Agroforestry Systems for Ammonia Abatement ... - Defra Science

24 downloads 90 Views 2MB Size Report
W.J. Bealey, A.J. Dore, U.D. Dragosits. 6.2. Assessment of the abatement potential of farm woodlands at the UK scale. A.J. Dore, WJ Bealey, U. Dragosits, and.
Agroforestry Systems for Ammonia Abatement AC0201 Final Report

W.J. Bealey, C.F. Braban, M.R. Theobald, D. Famulari, Y.S. Tang, A. Wheat, E. Grigorova, S.R. Leeson, M.M. Twigg, U. Dragosits, A.J. Dore, M.A. Sutton, E. Nemitz, B. Loubet, A. Robertson, A.D. Quinn, A. Williams, D.L. Sandars, G. Valatin, M. Perks, D. Watterson

Centre for Ecology and Hydrology, NERC, UK INRA, France University of Birmingham, UK Cranfield University, UK Forest Research, UK

1

Client: Defra and NERC CEH Client Project number: AC0201

CEH Project: SAMBA

Client Project Officer: Dr Daniel McGonigle CEH Project Officer: Bill Bealey Main authors: W.J. Bealey, C.F. Braban, M.R. Theobald, D. Famulari, Y.S. Tang, A. Wheat, E. Grigorova, S.R. Leeson, M.M. Twigg, U. Dragosits, A.J. Dore, M.A. Sutton, E. Nemitz, B. Loubet, A. Robertson, A.D. Quinn, A. Williams, D.L. Sandars, G. Valatin, M. Perks, D. Watterson

This report is a confidential document prepared under contract between Defra and the Natural Environment Research Council (NERC) for the UK Department for Environment, Food and Rural Affairs (DEFRA) and NERC. The work undertaken which this report summarises was funded by both organizations. It should not be distributed or quoted without the permission of both the Centre for Ecology and Hydrology (CEH) and Defra.

2

Full Author List with Affiliations

W.J. Bealey, C.F. Braban, M.R. Theobald, D. Famulari, Y.S. Tang, A. Wheat, E. Grigorova, S.R. Leeson, M.M. Twigg, U. Dragosits, A.J. Dore, M.A. Sutton, E. Nemitz, Centre for Ecology and Hydrology, NERC, UK B. Loubet, INRA, France A. Robertson, A.D. Quinn, University of Birmingham, UK A. Williams, D.L. Sandars Cranfield University, UK G. Valatin, M. Perks, D. Watterson, Forest Research, UK

Section Authorship attributions

3. Mechanistic modelling 3.1 Modelling of turbulence across complex shelter belts 3.2 Modelling of turbulence across complex shelter belts - Simulation of the wind-tunnel experiment 3.3 Modelling of shelter-belt and understory scenarios 3.4

Modelling NH3 volatilisation from sheltered slurry stores

B. Loubet B. Loubet, D. Famulari, C. F. Braban B. Loubet, W. J. Bealey, C. F. Braban, D. Famulari, and M.A. Sutton A. Williams

4. Measurements 4.1

Quantifying NH3 depletion by a woodland: a wind tunnel experiment

4.2

Experimental quantification of NH3 capture by an overhead tree canopy.

4.3

Field Case Studies of NH3 Concentrations downwind of Poultry Houses

5. Agroforestry ammonia abatement cost analyses Profitability analysis of trees and woody shelter belts on livestock farms 5.1 for NH3 abatement and carbon sequestration 5.2 Cost-effectiveness of Agroforestry options for NH3 Abatement as Climate Change Mitigation measure 6. National modelling 6.1 On-farm emission factor reduction 6.2

Assessment of the abatement potential of farm woodlands at the UK scale

3

D. Famulari, M.M. Twigg, A. Robertson, E. Grigorova, C. F. Braban, M. R. Theobald, M. A. Sutton, E. Nemitz D. Famulari, C. F. Braban, A. Wheat, M. Coyle, C. Helfter, E. Nemitz, M. A. Sutton Y.S. Tang, C.F. Braban, S. R. Leeson, W.J. Bealey, D. Famulari, M.A. Sutton D. Sandars G. Valatin

W.J. Bealey, A.J. Dore, U.D. Dragosits A.J. Dore, WJ Bealey, U. Dragosits, and M.A. Sutton

Table of contents

1

Project Overview ........................................................................................................................................................................ 5

2

Background ................................................................................................................................................................................ 6 2.1

3

References ......................................................................................................................................................................... 6

Mechanistic modelling ............................................................................................................................................................... 7 3.1

Modelling of turbulence across complex shelter belts ..................................................................................................... 7

3.2

Modelling of turbulence across complex shelter belts - Simulation of the wind-tunnel experiment ............................... 8

3.3

Modelling of shelter-belt and understory scenarios ......................................................................................................... 9

3.4 Modelling NH3 volatilisation from sheltered slurry stores: Windbreaks and slurry lagoons effects of wind and temperature, the Thermal-Volatilisation Effect (TVE) Model ...................................................................................................... 11 4

5

6

Measurements ......................................................................................................................................................................... 13 4.1

Quantifying NH3 depletion by a woodland: a wind tunnel experiment.......................................................................... 13

4.2

Experimental quantification of NH3 capture by an overhead tree canopy..................................................................... 14

4.3

Field Case Studies of NH3 Concentrations downwind of Poultry Houses ....................................................................... 15

Agroforestry ammonia abatement cost analyses .................................................................................................................... 17 5.1

Profitability analysis of trees and woody shelter belts on livestock farms for NH3 abatement and carbon sequestration 17

5.2

Cost-effectiveness of Agroforestry options for NH3 Abatement as Climate Change Mitigation measure ..................... 19

National modelling ................................................................................................................................................................... 21 6.1

On-farm emission factor reduction ................................................................................................................................. 21

6.2

Assessment of the abatement potential of farm woodlands at the UK scale ................................................................. 22

7

Summary .................................................................................................................................................................................. 24

8

Conclusions and Future directions ........................................................................................................................................... 25

Appendix 1: User Group meeting report .......................................................................................................................................... 26

4

Agroforestry Systems for Ammonia Abatement Abstract

Ammonia (NH3) emissions to the atmosphere increased significantly during the 20th century, largely due to the intensification of agricultural production. Ammonia is a soluble and reactive gas that is emitted by volatilization from various agricultural nitrogen forms including urea, uric acid and mineral fertilizers. Emissions are dependent on various meteorological inputs like temperature and wind speed, and are higher in warmer drying conditions, with smaller emissions occurring under cooler wetter conditions. Impacts of excess nitrogen can include eutrophication and acidification effects on semi-natural ecosystems that can lead to species composition changes. Agroforestry Ammonia Abatement (AAA) is a practical concept which uses both the dispersive effect of a barrier and the uptake of NH3 into the tree canopy to mitigate NH3 emissions. This work built upon the research carried out in Defra project AC0201, bringing together measurements, modelling and agroeconomic analyses to build an assessment of the potential benefits and drawbacks of applying AAA strategies both on a local and national scale. The project objectives were to assess the efficacy of farm woodland features for the recapture of agricultural NH3 emissions. The potential of farm woodlands for NH3 mitigation at a local and the UK scale were assessed. The combined modelling and measurement results from this project show that AAA carefully planned and implemented can lead to a significant decrease in NH3 concentrations downwind from sources and a moderate, up to 20% net decrease in emissions to the atmosphere. AAA systems could be used as a protective measure of downwind sensitive ecosystems where local concentration reductions can be higher. Use of existing woodland plantations and panting new forestry can both be used to mitigate emissions, though scrubbing of NH3 at source and reuse would also be a solution. UK scale modelling shows that targeted application of tree planting around agricultural installations would have a modest effect by modifying ‘on-farm’ emission factors, however when the approach is targeted in regions hot-spot emissions, significant effects on NH3 and N-deposition can be achieved. In many agricultural businesses there are no current economic advantages for converting valuable arable land to woodland without specific opportunity benefits (e.g. woodland egg price margins due to animal welfare considerations, carbon or nitrogen credits). However as the woodland egg example shows, when other considerations become relevant, AAA can be a useful approach. It is noted that mitigating ammonia with trees only addresses one nitrogen flow in the farming systems and the net effect on both the reactive and GHG N budgets over the landscape scale should be considered.

1

Project Overview

The aim of the Agroforestry Systems for Ammonia Abatement project was the following: (1) To assess the efficacy of farm woodland features for the recapture of agricultural ammonia (NH3) emissions, through a combination of wind tunnel studies and numerical modelling, and to optimise their design. (2) To quantify NH3 recapture in silvopastoral systems through targeted field measurements and emission, dispersion and recapture modelling. (3) To quantify the effect of upwind shelter for the reduction of NH3 emissions from farm sources through mechanistic emission modelling. (4) To provide accessible guidance and IT based design tools that can be used by farm managers to optimise farm woodland design and estimate effects in form of an updated “Farm Woodland Decision Support and Design Guide” and recapture parameterisations for a simple numerical dispersion screening tool. (5) To demonstrate the practical feasibility of implementing farm woodlands as NH3 abatement measures through case studies. (6) To quantify the potential of farm woodlands for NH3 mitigation at the UK scale.

