General enquiries on this form should be made to:Defra, Science Directorate, Management Support and Finance Team,Telephone No. 020 7238 1612E-mail: research.competitions@defra.gsi.gov.uk

SID 5 Research Project Final Report

SID 5 (Rev. 3/06) Page 1 of 26

NoteIn line with the Freedom of Information Act 2000, Defra aims to place the results of its completed research projects in the public domain wherever possible. The SID 5 (Research Project Final Report) is designed to capture the information on the results and outputs of Defra-funded research in a format that is easily publishable through the Defra website. A SID 5 must be completed for all projects.

This form is in Word format and the boxes may be expanded or reduced, as appropriate.

ACCESS TO INFORMATIONThe information collected on this form will be stored electronically and may be sent to any part of Defra, or to individual researchers or organisations outside Defra for the purposes of reviewing the project. Defra may also disclose the information to any outside organisation acting as an agent authorised by Defra to process final research reports on its behalf. Defra intends to publish this form on its website, unless there are strong reasons not to, which fully comply with exemptions under the Environmental Information Regulations or the Freedom of Information Act 2000.Defra may be required to release information, including personal data and commercial information, on request under the Environmental Information Regulations or the Freedom of Information Act 2000. However, Defra will not permit any unwarranted breach of confidentiality or act in contravention of its obligations under the Data Protection Act 1998. Defra or its appointed agents may use the name, address or other details on your form to contact you in connection with occasional customer research aimed at improving the processes through which Defra works with its contractors.

Project identification

1. Defra Project code BD1323

2. Project title

WETTING UP FARMLAND FOR BIRDS AND OTHER BIODIVERSITY

3. Contractororganisation(s)

Royal Society for the Protection of Birds (RSPB)Centre for Agri-Environmental Research (CAER), University of ReadingAllerton Research and Education Trust (ARET)Ponds Conservation (PC)     

54. Total Defra project costs £ 607,501(agreed fixed price)

5. Project: start date................ 01 August 2004

end date................. 31 July 2007

SID 5 (Rev. 3/06) Page 2 of 26

6. It is Defra’s intention to publish this form. Please confirm your agreement to do so...................................................................................YES NO (a) When preparing SID 5s contractors should bear in mind that Defra intends that they be made public. They

should be written in a clear and concise manner and represent a full account of the research project which someone not closely associated with the project can follow.Defra recognises that in a small minority of cases there may be information, such as intellectual property or commercially confidential data, used in or generated by the research project, which should not be disclosed. In these cases, such information should be detailed in a separate annex (not to be published) so that the SID 5 can be placed in the public domain. Where it is impossible to complete the Final Report without including references to any sensitive or confidential data, the information should be included and section (b) completed. NB: only in exceptional circumstances will Defra expect contractors to give a "No" answer.In all cases, reasons for withholding information must be fully in line with exemptions under the Environmental Information Regulations or the Freedom of Information Act 2000.

(b) If you have answered NO, please explain why the Final report should not be released into public domain

Executive Summary7. The executive summary must not exceed 2 sides in total of A4 and should be understandable to the

intelligent non-scientist. It should cover the main objectives, methods and findings of the research, together with any other significant events and options for new work.

SID 5 (Rev. 3/06) Page 3 of 26

Scientific Objectives

1. To review the delivery of resources for farmland birds, by different wet habitat features.

2. To use replicated experiments to measure the delivery of food resources and bird use of small-scale wetland features on the edges of arable and pastoral fields.

3. To make an agronomic assessment of the likely costs of a range of options for creating wet habitats in arable and pastoral farmland.

4. To synthesize results and make recommendations for provision of wet areas in arable land and intensive grassland within agri-environment schemes.

5. To disseminate results and practical advice to farmers and other stakeholders.

MethodsFour different types of wet feature were created or adapted to provide or enhance wet habitat availability in a mixed farming area of Leicestershire. These were: Bunded (dammed) drainage ditches in both arable and pastoral situations; ditch-fed Paired Ponds in field corners, Surface Scrapes in arable field margins; and livestock-excluded plots in existing wet areas of pasture (Waterlogged Areas). To experimentally assess the success of the features in terms of attracting farmland birds, providing enhanced invertebrate numbers for bird food and providing improved habitat for forging birds, the features were paired with nearby un-manipulated plots, to act as controls. Plots were created in autumn 2004 and fieldwork carried out in the period April 2005 to March 2007.

Bird use of plots and controls was assessed using timed watches. Emergent aquatic insects were sampled with fixed or floating emergence traps, terrestrial invertebrates with both pitfall traps and sweep-netting and botanical composition using a quadrat survey method. Birds were monitored year round, and other groups during spring and summer only. Emergent aquatic insect sampling was confined to the ditches only, terrestrial invertebrate and botanical surveys were undertaken at the ditches and scrapes.

Analysis of the results concentrated on establishing differences between the experimental and control plots, and where possible, suggesting how these differences could be exploited to enhance farmland biodiversity.

ResultsArable and Pastoral Bunded Ditches and Paired Ponds successfully met the primary aim of retaining water longer than the controls, and in many cases, year round. Stock exclusion in wet pastoral areas led to a reduction in surface water, possibly due to enhanced vegetative growth. Surface Scrapes were largely unsuccessful in retaining water during dry periods.

A wide variety of birds visited the experimental sites, which they exploited for feeding, drinking, bathing and territorial activity. However, visit rates of individual species was low, so it was necessary to group records of all bird species together for analysis.

For all feature types, excluding Surface Scrapes (where data were too sparse for analysis) bird visit rates were significantly higher in the experimental or wet plots than the controls. This differential was apparent year round, but most pronounced in the summer months. For the ditches and ponds there was no significant difference in bird use between 2005 and 2006 suggesting that some benefits accrue rapidly and do not require plot maturation. For the Waterlogged Areas, bird selection of the experimental features was greater in 2006, suggesting benefits due to management may increase over time.

Emergent aquatic insect biomass was significantly higher in the Bunded Ditches than the controls. Pastoral ditches produced a greater biomass of emergent aquatic insects than arable ditches.

Emergent aquatic insect biomass was greater where there were more extensive areas of bare, wet mud and lower levels of shading (from hedges). Most of the aquatic insects trapped were small or medium in size (less than 10mm). Increasing the number of larger insects may require adaptation of the features e.g. by making them larger.

SID 5 (Rev. 3/06) Page 4 of 26

For both Scrapes and Bunded Ditches, impacts on vegetation communities were consistent with an effect of soil disturbance, rather than water retention. The effect included increases in bare earth in year one which declined thereafter, as perennial vegetation recovered. Effects of water retention on perennial vegetation will likely take longer to accrue.

Bunding of ditches had a positive impact on surface active invertebrates, including adult Coleopteran and Diptera, which may reflect both wetness and availability of bare ground.

Bird visit rates to individual plots were positively correlated with aquatic insect biomass production, suggesting that birds actively seek out areas where foraging will be most productive. To a lesser extent, on the Pastoral Ditches only, foraging was also associated with the abundance of small sward-active invertebrates. However, there were no obvious links between treatment, invertebrate production and bird response.

The similar response of birds over the range of features considered suggests that providing wet habitat in farmland could lead to modest benefits, although the method of provision may be less important.

ConclusionsProviding wet features in farmland increases aquatic insect production, which seems to provide benefits for birds that feed on them. General biodiversity gains can also be expected, as habitat heterogeneity is increased at the local scale.

All features considered here would be suitable for inclusion in the existing English Environmental Stewardship Scheme, although whether at the Entry or Higher level still requires some clarification. Our preference would be to strive for inclusion at the Entry Level, where the benefits to biodiversity would be available for the widest uptake.

The experimental plots have been shown to a large number of land managers at various demonstration events. Their response has been extremely positive, with most prepared to incorporate such features on their farms, providing adequate funding is made available for capital works.

The current monitoring data cover a period of only two years, when the features are very new. To assess their full value requires continued monitoring over a longer time-frame as the features mature, including a period of remedial sediment removal.

Assessment of the biodiversity benefits from providing wet habitats in farmland would also be of considerable value. Particularly if these can be linked to other ecosystem services, such as flood mitigation and reduction of diffuse pollution from farming. Further investigation, particularly with regard to the scale of features required to deliver substantive results for a range of services, including biodiversity, is needed.

