BERA WP Draft Report August 25

Environmental impacts of eco-local food systems – final report from BERAS Work Package 2 Draft version last updated 25 August 2005

Editors: Artur Granstedt, Olof Thomsson and Thomas Schneider

Abstract Already established initiatives with Ecological Recycling Agriculture (ERA) and local food systems in the eight EU-countries in the Baltic Sea drainage area are evaluate during the project period. Surplus of nitrogen was 50 % lower per ha and year on Swedish and Finnish BERAS-farms compared to the average agriculture (36 kg compared with 79 for Swedish and 38 kg compared to 73 for Finnish agriculture with the same animal density of 0.6 au/ha) and it was no surplus of P. For all studied 36 farms in the eight countries, the average surplus of nitrogen was 36 kg compared to the average of 56 kg for the eight countries today, which include the low intensive agriculture in the Baltic countries and Poland. All BERAS-farms with an animal density below 0.75 AU per ha have a surplus below 50 kg N/ha and can be defined as ERA-farms with animal production based on minimum 85 % own fodder.

An estimated average loss of 30-40 % NH4 of the nitrogen exudates from the animal production gives a calculated reduction of nitrogen leaching with 70-75 % on BERAS-farms compared to the average Swedish agriculture farms (an average N- leakage of 7-9 kg compared with 28-30 kg). The same calculation made for the all eight BERAS countries gave a reduction of nitrogen leaching with 45-50 % on BERAS-farms compared to the studied average Baltic Sea agriculture which include regions with until now very extensive agriculture.

Two scenarios were calculated for the BERAS countries; 1) business as usual where the Baltic Countries and Poland convert to the same structure and use of resources as Sweden and Finland, and 2) the whole studied Baltic Sea agriculture convert to ERA similar to the BERAS-farms. The conventional scenario gave an increased pollution of Nitrogen and Phosphorus from agriculture. The Nitrogen leakage would increase with 50 %. The ERA scenario gave a reduction of nitrogen leakage from agriculture with 50 % and an elimination of the surplus of Phosphorus.

Four Swedish food basket scenarios are presenting results for Nitrogen surplus, Global warming impact and Consumption of primary energy resources. Scenario 2, where all the food is produced on ERA-farms, would give a global warming impact of 800 kg CO2- equivalents per capita and year compared to 900 kg in Scenario 1 with conventional agriculture. Scenario 3 with food from only ERA-farms and local processing and distribution gave a global warming impact of 700 kg CO2-equivalents per capita and year. Scenario 4 with food from only ERA-farms, more vegetable and less meat consumption and local processing and distribution gave a global warming impact of 500 kg CO2-equivalents per capita and year, a reduction by about 45 % compared to Scenario 1. Contents 1. Executive summary...... 3 BERAS – the project and cases...... 3 BERAS Work Package 2 – environmental assessment...... 4 Results and discussion ...... 7 General conclusions...... 10 2. Plant nutrient balance studies...... 12 Background and problems...... 12 Material and Methods...... 14 Results of plant nutrient balances in the BERAS countries ...... 16 Discussion and conclusions...... 28 References ...... 30 Evaluation of nitrogen utilization by means of the concept of primary production balance30 3. Effects of 100 % organic production on Funen, Denmark...... 35 Introduction...... 35 Methods...... 35 Results...... 37 References ...... 43 4. Nitrogen and phosphorus leakage in ecological recycling agriculture...... 45 Introduction...... 45 Methods...... 45 Results and discussion ...... 48 Conclusions...... 52 References ...... 52 5. Global warming and fossil energy use ...... 54 Methodology ...... 54 Ecological Recycling Agriculture...... 55 Transports of locally produced food in Järna...... 59 Small-scale food processing industries in Järna...... 67 Global warming impact and primary energy consumption in a local food chain in Järna... 69 References ...... 71 6. Organic waste management studies...... 73 Inventory in Juva...... 73 Biogas plant in Järna...... 84 Possibilities for developed recycling and renewable energy production in Juva and Järna. 86 7. Consumer surveys in Juva and Järna for identification of eco-local food baskets...... 87 Introduction...... 87 Methodology ...... 88 Results and discussion ...... 89 Conclusions...... 98 References ...... 98 8. Food basket scenarios...... 100 Food basket scenario, Järna Sweden...... 100 Food basket scenario, Juva Finland ...... 104 Appendix 1. Farm gate nutrient balances ...... 106

2 1. Executive summary

BERAS – the project and cases Artur Granstedt, Swedish Biodynamic Research Institute, Järna Sweden

Central objectives The aim of the BERAS project is to develop a knowledge base regarding possible means of significantly decreasing consumption of non-renewable energy and other limited resources and of reducing the negative environmental impacts on the environment based on practical examples of ecological, economical and sociological sustainability in the real life. The goal is to reduce of nutrient leaching to the water systems and emissions of reactive nitrogen and greenhouse gases to the atmosphere and protect the biological diversity of agricultural production, distribution, processing and consumption of food. At the same time, the measures identified should enhance landscape diversity and promote local economies and employment in rural areas of the Baltic Sea countries. The work is based on practical case studies, primarily in selected rural areas, complemented by selected reference farms in each country, where practical initiatives have been taken to bring about lifestyle changes through the whole of the food chain – from primary agricultural production, via processing, distribution and storage to final consumption – based on ecological production (agriculture and processing), recycling and a minimisation of transport systems.

Methodological approach The BERAS – project include five Work Packages (WPs). The first WP (1) promote selected, already established local ecological food initiatives and recycling farms in each country and exchange of experience with other initiatives in and between the project countries. Obstacles are identified and learning how solving such problems and developing initiatives further are promoted through confrontation with other similar initiatives and with the experts participating. The second WP (2) study what environmental benefits can be achieved through local ecological consumption, processing and ecological integrated recycling farming, in comparison with conventional food systems. Energy consumption, greenhouse gas emissions, surplus and emissions of reactive nitrogen (air/water pollution) and surplus of phosphorus compounds is quantified for the agriculture–society system and related to the food consumption. The main part of this work is done for Sweden and in more limited extent for Finland and in lower degree for the other BERAS countries related to the amount of economical resources for the project, but the results can be utilised for all countries. The third WP (3) evaluates market economic aspects, economic consequences at the societal level and consequences from a natural resource economy and the fourth WP (4) the consequences at the societal level including rural development and job opportunities. The fifth and final WP (5) will present an Agenda with recommendation for implementations and dissemination.

Locations and criteria of activities Studies of the whole food systems are located in: (1) Järna, Stockholm county in Sweden; (2) Vassmolösa, Kalmar county in Sweden; (3) Juva, Mikkeli county in Finland; (4) Fyn county, Denmark; (5) Bioranch Zempow Brandenburg in Germany; (6) Kluczbork and (7) Brodniza in Poland; (8) Raseiniai in Lithuania; (9) Organic farmer organisataion in Aizkraukle district in Latvia; (10) Pahkla Camphill village, Prillimäe in Estonia. Recently were 35 reference farms selected. They were mainly chosen for calculations of plant nutrient balances and were

3 selected to cover the main environmental and farming conditions and together also include a big part of the basic food production in the countries respectively. The criteria of the BERAS farms was that that they fulfil a high degree of plant nutrient recycling based on a balance between animal and crop production and a low part of purchased fodder.