5

2

Background

Ammonia (NH3) emissions to the atmosphere have increased during the 20th century, largely due to the intensification of agricultural production, which accounts for over 80% of NH3 emissions in the UK (Defra, 2002; Misselbrook et al. 2010). NH3 source activities in agriculture fall into four categories of management – emissions from housing, grazing, and storage and spreading of manure (Misselbrook et al., 2010). NH3 emissions vary greatly at the local scale and effects of dry deposited NH3 occur primarily close to the sources (Sutton et al., 2001b; Dragosits et al., 2002). NH3 is a reactive gas that is emitted by volatilization from various agricultural nitrogen forms including urea, uric acid and mineral fertilizers. Emissions occur from both animal housing and on fields. The extent of NH3 dispersion into the atmosphere is dependent on NH3 source type, climate (temperature, humidity, rainfall) and local meteorology (wind direction, topography). In general NH3 emissions are higher in warmer dry conditions, and lower under cooler wetter conditions. NH3 has a relatively short lifetime in the atmosphere as it is water soluble and readily deposits to surfaces including vegetation. Nitrogen sensitive ecosystems close to NH3 sources are at a high risk of negative impacts which can include eutrophication and acidification. These effects on semi-natural ecosystems can lead to species composition changes (Bobbink et al, 2010; Krupa, 2003; Pitcairn, 1998; Sheppard, 2008; Van den Berg et al, 2008; Wiedermann et al, 2009) and other deleterious effects. Species adapted to low N availability are at a greater risk; for example, many slower-growing lower plants, notably lichens and bryophytes. (Pearce, 2002; Bobbink et al, 1998). Since the inclusion of intensive farming to the Integrated Pollution, Prevention and Control (IPPC) process in 2005, pig and poultry installations above a certain size now have to be assessed and permitted to release NH3 into the atmosphere. For the pig industry, permits are required for farms with more than 2,500 pigs, while the number for poultry units is 45,000 birds. Similarly, other directives like the Habitats Directive ((Council Directive, 92/43/EEC)) provide a high level of protection to designated sites, and important BAP (Biodiversity Action Plan) species, by taking a precautionary approach to controlling polluting activities. Agricultural industries have to show that their emissions are not likely pose a significant threat to the integrity of the protected site. It is for this reason that the industry is increasingly interested in alternative abatement techniques that reduce effects on nearby protected sites. Asman et al (1998) showed that adverse effects on sensitive ecosystems caused by high N deposition can be reduced by lowering the emissions and, to a limited extent, also by removing sources in close proximity to the ecosystem to be protected. Trees are effective scavengers of both gaseous and particulate pollutants from the atmosphere (Beckett 2000; Nowak, 2000) with dry deposition rates to forest exceeding those to grassland by typically a factor of 3–20 (Gallagher et al, 2002; Fowler et al, 2004). This implies that the conversion of grassland and arable land to trees can be used to increase the removal of NH3 from the atmosphere near the source, thereby reducing the potential impacts on nearby sensitive ecosystems. Agroforestry NH3 Abatement (AAA) methods of mitigating the release and deposition of NH3 into the landscape from such sources have been studied. The adoption of AAA needs to be supported by quantitative theoretical and experimental evidence to underpin any policy development. Choice of tree species, canopy structure and planting area, need to be considered to maximize direct NH3 capture. In addition consideration of other effects such as screening, biodiversity, nature reserve extension and changing of pollutant pathways need to be assessed. This work has attempted to quantify the practical effectiveness of AAA experimentally using several methods: (i) modelling NH3/tracer emission, dispersion and recapture (ii) NH3/tracer release through trees in a wind tunnel (iii) NH3 release under a real woodland canopy and (iv) transect case studies of NH3 concentrations on farms with woodland downwind of poultry housing. The agro-economic and practical implications of AAA has been assessed and the potential of AAA at the UK scale to a first approximation derived. Overall we demonstrate potential and limits for AAA as an NH3 abatement measure and discuss the further constraints which would be necessary to develop it into an accredited NH3 emission reduction policy.

2.1

References

Asman et al. (1998) Ammonia: emission, atmospheric transport and deposition. New Phytologist, 139 (1998), pp. 27–48; Bobbink et al., 1998. The effects of airborne pollutants on species diversity in natural and semi-natural European vegetation. Journal of Ecology 86:717-738.; Bobbink et al. (2010) Global Assessment of Nitrogen Deposition Effects on Terrestrial Plant Diversity: a synthesis. Ecological Applications, 20, 30-59.; Defra (2002) Ammonia in the UK. Department for Environment and Rural Affairs, London 89 pp.; Dragosits et al. 2002. Ammonia emission, deposition and impact assessment at the field scale: a case study of sub-grid spatial variability. Environmental Pollution 117, 147 ; Fowler et al. 2004. Measuring Aerosol and Heavy Metal Deposition on Urban Woodland and Grass Using Inventories of 210Pb and Metal Concentrations in Soil. Water, Air and Soil Pollution: Focus 4 (2-3), 483, June 2004.; Gallagher et al. 2002. Measurements and parameterizations of small aerosol deposition velocities to grassland, arable crops, and forest: Influence of surface roughness length on deposition. Journal of Geophysical Research, 107.; Misselbrook et al. (2010). Inventory of Ammonia Emissions from UK Agriculture 2009, Inventory Submission Report, December October 200410, DEFRA contract AM0127AC0112.; Nowak, D.J., 2000. Impact of urban forest management on air pollution and greenhouse gases. In: Proceedings of the Society of American Foresters 1999 national convention; 1999 September 11–15; Portland, OR. SAF Publ. 00-1. Bethesda, MD: Society of American Foresters: pp. 143; Pearce and van der Wal 2002. Effects of nitrogen deposition on growth and survival of montane Racomitrium lanuginosum heath. Biological Conservation 104:83; Pitcairn et al. (1998): The relationship between nitrogen deposition, species composition and foliar nitrogen concentrations in woodland flora in the vicinity of livestock farms. Env. Pol. 102, 41 ; Sheppard et al. 2008. – Stress responses of Calluna vulgaris to reduced and oxidised N applied under "real world conditions". Environmental Pollution 154, 404; Sutton et al. A spatial analysis of atmospheric ammonia and ammonium in the UK. The Scientific World 1, 275e286.; Van den Berg et al. (2008) Reduced nitrogen has a greater effect than oxidised nitrogen on dry heathland vegetation. Environmental Pollution, 154, 359; Wiedermann et al. (2009) Ecophysiological adjustment of two Sphagnum species in response to anthropogenic nitrogen deposition. New Phytologist, 181, 208

6

3

Mechanistic modelling

3.1

Modelling of turbulence across complex shelter belts

3.1.1 Objective and methodology In order to assess the dispersion of ammonia through an area covered with trees, the air flow and air “turbulence” through the trees and the amount of NH3 taken up by the vegetation needs to be understood. This work package evaluated the effect of tree canopy structure on the air turbulence and on the deposition processes. This was achieved by coupling two pre-existing models: MODDAS and THETIS. THETIS is an Eulerian k-ε turbulence model designed for within-canopy turbulence as well as transfer of air from the canopy to the outside - the “planetary boundary layer”. MODDAS is a Lagrangian stochastic (LS) model for gas dispersion coupled with a multi-layer exchange model which can characterise the interaction of ammonia with vegetation (in this case tree leaves and bark) including a stomatal compensation point which is the point at which the plant stops taking ammonia up. THETIS outputs air turbulence which is used as input for MODDAS. Specifically: the horizontal and vertical components of the wind velocity (u and w), the standard deviation of the wind vectors, (σu and σw), the cross-correlation ����������) and the dissipation rate of the turbulent kinetic energy (ε). THETIS outputs the first three between the wind vectors, ( u’w’ parameters but does not give directly σu and σw. but outputs k which is defined in E1. To couple the two models an objective procedure to partition k into its horizontal and vertical components was developed based on the Lagrangian and Eulerian description of scalar dispersion statistics within the atmospheric boundary layer (see e.g. Csanady, 1980). 1

𝑘 = (𝜎𝑤 + 𝜎𝑢 + 𝜎𝑣 )2

E1

2

Hypothesis 1. Following Taylor (1921), the vertical diffusivities in the Lagrangian approach must be equal when t is much larger than the Lagrangian time scale TL: 𝐾𝑧𝐿 = 𝐾𝑧𝐸

when

𝑡

𝑇𝐿

L (Kz )

E

and the Eulerian approach (Kz )

→∞

Hypothesis 2. The ratio σu/σv is arbitrarily set to a constant value αuv in the whole domain. This hypothesis is true in neutral stratification and for a flat terrain. Since THETIS is only set for neutral conditions, it justifies (in a first approach) this assumption, which reads: σu / σv = αuv When the conditions of the two hypotheses are met, MODDAS and THETIS can be coupled with only two parameters (αw and αuv), of which αw can be constrained to 0.37, while αuv is empirically set to 1.25. The coupling was tested using a comparison of the MODDAS-Particle version of the MODDAS output and THETIS for a maize pollen experiment (see Jarosz et al., 2005). This showed the differences between the LS model and the Eulerian model for predicting dispersal. 3.1.2 Results Initial comparisons suggested that MODDAS predicted larger concentrations near the source. To analyse the actual sensitivity of MODDAS and to compare it to the Eulerian model, a random walk version of the MODAAS (RW) was written (it should behave exactly as Eulerian models, see Rodean, 1996). LS and RW were compared. The RW model diffuses more rapidly than the LS, but the two models match relatively well for x > xsource + 4 m (see Figure 1), i.e. the model is robust > 4 m downwind of x=0. Further sensitivity analysis showed that the RW model is much more sensitive to a change in ε than the LS model: so the embedding of an RW module into MODDAS provides a satisfactory sensitivity . Figure 1: Downwind of a 100 µg m-3 s-1 source located in the volume 0.2 x 0.2 m2 at x = 5 m and z = 2 m. Upper Panel: LS model Lower panel: RW model

3.1.3 Conclusions The coupling of the THETIS and MODDAS models was successful. The RW model is sensitive to changes in ε. The LS model is almost not sensitive to ε whereas it should be. The LS and the RW models give similar concentrations for t / TLi > 4 when ε = εAqui. The coupled models were useful in the subsequent parts of the SAMBA project. 3.1.4

References

Csanady G. T., 1980 : Geophysics and Strophysics Monographs, D. Reidel Publishing Company, Dordrecht, Holland,248; Jarosz et al. 2003, Atmos. Environ., 38:5555–5566; Rodean, H.C., 1996, Meterological monographs. American Meteorological Society, Boston, US, pp. 83

7

Modelling of turbulence across complex shelter belts - Simulation of the wind-tunnel experiment

3.2

B. Loubet, D. Famulari, C. F. Braban 3.2.1 Objective and methodology The MODDAS-AQULION coupled model (Section 4.1) was used to simulate the flows and concentration fields in the practical wind-tunnel experiments (Section 4.3). This was to validate the MODDAS-AQULION coupled model and understand the results from Section 3.2. The wind tunnel set up is shown in Figure 2 and is described more fully in 4.3 below. Two source heights were used at 0.1 and 0.5 m. 4 sonic anemometers were used to sample the space around and within the tree spaces. The wind tunnel 2 -2 flow was modelled with THETIS with Leaf Area Density (LAD) at heights (z) (Figure 3). The LAI was equivalent to 5m m . 2

z (m)

LAI = 5 m2 m-2

1.5

z ( m)

contour tree lines gas sampling source sonics

5

1

0.5

0 0

5

x (m)

10

15

20

0 0.00

0.20

0.40

0.60

LAD (m2 m-3)

Figure 2: Wind tunnel schematic

Figure 3: LAD (z)

3.2.2 Results Overall THETIS gave good correlation of the flow within the trees but gave a much larger wind speed vertical component above the canopy. THETIS reproduced the measured profiles relatively well. The modelled stress is much larger than measured and is larger at the canopy top indicating that the shear stress production by the mean flow is probably too large. This is probably because the simulation does not reproduce the roof, which forces the flow to be horizontal. The simulated inert tracer concentration in the wind tunnel is shown in Figure 4 (low flow conditions, source height 0.5 m) and shows that the tracer goes through the trees and disperses. Figure 5 shows NH3 release, dispersion and deposition onto the canopy. Figure 6 shows the NH3 deposition map indicating that the NH3 is depleted in the canopy due to dry deposition. The depletion is of the order of 10 to 15% for this case where relative humidity = 100%. The deposition was maximum at the canopy edge and source height, giving a wing-shape profile, with decreasing deposition rate as the concentration decreases, but also as the boundary layer resistance decreased.