Project Report to Defra8. As a guide this report should be no longer than 20 sides of A4. This report is to provide Defra with

details of the outputs of the research project for internal purposes; to meet the terms of the contract; and to allow Defra to publish details of the outputs to meet Environmental Information Regulation or Freedom of Information obligations. This short report to Defra does not preclude contractors from also seeking to publish a full, formal scientific report/paper in an appropriate scientific or other journal/publication. Indeed, Defra actively encourages such publications as part of the contract terms. The report to Defra should include: the scientific objectives as set out in the contract; the extent to which the objectives set out in the contract have been met; details of methods used and the results obtained, including statistical analysis (if appropriate); a discussion of the results and their reliability; the main implications of the findings; possible future work; and any action resulting from the research (e.g. IP, Knowledge Transfer).

SID 5 (Rev. 3/06) Page 5 of 26

Scientific Objectives (all fully met)

1. To review the delivery of resources for farmland birds, especially PSA target species, by different wet habitat features; including identification of the most appropriate forms and configurations of wet areas for introduction into arable and intensive grassland areas.

2. To use replicated experiments to measure the delivery of food resources and bird use of small-scale wetland features on the edges of arable and pastoral fields. The experimental work will focus on wetting-up existing ditches, an option with the potential for wide uptake. Ancillary work will be carried out on paired ponds and surface scrapes newly created on buffer strips, and waterlogged areas on buffer strips/grassland.

o To measure success in delivering open water, bare earth, sward heterogeneity and a diversity of terrestrial vegetation, as a measure of access to food resources for farmland birds. *

o To measure success in delivering obligate wetland invertebrates, especially those important as a food resource for farmland birds. *

o To measure success in delivering terrestrial invertebrates, especially those important as a food resource for farmland birds. *

o To measure use of wet areas in arable land and intensive grassland by foraging farmland birds.

* Due to financial constraints, not all wet features were surveyed for all taxa (see below).

3. To make an agronomic assessment of the likely costs of a range of options for creating wet habitats in arable and pastoral farmland.

4. To synthesize results and make recommendations for provision of wet areas in arable land and intensive grassland within agri-environment schemes.

5. To disseminate results and practical advice to farmers and other stakeholders, as results become available.

Methods & Results

Objective 1: Farmland birds and wet habitats – Literature Review

Methods

The initial review was carried out at the start of the project in winter 2004\05, submitted to the journal Biological Conservation for publication in April 2005 and published, after revisions in January 2006. The review attached in Appendix 1 has been updated to include relevant material published up to May 2007.

Literature searches were carried out using ISI Web of Science® (1981-present), checking the literature cited in the resulting collection of papers and reports to cover older publications. First, we searched for literature on UK farmland birds. Hypotheses cited by authors to explain variations in behaviour or demography, with respect to wet habitat resource provision, are collated and presented. We then searched the literature on resource protection and diffuse pollution from the UK and elsewhere, to identify consequences of drying out of farmland for resource protection and potential resource protection solutions with benefits for farmland birds. Decisions on inclusion of resource protection solutions were necessarily subjective, based on our consideration of whether they could provide the resources required by the farmland birds.

ResultsIn the UK, agri-environment schemes are key mechanisms for reversing the declines of birds and other farmland biodiversity, but recent reviews suggest that wet habitats might be a gap in provision by these schemes. Important resources provided by wet habitats include: (i) damp soil, for probing species, (ii) permanent water to provide water-dependent invertebrates, as a source of food, (iii) bare or sparsely vegetated ground in the draw-down zone, to improve access to food, and (iv) rank emergent vegetation for nesting. However, wet habitats have been lost from farmland as a result of loss of ponds and filling of ditches, as well as the effective removal of water from fields by surface run-off, itself affected by soil compaction, and extensive under-field drainage. The efficient removal of water from fields can cause problems downstream, both through flooding, and diffuse pollution. Regular farmland pollutants include pesticides, nitrogen, phosphorus and sediment, leading to environmental problems such as eutrophication and reduced quality of drinking water. Major new political instruments, such as the Water Framework Directive, will aim to reduce the impact of this diffuse pollution from agriculture. A variety of solutions to diffuse pollution, such as conservation tillage, buffer strips at field edges, and small constructed wetlands, could simultaneously provide some of the resources required by farmland birds. Future agri-environment schemes, to be truly multifunctional, could focus on bringing these diverse objectives together.

Objective 2a: The physical impact of feature creationFour different experimental treatments, consisting of 32 bunded ditches (16 arable, 16 pastoral), 8 paired ponds, 10 surface scrapes and 8 waterlogged areas, all paired with adjacent control areas, were created to investigate the biodiversity benefits of providing or adapting wet areas within agricultural land. Further details, including photographs of the different features, are provided in Appendix 2a. We specifically set out to:

Create a range of experimental plots that retained water beyond the period that would normally be expected.

Investigate the impact of the treatments over time on sedimentation, in order to assess maintenance requirements.

Compare the performance in terms of water retention between arable and pastoral sites. Assess the effects of stock exclusion on water availability in existing waterlogged areas within pastoral

catchments.

Methods

Site Selection, Construction and Management of Experimental Plots Experimental sites were selected in Autumn 2004, and the Bunded Ditches, Paired Ponds and Surface Scrapes were constructed using a JCB 360 tracked digger to excavate the sites and build bunds to retain water. Plastic pipes were incorporated at the top of the bunds to allow out-flow of water once the features became full. Control sites were selected upstream of the experimental features in the case of the ditches and in adjacent areas of field margin for the Paired Ponds and Surface Scrapes. For the Pastoral ditches only, all experimental and control plots were fenced to exclude stock prior to data collection. Bunds and Surface Scrape bases were seeded with a grass mixture to aid cohesion and minimise erosion damage. The Waterlogged Areas differed from the other experimental types as they were not physically created, but exploited existing wet areas in pastureland. Fencing to exclude stock was erected to incorporate wet and dry areas of around 100m2 and adjacent unfenced wet and dry areas of a similar size were selected to act as controls, giving four different plots; wet fenced, wet open, dry fenced and dry open. Following initial site construction, no further maintenance was undertaken at any of the features, other than normal farm management practices (e.g. hedge and margin trimming).

Recording MethodsA Standardised set of physical measurements relating to water retention within the different experimental components were collected twice monthly (April to October) and monthly (November to March), these are defined in Table 1. Additionally, sediment build up was measured on nine occasions, regularly spaced through the duration of the project, at experimental plots only, for Arable and Pastoral Bunded Ditches and Paired Ponds.

Table 1. Definition of physical measurements relating to water retention.Measurement DefinitionWater Depth Measured by fixed ruler at deepest point in ditches and ponds

and by portable ruler at deepest point in control plots.Ditch Water Extent Proportion of ditch length with standing or running water

(estimated in eighths).Ponded Water Extent Proportion of ditch length with standing (ponded) water

(estimated in eighths at experimental plots only).% Surface Water (Waterlogged areas only)

Percentage of total land surface within each treatment covered by surface standing water.

Statistical AnalysesGeneralised Linear Mixed Models (GLMM’s) were constructed using the statistical software package SAS (v9) to investigate the success of the bunded ditches in retaining water over time (Model 1), and to investigate differences in surface water availability in the wet sections of the Waterlogged Areas that could be attributed to stock exclusion (Model 2).

Model 1: Extent = Treatment + Period + (Treatment*Period)Model 2: Surface Water = Treatment + Period + (Treatment*Period)

“Period” refers to eight consecutive three-monthly periods beginning in April 2005 and ending in March 2007, “Treatment” is experimental (bunded) or control in Model 1, and wet fenced or wet open in Model 2. The analyses are described in more detail in Appendix 2a.

ResultsFor both Arable and Pastoral Bunded Ditches, there was a highly significant effect of “Treatment” (bunding) and “Period” on the extent of water within the ditches (P = <0.0001 in all cases). The interaction between Treatment and Period was non-significant. Adjusted Least Squares means calculated from the model are presented for Arable Ditches in Figure 1, to show where the main differences occurred.

Figure 1. Differences in water retention over time Figure 2. Difference in standing water availability between Arable Bunded Ditches and their between fenced and open wet sections of the equivalent controls Waterlogged Areas.