Dissemination Farmers involved in the project were continuously informed of results achieved and receive support for their activities but also dissemination through publications and adviser service organisations. Dissemination of results and promotion of further similar initiatives will be aimed at regional and local policymakers (special working with spatial planning), the food industry and distribution, and, at the grassroots level, at consumers, farmers and small-scale actors along the food chain, to encourage a transition to a more sustainable lifestyle in the agro-food sector.

BERAS Work Package 2 – environmental assessment

Common scope and aim BERAS – project is a research and developing project with the goal to promote reduced consumption of limited resources, emissions of nitrogen and phosphorus pollution to the Baltic Sea area and emissions of greenhouse gases and to promote the biological diversity of life within the food chain. This is realised through evaluation and demonstration of the potential of ecological recycling-based agriculture, local and regional processing, distribution and consumption. Already established initiatives in the countries involved are utilised and promoted during the project period.

The evaluation of the potential for reduced losses of nitrogen and phosphorus is in this project mainly based on plant nutrient balances. The nutrient balances give information about the potential risk for leaching and the potential risk of nitrogen emissions to the atmosphere when assuming a steady state in soil over a longer period of time (Granstedt et al. 2004). A reduction of the surplus of nutrient with more than 50 % in a longer perspective is necessary if we want reduce the leakage and also emissions of reactive nitrogen to the atmosphere at the same level.

For calculation of the efficiency of recycling, Pentti Seuri developed a method for primary nutrient efficiency in the BERAS project. The results of plant nutrient balances on farm level (i.e. farm gate balances) are complemented by this measure for nitrogen, which gives information of how much more nitrogen is harvested in the yield than is imported in external resources. These results are also supplemented with measurements of nutrient leaching to compare with the calculated values. This needs to be studied for a longer time after that this project period is finished. To give full results, a scenario study of a conversion of the whole agricultural system is done on the Danish island of Fyn to a purely organic and nutrient self- supporting system.

In the BERAS report 2 (Granstedt, et al 2004) it was emphasised that the consequences of mistakes in terms of agricultural specialisation must be taken into consideration in the new EU countries that are about to introduce changes in their agricultural sectors. Otherwise, there may, according to the final conclusions of the INTERREG IIIB BERNET project Baltic Eutrophication Research Network (Fyn’s Amt, 2002), be a dramatic increase in loadings also from these countries like Poland and the Baltic states,

4 The potential for reduction of losses from ecological recycling agriculture farms (ERA-farms) are studied under different agricultural conditions and production specializations in the eight BERAS countries. The studies have also included studies of local or regional processing, distribution and consumption. Results from Järna in Sweden and Juva in Finland are presented in this report, which have influences also on global warming+ and use of fossil energy resources. In this part of the BERAS study the possibility of recycling wastes from local food processing, distribution and big kitchens is essential for higher degree of recycling of nutrients and reduced emissions to water and also to the atmosphere.

Partners and researchers BERAS coordinator: Artur Granstedt, [email protected] WP2 coordinator: Olof Thomsson, [email protected]

Researchers: Assoc. prof. Artur Granstedt, PhD Thomas Schneider, PhD (plant nutrient emissions), [email protected] Olof Thomsson, PhD Winfried Schäfer, PhD Christine Wallgren, MSc Pentti Seuri, MSc Hanna-Riikka Tuhkanen, MSc Miia Kuismaa, MSc Annamari Hannula, MSc Tiina Lehto, MSc Jaroslaw Stalenga, PhD Assoc. Prof. Jozef Tyburski, PhD Holger Fischer, MSc Ib Sillebak Christensen, PhD Ole Tyrsted Jørgensen, MSc

Principles of ERA systems

The concept of ecological recycling agriculture (ERA) was presented in the BERAS- background report (Granstedt et al. 2004) as an alternative to conventional agriculture. ERA agriculture will produce food and other products on following basic ecological principles: 1. Diversity of life 2. Renewable energy. 3. Recyling of plant nutrients

In conventional agriculture we have the problem of damaging of the biological diversity through use of pesticides, consumption of energy resources, emissions of greenhouse gases and other input resource as a consequence, using up limited plant nutrient resources and to high emissions of plant nutrients in the environment .

The reduced recycling and increased surplus of nutrients caused of this separation increased after 1950 and culminated in Sweden, Finland and Denmark until 1980 and are after that on same level. This is illustrated through examples of nutrient balances in the background report as a more and more linear flows of nutrients. During the same more special forms of agriculture based on non renewable energy and use of pesticides was introduced.

5 An ideal frame for an ERA farm (or farms in close cooperation similar to one farm unit) is adapted to the human consumption. For our European food consumption it is about 85 % of farm land used for fodder production and animal production (animal density) limited to this own fodder production and about 15 % of farm land used for crop production for sale according to the example from Skilleby-Yttereneby farm in Järna presented in the background report (Granstedt et al 2004). The lower demand of external input of nitrogen depending on higher degree of internal recycling within the system is covered through biological nitrogen fixation of mainly clover/grass leys. For phosphorus and potassium it is only a limited deficit in the balance between input and output. Most of the amount the minerals are recycled with manure within the farm. The limited net export of phosphorus and other nutrients seemed to be compensated by the withering processes in most soils but a recycling of food residues could decrease these losses from the system further (Granstedt, 2000).

During practical condition we have a variation between farms and it is difficult to find the ideal ERA-farm in reality. It is necessary to set up realistic limits for inputs of external resources which can be realised in practise but that also fulfil the need of high internal recycling with the potential of reduced losses as a consequence. Even if the ideal concept is total self sufficiency with fodder some smaller inputs of seeds and fodder stuffs must be accepted under real conditions. A limitation of purchased fodder was set to 15 % of total used fodder for selection of the BERAS-farms. In addition to that it is also necessary that a part of the agricultural land is used for bread-grain or other cash crops for direct human consumption. Not all the selected farms fulfilled this condition during the whole project time (2002-2004) and this will be further discussed together with the results.

The principal difference between conventional agriculture and the ERA-farms on system level are that in conventional agriculture crop and animal productions are more or less separated between different groups of farms – on ERA-farms animal and crop production are integrated within each farm. Comparison have be done between these two systems (the separated conventional and integrated ecological system) but based on the same average animal density for all the farms together in the compared systems respectively and related to the human consumption of both animal and vegetable agricultural products. For Sweden this comparison was done with the help of a conventional food basket and an alternative food basket based on a more locally produced and vegetarian diet.

In these studies the potential of reduced nutrient surplus and losses of nitrogen based on ERA-farm criteria will be evaluated during the different conditions in the production areas in the eight BERAS countries. Depending on the closed relation between animal density and surplus of plant nutrient which result in losses, the comparing with the conventional agriculture system with separated crop and animal production will be based on the same average relation between animal and vegetable crop production.

6 Results and discussion

Nutrient balances

Influence of production systems (Chapter 2, 3, 4) Potential of ERA (ecological recycling agriculture).