Figure 4: Concentration of an inert tracer modelled by MODDAS-THETIS at a source strength of 100 µg m-1s-1. The top panel shows a contour plot of the concentration, the middle panel shows the vertical concentration profiles and the bottom panel shows the horizon.

Figure 5: Concentration of NH3 modelled by MODDAS-THETIS for the low flow rate of the wind tunnel for a source strength of 100 µg m-1s-1. (see Fig. 4 for panel details)

8

Figure 6: Deposition patterns in “housing” runs with and without backstop (LHS) and understorey” LAI = 1 (upper panel) and LAI = 6 (lower panel) (normalized by the source strength)

3.2.3 Conclusions The turbulence structure in the wind tunnel was evaluated by comparing against measurements. The modelled flow shows similar patterns to the measured ones but with larger magnitude turbulence. This overestimation is explained by the roof in the wind tunnel which is not present in the model, hence the modelled flow is less constrained in the vertical and shear stress is therefore larger. Subsequent experiments with a roof in the model show that the kinetic energy, vertical wind velocity and the shear stress magnitudes improve the results. The inert tracer experiments showed the canopy increasing the mixing efficiently and channelling the tracer below the canopy top. The modelled maximum NH3 deposition represented about 25% of emission, and depleted concentrations by 10 to 15%.

3.3

Modelling of shelter-belt and understory scenarios

3.3.1 Objectives and Methodology MODDAS-THETIS developed in Section 3.1 was used to look at theoretical but realistic scenarios in order to assess what shelter belt structures lead to the greatest NH3 recapture. The model was setup around a forest schema as shown in Figure 7, taking into account height of canopy (z), leaf area density (LAD), width of canopy (X), source strength (Q) and source width (Xs). LAD3 (z) LAD2 (z)

LAD0 (z) hc0

LADs (z) LAD1 (z) hcs Qs hc1 hs

hc3

hc2

xs xc0

xc1

xcs

xc2

xc3

Figure 7. General Scheme of the woodland and source geometry that will be tested in the scenarios.

3 NH3 sources scenarios were modelled:   

-1

a “housing” scenario: source (300 kg NH3-N yr ) emitting at h= 2-2.5 m; w =4-5m (Figure 8 LHS) -1 a “lagoon” scenario, source (393 kg NH3-N yr ) emitting at h= 0.1-0.2 m (Figure 9 LHS) -1 an “under-storey” scenario, source (625 kg NH3-N yr ) emitting at 0.1-0.2 m under the canopy, (Figure 8 RHS)

The scenarios were combined with a range of meteorological situations, noting that MODDAS-THETIS can only run for neutral -1 cases. The wind speed at 50 m was set to 5 m s . Five canopy characteristics were modelled and the deposition parameters used were such as to reproduce a maximum deposition.

Figure 8: LHS Housing and lagoon scenario; RHS Under storey source.

9

3.3.1.1 Results All the model scenarios and the resulting NH3 depositions are summarized in Table 1. The NH3 concentration fields in three “housing” runs are shown in Table 1 top. NH3 concentrations were maximum at the source, decreasing with distance away. For the symmetrical schemes, the canopies increase the vertical dispersion and the backward dispersion due to the increased turbulent kinetic energy (Table 1 top row). The asymmetrical scheme show downwind decreases in concentration inside the -1 canopy, but subsequent increases after the canopy due to a zone of calm air (wind speed below 1 m s ). The scheme with a wider main canopy and wider backstop leads to decreases in the concentration in the canopy. In the case of the lagoon, the same behavior is observed for the concentration with or without an upwind main canopy. For the “understorey” runs, the NH3 concentration can vary significantly depending on the canopy density (LAD and LAI). With LAI=1 the maximum concentration is similar to the maximum concentration in the “housing” scenario, but for LAI=6 the concentration reaches more than 4000 µg -3 NH3 m . This can be explained by the very low level of turbulence and wind speed in the canopy in the dense scenario. The deposition patterns in the housing runs follow the NH3 concentration patterns but are also affected by the LAD. For the housing scenario, a maximum deposition of 27% was seen with a 2-part treebelt of which the first part is “lollipop” type profile followed by a “Christmas tree” profile. A “Christmas tree” profile with 15-20% of the bottom of the trunk free of leaves is most beneficial for NH3 deposition. Having a wider backstop increases deposition from 16 to 25%. Table 1 Schematic housing (top) and under canopy (bottom) with tree structure scenarios and model results summary; % deposited before Canopy main height % (m) deposited canopy LAI

main Backcanopy stop width Height Main Canopy width (m) (m) LAD profile LAI (m)

design

% deposited in main canopy

% deposited in backstop

1 0.8

LAD-0

z/h

0.6 0.4 0.2

100

3

0

10

0.000

0.020

0.040

0.060

LAD / LAI

0

-

-

15%

0%

15%

0%

10

6

10

20%

0%

17%

4%

25

6

10

27%

0%

19%

8%

50

6

10

49%

0%

45%

4%

50

6

10

60%

0%

50%

9%

1 0.8

LAD-0

z/h

0.6 0.4 0.2

100

3

0

10

0.000

0.020

0.040

0.060

LAD / LAI

1 0.8

z/h

0.6 LAD-4

0.4 0.2 0

100

3

10

0

0.05

0.1

0.15

LAD / LAI

1 0.8

z/h

0.6 0.4 0.2

LAD-1

0

100

6

10

0.000

0.050

0.100

LAD / LAI

1 0.8

LAD-0

z/h

0.6 0.4 0.2 0

100

design

6

10

0.000

0.020

0.040

0.060

LAD / LAI

% deposited before main Back-stop Canopy main canopy width height % Height Main Canopy width LAI (m) (m) (m) deposited canopy LAD profile LAI

% deposited in main canopy

% deposited in backstop

1

z/h

0.8

LAD-0

0.6 0.4 0.2

30

6

0

10

0.000

0.020

0.040

0.060

LAD / LAI

0

-

-

16%

2%

14%

0%

0

-

-

17%

0%

17%

0%

25

6

10

16%

0%

5%

11%

50

6

10

25%

0%

5%

20%

50

6

10

27%

0%

15%

12%

1

z/h

0.8

LAD-0

0.6 0.4 0.2 0

30

6

0.000

10

0.020

0.040

0.060

LAD / LAI

1

z/h

0.8 0.6 0.4 0.2

25

3

LAD-2

0

10

0

0.05

0.1

0.15

LAD / LAI

1

z/h

0.8 0.6 0.4 0.2

LAD-2

0

25

3

0

10

0.05

0.1

0.15

LAD / LAI

1 LAD-10

z/h

0.8 0.6 0.4 0.2

50

3

10

0 0

0.05

0.1

LAD / LAI

0.15

The increase of the canopy height from 10 to 30 m with a constant LAI leads to a decrease in the deposition rates due to an increase in the turbulent mixing at the source (asymmetrical runs). Figure 9 LHS compares no back-stop (top panel) to a 50m back-stop (lower panel). In the understorey runs, the deposition increased from 15% to 37% for a backstop increasing from 0 to 50 m (LAI main canopy = 3, LAD main canopy = 99). The deposition pattern in the “understorey” runs varied a lot depending on the concentration levels and the LAI and LAD patterns. Figure 10 RHS illustrate this: a quite open canopy (LAI= 1), vs a dense main canopy (LAI=6). The deposition is only significant in the backstop for the less dense canopy (22% recapture) while is very large throughout the main canopy in the dense canopy scheme (60% recapture).

10

Figure 9 LHS Concentration field in “housing” (shelter belt black outline); RHS: Concentration field in “understorey”; LAI = 1 and 6 (upper and lower panel resp.)

Figure 10 Deposition patterns in “housing” runs with and without backstop (LHS) and RHS “understorey” LAI = 1 (upper panel) and LAI = 6 (lower panel) (normalized by the source strength)

3.3.2 Conclusions Maximum deposition rates were 28%, 19% & 60% in the housing, lagoon & understorey scenarios respectively. Deposition rate increased roughly proportionally with LAI if the LAI and the LAD are identical in the main and the backstop canopies. The increasing main canopy width does not proportionally increase the deposition rates. The canopy with a dense and homogeneous LAD favors deposition, while canopy with a dense crown and an open trunk space is less effective at recapture. When the source is close to the ground (lagoon and understorey setups) a dense canopy near the ground should be favored rather than a canopy with a dense crown. Taller canopies with identical LAI lead to smaller deposition rates. Under real conditions there is a potential for saturation of the surfaces exposed high loads of NH3, so it is stressed here that the reported deposition is a maximum. However the results are comparable between the scenarios and this is the first work to quantify which type of tree architectures could bring the greatest NH3 abatement benefits with specific types of agricultural activities.

3.4

Modelling NH3 volatilisation from sheltered slurry stores: Windbreaks and slurry lagoons effects of wind and temperature, the Thermal-Volatilisation Effect (TVE) Model

3.4.1 Objective and methodology In this work the Thermal-Volatilisation Effect (TVE) Model for understanding the effect of wind breaks on NH3 emissions from slurry lagoons is developed. There were two main approaches taken to develop the TVE model: 1) thermal model using a more mechanistic emissions formulation and 2) volatilisation modelling of temperature changes in lagoons that could drive part of the mechanistic NH3 emissions. The TVE model used was derived from that of Olesen and Sommer (1993). It was developed further by Webb et al. (2005) and assessed by Theobald et al. (2005). Once the thermal and volatilisation model developments were combined into the TVE model, the NH3 emissions were run over a year using the wind speed reduction factors of 10% to 100% coupled with the temperatures that were simulated (above) as input values. The main parameter values were (pH: 7.5; initial -3 o o TAN: 1 kg.m , depth 1m; initial lagoon T (August start): 17 C; initial lagoon T (November start): 8 C). 3.4.2 Results The TVE model was validated against two sets of experimental data for different lagoons in North Bedfordshire for which there were consistent data set for all required variables (Lagoon 1:~ 1000h 1 x 18 x 24m; Lagoon 2: 2780 h, 2 x15 x35, Scotford and Williams, 2001). The lagoon temperatures were measured were about 0.3 m below the surface. The TVE model generally simulated measure temperatures well. The overall effects of sheltering were to increase the simulated summer temperature -2 -1 and decrease the simulated winter temperature. The annual average emission rate was 1.3 g NH3-N m d , which is well within reported rates (0.4 to 5.7) and close to mean of 1.7 applied in the UK inventory by Misselbrook et al (2006). 11