Pastoral Ditches show a similar pattern, although both bunded and control plots generally remained wetter for longer than those in arable land. For the Waterlogged Areas, there was a highly significant difference (P = <0.0001) between periods and between treatments (fenced and unfenced) in surface standing water (Figure 2). Again, the interaction between Treatment and Period was not significant.

Comparative analysis of the Paired Ponds and Surface Scrapes was not applicable, as the control areas were field margins not expected to retain water. However, the Paired Ponds were generally successful in the primary aim of retaining water. During the driest periods (2 and 6), which represent July, August and September in 2005 and 2006 respectively, six of the eight Paired Ponds retained water in at least one of the two ponds. Surface Scrapes, in contrast to the other experimental features examined within this project, were largely unsuccessful in retaining water. Only one of the ten in 2005 and two in 2006 retained water longer than the surrounding land.

Sedimentation ratesSediment build-up over time within the experimental features has the potential to affect their efficiency in retaining water, and is likely to affect colonisation and use by a range of flora and fauna. Figure 3 shows how sediment accumulated in the Arable and Pastoral Bunded Ditches and the Paired Ponds over the period of the project. In terms of subsequent management, it is likely that ditches and ponds will require removal of sediment at regular intervals in order to maintain their primary function. Accumulated sediments are likely to be nutrient-rich, and spreading them on nearby land could provide benefits in the form of nutrient recycling.

Figure 3. Mean sediment build up over time for Arable and Pastoral Bunded Ditches and Paired Ponds. Confidence limits excluded, for clarity.

Objective 2b. Terrestrial vegetation structure

Methods

Sampling methodsIn order to characterise the ditch-bank vegetation, assessment of the cover of vascular plant species was carried out in 1m x 50 cm quadrats in July of 2005 and 2006. Quadrats were placed at two positions on the ditch-bank (bank-top and bank-side) and at three distances from the bund. A similar spatial structure of sampling positions was adopted in the control sections of ditch.

The second method was designed to characterise heterogeneity in vegetation composition and structure. A line of contiguous 50 x 50 cm quadrats was laid out in a transect along the top of the ditch-bank and through the centre of each Surface Scrape.

Analytical approachesDifferences in characteristics of the vegetation between Bunded and control ditches, and between Surface Scrapes and their controls, were tested using repeated measures mixed models.

The relationship between characteristics of the vegetation and the observations of birds foraging on the ditch-bank were explored using multiple regression techniques. First, the value of the vegetation variables as predictors of bird foraging were examined, then a block-wise addition approach was used to include the effects of the experimental treatments in the regression analysis.

The value of the experimental treatments in contributing to wider benefits for plant diversity were examined. Repeated measures mixed models were used to identify differences in plant diversity and redundancy analysis employed to investigate the effects of the treatments on plant community composition.

Results

0

10

20

30

40

50

60

2005 2006

Perc

enta

ge c

over

Bunded banksideBunded banktopControl banksideControl banktop0

102030405060708090

100

2005 2006

Perc

enta

ge c

over

Bunded banksideBunded banktopControl banksideControl banktop

y = 0.1887x + 1.4157R2 = 0.2791

0

5

10

15

20

25

0 10 20 30 40 50 60 70Water depth (mm)

Fora

ging

act

ivity

(bird

min

utes

)

BundedControl

Impact of the features on vegetation structureThe bunded Arable Ditches were found to have more extensive areas of bare ground than their respective controls, although the magnitude of the difference declined between 2005 and 2006. This pattern was repeated in the Pastoral Ditches. In addition, the bunded Pastoral Ditches had lower cover of graminoids, with the magnitude of the difference decreasing between years. In general, graminoid cover increased through the course of the study, with bare ground and forb cover declining. No treatment effects were found on the combined cover of forb species known to be important in the diet of farmland birds.

(a) Bare ground (b) Graminoid cover

Figure 4. Differences in the extent of bare ground and graminoid cover, in relation to experimental treatment (control vs. bunded) and quadrat position (bank-side vs. bank-top) for the Pastoral Ditches.

The vegetation of the Surface Scrapes differed significantly from that of the control areas. Bare ground was more extensive and graminoid cover lower in the Scrapes. The magnitude of the differences declined between 2005 and 2006. The vegetation of the Surface Scrapes was more heterogenous than that of the control areas. The Scrapes contained a higher diversity of vegetation types and greater variability in vegetation height. Again, the magnitude of these differences declined through the course of the study.

Relationship with bird foragingThe extent of bare ground had a significant relationship with the level of foraging by birds on the Arable Ditch-banks. The level of foraging was also affected by the depth of water in the ditch, primarily influenced by the construction of the bunds. A similar pattern was found for the Pastoral Ditches, with both bare ground cover and extent of water held in the ditch providing useful predictors of the level of bird use. In addition, a negative relationship between graminoid cover and the numbers of birds foraging on the ditch bank was found.

Figure 5. Relationships between depth of water in the Arable Ditches, extent of bare ground on the bank and the foraging activity of birds on the ditch-bank (NB data from bunded and control ditches were analysed together but are differentiated on the graph to aid interpretation).

Additional biodiversity benefitsFor both the Arable and Pastoral Ditches, the treatments resulted in significant changes in plant community composition. Vascular plant species richness was depressed by the treatments in 2005, especially on the bank-sides. The vegetation of the bunded ditch-banks was characterised by high cover of bare ground and cleavers

Treatment p<0.01Position p<0.001

Treatment p<0.05Position p<0.001

0

50

100

150

200

2005 2006

Num

ber p

er s

ampl

e

Bunded banksideBunded banktopControl banksideControl banktop 0

30

60

90

120

2005 2006

Num

ber p

er s

ampl

e

Bunded banksideBunded banktopControl banksideControl banktop

(Galium aparine), with longer-lived forb species such as welted thistle (Carduus crispus) and creeping buttercup (Ranunculus repens) increasing in the second year. The control sections were characterised by cow parsley (Anthriscus sylvestris), false oat-grass (Arrhenatherum elatius) and bramble (Rubus fruticosus).

The diversity of vegetation types on the Pastoral Ditches was increased by the experimental treatments. In the bunded ditches species such as annual meadow grass (Poa annua) and wavy bittercress (Cardamine flexuosa) increased in response to the treatments, whilst the control sections were characterised by perennial agricultural grasses such as crested dog’s-tail (Cynosurus cristatus) and perennial ryegrass (Lolium perenne).

Objective 2c. Terrestrial invertebrates

Methods

Sampling methodsTwo main sampling methods were used, designed to characterise the abundance of invertebrates active in the sward or on the soil surface. Sweep netting was carried out in the vegetation on the ditch-bank of the Bunded and control ditch sections, and in the Surface Scrapes and their respective controls. Pitfall trapping was used to sample surface-active invertebrate taxa. Traps were placed along the ditch-bank of the Bunded and control ditch sections, and in the Surface Scrapes and their respective controls. In the Arable Ditches, differences in beetle assemblages in relation to proximity to the ditch were also examined using a further set of traps. The invertebrates sampled were sorted into broad taxonomic groups and the numbers of individuals in different size classes counted for taxa known to be important in the diet of farmland birds. In addition, the beetles of certain families collected using the two sampling methods were identified to species level in order to allow investigation of the impacts of the treatments on the diversity of these groups. Leaf beetles (Chrysomelidae) and weevils (Apionidae, Curculionidae) were identified from the sweep net samples, and ground beetles (Carabidae) and rove beetles (Staphylinidae) from the pitfall traps.

For a sub-sample of the Arable and Pastoral Ditches, soil cores were taken in April and August each year in order to test for differences in the abundance and biomass of earthworms and other soil-dwelling invertebrates

Analytical approachesDifferences in invertebrate abundance and biomass between Bunded and control ditches, and between Surface Scrapes and their controls, were tested using repeated measures mixed models.

The relationship between overall invertebrate abundance and the observations of birds foraging on the ditch-bank were explored using multiple regression techniques. First, the value of the invertebrate abundance variables as predictors of bird foraging were examined, then a block-wise addition approach was used to include the effects of the experimental treatments in the analysis.

The value of the experimental treatments in contributing to wider benefits for invertebrate diversity were examined using data on surface- and sward-active beetles. Repeated measures mixed models were used to identify differences in beetle diversity and redundancy analysis employed to investigate the effects of the treatments on the composition of beetle assemblages.

