was excavated; and cobble and shingle fixed beds were installed just inside the up- and downstream confluences with the old canal. The inset figure shows the general layout of the site, as well as surveyed cross-sections.


Monitoring design

Monitoring of the project aimed to be a comprehensive programme sensitive to:

  1. Changes in geomorphology, looking at the evolution of physical habitat features.
  2. Changes in the hydrology and hydraulics of restored and adjacent reaches, to identify the impact on flood levels and enable analysis of in-channel hydraulic conditions within the restored reach.
  3. The ecological response within restored and adjacent reaches, to document how the biota adjust to the changing physical habitat and, via habitat suitability models, identify driving mechanisms.
  4. The ecological response of the surrounding landscape, particularly species in the floodplain.
  5. The drivers of changes in channel morphology, substrate composition and the establishment of flora and fauna, to compare the restored physical habitat with design aspirations.


The following datasets were collected:

Datasets collected


Note that this was partly an experimental programme, designed to investigate the best approaches to monitoring by attempting to understand interactions between monitored aspects. It is probably more comprehensive than is necessary for a typical individual UK restoration project.


Outcomes

Monitoring results led to the following basic conclusions, grouped according to the aims above:

  1. Survey detected small changes in channel shape in some areas, and that the greatest changes happened very quickly (< 1 yr). This confirmed the loop was evolving greater complexity of form.
  2. Cross-sections and flow data from downstream allowed modelling of hydraulic habitat, which was increasing in diversity in concert with increasing complexity in the cross-sections.
  3. Fish numbers and diversity appear to have increased post-construction, when controlled for trends up- and downstream. The scrape is being well colonized, suggesting more flooding. Macrophyte cover has increased steadily, but species number peaked soon after construction. Invertebrate data show no clear trends except a peak in diversity and numbers in 2006.
  4. Coarse-scale changes in floodplain vegetation can be detected via the fixed-point photography.
  5. The fixed cobble beds appear to be responsible for the greatest morphological changes.


Lessons Learnt

The project suffered from a lack of clear objectives and a formal protocol for reference. Though there are monitoring aims (listed above), the project objectives did not meet SMART specifications and were in fact decided upon after the work had started, for the purpose of the monitoring. Consequently, it is difficult to appraise the success, or otherwise, of the project with hard evidence. The fact that methods to be employed were not explicitly detailed meant that this ambitious monitoring program, in the absence of a consistent project manager, was greatly limited by apparently minor mistakes, misinterpretations and inconsistencies resulting from the involvement of many different members of staff. A significant amount of data had to be discounted from analyses as they were not comparable with previous and/or subsequent years.

The fact that this was a rather experimental exercise in monitoring dictated that the programme did evolve over time, with gradually more sampling introduced. However, this is to be avoided, owing to inconsistencies between years and the fact that newly collected data will lack baseline (or „before‟) control data. It is by far preferable to begin by sampling more points than can be sustained (ensuring a comprehensive baseline), and then eliminating those which may prove difficult to access in future, for example, or appear to be of less value or cannot be used with great confidence. The

Shopham program was not designed before works started, and so the baseline is incomplete. As such, most conclusions relate to changes after construction, rather than any improvements over the pre-project situation. The collection of data to control for effects outside the reach was fairly good, aided by the convenient location of standard surveillance monitoring stations for invertebrates and fish not far up- and downstream. Beyond these datasets, however, there was little indication of typical background dynamics of geomorphology or seasonal and successional progression of vegetation on nearby un-impacted parts of this river.

Finally, the selected methodologies were not capable of meeting all of the programme aims. No firm conclusions could be made about the evolution of specific features such as banks, bars and berms (mentioned in the full version of Aim 1). The ecological response of the floodplain and surrounding landscape was neglected (Aim 4), and there was insufficient information to meet all aspects of Aim 5, particularly with regard to substrate composition and the original design aspirations. Lack of baseline data prohibited firm conclusions as to the impact on flooding.