Surplus of nitrogen was 50 % lower per ha and year on Swedish and Finnish B-ERAS-farms compared to the average agriculture (36 kg compared with 79 for Swedish and 38 kg compared to 73 for Finnish agriculture with the same animal density of 0.6 au/ha) and it was no surplus at al of P. For all studied 36 farms in the eight BERAS countries the average surplus of nitrogen was 36 kg compared to the average of 56 kg for the eight BERAS countries which include the low intensive agriculture in the Baltic countries and Poland. All BERAS-farms with an animal density below 0.75 AU per ha have a surplus below 50 kg N/ha and can be defined as ERA-farms with animal production based on minimum 85 % own fodder.

An estimated average loss of 30 - 40 % NH4 of the nitrogen exudates from the animal production gives a calculated reduction of nitrogen leaching with 70 - 75 % on BERAS – farms compared to the average Swedish agriculture farms ( an average N- leakage of 7-9 kg compared with 28 – 30 kg). The same calculation made for the all eight BERAS countries gave a reduction of nitrogen leaching with 45 - 50 % on BERAS – farms compared to the studied average Baltic Sea agriculture which include regions with until now very extensive agriculture.

Two scenarios were calculated for the BERAS countries, business as usual where Baltic Countries and Poland convert to the same structure and use of resources as Sweden and Finland and the other alternative where the whole studied Baltic Sea agriculture convert to ERA agriculture like the BERAS farms. The conventional scenario gave an increased pollution of Nitrogen and Phosphorus from agriculture. The Nitrogen leakage would increase with 50 percent. The ERA scenario gave a reduction of nitrogen leakage from agriculture with 50 % and an elimination of the surplus of Phosphorus. Phosphorus leakage is specially related to farms with high animal density.

The conventional scenario would give an increased production of crops for direct human consumption of about 30 % and increased animal production of about 10 %. The alternative BERA Scenario with a reduced nutrient leakage should result in a decreased crop production of 20 % and decreased and more ruminant meat dominated animal production of 30 % in the BERAS-countries. The lower production in the BERAS scenario would reduce the overproduction in Europe. Using agriculture land in the Baltic countries that today is not used could compensate this production loss.

One scenario special for Fyn county was developed based on an earlier bigger study about converting the whole agriculture in Denmark to organic farming according ERA principles with total self-sufficiency with fodder and production related to the national food consumption. This gave, like the scenario for the whole BERAS study region, a reduction of the nitrogen surplus with 50 % but here related to Danish conditions. The balance contributed to 35 kg N-leaching per ha, around a 50 % reduction compared to the present situation in year 2002 in Denmark. In the scenario without external P-inputs the P-balance is about – 6 kg P/ha.

Global warming and energy use (Chapter 5)

The global warming impact, measured in Global Warming Potentials (GWP) as CO2 equivalents, is reported per kg products exported from the farm and per hectare. In both cases the GWP were lower for the average BERAS-farm than for the average Swedish agriculture (80 and 82 % lower respectively). The difference was reduced by the larger emission of methane from animals. The larger share of ruminant animals explains the difference on the BERAS-farms, compared to average Swedish agriculture having a larger part of monogastric animals which emit very little methane.

The consumption of primary energy counted per kg product and per ha is substantially lower on the BERAS-farms in average compared to average Swedish agriculture (53 and 57 % respectively). The most important reason is the lesser use of fire oil (for drying of grain), no use of imported fertilisers and lower use of electricity.

One sub-study describes how locally produced and consumed food in the food system case Järna, Sweden is transported, and how much fossil energy this transportation has used. The study comprises vegetables, potatoes and root crops from four local growers; milk and dairy products from Järna Dairy; bread from Saltå Mill & Bakery; and meat from animal keepers in the vicinity of Järna. The products are collected at the different producers and delivered to stores, schools and other large kitchens in both the Järna area and in Stockholm (about 60 km away). Counted per kg bread both the GWP and use of primary energy were lower for the average local bread distribution from Saltå Mill and Bakery (30 and 38 % respectively). Counted per kg vegetables both the GWP and use of primary energy were lower for the average local distribution from the gardens and farms (96 and 92 % respectively). In contrast counted per kg meat from animal keepers in the vicinity of Järna, depending on longer distances to the regional slaughters, both the GWP and use of primary energy were higher. Counted per kg milk products both the GWP and use of primary energy were lower for the average local distribution from Nibble dairy in Järna (77 and 19 % respectively).

Improving nutrient balances by waste management (Chapter 6) An inventory was conducted at Juva municipality in Finland to determine the possibilities for recycling biowaste (organic waste) produced by local food actors back into agriculture. The examined food actors belong to the food chain after primary production (agriculture): food processors, grocery stores, schools, municipal kitchens, and private consumers. The research methods used are waste flow and substance flow study.

Alternative treatment processes to recycling of nutrients and humus of biowaste are composting and biogas treatment (centralized or small-scale treatment or co-digestion plant). The conventional treatment of biowaste and wastewater cause nutrient losses and the most nutrients in the treated biowaste do not become available to plants. The composting process mainly affects the amount of nitrogen. The loss of nitrogen in composting can be as much as 50%.

From the environmental viewpoint, separate urine collection would be preferable to the conventional system. It would increase the recyclable nutrient amounts to fields and at the same time decrease the emissions into water systems. In addition, phosphorous would be in

8 more usable form for plants and risk of heavy metal contamination would be less compared to sewage sludge.

The small scale biogas plant on farm level established as a prototype in Järna seemed to be the best solution for recycling of the solid fractions of nutrients from the human food sector (local processors, ecological public kitchen), to reduce emissions of greenhouse gases and also have a good control against contaminations of pathogens and harmful substances.

Influence of consumption patterns (Chapter 7, 8) Results from food basket surveys in Sweden and Finland

Our food habits are, unquestionably, important both for the health and for the environment. That is recognised as one of the starting points in BERAS.

The main objective of the consumer surveys was to put together realistic food baskets (consumption profiles) for a Swedish and a Finnish case, containing mainly locally and ecologically produced foodstuffs. The case studies were performed in Juva, Finland and Järna, Sweden – the same sites used for many other studies in the BERAS project. For a more extended presentation of the sites see Seppänen (ed. 2004). Both the municipality of Juva and the community of Järna have around 7500 inhabitants. We conclude that the consumption profile of the participating households in Sweden differs more from the average Swedish consumers than was the case in the Finish study. However, also the Finnish households participating in this study buy more organic products than ordinary Finnish households.

The food consumption profile of the Järna households seems to follow the diets suggested in the Nordic Nutrition Recommendations (NNR, 2004) and in the S.M.A.R.T. project (CTN, 2004). These households buy a larger share of vegetables (less meat), less “empty” calories, more ecological food, “right” vegetables (e.g. more legumes and root crops, and less lettuce and cucumbers) and less transported food, compared to the national average food.

Scenarios are presenting results for Nitrogen surplus, Global Warming Potentials and Consumption of primary nutrient resources. The scenarios are: 1. Average Swedish food consumption, Average Swedish agriculture 2002-2004, and Conventional food processing and transports 2. Average Swedish food consumption, ERA farms, and Conventional food processing and transports 3. Average Swedish food consumption, ERA farms, and Local (small-scale) food processing and transports 4. Eco-local food consumption (presented in Chapter 10), ERA farms, and Local (small- scale) food processing and transports

One scenario based on BERAS farms and with the same total meat consumption (but with a higher share of ruminant meat) will have an unrealistic acreage of arable land in Sweden as a consequence of the lower gross productivity on the studied BERAS farms. The conventional system use imports from farms outside Sweden but this aspect is not included here. However, in a study based on Järna type alternative consumption pattern and BERAS farms the need of agricultural arable land would instead decrease with almost 40 % compared with the situation of to day (figure 8-1, 8-2 and 8-3).