Results

Impact of the features on invertebrate abundanceNo difference was found in the numbers of sward-active invertebrates between the control and Bunded sections of Arable and Pastoral Ditches. This was the case for each of the 11 invertebrate groups identified as being important in the diet of farmland birds and also the total numbers and biomass of invertebrates over 2 mm body length.

(a) Coleoptera adults (b) Diptera adults

Treatment p<0.05Position p<0.01Year p<0.001

Treatment nsPosition nsYear p<0.001Treatment.Year p<0.01

y = 0.0209x + 1.4429R2 = 0.1099

0

2

4

6

8

10

12

0 50 100 150 200 250Abundance of small, sward-active invertebrates

(number per sample)

Fora

ging

act

ivity

(n

umbe

r of b

irds)

BundedControl

Figure 6. Total numbers of surface-active Coleoptera adults, Diptera adults and larvae and Araneae in the Arable Ditches and their respective controls.

However, significant treatment effects were found amongst the surface-active taxa. In the Arable Ditches, numbers of adult Coleoptera were higher in the controls than the bunded sections. The abundance of Diptera adults was higher on the banks of the Bunded ditches in 2005, but the difference was not found in 2006. Diptera larvae were more abundant on the bank-tops of the bunded ditches, but differences in bank-side samples were not apparent. For the Pastoral Ditches, adult Diptera were more abundant in the samples on the control ditch-banks than those of the Bunded ditches, whilst adult Hemiptera showed a similar pattern in 2006 only. In contrast, Lepidopteran larvae were more abundant in the Bunded ditches in 2006.

The numbers of sward-active invertebrates of all body size classes were lower in the Surface Scrapes than in the control areas. For the surface-active taxa, adult Diptera were more abundant in the Scrapes than their controls, whilst the reverse was true for Diptera larvae. Lepidoptera larvae were more abundant in the control areas, but only in 2005, no difference being found in 2006.

Relationship with bird foragingA significant relationship was found between the level of bird foraging on the Pastoral Ditches and the abundance of small, sward-active invertebrates (body size 2-5 mm). This invertebrate abundance variable maintained its explanatory power when the extent of water in the ditch, which also had significant power in explaining bird foraging, was included in the regression equation. For the Arable Ditches, no variables relating to the overall abundance of sward- or surface active invertebrates had a significant relationship with the level of foraging by birds on the ditch-banks. However, the level of foraging was affected by the depth of water in the ditch, primarily influenced by the construction of the bunds.

Figure 7. Relationship between the abundance of small sward-active invertebrates on the banks of the Pastoral Ditches and the foraging activity of birds on the ditch-bank. (NB data from Bunded and control ditches were analysed together but are differentiated on the graph to aid interpretation).

Additional biodiversity benefitsThe abundance and diversity of sward-active leaf beetles and weevils was higher in the Bunded Arable Ditches than their respective controls. For the Pastoral Ditches, abundance was higher in the controls and no effect on the diversity of sward active beetles was found. No differences were found in the abundance of sward-active beetles between the Surface Scrapes and their controls, although their diversity was depressed by the Scrape treatment.

No effect of the treatments on abundance or diversity of ground and rove beetles was found, neither in relation to bunding of ditches or Surface Scrapes. There was a general increase in the abundance and diversity of these beetle families on the Arable Ditch-banks between 2005 and 2006, whilst the reverse was true for the Surface Scrapes and their controls. Small but significant effects of the Arable Ditch and Surface Scrape treatments on the community structure of surface-active beetles were found.

Objective 2d: Obligate wetland invertebrates

MethodsAquatic insects emerging from water or wet mud were collected using pyramidal emergence traps with a surface area of 0.1 m2. The traps were run between April and July in 2005 and 2006. Five traps were set in each bunded and control ditch (see Figures 2d.1 and 2d.2 in Appendix 2d). A range of physico-geographic data was also collected from each ditch (Appendix 2d, Table 2d.1). A more detailed description of the methods used is given in Appendix 2d.

ResultsLarge numbers of insects emerged from the wet or damp mud of the ditches. On average approximately 900 individuals per m2 were recorded from the control ditches compared to 1,400 individuals per m2 from the Bunded ditch features. At all sites, true flies (Diptera) dominated both in terms of abundance and biomass. Other taxa: Caddis flies (Trichoptera), stoneflies (Plecoptera), mayflies (Ephemeroptera) and dragonflies and damselflies (Odonata) were recorded more sporadically and produced a much lower proportion of the total biomass (Figure 8). In terms of the size of emerging insects, the greatest biomass was produced by small or medium sized flies (2-5 mm and 5-10mm respectively). In the first year (2005), medium-sized flies were particularly abundant in the Bunded Ditches and added significantly to the biomass produced.

Difference between Bunded Ditches and controlsOverall, the Bunded Ditches produced a significantly greater biomass of emerging insects per square meter than the control ditches (Figure 9). This difference was significant in both years, though most pronounced in the first year (2005, P<0.001; 2006, P =0.01). There was an apparent decline in the mean biomass per square meter produced by Bunded Ditches between 2005 and 2006 but this was not statistically significant. The mean biomass produced by the control ditches increased significantly between 2005 and 2006 (P=0.03). Since the Bunded Ditches were typically wider than the controls, they also had a more extensive wet area. This added to the total value of the Bunded Ditches in terms of insect production. Thus, on average, the Bunded features produced nearly four times the biomass of emerging insects produced by a similar length of ‘unbunded control ditch.

0

1

2

3

4

5

True flies Stoneflies Caddis flies Mayflies Dragonflies

Dry weight (g/m2

)

ControlBunded

Figure 8. Proportion of biomass produced by different faunal groups (two year site mean).

Figure 9. Biomass from bunded and control ditches.

0

350

700

1050

1400

2005 2006

Dry weight (mg/m2)

Control

Bunded

Differences between arable and pastoral ditches

Overall, the Pastoral Bunded Ditches produced a significantly greater biomass of insects than the Arable Bunded Ditches (P<0.001, Figure 10). This relationship held true in both 2005 and 2006 (see Appendix 2d, Fig.2d.9). The pastoral control sites also produced a greater average biomass than arable control sites, but this relationship was only statistically significant in 2006 (Appendix 2d, Figure 2d.9).

Seasonal differences in biomass In the first summer after bund construction (2005), the biomass of emerging insects produced by both the Bunded ditches and controls peaked in late May and early July (Figure 11). The Bunded Ditches had significantly greater biomass than the control ditches at this time and biomass remained significantly higher than the control sites until late July, when many controls were dry. During this period the biomass per square meter collected from the Bunded Ditches was, on average, 1.9 to 2.5 times higher than that produced by the controls. In 2006 the pattern of insect emergence was bimodal in both the control and Bunded Ditches, with a relatively high biomass produced in April and a second peak in late June. On average, in 2006, the bunded sites produced a greater biomass than the control sites throughout the summer, but this difference was only statistically significant at the end of the season in late July (Appendix 2d,Table 2d.3).

Explaining biomass differences between the sitesThere were considerable differences in the biomass produced by individual ditches (Appendix 2d, Figure 2d.5). Thus the most productive Bunded Ditches and their controls produced over five times the biomass of the least productive sites. Knowledge of the possible causes of this variation is valuable because it has implications for the optimal design of Bunded Ditches to maximise insect biomass production. The multilevel modelling programme MlwiN was used to model the effects of different environmental variables on biomass produced by the ditches. The results showed that:

1. The most important factors associated with aquatic insect biomass production in both the Bunded and control ditches were (a) the percentage of bare mud, and (b) hedge shade. Thus, biomass was likely to be particularly high between late May and July in any ditch (control or bunded) that had a high percentage of bare mud in the ditch-bottom, and relatively little hedge shade.

0

600

1200

1800

2400

28th Mar-12th April

26th April-10th May

24th May-8th June

22nd June-6th

July

20th July-2nd Aug

28th Mar-12th April

26th April-10th May

24th May-8th June

22nd June-6th

July

20th July-2nd Aug

2005

2006

Mean dry weight (mg/m2

) Control Bunded

Figure 11. Seasonal variation in biomass from bunded and control ditches

Figure 10. Biomass from arable and pastoral ditches.