The effects of wind breaks on volatilisation rates were calculated by running the model with the simulated wind speeds and lagoon temperatures. The crucial observation is that the use of wind breaks still reduces the RAER when both wind speed and temperature effects are integrated into one model. The minimum RAER was 39%, (61% reduction in emissions).Note that this would not apply to the full lagoon surface, hence is an upper estimate. Table 2 Effect of wind break in relative NH3 emission rate (RAER) on annual basis (n.b. 100% is normal, without wind break) Wind speed reduction factor

RAER compared with normal

Fitted curve

10%

39%

39%

25%

62%

61%

50%

82%

83%

75%

93%

94%

100% 100% 99% The fitted curve for RAER against wind speed reduction factor (w) took the following form with parameter values in ER = A + B*exp (k w)

The modelled changes in wind were integrated with the RAER effects. Wind break heights of up to 10 m were analysed with an average lagoon width of 50 m (Williams, 2002). Note RAER estimates were applied for up to 120 m. Effectiveness increased with height and decreased with distance across the lagoon (e.g. Figure 11 LHS). The effect for a 50 m wide lagoon ranged from a reduction in NH3 emission rates in the range from 7% to 26%. The overall NH3 reduction effect is somewhat larger than was previously estimated (AMBER).The error is quite high, with an average coefficient of variation (CoV) of 44%, indicative of range for an unknown windbreak. TVE model errors were examined using Monte Carlo simulations leading to an error estimation for the models of CoV = 7.1%. The errors estimated do not include those from pH or surface crusting. 30%

40% v loose loose

30%

medium dense V dense

20%

phi = 50% phi = 34% phi = 20%

10%

phi =0 %

Mean reduction in ammonia emissions, %

Reduction in ammonia emissions, %

50%

Wind break height

20%

1m 2m 5m 7.5 m

10%

10 m

0%

0% 0

30

60

90

120

0

20

40

Lagoon width, m

60

80

100

120

Lagoon width, m

Figure 11: LHS:Effect of a 7.5 m high wind break on relative NH3 emission reduction with integrated thermal and emissions model; RHS: Averaged effects of wind breaks of all porosities on relative NH3 emission rates across slurry lagoon

3.4.3 Discussion The revised modelling of NH3 volatilisation from sheltered slurry stores presented here gives a nuanced interpretation of the effects of windbreaks. Temperature clearly affects the seasonal emission rate and in turn is affected by wind speed. The large wind speed effect previously observed was larger again in the TVE model due to a much more mechanistic understanding of the relationships between wind speed temperature and the mass transfer of NH3. The results are plausible and intuitive but they have not been independently tested against emissions measurements and that instantaneous NH3 emission rate is the surface pH which is not easy to predict. The range of benefits is reasonably large, reflecting the potential range of heights and porosities that a windbreak may have. The effectiveness will tend to increase over time as trees grow and it is essential not to assume the largest benefits immediately. The results in Table 3 indicate that the benefits of a windbreak roughly double as it grows from 1 to 5m height and then doubles again at 10m. One other benefit of windbreaks is that the source of emissions of not only NH3, but also malodours is masked. This simple physical screening is likely to reduce the public perception of a source of odour, although we make no claims about the potential for mitigating malodours per se. Effectiveness is highest for short lagoons, higher trees and higher densities of trees, however it is important not to assume the best case, but the average and to consider the other impacts and implications the windbreak may bring to the farm environment. Table 3 Summary of effect of windbreaks of different heights and densities on the reduction in NH3 emissions from a 50 m wide lagoon Height of wind break 1 2 5 7.5 10

Nominal density of wind break (average of three simulations) Low Medium 5% 8% 7% 12% 10% 18% 13% 24% 12% 23%

12

High 8% 14% 21% 29% 27%

3.4.4

References

Heisler G.M.; Dewalle D.R. (1986) Effects of windbreak structure on wind flow. Agriculture, Ecosystems and Environment 22/34, 41-69; Misselbrook, T.H. et al.. (2006) Inventory of NH3 Emissions from UK Agriculture 2004 DEFRA Contract AM0127 Inventory Submission Report March 2006; Naegeli (1946) Weitere Untersuchungen fiber die Windverh/iltnisse im Bereich von Windschutzanlagen. (Further investigations of wind conditions in the range of shelterbelts.) Mitt. Schweiz. Anst. Forstl. Versuchswesen, 24: 660-737. Cited by Heisler and Dewalle (1986); Olesen & Sommer (1993). Modeling Effects of Wind-Speed and Surface Cover on NH3 Volatilization from Stored Pig Slurry, Atm Env, 27, 2567-2574. ; Scotford, I.M.; Williams, A. G. (2001) Practicalities, Costs and Effectiveness of a Floating Plastic Cover to reduce NH3 Emissions from a Pig Slurry Lagoon. Journal of Agricultural Engineering Research 80 (3), 273-281 doi:10.1006/jaer.2001.074; Theobald M.R. , et al., WA0179, Final report to Defra, available at http://randd.defra.gov.uk, 2003; Theobald, M.R et al. . (2005) An assessment of how process modelling can be used to estimate agricultural NH3 emissions and the efficacy of abatement techniques. Final report to Defra on project AM0130.; Webb, et al.. (2005) The influence of store type and climatic variation on nitrogen dynamics in stored farm slurries. Final report to Defra on project ES0117; Williams, A.G.( 2002) Covering slurry lagoons under IPPC. Part 1 – Scoping study. Final report to Defra on project AM0117; Williams and Scotford, (1999) Covering a farm scale lagoon of pig slurry. Final report to MAFF on project WA0708; Zhang (1994) A computer-model for predicting NH3 release rates from swine manure pits. Journal of Agricultural Engineering Research 58, 223-22

4

Measurements

4.1

Quantifying NH3 depletion by a woodland: a wind tunnel experiment

4.1.1 Objective and methodology The objective of this study was to assess farm woodlands for the recapture of agricultural NH3 emissions, which for the first time, addresses the quantification of NH3 recapture by trees in under controlled conditions, simulating free range livestock in a shelterbelt, and a livestock farm building. To quantify the NH3 recapture by woodland, a line source and trees were set up in a wind tunnel. NH3 and CH4 were released simultaneously and measured both upwind and downwind of the source. The depletion of CH4 tracer with increasing distance from the source should only be due to dispersion, whereas the NH3 depletion will also depend on interaction with the trees stomata. The differences between NH3 and CH4 indicates NH3 uptake. The wind tunnel is at the Air Flow Laboratory, Silsoe (Cranfield University). The tunnel dimensions are 29.5mx 5.8mx 4.8m. The turbulence was measured by one sonic anemometer downwind and 4 mobile ultrasonic anemometers. Temperature and relative humidity (RH) were measured. A line source of NH3 and CH4 released at 0.5 m or 0.1 m height, to simulate a farm (1:4 source: canopy ratio) and roaming animals (1:20). 28 Picea Abies (Norway spruce, 2m tall) were placed in 5 rows, and straw bales made a floor surface. Growth lights ensured photosynthetic activity. Leaf area index (LAI) was measured. 9 inlets along the centre of the tunnel, and 4 inlets set at logarithmically increasing heights were connected to the chemical analysers. The conditions were representative of a UK winter climate. Runs were carried out with and without trees, under varying turbulence conditions and with wetted leaves (see full report for details).

Figure 12: LHS: “empty” wind tunnel; RHS: filled tunnel

4.1.2 Results and Discussion The NH3 uptake onto the trees cannot be measured directly, so a “depletion factor” ∆𝐶 is established that is directly proportional to actual uptake by the trees: 𝑐𝑛 =

𝑐−𝑐0

∆𝐶 =

E1

𝑐1

𝑐𝑛 (𝐶𝐻4 )−𝑐𝑛 (𝑁𝐻3 ) 𝑐𝑛 (𝐶𝐻4 )

E2

The data collected in this study are representative of a winter climate in the UK (low variability of T and RH) and conifer trees. Depletion of NH3 relative to CH4 was observed in the empty tunnel and is due to walls of the tunnel and the straw bales which made the “floor” of the trees. This meant that the lower source height had much more of a depletion effect due to interaction with the straw bales. Similar ground effects may be observed in the environment due to soil surface, leaf litter etc. under trees. Depletion factor (ΔC) linearly proportional to distance from source, i.e. number of trees NH3 is exposed to. Trees wetted with weak acidic solution took up slightly more NH3 than unwetted trees. The results from all of the runs show that significant NH3 was recaptured. The source height is key to effectiveness of tree belts as a mitigation measure. ΔCaverage was 22% of NH3 from all the runs in the wind tunnel study. More NH3 was found to be recaptured under wetter conditions (up to 43%). An example of an experiment profile shown in Figure 13 LHS, and Figure 13 RHS of Figure 13 shows ΔC as a function of distance from the line 13

source for wet trees, dry trees and the empty tunnel under the same source release and turbulence conditions (RH ~85%, T = o 10 C). Overall Results are summarised in Table 4. It is noted however that the effects of the trees being in a tunnel with roof and wall effects means that the results may not directly be replicated in field measurements.

Figure 13: LHS: Example of CH4 and NH3 concentration profile down wind tunnel length; RHS: Depletion factor as a function of distance from line source for wetted trees (blue), unwetted trees (red) and the empty tunnel (yellow) Table 4. Summary of Wind tunnel experiment results. NH3 release level (ppbV) 250 250 250 180 750

4.2

Wind speed (m/s)

Number of trees

Source height (m)

2 2 2 5 5

0 0 28 28 28

0.5 0.1 0.1 0.1 0.1

Depletion “just trees” % 28 30 15

Depletion % 8 24 52 39 30

Experimental quantification of NH3 capture by an overhead tree canopy.

4.2.1 Objective and methodology The objective of this study was to assess a woodland recapturing NH3 after release of NH3 underneath the canopy. A Forest Research plot within the Forest of Ae, near Dumfries, was identified as having suitable spacing and size for a release experiment (Figure 14). The hybrid larch plantation is within the Ae Forest with SW prevailing winds. A network of ½”tubing, 0.1 mm pinholes at 1m spacing (400 release points, Figure 15 ) was laid across the woodland floor. Pressure drop calculations showed that the flow through each pinhole should be within error the same, and indicative flow measurements validated this. NH3 and CH4 were co-released. Atmospheric NH3 was measured using a photoacoustic instrument (Nitrolux, Pranalytica) and CH4 with a tuneable diode laser spectrometer (TDL). Methane in this experiment is used as a tracer gas, as in Section 3.2. In addition the wind profile, turbulence above and within the canopy (using a moveable micro-sonic anemometer), ambient meteorological conditions, leaf area index (LAI) and leaf wetness were monitored. Average LAI values ranged from 4.95 to 4.37.