Nutrient balances in Scenario 2 and 3 with food from only ERA-farms gave a nitrogen surplus of 14 kg and year compared with 22 kg for average agriculture counted per capita or 26 kg compared with 80 kg counted per ha (a reduction with about 35 % and 65 % respectively). Scenario 4 with reduced agricultural arable land gave a surplus of 8 kg counted per capita and 42 kg counted per ha and year (a reduction with about 65 % and 45 % respectively). The losses of ammoniac from manure is about the same in the compared systems and results in lower losses and leaching of nitrogen from fields from the ERA farms compared to the conventional system than the above figures.

Global warming calculations in scenario 2 with food from only ERA-farms would give a global warming potential of 800 kg CO2 equivalents per capita compared with 900 kg CO2 equivalents per capita and year in scenario 1 with conventional consumption. Scenario 3 with food from only ERA-farms and local processing and distribution gave a global warming potential of 700 kg CO2 equivalents per capita compared with 900 kg CO2 equivalents per capita and year. Scenario 4 with food from only ERA-farms, more vegetable and less meat consumption and local processing and distribution gave a global warming potential of 500 kg CO2 equivalents per capita compared with 900 kg CO2 equivalents per capita and year (a reduction with about 45 %).

Primary energy calculations in scenario 2 with food from only ERA-farms and 3 with also local distribution and processing gave a consumption of primary energy resources of 5 GJ per capita compared with 8 GJ in the average food system (a reduction with about 40 %. Scenario 4 with food from only ERA-farms and local processing and distribution and eco local (more vegetable and less meat consumption) consumption gave a consumption of primary energy resources of 3 GJ per capita and year (a reduction with about 60 %).

General conclusions

Already established initiatives with Ecological Recycling Agriculture (ERA) and local food systems in the eight EU-countries in the Baltic Sea drainage area are evaluate during the project period. Surplus of nitrogen was 50 % lower per ha and year on Swedish and Finnish BERAS-farms compared to the average agriculture (36 kg compared with 79 for Swedish and 38 kg compared to 73 for Finnish agriculture with the same animal density of 0.6 au/ha) and it was no surplus of P. For all studied 36 farms in the eight BERAS countries the average surplus of nitrogen was 36 kg compared to the average of 56 kg for the eight BERAS countries which include the low intensive agriculture in the Baltic countries and Poland.. All BERAS-farms with an animal density below 0.75 AU per ha have a surplus below 50 kg N/ha and can be defined as ERA-farms with animal production based on minimum 85 % own fodder.

10 studied average Baltic Sea agriculture which include regions with until now very extensive agriculture.

Scenarios are presenting results for Nitrogen surplus, Global Warming Potentials and Consumption of primary nutrient resources. Global warming calculations in scenario 2 with food from only ERA-farms would give a global warming potential of 800 kg CO2 equivalents per capita compared with 900 kg CO2 equivalents per capita and year in scenario 1 with conventional consumption. Scenario 3 with food from only ERA-farms and local processing and distribution gave a global warming potential of 700 kg CO2 equivalents per capita compared with 900 kg CO2 equivalents per capita and year. Scenario 4 with food from only ERA-farms, more vegetable and less meat consumption and local processing and distribution gave a global warming potential of 500 kg CO2 equivalents per capita compared with 900 kg CO2 equivalents per capita and year (a reduction with about 45 %).

11 2. Plant nutrient balance studies Artur Granstedt (where nothing else is stated)

Background and problems This project focuses on the potential of reducing nitrogen and phosphorus load to the Baltic Sea by increasing the efficiency of recycling within the agricultural system, according to the principles of organic farming with an integration of crop and animal production with self- sufficiency of fodder. The BERAS background report, Effective recycling agriculture around the Baltic Sea (Granstedt, et al 2004) gives an overview of the main objectives of the current project and the problems to be solved.

The Baltic Sea countries have undertaken internationally efforts, within HELCOM and OSPARCOM, to halve their discharges of nitrogen and to reduce their discharges of phosphorus from human activities. The countries have not achieved these goals during the target period 1987 – 1995 and no improvement has been observed until the year 2000 according to the Executive Summary of the Fourth Baltic Sea Pollution Load Compilation. Measurements in streams to the Baltic Sea show no significant decrease of the total load (HELCOM, 1998; 2004).’

BERAS countries in this project include all countries within the Baltic Sea drainage area which are EU-member and exclude the regions belonged to Russia and Belarus. The focus is to reduce the loadings of nitrogen and phosphorus to surface waters and the Baltic Sea. Well established ecological recycling model farms focusing on different production areas has been used to document to what extent emissions can be reduced and to serve as possible examples in the eight BERAS countries. The measures were taken account present growing conditions, the risks of nutrient losses, and the proximity of watercourses, lakes and sea areas.

The ultimate aims is also to finally analyse possible changes in agricultural structures which will maximize recycling and minimize losses of nitrogen and phosphorus compounds on regional and country level and for the whole Baltic Sea drainage area. This is the reason way we in this introduction analyse the present and very different situation in the countries included in this project

Nitrogen and phosphorus load in the Baltic Sea drainage area Total nitrogen input to surface waters within the Baltic Sea catchment was 822 kt in the year 2000 of which diffuse sources (mainly agriculture) contributed with 58% according to HELCOM (2004). Highest inputs were in Poland, Sweden and Finland (Figure 2-1 A). However, nitrogen load per capita is highest in Denmark, Sweden and Finland (Granstedt et al, 2004). Even nitrogen surplus in agriculture is higher in these countries due to higher fertilizer usage and livestock density compared to countries like Latvia, Lithuania or Estonia (Figure 2-1 B).

12 230

) 250 r A a

e 200 176 y / t 147 k (

150 r e t a w

100

e 62 c 54 a f

r 36 33 32

u 50 s

N 0

) 160

r 139 a e

y B

120 a

h 78 / 80 75 g 56 56 k (

s u l

p 40 25 r

u 12 s

N 0 -7 k a y d n d a a i r i i n n e n v n n a t a d a a l l a o a t e m o n u m i L s n r w h P F t e E e i S D L G Figure 2-1. Total nitrogen input to surface water (A) and nitrogen surplus in agriculture (B) in countries participating in BERAS. Data in (A) from HELCOM (2004) and data in (B) from Granstedt et al (2004). The distribution of sources differs slightly in the different countries but the pattern is similar. In Sweden about 60% of total nitrogen load to the Baltic Sea is anthropogenic and about half of the anthropogenic load originates from agriculture (Figure 2-2) (Brand and Ejhed 2002).