0

400

800

1200

1600

Arable Pastoral

Mean dry weight (g/m2)

Control Bunded

2. Given that high biomass controls were often associated with high biomass Bunded Ditches, the control:bunded ratio was used to explore where the bunded features gave “added value”. Percentage of bare mud was the only variable which significantly explained the variation in this ratio (Figure 12). Thus, Bunded Ditches were likely to increase emergent insect biomass production most successfully if they provided a high proportion of bare mud in the ditch bottom.

Objective 2e: Responses of farmland birds to the wet features

Field MethodsA “timed watch” protocol was adopted for recording birds. Each experimental plot and paired control plot was observed for a period of 45 minutes from a portable hide. During the observation period, all birds entering and leaving the study plot were identified to species, categorised according to activity and micro-habitat visited and recorded on standardised fieldwork sheets. Each individual plot was visited for bird recording twice per calendar month in April to October, and once per month in November to March, excepting the Surface Scrapes, where watches were reduced to once-monthly, once it became apparent that they had not met with their primary aim of retaining water.

Analysis MethodsGeneralised Linear Mixed Models (GLMM’s) were used to analyse the bird response to the different wet features with respect to “Treatment” (experimental or control) and to two different interpretations of time. Firstly to examine possible effects of plot maturation, a fixed factor “Period” (eight sequential 3-month measurement periods from the start of the project) was specified. Second, “Season” and “Year” were used as fixed factors within a similar model to investigate the seasonal and annual response to treatment (see Appendix 2e for full details). For the Waterlogged Areas only, an additional third analysis was applied, to examine bird response to the two components of treatments; wet v dry and fenced v open. No analyses have been undertaken on data from the Surface Scrapes.

Model 1: Bird visits = Treatment + Period + (Treatment*Period)Model 2: Treatment + Season + Year + (Treatment*Season) + (Treatment*Year) + ( Season*Year)

Model 3: Bird visits = Dry/Wet + Fenced/Open + Period

ResultsDuring the bird observation watches, a wide range of information was collected on species visiting the plots, including their activity whilst present, and within-plot micro-habitat use. For the ditches and ponds the most regularly recorded species were very similar, with Wren, Robin and Dunnock between them accounting for 39% (Arable Ditches), 32% (Pastoral Ditches) and 21% (Paired Ponds) of all records. Waterlogged Areas, with their generally more open aspect, attracted a different sub-set of species. In total, 62 species were recorded across the range of experimental features over the two study years (see Appendix 2e).

Within the experimental plots, birds visited the hedge (where present) more frequently than any other defined micro-habitat, accounting for up to 55% of records (Arable Ditches). The most frequently recorded activity of birds whilst visiting each experimental feature type was foraging for self, which accounted for 43 to 54% of all

0.4

1.3

2.2

3.1

4.0

0% bare mud

Mean biomass ratio

22nd June–6th July

28th March –12th April

26th April – 10th May

24th May–8th June20th July–2nd Aug

0 25 50 75 100

Figure 12. Estimated effect of bare mud on insect biomass ratio by visit.

records across the ditches, ponds and waterlogged areas. Other activities (e.g. bathing, drinking, territorial behaviour, provisioning young) individually made up no more than 18% of the total records.

For the purposes of formal statistical analysis, it was decided to exclude any bird records from the hedge, on the grounds that these could not be attributed with any certainty to use of resources associated with the wet feature. This substantially reduced the amount of data available for analysis. This aggravated another problem of data scarcity – i.e. due to the small size of the features, actual visit rate to plots was very low (typically 0-2 visits in a 45 minute watch). While we do not consider this low actual visit rate to be a problem per se, as it is the difference in visit rate between control and treatment plots which is of interest, it did nevertheless lead to constraints on the analysis. Specifically, we found that there were insufficient data to analyse responses of individual species, or of guilds of species. Therefore, all results reported below refer to analyses of visits of all bird species combined. While many of these visits are by small insectivorous passerines (eg Robin, Wren and Dunnock), there is also reasonable representation of, for instance, small granivorous, red-listed passerines (eg Yellowhammer and Tree Sparrow).

For all experimental feature types, “Treatment” and “Period” were significant predictors of bird use (P=<0.05; Table 2), as were “Treatment” and “Season” (Table 3). Year very closely approached significance (P= 0.0518) for the Waterlogged Areas in Model 2, but not for the ditches or ponds. Interaction terms were not significant in any model. Adjusted least squares means were calculated from the models to show where the main differences lie (see Appendix 2e).

Table 2. Results from GLMM analysis to investigate the effect of treatment and period on the number of bird visits per timed watch.

Arable Ditches Pastoral Ditches Paired Ponds Waterlogged areasFactor df F-

valueP-value df F-

valueP-value df F-

valueP-value df F-

valueP-Value

Treatment 1,708 18.67 <0.0001 1,652 15.53 <0.0001 1,354 48.98 <0.0001 3,554 8.83 <0.0001Period 7,709 2.22 0.031 7,654 2.36 0.028 7,354 3.19 0.003 7,555 4.99 <0.0001Treatment*Period 7,701 1.6 0.132 7,645 1.33 0.235 7,347 1.01 0.425 Model failed

Table 3. Results from GLMM analysis to investigate the effect of treatment in each “season” and year on the number of bird visits per timed watch.

Arable Ditches Pastoral Ditches Paired Ponds Waterlogged areasFactor df F-

valueP-value df F-

valueP-value df F-

valueP-value df F-

valueP-Value

Treatment 1,711 18.38 <0.0001 1,679 17.02 <0.0001 1,357 49.61 <0.0001 3,557 8.92 <0.0001season 3,713 3.85 0.0094 3,680 5.35 0.0012 3,357 6.52 0.0003 3,558 9.6 <0.0001year 1,712 0.66 0.415 1,687 0.15 0.697 1,357 1.15 0.285 1,558 3.8 0.0518Treatment*Season 3,704 1.7 0.165 3, 72 2.39 0.068 3,350 1.15 0.33 9,542 0.65 0.758Treatment*Year 1,704 0.5 0.479 1,672 1.55 0.214 1,350 0.39 0.534 3,542 0.81 0.491Season*Year 3,705 1.15 0.33 3, 73 0.55 0.645 3,350 1.03 0.379 3,543 1.01 0.389

Figure 13. Comparison of overall effect of treatment for Arable and Pastoral ditches and Paired Ponds.

Figure 14 shows temporal trends in use of Arable Ditches only, but the pattern of use was similar for both Pastoral Ditches and Paired Ponds (see Appendix 2e). There was no obvious indication of a directional change in use over the course of the study.

.

Figure 14. Mean number of bird visits to experimental and control areas of the Arable ditches in each 3-month period of the study.

There were significant effects of season on the number of bird visits for all feature types. Again, the Arable Ditch results are presented as an example, showing the greatest mean number of visits and the greatest differential between experimental and control plots during summer (Figure 15). Similar patterns were found in the other feature types (see Appendix 2e).

Figure 15. Mean number of bird visits to experimental and control areas of the Arable ditches during each “season”.

For the Waterlogged Areas, bird visit rate was generally lower than for the other feature types. However there were significant differences between treatments, and between periods (Table 4). Unlike the other experimental features there appeared to be a year effect, with a trend towards more visits in the second than the first year, possibly indicating that bird benefits are increasing as the plot matures (Table 3). Combining all four treatments, the “spring” and “summer” periods (1 and 2, 5 and 6) had more bird visits than “autumn” and “winter”, and there was a trend towards a higher visit rate in the second year (5 to 8). This is illustrated in Figure 16.

Figure 16. Mean number of bird visits to waterlogged features (all treatments combined) in each 3-month period of the study.

Further analysis of the difference between treatments (Model 3, Table 4) shows that within the different plot types, wet plots attracted more bird visits than dry, and open plots more than fenced (Figure 17).

Table 4. Results from GLMM analysis to investigate the effect of treatment and period on the number of bird visits per timed watch for waterlogged areas.

Factor df F-Value P-ValueDry\wet 1, 555 21.06 <0.0001Fenced\open 1, 555 5.14 0.0237Period 7, 556 4.99 <0.0001

0

0.5

1

1.5

2

spring summer autumn winter

bird

vis

itsControlExperimental

Figure 17. Mean overall number of bird visits to the four Waterlogged Area treatments.