P1 CH4

NH3 1.5m 1.5 m

30 m Figure 14 : Forest of Ae site; release network tubing is visible running across the woodland floor.

Measurement point

Figure 15 Schematic Layout of Undercanopy release experiment

4.2.2 Results The proposed experiment led to significant logistical and scientific issues. The TDL was not functional for NH3 measurements so the lower time resolution photoacoustic instrument was used which meant quantitiative comparison of NH3 and CH4 data was 14

not possible. Experiments were in September – October 2009 during which conditions and equipment performance were very variable including periods of complete calm and periods of highly variable wind direction. Several experiments to release NH3 and CH4 were attempted, e.g. Figure 16. Due to the issues it was not possible to calculate NH3 uptake fractions from the data. However the 5 week deployment gave a good set of parameter data: LAI data were analysed and used for the LAD profiles in Section 4.3. The turbulence data above and within the canopy and the wind profile and meteorological data were also used in Section 4.2 and are a valuable modelling resource as the presence of the canopy changes radically the turbulence patterns.

NH3 concentration / ppbV

CH4 concentration / arb units

Date and Time

Figure 16 NH3 and methane release experiment. Colours represent measurements made at different heights through the canopy

4.3

Figure 17 Wind rose for Ae field measurement period

Field Case Studies of NH3 Concentrations downwind of Poultry Houses

4.3.1 Introduction NH3 release from point sources represents a significant portion of the NH3 input into the environment. One potential route for mitigating local impacts of the NH3 and enabling recapture of the NH3 is the use of trees. Agroforestry is a small but significant part of UK agriculture and in particular the popularisation of free range chickens with access to tree cover increased significantly st in the first decade of the 21 century, due to animal welfare improvements within the egg and poultry meat production sectors. In this study the potential for mitigation of NH3 concentrations and emissions from chicken farms were studied by monitoring the concentration of NH3 upwind and down wind at three case study farms. The three farm sites were selected where there were trees in close proximity to the poultry housing, which in this case actually used some of the tree cover for the birds to use as part of their range, i.e. the trees were their for animal welfare purposes rather than for an abatement purpose a priori. NH3 is a reactive gas and concentrations decrease rapidly as a function of distance from the NH3 source. Under the simplest approximation NH3 concentrations downwind of the chicken sheds with and without trees in the same location can be compared in order to assess NH3. 4.3.2 Methodology and Sites Monthly average NH3 concentrations were measured along a transect following the prevailing wind for a period of ~7 months (Adapted Low-cost Passive High Absorption (ALPHA) samplers; (CEH, Tang et al., 2001). One NH3 sampler was placed upwind of the chicken shed and up to 5 samplers downwind. Three farms were found for Case study sites, the details of which are summarised in Table 5 and images of the farms from GoogleEarth images shown in Figure 18 (note the GoogleEarth images predate the current study by several years therefore the images aer inidicative of the situation rather than accurate records. See full report for more details. Table 5 Summary of three case study farms Farm FAI Farms, Wytham Din Moss, Fife Freuchie Mill, Fife

Farm type Mixed animal farm Poultry/ low intensity sheep Poultry

Poultry type Free range Free range Free range

Housing type Arks Sheds Sheds

15

Bird #s/shed 700 3,000-20,000 5000

Comment Research plot1 Cleared conifer plantation Woddland egg scheme new trees plant

Figure 18 GoogleEarth© images of three case study farms (note scales different on each image). Pink dots: Open transect; Red dots: wooded transects

2.5

average [NH3]wooded/[NH3]open (all 8 periods)

4.3.3 Results and Discussion FAI At the FAI Farm, chicken are kept in mobile arks which are moved periodically. Though the original experiment finished prior to this experiment, FAI continues to use the arks on the site where patches are wooded and other areas are not. The wooded patches are quite densely vegetated and are approaixmately 3-4 m. in height. Systematically the concentration downwind of the wooded transect was lower than the unwooded transect. When the measured NH3 concentrations are normalised to the same distances and concentrations ratioed (ΔC) calculated it can be seen that there is 10-25% lower NH3 concentration downwind of the wooded area compared to the open transect at distances beyond the trees. Overall the average ratio ΔC (30m) = 25% for 7 month measurement period

2.0

1.5

1.0

0.5

0.0 0

10

20

30

40

50

Distance from shed Din Moss Din Moss and Saline Farms are in central Fife and are owned and operated by Noble Foods plc (UK). Two of the poultry sheds were put into a clear Figure 19: Ratio of average NH3 concentrations cut part of an established conifer plantation in 2004 and then the surround of the between wooded and open transects for the FAI Farm case study as a function of distance (m) NE sheds planted with conifers spaced according to the Woodland Trust Woodland from poultry house. egg programme guidelines.The new plantation trees are ~ 5 -10 years old and are ~2 m spaced and birds are free to forage underneath these trees. The new tree belt surrounds the houses to a width of ~100 m both upwind and downwind and the trees are ~2-4 m tall. Further up the hillside is the remaining more mature conifer plantation (h ~ 10m). Due to the compelxity of the farm site, the two sets of measurements are not directly comparable, however the results are indicative of the differences between the two houses. Application of the simple screening model (SCAIL, http://www.scail.ceh.ac.uk) shows that the measured NH3 concentrations downwind of the trees are lower than would be predicted in the absence of trees: ΔC (30m) = 43%. However there are large uncertainties assoicated with this value and more detailed modelling of the complex site would be useful.

Figure 20 Din Moss wooded and open transects: Mean NH3 concentration over full measurement period

Freuchie Mill Freuchie Mill Farm is an example of a poultry farm which became part of the Woodland Egg programme by having new trees planted around the pre-existing poultry sheds. The trees are widely spaced (~3 m.) and planted in lines. The trees are a deciduous mix and are ~ 5-10 years old. Southwest of the two sheds on the west of the far there is ~100 m tree belt downwind of housing, as well as trees up wind. Birds are free to range under the trees. The spacing and age of trees mean that the trees act as an effect tree belt, and thus are thought to have a minimal recapture potential. The site illustrates that when the tree planting is undertaken the full range of effects - animal welfare, aminity, wood fuel etc. could also have further mitigation potential if that aim is taken into account at the planning and planting stage of the projects. Figure 21: Trees at Freuchie Mill Farm (with NH3 samplers in foreground)

16

4.3.4 Overview This short report summarises briefly the measurements made and the results found at the three case study farms. At two sites, evidence for significant decreases of NH3 concentrations downwind of tree belts was observed. This shows the direct potential for woodland tree belts to act in a protective capacity for relatively small sensitive areas which are in the vicinity of NH3 sources such as poultry houses. The concentration measurement do not show what fraction of the NH3 decrease at these sites is due to recapture by the trees and other vegetation and how much is due to increase dispersion at the front of the tree belt. As shown in the modelling studies of Section 1 the process is likely to be a combination of the two. The case studies illustrate that animal housing can either be proactively placed in forestry areas or have trees planted in the vicinity of the housing, but that planning of the geometry, tree type and management are key to ensuring that all the uses of the trees aspired to, are met. Future work for this project is to do more detailed modelling for the complex sites to understand the net changes the trees bring to the NH3 concentration fields.

5 5.1

Agroforestry ammonia abatement cost analyses Profitability analysis of trees and woody shelter belts on livestock farms for NH3 abatement and carbon sequestration

5.1.1 Introduction In WA0179, Theobald et al. (2003) indicated that a 30-60m woody belt around strong NH3 point source, such as intensive livestock housing was the most cost effective compared to other designs. This work revisits the potential profitability of such shelter or woody belts given better knowledge of design and performance, with recent price and grant information, comparing original and improved designs. In determining the potential profitability there are two questions: 1) What is the farm giving up when land is taken for the new purpose 2) what is the farm gaining by adopting this new enterprise”? The default management model is assumed to be one of grant-aided establishment followed by least-cost maintenance and harvesting. Establishing woodland is a long-term proposition. Near future costs have to be covered by uncertain future returns. Discounted Cash Flow (DCF) is applied (Warren,1982). At the end of 40 years, the belt will be in a state of equilibrium management. The financial life cycle of a tree or woody crop can be analysed in three stages; establishment, maintenance and harvesting. The 2009/2010 prices and costs have been obtained (Nix, 2010, ABC, 2010). Establishment covers:1) obtaining the land, 2) ground preparation, 3) planting 4) tree protection. The Silsoe Whole Farm Model (SFARMOD, Annetts and Audsley 2002) was used to quantify the opportunity cost of losing one hectare of productive agricultural land. Two farm type cases exist: arable providing land for pig and poultry and grass-arable farm types providing land for dairy. Annual rainfall and soil types have been evaluated and farming systems with high value crops incur the greatest financial penalties. Combining results with WA0178 results to current cost terms gives an appreciation of how relative farm profitability changes over time and thus perhaps a fairer reflection of the opportunity costs of tying land into the woody belts over a very long term (Table 6). Preparation cost might include spraying, sub-soiling, ploughing and seed bed cultivation. The model allows for ground work, weed control, fertilising, liming, grazing protection. Maintenance allowances are made. Table 6 A comparison of the marginal cost of one ha in 2010 terms for arable (2002 & 2010) and arable with grassland (1995, 2002, & 2010) Arable Annual Rainfall, mm Soil 600 Light 544 (516 to 572) Medium 595 (485 to 705) 601 (463 to 740) Heavy Dairy Annual Rainfall, mm 600 Light 723 (423 to 900) Medium 755 (381 to 984) Heavy 781 (325 to 1106)

900 580 (554 to 607) 618 (577 to 660)

1200 602 (521 to 682) 633 (600 to 665)

610 (556 to 665)

593 (585 to 602)

900 772 (366 to 1074) 822 (334 to 1070) 846 (289 to 1161)

1200 754 (313 to 1218) 829 (284 to 1213) 813 (256 to 1160)