200 184 Gross load

) Net load r 135 a

e 114 y / t

k 100 79 (

61 d a o l 39 N 0 Total Anthropogenic Agriculture Figure 2-2. Mean annual nitrogen load in Sweden for the period 1985-1999 (Data from Brand and Ejhed, 2002). Total – includes total load to surface water and directly to the Baltic Sea, Anthropogenic - includes all anthropogenic emissions (incl all diffuse and point sources). Gross load is total load to surface water and directly to the Baltic Sea and net load is nitrogen load to the Baltic Sea after retention. The figures of nitrogen leakage from agriculture can be compared with calculated surplus of nitrogen calculated according to the farm gate method. Swedish agriculture produces a nitrogen surplus of 78 kg/ha year which corresponds to 184 kt/year (Granstedt et al., 2004). Notice that N-surplus from agriculture in Sweden (Figure 2-1B) is as large as gross load of nitrogen to surface water and directly to the Baltic Sea (Figure 2-2).

Gross leakage from Swedish agriculture to surface water is 61 kt/year of N-surplus (33%). The remaining 123 kt/year consist of ammonia losses from manure and animals (56 kt/year, 31%) (Granstedt et al., 2004), leakage to groundwater (12 kt/year according to Brink, 1990). Gaseous losses due to ammonia evaporation at the soil surface and denitrification processes in the soil equal the remaining 55 kt/year (36%) assuming steady state of mineral and organically fixed nitrogen in the soil.

13 Although calculated in a slightly different way, these results agree well with the results of the Swedish Statistics Agency (SCB 2000. 2002). SCB calculated nitrogen surplus from farm gate and field balances (surface balances) to 186 kt/year and N-leakage to 63 kt/year. N-losses from animal and manure can be calculated from the SCB data to 56 kt/year, as the difference between fodder production on one side and the production of animal products and manure on the other side by taking into account fodder production losses of 5%.

An overall reduction of the nutrient load to the Baltic Sea by 50% is one of the national and international agreed environmental goals for the Baltic Sea Region (HELCOM, 2004). This implies different strategies for the different countries. In countries with nutrient intensive agriculture like Sweden, Finland or Denmark loads have to be decreased, whereas in countries with today nutrient extensive agriculture, like Estonia, Latvia and Lithuania the development of agriculture towards more nutrient intensive methods has to be prevented. The overall goal of the nutrient studies in the BERAS-project is to investigate the potential of ERA to reach the proposed reduction of nutrient leakage from agriculture.

Material and Methods The study is based on plant nutrient balance calculations for agriculture in the whole countries respectively in total figures and per ha. These values are representative for the average in the countries respectively and were compared to the selected BERAS farms in the countries respectively.

Selected BERAS-farms ERA farms were selected in all BERAS countries to evaluate the potential to reduce nutrient surplus and losses from agriculture in the Baltic Sea drainage area. The test-farms were selected so they cover the main agricultural conditions and drainage regions in the area (Figure 2-3). Characteristics of the farms are presented in Appendix 1. The initially 35 farms were complemented with farms in Finland (the region of Juva close to Mikkeli) and in Järna in Sweden for more detailed studies and to cover different type of agricultural production so it finally was 50 farms included. During the project time some of the farms did not fulfil the preconditions for ERA farms and the consequences of that is discussed together with the evaluation of results.

Figure 2-3. The Baltic Sea drainage basin with locations of farms included in BERAS.

15 Field and farm gate balances The methods for calculating nutrient balances follow the concept described in earlier publications (Granstedt, 2000; Granstedt et al 2004) and are summarized below. The difference between this input and output of plant nutrients is here defined as surplus of plant nutrients and is the same as potential losses. For estimation of potential nutrient losses, plant nutrient balances can be calculated at field level (field balances) or for whole farms (farm gate balances).

Farm gate balances are based on the difference between the import (input) of fertilizers, imported fodder, precipitation of atmospheric nitrogen and the export (output) of agricultural products from the farm. This method can also be used to calculate balances for larger systems such as administrative regions or drainage areas, e.g. the Baltic Sea drainage area.

Field balances are based on the difference between input and output of plant nutrients on field level with the use the amount of manure and fertilizers for input data and the amount of harvested crops for output data.

SCB (2000. 2002) in Sweden has present regularly both type of balances, farm gate balances for the whole Sweden based on agricultural according similar methods used of Granstedt et al for the eight countries around the Baltic Sea and field balances (of SCB called surface balances) based on collected information direct from the Swedish farms.

Nitrogen fixation and input trough atmosphere deposition En important part of the nutrient balances is the input through nitrogen fixation. The nitrogen fixation on the farm in Sweden, Finland (F-BERAS farms), Estonia, Latvia Lithuania and Poland was estimated with the used calculation programme for nutrient balances Stank 2:1 (Jordbruksverket, 2000) and the collected data on yield level an clover percent in the clover grass fields. The clover percent was estimated in field combined with cutting of proofs for calibrating of this estimation and done of the same person in the above countries during the actual years in the countries respectively. In estimating the contribution of nitrogen fixation on country level it was assumed that 25 % of the ley area in the form of a first-year ley with legumes producing 100 kg fixed nitrogen per ha. Fodder peas and beans were assumed to fix 50 kg nitrogen/ha. The figures for nitrogen deposition are based on measurements of wet and dry deposition made by the Environmental Research Institute for the years respectively according to Granstedt (2000).

The estimation of nitrogen fixation in Denmark and Germany is done according the there used methods and the estimated surplus of nitrogen can deviate from the other countries of this reason. The given fixation in Denmark is calculated as in Nielsen & Kristensen (2005): Results of plant nutrient balances in the BERAS countries

Sweden The average results of nutrient balance calculations 2002-2004 on farm level is presented in Figure 2-5. For nitrogen, the results are presented for all the three years separately in Figure 2-4. For more detailed results, also for separate years, see Appendix 1. The average nitrogen surplus on the BERAS farms was in the range 35-37 kg N per ha and year during the study period (Figure 2-4). Calculated all the 12 farms together and divided with total area of the BERAS-farms, the average surplus of N was 39 kg per ha and year. This value can be

16 compared to the average for Swedish agriculture which was calculated to 79 kg per ha and year for 2000-2002 (Figure 2-6).

90 90 90 2002 Surplus 2003 Surplus 2004 Surplus 80 80 80 Anim Anim Anim 70 Veg 70 Veg 70 Veg 60 60 60 12 12 15 50 50 50

40 78 40 74 40 77

N (kg /ha year) 30 N (kg /ha year) 30 N (kg /ha year) 30 49 48 47 20 20 20 16 37 14 35 14 37 10 10 10 8 11 12 0 0 0 Input Prod Export Surplus Input Prod Export Surplus Input Prod Export Surplus

Figure 2-4. Average input, plant production, export and surplus of nitrogen (N) on the BERAS-farms in Sweden for the years 2002, 2003 and 2004.

120 10 40 N Surplus P Surplus K Surplus 35 100 Anim 8 Anim Anim Veg Veg 30 Veg 80 6 25 9 60 4 20 37 13 3 15 P (kg /ha year) K (kg / ha year) N (kg /ha year) 40 76 2 3 2 10 48 1 20 15 36 0 6 3 -1 5 10 3 4 0 -2 0 2 Input Prod Export Surplus Input Prod Export Surplus Input Prod Export Surplus

Figure 2-5. Average input, plant production, export of farm products and surplus of nitrogen (N), phosphorus (P) and potassium (K) on the BERAS-farms in Sweden 2002-2004.