In summary:

  • Primarily, a lack of SMART objectives meant…
  • methods selected were not always appropriate for evaluating success, and…
  • insufficient control data were collected.
  • Also, a lack of formal protocol definition and planning …
  • led to mistakes in data collection and abandonment of many data.


An improved protocol

With the benefit of hindsight, taking on the lessons above and more specific issues relating to each dataset, the following outlines a suggested revised protocol to meet the (somewhat 'woolly') aims stated earlier for Shopham.

1 Geomorphology

Cross-sections should be surveyed only at key areas of expert-predicted channel adjustment in the design (including immediate down- and upstream reaches and where change may be a particular problem) and defined by clearly visible, fixed and surveyed in markers well up on the bank (e.g. painted 2 m stakes driven into the ground). An evenly horizontally distributed number of elevation measurements should be taken before works begin, immediately after construction (i.e. before any flow is allowed in the loop) and in the 1st, 2nd, 4th and 10th years after completion, or more or less frequently, as appears necessary. Measurements should be taken when vegetation is limited but flows are not too great to prohibit access. It is also suggested that the surveys extend to the canal cut, which remains as a backwater and assumed sediment sink. It is key that the surveys are done in exactly the same way each time. See also point 5 below.

Physical biotope (habitat) mapping based on the RCS methodology is a more direct way to achieve Aim 1. It is suggested that this is done at the same times as the topographic survey, or in conjunction with channel ecology data collection if this is done more frequently, so that interactions can be investigated as part of Aim 5, and that the maps are digitized to enable further analysis. See also point 5 below.


2. Hydrology & hydraulics

The water level recording set-up (one sensor in the loop and two more a few hundred meters up- and downstream) was sufficient, in conjunction with floodplain surveys, for determining impacts on flood levels and frequency. The sensors (pressure transducers) must be properly protected from peak flow events and vandalism, as well as their elevation accurately surveyed at installation. A simpler design might be to have one or two pressure sensors at points of interest in the floodplain, giving a more clear-cut definition of flooding and flood depth, and not necessarily requiring surveying-in. In-channel measurements aid the verification of hydraulic modelling, however (see below).

Hydraulic conditions and how they change may largely be inferred from interpretation of the geomorphological monitoring, but modelling based on the topographic survey will be informative. Data requirements for modelled sections or more intensively surveyed reaches (see point 5, below) are likely to include flow data (from nearby gauging stations, perhaps adjusted in light of locally recorded water levels or velocity measurements) and observations of the channel vegetation and substrate, all recorded at the same time as the surveys.


3. Channel ecology

The sampling area of all ecological monitoring should be extended to include the canal cut backwater, which represents an entirely novel set of habitats and another area of great change due to the project.

Sampling times of invertebrates and fish should be matched to up- and/or downstream surveillance monitoring as closely as possible, to control for wider patterns of natural variation.

Kick sampling should be performed using a standard method to facilitate analysis, but specifically where and when depends on monitoring objectives and available resources. If a representation of the whole loop is required, and many samples may be taken, a randomly distributed design may be suitable. If fewer samples may be taken, it is advisable to take these at fixed points. It is suggested that sampling should again be contemporaneous with the geomorphological surveys, though to account for high variability in invertebrate communities, sampling should be more frequent if possible. In any case, if nearby surveillance monitoring data are not available for any sampling period, further control data should be collected. Furthermore, one should be aware that the limited spatial extent of surveillance monitoring may not make it comparable with the more extensive loop monitoring. If interested in the colonization of the new habitats, this may be rapid and so sampling frequency should be increased in the first year. See also point 5, below

Electro-fishing should again be contemporaneous and standardized with wider surveillance monitoring. It‟s not essential that this sampling is performed at the same time as other data are collected, but the frequency of the topographic surveys may be a good guide (minus the „as-built‟). Main known spawning periods should be avoided. One of the few specific project objectives was fishery enhancement, and so it is recommended that size (or weight) data be collected, and larval fish be explicitly included, to capture population dynamics.