Grants are available for the planting of farm woodlands and are regionally devolved in the United Kingdom, details change over time. The grant values used in this analysis were the English Woodland Grant Scheme. Currently, there are sources of support for the woodland, but they are regionally devolved and subject to change. The English Farm Woodland Payment Scheme (FWPS) is one source and with it the Single Farm Payment (SFP). As with all grants the eligibility and rules need to be checked. Removal of trees and other biomass could be modelled as a maintenance cost rather than a yield. However, Nix (2010) does quotes values for thinnings and we assume similar costs. 5.1.2 Scenarios Three basic designs that have been looked at: The first design stems from WA0178 (4 zones: 1 shrubby intake, a broadleaf, a conifer and a tight hedge back stop). The second design is a 1 brashed broadleaf zone woody one unbrashed conifer zone. The third design is a conifer backstop to a woodland animal enterprise. The first design has two depths (30 and 60 m). The second design is to enclose point sources such as existing poultry houses and slurry tanks. The size of the protected facility changes the 17

relative amounts of the broadleaved and conifer components because the conifer component wraps around the edges to provide the back stop effect. We have considered facilities that are 50m, 100m 200m, and 400m long and are protected through 180 degrees on the downwind side. There is an additional design choice of the depth of conifer back stop with choices being 5m, 15m, 25m, and 50m. The third design offers the same four depths of back stop choices as the second design.The Net Present Value (NPV) results are shown in Table 7. All scenarios return negative NPVs, which shows the trees by themselves would not be financially viable. When analysing the biggest term is typically the opportunity cost of the land, typically representing 62% of costs. Thus the exact value of such land to the farm and farmer is key. One challenging term in any DCF is the discount rate as it contains a subjective component. Higher discount rates imply a lower NPV, which are significant here. Often for Green investments there is high initial outlay followed by long term benefits; in this case a small discount rate is more flattering. That is not the case here. Table 7 Net present values of all scenarios, £ha planted Wood belt depth

Design 1 Design 2

Backstop depth

30m -£12,966

60m -£11,735

5m

15m

25m

50m

50m building

-£12,299

-£13,883

-£15,154

-£12,498

100m

-£12,185

-£13,525

-£14,525

-£12,326

200m

-£12,126

-£13,292

-£14,022

-£12,137

400m

-£12,095

-£13,156

-£13,686

-£11,972

Design 3 (backstop only)

-£15,832

-£16,913

-£14,011

-£11,738

-40000

-30000

Net present value, £/ha -20000 -10000 0 10000

20000

30000

Woodland Grant Scheme

Figure 22 Breakdown of Net Present Value of four designs options, showing that the opportunity cost of land is a major compenent

Farm Woodland Payment Scheme Design 1 (30m deep)

Single Farm Payment Timber Income Ground preparation

Design 1 (60m deep)

Fetilising Spraying Tree protection

Design 2 (25m deep belt around 3 sides of a 50m point source)

Planting costs Managerial oversight Backstop maitenance Brashing/ thinning

Design 3 (25m deep backstop around 1 ha of woodland chickens)

Opportunity Cost

5.1.3 Discussion The results here suggest that woody belts are not economically feasible in purely financial terms. On a case by case basis there might be a different story. There are several factors for and against: 1. 2. 3. 4. 5. 6. 7.

Land opportunity costs: These are local to the farm. We have assumed that the land will be commercially attractive lowland with good access and farming potential. Commercial rates for labour and machinery costs are assumed, which may be available at marginal cost The establishment, over 40 years, of a steady state uneven-aged woody belt to provide long-term continuity of the NH3 abatement is an ideal assumption in many ways. There is non-market and hard to value benefits from tree planting, such as privacy, landscape character. There are some theoretical negative factors e.g. drawing in predators and wild avian species There may be numerous public policy benefits, such as carbon sequestration, biodiversity, rural aesthetics We have not considered alternative NH3 mitigation investments e.g. flue gas scrubbers.

5.1.4 Conclusions In summary for a farmer the decision boils down to being prepared to invest in a case by case way in the woody belt to achieve a mixture of public and private benefits. These tradeoffs are likely to be favourable if the NH3 emissions are very strong, the vulnerable habitats are known to be very vulnerable and close to the NH3 sources, and there is a convincing privacy and landscape character/value argument. Public financial recognition of any public benefits would of course help mitigate opportunity costs. 5.1.5

References

Agro-Business Consultants (ABC); (2010); The Agricultural Budgeting and Costing Book, 71th Edition, ABC, Leicestershire, 401pp.; Annetts, J. E., & Audsley, E. (2002). Multiple objective linear programming for environmental farm planning. J. Op. Res. Soc., 53(9), 933-943.; Audsley, E. An arable farm model to evaluate the commercial viability of new machines or techniques. J. Agric. Eng. Res., 1981, 26 (2) 135; Audsley, E., Archer, J.E. (1989) The profitability of an arable wood

18

crop for electricity. Conference on Arable Wood Crops for Electricity, 1989, NAC, Stoneleigh.; Blyth, J, (2003) Personal Communication. Senior Lecturer in forestry, University of Edinburgh. ; HM Treasury (2003) Green Book, Appraisal and Evaluation in Central Government http://greenbook.treasury.gov.uk/; Nix, J.; (2010); The John Nix Farm Management Pocket Book, 41st Edition, Agro Business Consultants Ltd, Melton Mowbray, UK. 289pp.; Sells, J.E., Audsley, E. (1989) The profitability of an arable wood crop for electricity. DN 1549, AFRC Inst. of Eng. Res., Silsoe, 18 pp; Sells, J.E., Audsley, E. (1991) The profitability of an arable wood crop for electricity. Journal of Agricultural Engineering Research, 48, 273-285; Slachter, C. (2003) Personal communication. Silsoe Research Institute member of staff expert of management of trees in an amenity setting; Theobald M.R. , et al., WA0179, Final report to Defra, available at http://randd.defra.gov.uk, 2003; Warren, M. F.; (1982); Financial Management for Farmers. Hutchinson, London, 306pp

5.2

Cost-effectiveness of Agroforestry options for NH3 Abatement as Climate Change Mitigation measure

5.2.1 Introduction The cost-effectiveness of UK forestry measures aimed at climate change mitigation has been a focus of several recent studies (1, 2, 3). Fewer studies have examined this for options aimed primarily at other objectives (4) or as part of a broader ecosystem services assessment (5). The current study is one such case. It complements previous results by focusing upon the costeffectiveness of agroforestry measures motivated by ammonia abatement. The cost-effectiveness analysis focuses on similar scenarios to those considered elsewhere in this report. The first (option 1) involves planting 0.5 ha of trees (0.125 ha of broadleaves fringed on three sides with a 0.375 ha conifer backstop) downwind of a barn or slurry lagoon. The second (option 2), a free-range woodland chickens scheme, involves planting the equivalent of 1.675 ha of trees (0.8 ha of broadleaves and 0.875 ha of conifer backstop) over an area of 1.875 ha downwind of poultry housing. Carbon estimates were obtained from CEH’s C-FLOW model for planting beech yield class (YC) 6 (with intermediate thinning) and sitka spruce YC 12 (with no thinning). These were chosen to represent broadleaf and conifer components respectively of the two agroforestry options. The estimates were obtained for a 100 year time horizon, upon which the cost-effectiveness estimates (3) reported in the Read Report (6) were based. However, climate change cost-effectiveness is also analysed over an initial 40 year period. Current government guidance on estimating cost-effectiveness in appraisal and evaluation (7 p.29) recommends deriving the cost-effectiveness of a measure by dividing its net present value (NPV) excluding the present value of the carbon benefits by (the negative of) the total tonnes of carbon dioxide equivalent saved. Whether a measure is cost-effective is then determined by comparing the cost per tonne of carbon dioxide equivalent abated with the relevant cost comparator based upon estimates of the social value of carbon. The ammonia abatement potential is assumed to increase linearly from zero to a maximum after 26 years in each case. Maximum abatement potentials of 1.74 and 0.11 tonnes of ammonia per hectare per year are assumed for central estimates for option 1 and option 2, respectively. This abatement is then valued by following current Defra guidance (8) on valuing the benefits to society of avoided air quality damage costs per tonne of pollutant, which includes a central estimate of £1,972 per tonne of ammonia at 2010 prices. Ideally cost-effectiveness estimates should also take into account wider impacts on the provision of ecosystem services associated with land use change. A review (9) of the limited evidence available suggests that the non-use value of woodland biodiversity and cultural aspects of woodlands may range from £30 to £300/ha/yr depending on the priority status of the woodland. This range is assumed in the current study. Table 8: Cost Assumptions (£ per ha of project area per year at 2011 prices) Option 1 (housing/lagoon shelterbelt) Agricultural Opportunity Cost (p.a.) £595† Establishment Cost (yr 0) £8,635 Management Costs (yr 1 onwards) £22 Fertiliser and spraying Costs (yrs 1-4) £93 Fencing Costs (yr 4 onwards) £84 Backstop Maintenance Costs (yr 5 onwards) £10 † central estimate (ranged for sensitivity analysis from £514/ha/yr to £760/ha/yr).

Option 2 (woodland chickens) £311 £6,182 £22 £93 £38 £10

5.2.2 Results Table 9 summarises the present values of the cost and benefit estimates associated with the two scenarios for the 40 year time horizon, along with corresponding NPV estimates (including carbon benefits). The positive NPVs for option one indicate that even over the shorter 40-year time horizon this scenario offers positive net benefits from a societal perspective. This also holds for central and high estimates for the woodland chickens scheme (scenario 2), although not if woodland management is assumed to subsequently revert to 40-year rotations (low estimate). The latter conclusion might be reversed, however, were agricultural opportunity cost estimates further refined to exclude subsidies, or wider ecosystem service impacts covered.

19

Table 9: Present Values over 40-year time horizon (£/ha at 2012 prices) Option 1 (housing/lagoon shelterbelt) Low Central High -£28787 -£25038 -£23171

Forestry Costs Wood Production

Option 2 (woodland chickens) Low Central -£15262 -£15262

High

-£15216

£953

£980

£80

£825

£871

£137

£27323

£70068

£119438

£2379

£4576

£6933

Habitat and non-use

£192

£2363

£5427

£172

£2111

£4869

Carbon Sequestration

£2639

£22599

£43240

£2288

£18892

£36149

NPV

£2321

£70973

£145015

-£9598

£11189

£32873

Ammonia Abatement

The cost-effectiveness results are presented in Table 10. These suggest that the first agroforestry measure (option 1) is highly cost-effective from a climate change mitigation perspective. Indicative estimates over a 40 year time horizon ranging from £186/tCO2 to £3/tCO2 compare very favourably with cost-effectiveness comparators that range from £47/tCO2 to £53/tCO2 based upon discounted social values of carbon. For option 2 the results imply indicative cost-effectiveness estimates over a 40 year time horizon ranging from £7/tCO2 to £123/tCO2, with a central estimate of £21/tCO2. As the central and high estimates are below associated cost-effectiveness benchmarks, these are also considered cost-effective. Extending the time horizon increases cost-effectiveness in each case. Table 10: Cost-Effectiveness (£ per tonne of carbon dioxide at 2012 prices) Time frame 40 years