120 12 N Surplus P Surplus 100 Anim 10 An.prod. Veg V.prod. 80 8 4

92 60 6 11 4 P (kg /ha year) N (kg /ha year) 40 79 4 70 6 20 18 2 3 3 18 13 0 0 Input Prod Export Surplus Input Prod. Export Surplus

Figure 2-6. Average input, production, export and surplus of nitrogen (N) and phosphorus (P) for Swedish agriculture 2000-02 (Granstedt et al 2004).. The calculated crop production in terms of N is higher but food production is about 25 % lower. This can be explained by the higher part of ruminant clover grass based animal production compared to conventional agriculture with domination of more grain converted meat production. It is a smaller rate between input of fodder protein and output of animal protein in grain converted meat production compared with clover grass based ruminant meat production. This only partly compensates the high yield in producing clover grass compared with grain. The grain production is significant lower on BERAS-farms compared with conventional agriculture. Surplus of nitrogen is in this comparisons more than 50 % lower on

17 Swedish BERAS-farms, with the same animal density of 0.6 au/ha but a little lower share of animal products of the total production in terms of N.

The variation between the farms was rather high with lower surplus on farms with low animal density and higher on farms with high animal density (Figure 2-7). All BERAS-farms with an animal density below 0.75 AU per ha have a surplus below 50 kg N/ha and can be defined as ERA-farms with animal production based on minimum 85 % own fodder.

The lowest values is the calculated surplus on the more extensive producing meat and cereal farms Oxsätra and Håknäs and the highest values for the intensive dairy farm Skogsgård which not fulfilled the conditions for an ERA-farm. 70

(kg / ha year) 30 surplus

N 20

0 0,0 0,5 1,0 Animal density (au/ha)

Figure 2-7. Average animal density and N-surplus 2002-04 for the Swedish BERAS-farms. The External Fodder Rate (EFR) and surplus of nitrogen per ha and year for the three years period for the BERAS-farms is presented in Figure 2-8. Two farms, one of the dairy farms (Skogsgård) and the pig farm (Davidsta), bought more than 15 % of the fodder (but less than 20 %) and have a surplus of 66 and 44 kg N respectively and diverge more or less from the average value 36 kg N/ha for all farms. 70

(kg / ha year) 30 surplus

N 20

0 0 0,05 0,1 0,15 0,2 External fodder rate

18 Figure 2-8. External fodder rate and N-surplus 2002-04 for the Swedish BERAS farms.

The surplus was highest for the dairy farm Skogsgård which already in the beginning of the study period had too high animal density in relation to the criteria for an ERA-farm. During the study, the farm increased the animal density and the part of purchased fodder further to an average animal density of 0.9 AU per ha and an average surplus of 70 kg N per ha (Figure 2-9). Thus, it does not fulfil the condition for an ERA-farm.

The pig raising farm Davidsta developed in the opposite way and decreased the animal density and need of purchased fodder during the study period – with a decreased surplus of nitrogen as a consequence (Figure 2-9).

1,2 90

80 1 70 Skogsgård 0,8 60

50 0,6 40

0,4 30 N surplus (kg / ha year)

Animal density (au / ha) Davidsta 20 0,2 Animal density 10 N-surplus 0 0 2002 2003 2004

Figure 2-9. Surplus of nitrogen on two farms with different animal densities. Increased N-surplus on the dairy farm Skogsgård due to increased animal density and increased use of purchased fodder (EFR: 0.8; 0.20; 0.21) and decreased N-surplus on the pig raising farm Davidsta due to decreased animal density and decreased use of purchased fodder (EFR: 0.24; 0.16; 0.18).

The relation between animal density and surplus of nitrogen is illustrated in Figure 2-10. On the BERAS dairy farms with low animal density (0.65 AU/ha), the surplus was average 42 kg N per ha. On more specialised conventional dairy farms, illustrated by the average of 608 dairy farms (Myrbeck 1999), the average surplus was 131 kg N per ha. Not published data from the Swedish action program “Greppa Näringen” with about 6000 nutrient balances give nearly the same information with a surplus of 50 kg N/ha on crop farms and 133 kg N/ha on dairy farms.

19 160

140 CA ERA 120

100

80 (kg / ha year) 60 surplus

N 40

0 0,0 0,5 1,0 1,5 2,0 Animal density (au/ha)

Figure 2-10. The low animal density on the BERAS dairy farms (ERA) resulted in average 42 kg N surplus per ha from this study, compared to average 131 kg N surplus per ha on 608 conventional dairy farms (CA) published of Myrbeck (1999). The study of Myrbeck (1999), with an evaluation of more than 1000 conventional farms (dairy and specialised crop farms), is until now the largest study performed and it is representative for the Swedish agriculture also for the three BERAS-study year 2002-2004 since the total surplus of plant nutrients is on the same level. Figure 2-11 shows the lower surplus of nitrogen on four groups of BERAS-farms compared to the corresponding four groups in conventional agriculture. The lower surplus of nitrogen on BERAS-farms is associated with described lower animal density on the animal farms, i.e. lower degree of specialisation and higher degree of integration of crop and animal production. 140 1.2 CA ERA 120 1 100 0.8 80 0.6

(kg / ha year) 60 0.4

surplus 40 N 20 0.2 Animal density (au / ha)

0 0 A B C D Farm type

Figure 2-11. Surplus of nitrogen in four groups of BERAS-farms (ERA) and corresponding conventional farms (CA) (Myrbeck 1999). Dots show animal density. A - Mixed production with vegetables (2 farms), B - Milk, meat and cereals (6 farms), C - Cereals and ruminant meat (2 farms), D - Pork, poultry, egg and cereals (2 farms). Based on the above presented results from the BERAS farms are field balances (Field balance = N-manure + N-fertilizers + N-fixation + N-deposition + N-fixation + N-deposition - N-crop production) and possible losses trough denitrification and leakage calculated with the assumption of a steady state of stored amount nitrogen in soil. The calculation is done with

20 two alternative levels of NH4 emissions from the animal production and manure management. One calculation is based on an estimated amount of N in manure based on an assumed emission of 30 % and one calculation is based on an assumed emission of 40 % of the difference between fodder consumption and animal production (Figure 2-12). This calculations gave a surplus of nitrogen to the fields of 20 and 14 kg N per ha respectively which can be compared to the same assumptions made for calculation for the whole average Swedish agriculture which give 68 and 58 kg per ha in surplus respectively. An assumed drainage leakage of 48 % the field balance surplus give a theoretical nitrogen leakage from the fields of 9 and 7 kg N per ha respectively according the discussion on page 10. This can be compared to the same assumptions made for calculation for the whole average Swedish agriculture which give 30 and 28 kg per ha in leakage respectively. This will give a reduction of nitrogen leaching with 70 and 75 % respectively on BERAS – farms compared to the average Swedish agriculture farms. 90

80 NH4-loss 16 Denitrif. 70 22 60 Leakage

50 33 29 40

30 16 22 20 Nitrogen losses (kg / ha year) 30 28 11 10 7 9 7 0 CA 30 % CA 40 % BERAS 30 % BERAS 40 % Types of NH4-losses

Figure 2-12. The distribution of N losses of the calculated N surplus for Swedish average agriculture and the BERAS-farms with two different manure handling systems (resulting in 30 and 40 percent ammonia losses from the animal exudates respectively).