The response of birds to invertebrate biomass availabilityAlthough there are a variety of potential routes through which wet areas can provide benefits to birds (see Appendix 1), perhaps the most compelling and potentially wide ranging is through increasing the amount and availability of invertebrates on which birds can feed. Information from emergent aquatic insect sampling established that Bunded Ditch plots produced a greater biomass of emergent aquatic insects than control plots (Appendix 2d).

For neither Arable nor Pastoral Ditches was there any significant interaction effect of biomass and treatment on bird occurrence (Arable Ditch; F1,13 =0.18, P=0.6753: Pastoral Ditch; F1,13=0.76, P =0.3996). For Arable Ditches, neither biomass (F1,14=2.09, P=0.1699) nor treatment(F1,14 = 2.98, P= 0.1062) were significant when included together in the model. However, both were significant on removal of the other variable (biomass; F1,15= 7.07, P=0.0178; treatment; F1,15 = 6.87, P= 0.0193). For Pastoral Ditches, treatment was not significant ( F1,14 = 1.55 P= 0.2339) in the presence of biomass. Both biomass (F1,15 = 6.37, P=0.0234) and treatment F1,15 = 6.05, P= 0.0265) were significant when included in the model alone. The response to aquatic invertebrate biomass was positive – the greater the biomass present, the higher the probability of recording a bird during the watch.

Objective 3: Agronomic Assessment

MethodsTo determine the impact on farm management, the analysis focused on: (i) the financial costs of establishing the wet habitat features in terms of capital investment, (ii) the financial costs associated with ongoing maintenance and repair of features, and (iii) the costs associated with potential detrimental impact on farm income.

ResultsThe main cost of establishment for the bunded ditches, paired ponds and scrapes is the time, labour and machinery costs associated with the use of a ditch excavator. Additionally, for the bunded ditches and paired ponds, piping is required as an outflow in the bunded ditches and from the second of the paired ponds. Relevant costs are given in Table 5. The following assumptions are made. Labour use with regard to the creation of wet habitat features is based upon experience within the project. First, it would take one man and a digger one day (7.5 hours) to create three bunded ditches, that is about 2.5 hours per ditch. Second, in one day one man could create a paired pond. Finally, between eight and 10 scrapes could be created in a day using one man, that is about one scrape per hour. The equipment used at Loddington was a 13 tonne JCB 360° excavator, which would typically be used for ditch maintenance on many farms. Contractor costs would typically be £38 per hour for the hire of the excavator and £10 per hour for labour (Nix, 2006). Piping costs depend on diameter and length. Pipes of 3m in length and 150 mm diameter were used on both the bunded ditches and the paired ponds. The cost per pipe is assumed to be £4.25 per metre.

Table 5. Cost of Feature Establishment a

Item Ditch Paired ponds ScrapesDigger @ £38/hr £95.00 £285.00 £38.00Labour @ £10/hr £25.00 £75.00 £10.00Pipe at £4.25/m £12.75 £12.75 n/aTotal cost £132.75 £372.75 £48.00a Data based upon standard figures from Nix, 2006.

Following on from establishment, the other main cost is that of ongoing maintenance and repair of features. The primary concern here is that of the arable and pastoral ditches. Over time, and as expected, the ditches accumulate sediment, more so in the case of arable ditches and following periods of heavy rainfall. Assuming current rates of accumulation, the need to clear the ditches of sediment would be required about once every four years. To some extent this is dependent on soil type and typography. Within a 10 year period, it might be the case that ditch maintenance would be required two years after initial establishment of the bund feature and associated disturbance but then would not need to be done again for five years. The same equipment used for creation of the feature would be used for ditch maintenance with a similar hourly rate.

All of the features were established within existing field boundaries including set aside and stewardship margins. For the purposes of this exercise, the assumption is made that the land upon which the features are created is therefore not used for productive purposes and that no estimate of income foregone is required. However, it is possible that land adjacent to the features may face reduced yields due to the increased wetness of the area within the vicinity of the features. To establish whether this was the case, arable yields for a range of crops in 2006 were taken from a single grid reference point nearest to the features and also their controls. Only data for the latter period of the project data was collected to allow time for the wet habitat features to have established and, at the same time, to exclude the first year and potential detrimental impacts associated with feature creation works undertaking during the first autumn of the project. The final data set is shown in Table 6. Roughly half the experimental plots showed an increase in yield and the other half showed a decrease. This was irrespective of crop type. The paucity of data makes it difficult to undertake thorough statistical analysis, however, certain inferences can still be drawn, primarily that the wet features do not appear, at least in the first years of establishment, to be having a detrimental impact on adjacent crop yield.

Table 6. Impact on Yields of Adjacent Crops

Feature Adjacent crops Yield impact t/haBunded ditch Wheat (4), spring beans (3), oilseed rape (1) +1.48 to -0.85Paired pond Wheat (1), oilseed rape (1) +1.70 to -1.18Scrape Spring beans (2), oilseed rape (8) +0.85 to -1.95

Discussion (objective 4, synthesis of results)From the literature review (objective 1), there is ample evidence that a variety of farmland birds respond to a range of resources associated with wet habitats. These may include: damp soil, to benefit soil invertebrates and increase penetrability for probing species; permanent water to provide water-dependent invertebrates, as a source of food; bare or sparsely vegetated ground in the draw-down zone, to improve access to food; rank emergent vegetation for cover and nesting. We were therefore confident that at least some of the features created in this project would be attractive to farmland birds.

Feature Construction, Development and MaintenanceFor Bunded ditches (both Arable and Pastoral) and Paired Ponds, the primary aim, of retaining water longer than would otherwise occur, has been successfully achieved. Within the individual plots there was, however, considerable variation in performance and by no means all managed to retain water year-round (see Table 2a.3). Maintaining standing water year- round, while not the only aim of this project, is likely to be important to some aquatic invertebrate species whose life cycles require this. Also, it is likely that in the driest periods, providing an additional source of water will benefit other species including birds, which require water for drinking and bathing. Other benefits may be available from plots that are drying, or have recently dried, in the form of damp conditions that appear to stimulate aquatic insect emergence (see Appendix 2d) and favourable foraging conditions for birds. Similar periodicity in water retention between both sets of ditches and the Paired Ponds suggest that they could be equally successful as measures to introduce distinct, new, wet habitats to farmland. With the experience gained from this project and careful attention to siting it should be possible to maximise the chances that newly created features will retain water, and produce a high aquatic insect biomass.

Sedimentation of the ditch and pond features over time reduces their capacity for holding water, and although a certain amount of sediment appears beneficial to aquatic invertebrate production (see Appendix. 2d), continued build-up will require management by dredging if the features are to be maintained in the long-term. Sediment build up in the Paired Pond Ditches, for example, had reached a mean of 74% of original water depth by the end of the project (Table 2a.5). Sediments accumulating from agricultural catchments are likely to contain high levels of nutrients, some of which may be recycled if dredged material is spread back onto surrounding land. However, if further research were to suggest that clearing has a short-term impact on exisiting biodiversity, then, where features exist in close proximity to one another a rotational approach to dredging is likely to prove environmentally beneficial.

Surface Scrapes were largely unsuccessful in retaining water much longer than the surrounding land. In the few cases where they were successful, however, the habitat created included many of the variables deemed desirable, including water, wet mud and sparse vegetation (see photograph in Appendix 2a, Supp. Info.). Further encouragement is provided by the fact that two Scrapes retained water into summer 2006, compared to only one in 2005, suggesting that water retention may improve with time, possibly due to the process of “self-sealing”, whereby silt accumulation blocks porous cracks within the base. Overall, the value of Surface Scrapes is

unproven. More thought is needed as to the location of these features to ensure they retain water (e.g. at the base of field drains), and they may need a longer period of establishment before they function effectively.

Stock-proofing sections within extant wet areas of pastureland appears to reduce surface water availability compared to open areas (Fig.2). Some possible explanations for this observed effect are given below:

i) Some aspect of the treatment caused the fenced plots to dry out at a faster rate than the open plots. A possible cause could be increased evapo-transpiration from the greater volume of vegetation. If this was the only cause however, it could be expected that the differential between the plots would be substantially higher in summer (greater transpiration rate coupled with a greater soil moisture deficit) than in winter. This effect is not substantiated by the model results shown in Figure 2 (although bear in mind the substantial and sometimes overlapping confidence intervals).

ii) Some aspect of the treatment caused the open plots to remain wetter than the fenced plots. A possible explanation for this could be soil compaction and poaching by livestock. This “puddling” effect can create areas where infiltration rates are reduced, as frequently observed around stock feeders and in gateways.