100 years

Basis Low

Option 1(housing/lagoon shelterbelt) Estimate Comparator £3 £47

Option 2 (woodland chickens) Estimate £123

Comparator £47

Central

-£113

£53

£21

£53

High

-£186

£53

£7

£53

Low

-£289

£47

£121

£47

Central

-£340

£53

£9

£53

High

-£266

£50

-£7

£49

5.2.3 Conclusions The results suggest that of the two agroforestry scenarios, option 1 (planting a shelterbelt downwind of a barn or slurry lagoon) is highly cost-effective from a climate change mitigation perspective. While option 2 (woodland chickens) is cost-effective under central and high estimates, whether the woodland planted is subject subsequently to 40-year rotations (assumed under the low estimate) is critical to judging the cost-effectiveness of this scenario. Inclusion of other ecosystem service benefits associated with woodland planting, as well as carbon substitution benefits, would be expected to increase the cost-effectiveness of both options. 5.2.4

References

Radov, D., et al. (2007). Market mechanisms for reducing GHG emissions from agriculture, Forestry and Land Management, NERA Economic Consulting, London, Defra., Moran, D. et al. (2008) UK Marginal Abatement Cost Curves for Agriculture and Land Use, Land-use Change and Forestry Sectors out to 2022, with Qualitative Analysis of Options to 2050, Final Report to the Committee on Climate Change, London http://www.sac.ac.uk/research/projects/landeconomy/featured/macc/; ADAS (forthcoming). Analysis of Policy Instruments for Reducing Greenhouse Gas Emissions from Agriculture, Forestry and Land Management – Forestry Options (draft report to FCE dated June 2009); Nisbet, T. et al. (2011). Slowing the Flow at Pickering. Final Report to Defra FCERM Multi-objective Flood Management Demonstration project RPM5455, Forest Research, http://www.forestry.gov.uk/fr/INFD-7ZUCQY#final1.; Valatin, G. and Saraev, V. (2012). Natural Environment Framework: Woodland Creation case study. Report to Forestry Commission Wales, Forest Research, Edinburgh, http://www.forestry.gov.uk/fr/INFD-8YAECD; Matthews, R.W. and Broadmeadow, M.S.J. (2009). The Potential of UK forestry to Contribute to Government’s Emissions Reduction Commitments. In Read, D. et al. (eds) Combating climate change – a role for UK forests – An Assessment of the potential of the UK’s trees and woodlands to mitigate and adapt to climate change, the Stationery Office, Edinburgh. HM Treasury and DECC (2012). Valuation of energy use and greenhouse gas (GHG) emissions. Spplementary guidance to the HM Treasury Green Book on Appraisal and Evaluation in Central Government. HM Treasury and Department for Energy and Climate Change, London, http://www.decc.gov.uk/en/content/cms/about/ec_social_res/iag_guidance/iag_guidance.aspx.; Defra Guidance Documents: http://www.defra.gov.uk/environment/quality/air/air-quality/economic/damage/; Eftec (2010a). The economic contribution of the public forest estate in England. Report to the Forestry Commission

20

6

National modelling

6.1

On-farm emission factor reduction

6.1.1 Objectives and methodology A set of revised emission factors for all UK livestock types were developed. These reduced emission factors can be described as ‘on-farm’ emissions as the tree recapture element can be seen as being part of the total farm emissions. The FRAME (Fine Resolution Atmospheric Multi-species Exchange) model (Singles et al.,1998; Fournier et al., 2003; Dore et al., 2012) was applied at a 1 km resolution across the British Isles to assess the scenarios with these changes. Modelling results from Section 3.3 were used to optimise the tree canopy structures for capturing NH3 based on three farm management practices – canopy structure to capture NH3 from housing, slurry lagoons, and livestock under the tree canopy. By changing model parameters (canopy width, LAI and LAD) optimal tree structure configurations were established. From this a set of average NH3 capture percentages were used to calculate the reduction in the 2008 livestock emission factors: • 20% NH3 capture efficiency for housing emissions • 20% NH3 capture efficiency for storage emissions • 45% NH3 capture efficiency for livestock under-canopy farming systems (i.e. grazing emissions) For each livestock type emission reduction percentages were calculated, based on various management systems: ‘sheltering’ housing, storage lagoons with trees, and grazing livestock under the trees (Table 11). Emission factor reduction for livestock types using two tree planting scenarios of 45% for grazing under trees, and 20% for sheltering housing units. From this percentage emission reductions were calculated and 8 scenarios designed for FRAME model runs (Table 12). Table 11: Modification of management and resultant percentage emission changes for UK livestock types

Livestock Type

Management System

Layers Layers Layers Sows Other pigs >80-110 kg Other pigs >50-80 kg Other pigs >20-50 kg Dairy cows & heifers

In housing upwind of tree belt, no ranging In housing upwind of tree belt + 25% ranging under trees In housing under tree canopy (arks) + 25% ranging under trees Double the number of sows outdoors (currently 36%) + ranging under trees Increase to 15% the herd outdoors (currently 0.01%) + ranging under trees Increase to 15% the herd outdoors (currently 0.01%) + ranging under trees Increase to 15% the herd outdoors (currently 0.01%) + ranging under trees In housing upwind of tree belt, no ranging + slurry store with trees downwind

% emission reduction 12% 27% 38% 46% 9% 9% 8% 13%

Table 12 Percentage change in total nitrogen deposition from each emission reduction scenario SCENARIOS CURRENT TREND SCENARIO 1 POULTRY - 50% of all poultry houses sheltered SCENARIO 2 POULTRY - housing sheltered and foraging under trees SCENARIO 3 POULTRY - Birds ranging under trees, 70% houses sheltered, 30% in arks under trees SCENARIO 4 POULTRY - broilers (60% houses sheltered, 10% forage under trees) SCENARIO 5 POULTRY (combination of Runs 1-4) SCENARIO 6 Dairy+ Beef (20% of cattle houses and slurry stores sheltered) SCENARIO 7 PIGS (72% of sows and 15% of other pigs foraging under trees) SCENARIO 8 COMBO ( SC5 Poultry, SC6 Cattle and SC7 Pigs)

6.1.2

Scenario 0 1 2 3 4 5 6 7 8

% reduction in total N (grid average) 0.6% 0.4% 0.5% 0.3% 1.3% 2.0% 1.30% 4.5%

Conclusions

Scenarios 1-4, covering the poultry sector, show small reductions in total nitrogen deposition even though the woodland systems were applied to over half, in some cases, of the total UK flock. This is due to the low emission factor for poultry as a whole, even though there are over 160 million birds in the UK. However, Scenario 5 has a higher reduction of 1.3% when all scenarios 1-4 are included. For the cattle sector a total reduction of 2% is achievable with placing woodland structures around 20% of the cattle housing around the UK and 20% of the slurry stores. Doubling the number of outdoors sows together with foraging under trees (36% to 72%) and putting a smaller percentage (15%) of other pigs under trees reduces deposition by 1.3%. The best reduction in total nitrogen deposition is the combination of all the other scenarios producing a reduction of 4.5%. 6.1.3

References

Dore, A.J., Kryza, M., Hall, J. Hallsworth, S., Keller, V., Vieno, M. & Sutton, M.A. (2012) The Influence of Model Grid Resolution on Estimation of National Scale Nitrogen Deposition and Exceedance of Critical Loads Biogesciences (in press) Fournier, N.; Pais V.A.; Sutton M.A.; Weston K.J.; Dragosits U.; Tang S.Y. and Aherne J. (2003) Parallelisation and application of a multi-layer atmospheric transport model to quantify dispersion and deposition of NH3 over the British Isles. Environmental Pollution, 116(1), 95-107.

21

6.2

Assessment of the abatement potential of farm woodlands at the UK scale

A.J. Dore, W.J. Bealey, U. Dragosits, and M.A. Sutton 6.2.1 Objectives and methodology In order to assess the potential changes due to abatement ammonia with farm woodlands at the UK scale, the FRAME model was applied at a 1 km resolution across the British Isles. This can assess the influence of national scale re-afforestation on NH3 concentrations in air and the deposition of reduced nitrogen. To assess the influence of afforestation on recapture of NH3, three land cover scenarios were generated: a baseline scenario (0) an increases of (1) 25% total UK forest cover and (2) 50% total UK forest cover. Tree planting was targeted near emission sources where NH3 concentrations are highest to maximise the effects. Trees were only planted on arable and grassland, with the other land cover categories (semi-natural ecosystems (excluding woodland) and urban) remaining unchanged (see Table 13 ). Tree coverage was increased by scaling existing forest cover to the NH3 emissions (or by adding new forest in grid squares with no tree coverage). Forest cover for the baseline scenario and the change between the baseline and the +50% scenario 2 are shown in Figure 23. Table 13 Percentage of land cover types for the baseline and 25% and 50% afforestation scenarios. 0. Baseline 1. + 25% 2. + 50%

arable 23.0 21.7 20.4

forest 11.7 14.7 17.6

grass 22.3 20.6 19.0

semi-natural ecosystems 33.8 33.8 33.8

urban 6.6 6.6 6.6

water 2.6 2.6 2.6

Figure 23 Forest distribution in the UK. Percentage of land cover which is woodland for the baseline scenario (left); Percentage of land which is new woodland for the +50% scenario (right)

6.2.2 Results The results for the baseline scenario for NH3 concentration and deposition of reduced nitrogen are illustrated in Figure 24 . Agricultural NH3 concentrations in the UK are highest across cattle farming areas in western UK, as well as in localised hot spots around intensive pig and poultry farms. This pattern is reflected in the dry deposition of reduced nitrogen map, which is driven by locally emitted NH3 gas. Reduced nitrogen wet deposition is from ammonium aerosol formation and resulting long range transport. Wet deposition is highest in the upland areas of Wales and the Pennines. Scenarios 1 and 2 increase NH3 dry deposition near the emission sources due to the lower forest canopy resistance compared to grassland and arable. This effect leads to decreases in wet deposition of reduced nitrogen and of dry deposition to semi-natural land. Figure 25 illustrates the decrease in reduced nitrogen deposition resulting from implementation of scenario 2 (50% national increase in forest cover). Significant reductions in nitrogen deposition were achieved with this scenario. In areas of high wet deposition (the Pennines and Wales), the change in wet deposition was up to 0.5 kg N ha-1. Higher decreases of up to 2 kg N ha-1 for dry deposition were achieved for large areas of semi-natural land and forest. While the deposition per unit area of forest decreased, it is important to note that total mass of reduced nitrogen deposited to forest increased due to the national increase in forest area. This is generally considered to be beneficial, as new deposition would be directed to plantation forests in agricultural areas, reducing the impact on established natural forest ecosystems.