Finland Artur Granstedt and Pentti Seuri

The Finnish study was based on five BERAS-farms (called F-BERAS-farms); two in the cereal-dominated south part of Finland, one in central part of Finland (Tampere), one in the animal-dominated north-west part of Finland (Österbotten) and one farm in east part of Finland in Juva. This is extended with a deeper study of 8 organic farms located in the Juva region (called J-BERAS-farms).

The average nitrogen surplus on the five F-BERAS-farms range between 32-43 kg N per ha and year during the study period (Figure 2-13). For more detailed results, see Table A1- and Table A1- in Appendix 1.

21 100 14 25 N Surplus P Surplus K Surplus Anim 12 Anim Anim 80 20 Veg Veg Veg 10 60 15 8

16 6 40 10 20 P (kg /ha year) K (kg / ha year) N (kg /ha year) 4 4 9 49 20 35 38 5 2 8 2 3 5 9 2 1 0 4 0 1 0 0 1 Input Prod Export Surplus Input Prod Export Surplus Input Prod Export Surplus

Figure 2-13. Input, plant production, output of farm products and surplus of nitrogen (N), phosphorus (P) and potassium (K) on the F-BERAS-farms in Finland 2002-2004.

100 14 10 N Surplus P Surplus Anim 12 Anim 80 3 Veg Veg 10 60 8

90 6 40 73 10 10 P (kg /ha year) N (kg /ha year) 54 4 3 8 20 14 2 13 2 0 0 Input Prod Export Surplus Input Prod Export Surplus

Figure 2-14. The input, production, export and surplus of nitrogen (N) and phosphorus (P) in Finnish agriculture, averages per ha and year 2000-2002 (Granstedt et al 2004). Mean animal density was 0.6 au/ha. The average nitrogen surplus on the five Finnish F-BERAS-farms range between 32-43 kg N per ha and year during the study period and a range between the farms 27- 52. The average surplus of 38 kg N per ha and year can be compared to the average for Finnish agriculture which was calculated to 73 kg per ha and year for 2000-2002 (Figure 2-14) showing the surplus of nitrogen is twice the surplus on Finnish F-BERAS-farms with the same animal density (0.6 au/ha).

The calculated nitrogen fixation with deposition included was average 30 kg N per ha and the. calculated fodder production is 49 kg N per ha.

Surplus of P is 3 kg per compared with average 8 kg for the whole Finnish agriculture (Figure 12 c and 12 d. Most of the Finnish BERAS farms have like in Sweden a deficit for P in the balance and the potential for P is reduced compared to main stream agriculture.

The crop production in terms of N is only some percent lower but food production (vegetable and animal products) is about 40 % lower. This can also for Finland be explained by the higher part of ruminant clover grass based animal production compared to conventional agriculture with domination of for effective grain converted meat production.

The variation between the farms was rather high with lover surplus on farms with low animal density and higher on farms with high animal density (Figure 2-15). All BERAS-farms with an animal density under 0.7 AU per ha show a surplus lower than 52 kg N/ha.

22 60

30 (kg / ha year)

surplus 20 N

0 0 0.5 1 Animal density (au/ha)

Figure 2-15. Average animal density and surplus of Nitrogen per ha and year for the three years period for all the Finnish F-BERAS-farms.

The Baltic countries The selected BERAS-farms in the Baltic countries were studied during three years in Estonia and two years in Latvia and Lithuania. Nitrogen and phosphorus balances on these are presented in Figure 2-16. The surplus of nitrogen is here higher than average agriculture in Estonia (Figure 2-17). The common agriculture is in Estonia like also in Latvia and Lithuania very extensive and use of artificial fertilisers very low compared to Sweden and Finland. The statistic for nutrient balances in Latvia and Lithuania probably give very uncertain values with partly not used agricultural land and, therefore, not presented here. Also the organic agriculture are producing under extraordinary circumstances with areas of land, not optimal used and with a week relation between field production and harvested yield and finally production of sold animal and vegetable products from the farms. Several technical problems and also social problems can be the reason for that and need further developing work. Some farms just harvesting what land and soil can give.

The relation between animal density and surplus on the eleven studied farms is presented in Figure 2-18 and show a tendency with a higher surplus on farms with higher animal density. The animal density is on the organic farms low like in the average of the countries respectively rather low. It was a dramatic decrease of the animal production after 1990 in all this countries depending on extraordinary low prices of agricultural product after the collapse of the Soviet Union which earlier was the most important market.

If we exclude three of the BERAS farms with the most extreme surplus of nitrogen depending on high calculated nitrogen fixation in relation to utilized yield, the relation between surplus and animal density is more representative for the countries (the lower regression line in Figure 2-18). The surplus is then 31 kg N per ha and the animal density 0.3 animal unit per ha. That surplus is also more like the surplus for the BERAS-farms in Sweden and Finland.

23 60 7 25 N Surplus P Surplus K Surplus 6 50 Anim Anim 20 Anim 5 Veg 5 Veg Veg 40 15 4 22 30 3 6 10 45 2 P (kg /ha year) K (kg / ha year) N (kg /ha year) 20 41 5 36 2 1 2 2 3 2 10 0 0 0 0 0.5 7 -0.7 -0.5 0 3 -1 -5 Input Prod Export Surplus Input Prod Export Surplus Input Prod Export Surplus

Figure 2-16. Input, plant production, output of farm products and surplus of nitrogen (N), phosphorus (P) and potassium (K) on the BERAS-farms in the Baltic countries 2002-2004.

60 8 N P Surplus 7 Surplus 50 Anim Anim Veg 6 Veg 40 5 12 30 4 2

43 3 6 P (kg /ha year) N (kg /ha year) 20 2 27 1 10 9 23 3 1 2 7 1 0 0 Input Prod Export Surplus Input Prod Export Surplus

Figure 2-17. The input, production, export and surplus of nitrogen (N) and phosphorus (P) in Estonian agriculture, average per ha and year 2002.

100

90 80 70 60 50 (kg / ha year) 40 surplus

N 30 20

10 0 0.0 0.2 0.4 0.6 Animal density (au/ha)

Figure 2-18. Average animal density and surplus of nitrogen per ha and year for the three years period for the Baltic countries’ BERAS-farms. The red circles represent three farms which are excluded in the blue (lower) regression line.

24 Poland Artur Granstedt, Jozef Tyburskij and Jaroslaw Stalenga

In Poland 7 farms were selected, covering the main different farming conditions in the country during two years, 2003-2004. Only during the last year 2005 it was possible to visit all the farms to collect data and make fixation estimations depending of the extra economical resources BERAS-Poland got this year. But the farms were very well-known and representative for the farming conditions respectively. A special study on 20 organic farms in the district of Brodowice was done during 2002. The results were published in Polish.