Whilst it is not possible, without further and more detailed experimental work, to ascertain whether either or both of the above hypotheses played a part in the observed difference in standing water levels it appears that, if fencing is restricted to a portion of the wet site only, it is unlikely to severely compromise local water availability.

Terrestrial vegetation Overall, the data on the ditch-bank vegetation from the Arable Ditches show trends consistent with the disturbance associated with creation of the features in autumn 2004. Disturbance of the soil during construction of the bunds is the likely cause of the increased extent of bare ground, with decreased plant species richness caused by the loss of extant vegetation. The increase in grass dominance in both Bunded and control ditches is also likely to reflect recovery of the ditch-bank vegetation after initial disturbance during the establishment of the features. Few direct effects of the increased water levels in the bunded ditches were found during the course of the study, with the cover of species typical of damp soils showing no significant increases. However, the vegetation was only sampled in the two summers following the installation of the bunds in autumn 2004, and it is likely that changes in the predominantly perennial vegetation of the ditch-banks will take longer to respond to the changes in water level.

The creation of the Scrapes created open areas and a heterogeneous early-successional vegetation. Differences between the vegetation of the scrapes and that of the control areas were most pronounced during the first year after creation and declined as vegetation developed on the scraped areas. The water held in the surface scrapes was not sufficient to halt the colonisation of the bare soil by perennial species of grass and forb.

Terrestrial invertebrates

Overall, few significant differences were found in the abundance of sward-active invertebrates in response to the bunding of the Arable and Pastoral Ditches. None of the taxa important in the diet of farmland birds were more abundant in the sweep net samples from the Bunded Ditches than their respective controls. However, the abundance and diversity of leaf beetles and weevils was higher on the banks of the Bunded Ditch sections, probably as a result of changes in the vegetation resulting from the construction of the features.

In contrast, the abundance of surface-active invertebrate taxa, sampled using pitfall traps, did differ between treatments and controls. For certain groups, such as adult Coleoptera, this may reflect the effect of increases in the amount of bare ground in the Bunded Ditches, since pitfall traps measure activity-abundance rather than absolute abundance. However, the observed responses of some groups, such as the Diptera, may reflect a direct effect of the increased water levels in the ditches.

For the Surface Scrapes, the abundance of sward- and surface- active taxa can be explained with reference to the resulting changes in the extent of bare ground and the characteristics of the vegetation. Considerable variability in invertebrate numbers was found between blocks and between years, which is likely to have limited the probability of identifying the effects of the treatments. In spite of the paucity of treatment responses, especially amongst sward-active taxa, the abundance of small sward-active taxa had some power in explaining the level of bird foraging in the Pastoral Ditches, in addition to the effect of the treatment itself.

Aquatic invertebratesAlthough it is apparent that ditches in both arable and pastoral landscapes routinely produce large numbers of emergent invertebrates, by introducing bunded sections it is possible to significantly increase the overall abundance, and particularly with regard to producing a food source for farmland birds, the overall biomass of invertebrates (Fig 2c.6).

The overall effect of bunding on both aquatic insect abundance and biomass was positive (Fig 2c.4). However, there was substantial variation both within and between years (Fig 2c.8). While it is possible that the between year differences reflect plot maturation, there is some evidence that insects are responding to factors such as climatic variables (Fig 2c.9). A longer data-run is required to enable the relative importance of such factors to be distinguished.

In both years, the pastoral bunded ditches produced a significantly greater biomass of aquatic insects than the arable bunded ditches, and in 2006, (but not 2005) the pastoral control plots produced more than the arable controls. For most faunal groups the proportions within the samples were broadly similar between the two landscape types, although the biomass of individual taxa was often greater at the pastoral sites. The proportion of different faunal groups was also relatively similar between years. A notable exception occurred in 2006 with a large and sustained emergence of stoneflies within both bunded and control pastoral plots.

Within the physical variables measured during the project, two stand out as predictors of biomass production. The first is percentage of hedge overhang (equivalent to shading), with higher biomass produced in the least shaded plots, this effect was noted at both bunded and control sites (Fig. 2c.13). The second is the percentage of bare mud (as measured beneath the aquatic emergence traps).This variable was positively linked to aquatic biomass production at both bunded and control sites, and was also the factor which best explained the additional increase in aquatic insect biomass produced by bunding ditches. Put simply, in the Loddington area, all ditches (bunded or not) produced more emerging aquatic insects if they were little shaded and had lots of bare wet mud. By bunding the ditches, we systematically created more of the wet bare mud habitat, and produced a greater biomass of insects in these areas.

An implication of this finding is that other ditch management techniques that reduce shade (e.g. hedge management), or increase the proportion of water and wet mud, may also have the potential to increase emergent insect biomass production.

Birds A wide variety of bird species visited both the experimental and control plots of all feature types. There were many similarities in the most regularly recorded species across the ditches and ponds (Appendix 2e Supp. Info.), largely reflecting their relative abundance within the landscape. Robin, Wren and Dunnock were the most common in each case and all are species that breed regularly within the hedgerows around the experimental sites and exploit the ditch\hedgerow habitat year-round. Being primarily insectivorous, they are highly likely to benefit from the increased invertebrate abundance resulting from the bunded ditches and ponds. There are few real surprises within the most regularly recorded species. Whitethroat and Swallow are both summer migrants, and their presence in the top ten indicates a proportionately higher level of use during their period of residency than their position on the list might suggest, as all the rest are resident species and hence potentially observable on all visits.

Water in different situations will attract different species of birds, depending on their habits and preference for different habitat types. The majority of experimental plots considered within the scope of this project for example, were situated in positions adjoining hedgerows, and this is reflected in the composition of the most frequently recorded species (Appendix 2e Supp. Info. Figs 2e.S1 – S4). By creating plots in open aspects, the range of species using them could be very different. This is demonstrated by examining the most frequently observed species visiting the waterlogged areas, which contain several not regularly seen at the ditches and ponds. Also, at Loddington, open buffer strip pools fed by ditches support species such as Reed Bunting and Whitethroat, and most recently Sedge Warbler and Grasshopper Warbler. Further anecdotal evidence of this effect comes from RSPB Grange Farm, where one (albeit more substantial at around three times the size of our largest plot) bunded ditch in an open arable situation regularly attracted Lapwing, Yellow Wagtail, Meadow Pipits, Corvids, Waterfowl and some Waders (Ian Dillon pers. comm.).

For both ditches and ponds, birds were most frequently recorded in the hedge. This is no surprise and likely reflects the many ways in which hedges are exploited by birds in farmland. Records of birds in the hedges encompassed a wide range of activities such as perching, singing, foraging, shelter, and as a route for moving around. Birds recorded in the other micro-habitats were likely to be exploiting them for a specific reason such as foraging, drinking and bathing, meaning that visits to them were less frequent and often for a very limited time period. As many of the hedge records were likely to be unrelated to the habitat manipulations undertaken within this project, the main analyses in Appendix 2e have been restricted to records from the other micro-habitats only. This means our estimates of bird use will be highly conservative, as an unquantifiable proportion of birds using the hedges will be exploiting benefits (such as increased invertebrate emergence) resulting from our habitat manipulations.

Bird visits to individual plots occur relatively infrequently, often at a level of only one or two per watch. However, it is the ratio of visits to experimental and control plots that is important, and the fact that this is positive in virtually all periods and seasons suggests that the experimental plots have the potential to provide substantial benefits to birds. Unfortunately, the overall sparse occurrence of bird records during the timed watches has meant that we have had to group all bird species together for analysis, rather than being able to analyse them individually or by guild.