22

Figure 24 Modelled concentration of NH3 in air (left); Dry deposition of NHx (middle); Wet deposition of NHx (right)

Figure 25 Reduction in deposition of reduced resulting from a 50% increase in forest cover: Wet deposition (left); deposition to semi-natural land (centre) ; forest deposition (right)

The FRAME model was also used to calculate a budget of the total mass of nitrogen entering and leaving the domain of the United Kingdom. The national reduced nitrogen budget for the three scenarios is illustrated in Table 14 and Table 15. The two tree planting scenarios result in significant changes to the fate of emitted NH3: significant increases in dry deposited reduced nitrogen and decreases in wet deposited reduced nitrogen and decreased export of reduced nitrogen in air leaving the UK (which contributes to the long range transport of air pollution in Europe). In Table 15, changes in NHx deposition and export for tree planting scenarios 1 and 2 are expressed as percentages relative to the baseline scenario. It can be seen that a 50 % increase in forest cover targeted at high NH3 emission areas would result in a 19.5% increase in dry deposition, a decrease of 4.5% in wet deposition and a 6.9% decrease in the export of reduced nitrogen. Table 14 The UK mass deposition and export budgets for simulations 0, 1 (+25%) and 2 (+50%) Gg N-NHx Dry Deposition Wet Deposition Total Deposition Export

0.

BASELINE 61.5 81.1 142.6 121.4

1. + 25% forest 68.0 79.1 147.1 116.9

2. + 50% forest 73.5 77.4 151.0 113.1

Table 15 Changes to the UK mass deposition and export budgets for scenarios 1 (+25%) and 2 (+50%) (Gg N-NHx) Dry Deposition Wet Deposition Total Deposition Export

1. (Gg N-NHx) 6.4 -1.9 4.5 -4.5

2. (Gg N-NHx) 12.0 -3.6 8.4 -8.4

1. (%) 10.4 -2.4 3.2 -3.7

23

2. (%) 19.5 -4.5 5.9 -6.9

7

Summary

Understanding the physical and chemical process by which NH3 is emitted into the atmosphere and then recaptured by trees was the subject of two parts of this project. This project delivered two model techniques and detailed estimates for many scenarios under which AAA could be applied. The desktop coupled model MODAAS-AQULION was developed allowing air turbulence calculations to be provided as input for the gas dispersion and recapture model allowing real atmospheric data to parameterise model runs. This model allowed the testing of multiple tree belt scenarios for AAA assessment. By applying different LAIs, LADs and widths of backstop AAA can result in percentage NH3 recapture of up to 20% for sheltering housing and up to 45% for livestock under trees for realistic densities of vegetation. A slurry store model was developed to have both wind speed and temperature as co-variables. Application of the model indicated both wind speed and temperature are important in determining the effectiveness of a tree belt. AAA effectiveness was found to increase with short length lagoons, taller trees and higher tree densities. Measurements were made with a tree belt with a line source was constructed in a wind tunnel. Micrometeorological measurements were made and NH3 recapture measured. Results showed that the tree belts can result in percentage NH3 concentration reduction of up 10-25% depending on the structure of the trees. From model intercomparison this was seen to be a function of both dispersion and recapture. Three case study farms had NH3 monitoring transects set up for 7 months, resulting in a detailed dataset. Results in two of the case studies where an “open” transect was set up in parallel to the wooded monitoring transect significantly lower concentrations were measured beyond the tree belt, of up to approximately 40%. The case study sites illustrated that tree belts are being used on UK farms for many purposes including silvopastoral applications and therefore AAA can be achieved as a side benefit to those purposes if the tree planting density and geometry is applicable to AAA. In order to assess the potential impact on a UK scale, the results from the modelling and measurements were applied at a national level using the FRAME model (1 km resolution) A set of revised emission factors for all UK livestock types were developed (using the 20% reduction in NH3 for housing and the 45% reduction for livestock under trees) which theoretically reduced the emissions factors from farm activities due to AAA. These reduced emission factors can be described as ‘on-farm’ emissions as the tree recapture element can be seen as being part of the total farm emissions. Scenarios covering the poultry sector show small reductions in total nitrogen deposition even though the woodland systems were applied to large fraction of the total UK flock due to the low poultry emission factors. For the cattle sector a total reduction of only 2% was achievable with placing woodland structures around 20% of cattle housing and slurry stores. The potential changes in NH3 concentrations in air and the deposition of reduced nitrogen due to AAA via re-afforestation on farms at the UK scale was assessed targeting hot-spots of NH3 around the UK. In scenarios where the UK forest was increased by 25% and 50%, NH3 dry deposition increased near the emission sources. This is due to the lower forest canopy resistance compared to grassland and arable. The net effect was decreases in wet deposition of reduced nitrogen and of dry deposition to -1 semi-natural land. In the 50% scenario, changes in wet N deposition were up to 0.5 kg N ha (~12%). Higher decreases of up to 2 -1 kg N ha (~30%) for dry deposition were achieved for large areas of semi-natural land and forest. The potential profitability of tree belts was assessed given knowledge of designs, performance, and recent price/grant information. In determining the potential profitability there are two key questions: 1) What is the farm giving up when land is taken for the new purpose 2) what is the farm gaining by adopting this new land-use. The analysis results suggest that overall tree belts are not currently economically feasible in purely financial terms, however on a case by case basis there might be many different reasons for making the development of AAA approaches feasible. There are several factors for and against AAA development, including against: Land opportunity costs, commercial rates for labour and machinery, establishment timescales, drawing in predators and wild avian species and for: non-market, hard to value benefits e.g. silvopastoral agriculture, welfare considerations, public policy benefits, biodiversity/ecosystem service benefits, odour mitigation. For a farmer the decision to invest in AAA would be a case by case process to assess the ways in which the tree belt achieves a mixture of public and private benefits. Tradeoffs are likely to be favourable if the NH3 emissions are very high, vulnerable habitats are nearby and if there is a convincing privacy and landscape character/value argument. Public financial recognition of any public benefits would of course help mitigate opportunity costs. NH3 abatement as climate change mitigation is an area which had not previously been considered. The cost-effectiveness of UK forestry measures aimed at climate change mitigation was assessed in a context in which implementation is motivated by NH3 abatement. Planting a shelterbelt downwind of a barn or slurry lagoon can be considered highly cost-effective from a climate change mitigation perspective. Woodland chicken farming would be cost-effective under some estimates. The choice of time horizon and how woodland planted is subsequently managed are critical issues in assessing cost-effectiveness under the low estimate. Inclusion of other ecosystem services (e.g. water quality, amenity and health benefits) associated with tree planting, as well as carbon substitution benefits, could be expected to further increase the estimated cost-effectiveness of the agroforestry 24

options as climate change mitigation measures, however it is noted, that from the on-farm modelling (5.2) without the carbon credit based positive costings, the tree-belts would not be economically feasible per se.

8

Conclusions and Future directions

The combined modelling and measurement results from this project show that AAA carefully planned and implemented can lead to a significant decrease in NH3 concentrations downwind from sources and a net decrease in emissions to the atmosphere. As such AAA systems could be used as a protective measure of downwind sensitive ecosystems. Use of existing woodland plantations and panting new forestry can both be used to moderately mitigate emissions, though scrubbing of NH3 at source and subsequent re-use of scrubbed nutrients would also be an effective solution. UK scale modelling shows that targeted application of tree planting around agricultural installations would have a modest effect by modifying ‘on-farm’ emission factors, however when the approach is targeted in regions hot-spot emissions, significant effects on NH3 and N-deposition can be achieved. In many agricultural businesses there are no current economic advantages for converting valuable arable land to woodland without specific opportunity benefits (e.g. woodland egg price margins due to animal welfare considerations, carbon-credits). However as the woodland egg example shows, when other considerations become relevant, AAA can be a useful approach. Future work should consider: 1) Nitrogen flows in farm systems with AAA implemented and the net effect on both the reactive and GHG N budgets 2) Ecosystem service approaches for making AAA cost effective. 3) It would also be possible to modify the MODDAS-THETIS model into a relatively simple on-line model for users to model woodland developments

25

Appendix 1: User Group meeting report

th

A single User Group meeting was held at the FAI Farms Conference facilities on 24 November 2010. The Agenda for the meeting is shown in Table 1 and the Attendee list in Table 2. Table 4 User Group meeting Agenda 10:30 – 10.50

Defra Policy Perspectives

Dan McGonigle (Defra - Farming and Food Science) & Peter Coleman (Defra - Science and Evidence)

10.50 – 11:10

Introduction to the SAMBA project, NH3, and recapture by trees

Mark Sutton (Centre For Ecology & Hydrology)

11.10 – 11.35

SAMBA Field Work

Daniela Famulari (Centre For Ecology & Hydrology)

11.35 – 12.00

Modelling the recapture efficiency of farm woodland structures

Bill Bealey (Centre For Ecology & Hydrology)

12.00 – 12.10

Modelling sheltered slurry stores

Adrian Williams (Cranfield University)

12.10 – 13.30

Farm Tour and Hot Buffet Lunch

Paul Cook (FAI)

13.30 – 13.55

Agro-economics of tree-belt and silvopastoral systems

Daniel Sandars (Cranfield University)

13.55 – 14.15

Carbon storage and agro-forestry potentials

Bill Bealey (Centre For Ecology & Hydrology)

14.15 – 14.35

Abatement potential of farm woodlands at the UK scale

Tony Dore (Centre For Ecology & Hydrology)

14.35 – 14.50

Round-up overview, future abatement potentials and policy implications

Mark Sutton (Centre For Ecology & Hydrology)

14.50 – 15.30

Group Discussion

Chaired by Mark Sutton (Centre For Ecology & Hydrology)

15.30 – 15.40

Final Words

Dan McGonigle (Defra - Farming and Food Science)

Table 5 User Group meeting Attendee list Dan McGonigle Peter Coleman Colin Powlesland Nicola Barnfather Alexa Varah Keith Kirby Mark Broadmeadow Dominic Lamb Martin Wolfe Jo Smith Deborah Carlin Sirwan Yamulki Elena Vanguelova Tom Judd

Defra Defra Environment Agency Environment Agency Reading University Natural England Forestry Comission South Oxfordhsire Council Organic Research Centre Organic Research Centre Noble Foods Forestry Comission Forestry Comission Environment Agency

Paul Burgess Bill Bealey Mike Colley Tony Dore Daniela Famulari Daniel Sandars Paul Cook Mark Sutton Adrian Williams John Tucker Jamie McGeachy John Newman Mark Wilson Sarah Eddington

26

Cranfield University Centre for Ecology & Hydrology Food Animal Initiative Centre for Ecology & Hydrology Centre for Ecology & Hydrology Cranfield University Food Animal Initiative Centre for Ecology & Hydrology Cranfield University Woodland Trust Scottish Environment Protection Agency Abbey Home Farm Abbey Home Farm 2 Sisters Food Group