Macrophyte data collection actually implemented (species and total percentage cover) was fairly fit for purpose, but cover estimates for individual taxa would be informative, and pre-project data for the original course of the river (i.e. the canal, but also the loop if possible) should be collected. Only staff capable of identifying confidently to species level should collect these data, as apparent mis- or incomplete identification were an issue in the actual implementation. See also point 5 below.


4. Landscape and floodplain

Fixed point photography should aim to cover the full site, particularly focusing on predicted areas of change. Many points should initially be established, and if necessary, some may be discarded at a later date. These points should be clearly marked with stakes, in a similar way to those demarcating cross-sections, and additionally marked with a direction for the central point in the image. Ideally the same digital camera and lens settings should be used year on year, to facilitate any digital image processing and overlaying (35 mm slide film was used in reality here). Certainly, data about the focal length (usually recorded automatically as file metadata with most digital cameras) and field of view should be noted to allow for later correction if necessary. Pre-project photos absolutely must be taken, as should „as-built‟. Photos should be taken at least once during winter, and preferably not during high flows (so that the channel is most visible), and once during summer (when the vegetation is most visible). As succession of vegetation is ongoing, sampling should occur every year for the first 5 years, and then perhaps every 3-5 years. Coverage of other river sites is advised.

Quadrat surveys of the floodplain will capture changes in plant species present and relative cover. A campaign of 10 – 20 x 1 m2 quadrats, randomly distributed, pre- construction and then conducted towards the end of spring in each year photographed should be sufficient.

Pre-project bird surveys are already available for the immediate upstream reach. It is suggested that increased flooding and vegetation change is likely to affect the avian community at this site, and that these are followed up, perhaps in years 1, 2, 4, 7 and 10. Care should be taken to ensure that the same sampling effort and methods are employed. Ideally, pre-project data should be collected for the exact restoration site, but owing to inter-annual variability, these might have to be collected for several years to get a clear picture of the species present. Involvement of community groups or local ornithological societies in these surveys is a good way of proceeding.

Pit-fall traps would also be valuable for detecting changes in the floodplain invertebrate and small mammal communities. Around 20 of these, well distributed around the site, could be installed semi-permanently and sealed with lids when not actively sampling. Species and abundance data could be collected, over a period of a week or so, pre-project and once or twice (not during winter) yearly at the same frequency as suggested for the bird surveys. It is important that traps are opened at the same times each year, and that they are checked daily to minimize trapped organisms predating each other.


5. Process drivers

The rather ambitious goal of investigating the drivers of changes would require a large dataset for statistical analysis, and thus considerable further monitoring.

The topographic survey approach could be modified in this case.

Still focusing just on key areas where change is expected or important, channel elevation (bathymetry) data could be collected, distributed either evenly (gridded) or in a properly randomized manner in the x and y directions. This represents a more intensive, smaller-scale and more evenly distributed version of the approach illustrated in the inset figure, and would allow more powerful statistical and modelling analyses and importantly, reliable interpolation (kriging) between the measured points, building a 3D surface model of the channel (coloured layer in the figure). Simple cross-sections may be extracted from anywhere on this surface. Even in the sparsely distributed, large- scale example illustrated, it‟s clear that this approach also aids identification of morphological features such as pools and meander bars.

During biotope mapping, special attention should be given to recording the substrate present. Dependent on specific details of objectives, samples may be taken for particle size analysis.

Point velocity measurements across surveyed sections or reaches would help validate the modelling mentioned in point 2, but also allow more detailed interpretation of how hydraulic factors are influencing or associated with physical habitat development and the establishment of vegetation and other organisms.

Kick samples stratified by physical habitat, will allow the investigation of associated or causative factors in biota establishment. For a complete picture, at least some of these biotopes should have their bathymetry surveyed. This stratification was actually done, though without formal biotope definitions there was little comparability between different sampling personnel‟s interpretations of the meso-habitats.

Macrophyte mapping would also facilitate investigation of the mechanisms of morphological change, as channel plants are obvious candidates as both drivers and responders. It is recommended that these data are recorded digitally as areas of coverage to facilitate integration with other datasets.