The overall similarity in bird response over time to the different experimental features strongly suggests that birds are responding to a factor or factors that are common across the range of features. Presence of water is perhaps the most likely explanatory variable, as the differential between bird visits to experimental and control plots fits closely to that of water retention (see Appendix 2a), with the greatest differential in all cases being during the summer periods (e.g. Fig 2e.2). However, the fact that the differential appears to favour the experimental plots over the controls in virtually all periods over the range of features albeit at a reduced level compared to the summer, suggests that the features are meeting some other requirements for attracting birds. Possibilities include increased invertebrate food due to year round retention of water in some plots; increased accessibility to food, possibly through underlying alteration of the surrounding vegetation; and generally increased heterogeneity making them attractive to birds occupying both summer and winter territories. High emergent invertebrate biomass production has been shown to increase the probability of recording birds at both arable and pastoral ditch plots in the spring and summer months (App. 2e) and is a plausible explanatory variable for other periods and feature types, although we cannot confirm this with the data collected during this project.

As there was no significant difference in bird use between 2005 and 2006 for the Arable or Pastoral Bunded Ditches or the Paired Ponds, it would appear that maturation of the plots over time (at least within the scale of this project) is not influential. For the Waterlogged Areas however, time does appear important, with more bird visits in 2006. A primary effect of excluding livestock, as expected, is in vegetative growth (see photograph in App. 2a. Supp. Info). This leads to increased heterogeneity at the local scale. Providing different habitat types in close proximity is likely to lead to biodiversity benefits in terms of species variety and abundance. For birds in particular, having short swards that promote easy foraging close to longer swards for cover and nesting habitat has been shown to be beneficial in other studies.

Main implications of findings (objective 4, recommendations)The biodiversity benefits of incorporating small-scale wet habitats, as created within this project, suggest that they merit consideration for funding within agri-environment schemes. Consideration would need to be given to guidance on feature establishment, including the most appropriate period within which to create the features, bearing in mind both farm management and environmental requirements.

It is evident that there are significant costs associated with the establishment of such features, though their subsequent impact on the farming system is minimal. This would suggest the need for some form of grant funding for the capital works. The bunded ditches, paired ponds and surface scrapes considered here would most likely fit within the current English Environmental Stewardship Higher Level Scheme (HLS), which has provision of payments for capital works. The fencing of existing waterlogged areas, if deemed appropriate, would perhaps come under the Entry Level Scheme (ELS).

However, considering that HLS already contains options for creating wet habitats, and given the small scale of the features investigated here, coupled with their relatively low cost compared to many HLS options, they may be better suited to an ELS type approach. At this level, a broad uptake could lead to substantive rather than site-specific biodiversity gains. Further work is required to compare the potential biodiversity benefits of the features considered here with other Stewardship options to establish a relevant “payment”. Although the current format of

ELS does not allow for capital funding, land managers are expected to use payments to fund environmentally beneficial measures, and other ELS options require capital input, for instance to purchase seed for pollen and nectar mixtures. The establishment costs calculated for objective 3 assume that Contractors are employed to undertake the work, whereas in many cases land managers would have the necessary equipment and skills to undertake the construction themselves, thus reducing costs. The minimal productive land-take required for the features investigated here may also be attractive to many, especially set against the current back-drop of increasing commodity prices. Thus, if a sensible point allocation could be arrived at (dependent upon comparative biodiversity benefits), the positive response received at the various demonstration sites suggests that these options could befit inclusion in the current form of ELS.

Possible future workThe results of the project suggest the need for both a limited extension to the current monitoring, and additional follow-up work that builds on the lessons learnt.

BD1323 started in August 2004 and finished in July 2007. Hence, field data was collected in only two summer field seasons (the baseline year of 2005, plus 2006). Given that the features created in this project will mature and change in value over time, two years is insufficient time to properly appraise their usefulness. It would be particularly useful to monitor the features through a dredging cycle. A straightforward extension of the current project would therefore be valuable.

In the current project, the visit rate of birds to the features was low, compromising our ability to collect data at a resolution sufficient to analyse effects on individual species. This could be addressed in two ways. Firstly, we could consider changes in bird recording methodology. For instance, a regular (daily?) transect past all the features, with recording of birds flushed, may prove a more fruitful way of collecting data. Secondly, the size and structure of the features could be amended to (i) increase the probability of recording a bird, as a simple function of feature size, and (ii) to make the features more attractive, for instance by providing wider edge habitats. Similarly, the project showed there was clear value in having emergent insect data from the bunded ditches. In future monitoring, it would be valuable to collect similar data for the other features. This would be particularly useful for the Paired Ponds, where insect emergence may provide evidence of the impact of water quality improvements in the second of the two linked ponds.

Given the increasing focus on ecosystems approaches and provision of multiple ecosystem services from the same piece of land, it would be appropriate to consider delivery of these attributes against objectives other than just bird targets. These could potentially include retention of diffuse agricultural pollution and mitigation of flood risk.

In addition, the effect of dredging on the biodiversity within and around the ditches and ponds has not been investigated here, and should be considered in further work, as should the cost and practicality of redistribution of dredgings. The fact that substantial sediment build-up occurs demonstrates a potential for these types of feature to reduce sediment loads in water exiting farmland.

Finally, another larger feature that could be considered would be based on a combination of the concepts of ‘horseshoe wetlands’ and ‘multi-ponds’ within a buffer strip. This would provide a series of permanent or temporary pools for wildlife, and would serve to strip sediment and nutrients from both run-off and tile drain water before it entered the ditch system. This would be appropriate for areas that, with decoupling, farmers may want to take out of production in the medium to long term. A 10m headland of a sloping field, drained by a herring-bone system of tile drains, would be taken out of production and a series of horseshoe-shaped excavations dug within the vegetated strip. Field drains would then be opened at the point where the crop meets the set-aside, so that water flows over the surface of the buffer strip vegetation. Water would then flow across the surface, being intercepted by the horseshoe depressions at times of peak flow.

Action resulting from research (eg IP rights, knowledge transfer - objective 5)The bunded ditches and paired pond wet features at Loddington have been used as a focus for discussion with farmers about broad issues including those associated with providing foraging habitats for farmland birds. This has included over 200 farmers participating in the VTS ‘Pathfinders’ project as well as other events organised by the Allerton Trust. Farmers from low-lying areas, where there is a need to maximise drainage, were generally against the installation of the features on their farms. Farmers from mid and upper catchments found both options acceptable if there was to be sufficient payments for capital costs. A number of concerns were raised, including the potential interference with field drains, the need to know about and compensation for frequency of dredging,

and the need for guidance on the disposal of sediment. These could easily be addressed in policy guidance. Finally, feedback from the neighbouring farmers whose land has been used to install some of the wet habitat features recognises that they do not interfere with their agricultural operations. All farmers felt that they, and others, would be willing to create bunded ditches on their own farms if they were funded within a Stewardship scheme.

Three bunded ditches, constructed in an identical manner to those considered here but around three times the size, were constructed at RSPB Hope Farm in Cambrigeshire in 2006. These have also been used to solicit responses from large numbers of visiting farmers, farm advisors and other interested parties. Generally, responses have been extremely positive (D. Moorcroft, C. Bailey pers. comm.). Any concerns raised were usually specific and similar in nature to those expressed above. A few have been sceptical of the biodiversity benefits that could be provided by features at this scale, but most have said that they would consider installing them at their own sites if they were fully funded.

An example bunded ditch was created within the Velcourt\RSPB demonstration stand at Cereals 2007. A large number of visitors passed through the demonstration area over the three show-days, the majority being farmers and land managers. The bunded ditch attracted considerable interest, and feedback from the demonstrators suggested responses were extremely positive.

References to published material9. This section should be used to record links (hypertext links where possible) or references to other

published material generated by, or relating to this project.

Bradbury, R.B., Kirby, W.B., 2006. Farmland birds and resource protection in the UK: cross-cutting solutions for multi-functional farming? Biological Conservation, 530-542.

Bailey, A., Aquilina, R., Bradbury, R., Kirby, W., Lawson, C., Mortimer, S., Stoate, C., Szczur, J., Williams, P., Woodcock, B. 2007. Farm management implications of providing wet habitats to improve biodiversity. In: O’Reilly, S., Keane, M., Enright, P. (Eds.), Proceedings of the 16th International Farm Management Association Congress - A Vibrant Rural Economy – The Challenge For Balance. Snap Printing, Cork, Ireland. pp. 665-673

Aquilina, R., Williams, P., Nicolet, P., Stoate, C., Bradbury, R. 2007. Effect of wetting up ditches on emerging insect numbers. Aspects of Applied Biology 81, 261-262.