The average surplus of the Polish BERAS-farms is presented in Figure 2-19 which can be compared to the figures for the average Polish agriculture in Figure 2-20. The calculated nitrogen surplus of 32 kg N per ha was 45 % lower compared the Polish nitrogen surplus of 57 kg N per ha. The export of phosphorus in agricultural was two kg lower than the input compared to the average Polish agriculture with a surplus of 19 kg P per ha. Big areas of the Polish agriculture is very extensive and other parts special in the north west part of Poland more intensive and more like agriculture in Western Europe. The high use of P-fertilisers is remarkable. Such an over-optimal use of P-fertilizers was earlier used also in Sweden and Finland but has decreased the last 20 years. 90 16 25 N P K 80 Surplus Surplus Surplus 20 Anim Anim Anim 70 12 Veg Veg 15 Veg 60 8 22 50 5 10

40 5 4 1

P (kg /ha year) 6

1 K (kg / ha year) 4 N (kg /ha year) 30 59 48 3 0 0 20 -5 9 32 0 0 -5 10 -3 12 0 -4 -10 Input Prod Export Surplus Input Prod Export Surplus Input Prod Export Surplus

Figure 2-19. The input, production, export and surplus of Nitrogen, Phosphorus and Potassium on the seven BERAS-farms in Poland 2003 – 2004. Animal density 0,6 au/ha.

90 16 N Surplus P Surplus 80 3 Anim 70 12 2 Anim Veg Veg 60 4 8 50 15 11 40 76 4 7 P (kg /ha year) N (kg /ha year) 30 57 57 20 0 13 -2 10 9 0 -4 Input Prod Export Surplus Input Prod Export Surplus

Figure 2-20. The input, production, export and surplus of Nitrogen, Phosphorus and Potassium in conventional agriculture in Poland 2000 – 2001. Animal density 0,6 au/ha.

25 Germany Artur Granstedt and Holger Fischer Plant nutrient balances of the two selected BERAS-farms in the Oder drainage area in Germany are presented in Figure 2-21. These figures can be compared with the diagram for Märkisch Oderland in Figure 2-22. The very low surplus on the two German BERAS farms of 16 kg N per ha can be evaluated in relation to the very low animal density representative for this regions. The surplus compared with conventional agriculture in the region is only 22 % of the conventional surplus of 74 kg N per ha in the region but also the agricultural production seem to be rather extensive compared with the conventional production with only15 kg N per ha in agricultural products compared with the conventional production of 67 kg N per ha. Like for the most other BERAS-farms is the P balance negative with higher export of P in agricultural products compared near zero input of P. 160 16 25 N Surplus P Surplus K Surplus 140 20 Anim 12 Anim Anim 120 Veg Veg 15 Veg 100 8 22 10 80 5 1 60 4

P (kg /ha year) 6

1 K (kg / ha year) 4 N (kg /ha year) 0 0 40 3 0 0 0 -5 20 -3 -5 31 30 3 12 16 0 -4 -10 Input Prod Export Surplus Input Prod Export Surplus Input Prod Export Surplus

Figure 2-21. The input, production, export and surplus of Nitrogen, Phosphorus and Potassium on the two BERAS-farms in Germany 2002 – 2004. Animal density 0,35 au/ha

160 16 N P 140 Surplus Surplus 22 Anim 12 2 Anim 120 Veg Veg 100 4 8 15 80 11 4 60 117 8 7 P (kg /ha year) N (kg /ha year) 40 74 67 58 0 20 -2

0 -4 Input Prod Export Surplus Input Prod Export Surplus

Figure 2-22. The input, production, export and surplus of Nitrogen, Phosphorus and Potassium in the region of Märkisch-Oderland in Germany 2002 – 2004. Animal density 0,4 au/ha

Denmark Ib Sillebak Christensen, Ole Tyrsted Jørgensen and Artur Granstedt

The BERAS-farms in Denmark is in this report only represented of one farm situated on Fyn. The surplus on this farm of 68 kg N per ha (Figure 2-23) is higher than on other BERAS- farms presented but 47 percent lower than the average Danish agriculture with a average surplus of 129 kg N per ha (Figure 2-24). The P balance is negative on the BERAS-farm which can be compared to an average surplus of 8 kg P per ha for the here presented year 2002 for the whole agriculture in Denmark. This is a consequence of input of both P in fodder and fertilizers higher than the output from agriculture in form of agricultural products in Denmark.

26 The surplus of nitrogen in a scenario of ERA agriculture on the whole of Fyn, presented in Chapter 5, shows a surplus of 57 kg N/ha and, like the most other BERAS-farms, no surplus of P. Total area: 24 40 200 N Surplus P Surplus K Surplus 20 Anim Anim Anim 30 Veg 16 Veg Veg 150 12 22 20 39 100 8 P (kg /ha year) K (kg / ha year) N (kg /ha year) 10 4 3 50 112 10 3 2 1 62 52 0 0 -2 11 7 6 0 12 -4 0 0 1 Input Prod Export Surplus Input Prod Export Surplus Input Prod Export Surplus

Cultivated area 35 60 200 N P Surplus K Surplus 30 Surplus Anim Anim 50 Anim Veg 25 Veg Veg 150 40 20

15 30 30 100 56 16 < 158 10 P (kg /ha year) K (kg / ha year) N (kg /ha year) 20 4 50 5 82 4 4 10 2 15 68 0 -2 11 10 20 0 -5 0 2 Input Prod Export Surplus Input Prod Export Surplus Input Prod Export Surplus

Figure 2-23. The input, production, export and surplus of Nitrogen, Phosphorus and Potassium on the two BERAS-farms in Denmark 2002 – 2003.

28 200 N Surplus P Surplus 24 Anim Anim 84 Veg 20 Veg 150

16 15 27 100 12 9 P (kg /ha year) N (kg /ha year) 120 126 129 8 50 46 4 8 8 6 29 0 0 Input Prod Export Surplus Input Prod Export Surplus Total area: 40 200 N P Surplus Surplus 35 Anim Anim 84 Veg 30 Veg 150 25

20 100 15 15 P (kg /ha year) N (kg /ha year) 27 129 120 126 9 50 46 10 5 29 8 6 8 0 0 Input Prod Export Surplus Input Prod Export Surplus Cultivated area: Figure 2-24. The input, production, export and surplus of Nitrogen and Phosphorus in Denmark 2002.

27 Discussion and conclusions Nitrogen and phosphorus surpluses, calculated ammonic losses and based on that calculated field losses and leakage for all BERAS-countries and the BERAS-farms are presented in Table 2-1 and Figure 2-25. The average nitrogen and phosphorus surplus is 56 and 11 kg per ha respectively in the BERAS-countries. On the selected BERAS-farms, the average nitrogen surplus was 35 kg N per ha which is 35 % lower and the calculated field losses was 50 % lower. Table 2-1. Total Load (according to HELCOM) and in the project calculated total surplus, field losses and leakage of N and P in average agriculture and BERAS agriculture based on BERAS farms presented in total figures, calculated per ha for the countries respectively and the total agricultural land area.

BERAS countries Average Agriculture BERAS Agriculture

Nitrogen Phosphorus Aver NH4 loss Nitrogen Phosphorus Year 2000 Arable landLoads 1000 haN t/a Loads of P T7aSurpl/ha- Suplus tot Surpl/ha-Suplus tot. per ha total Surpl/haSupl tot Surpl/ha- Suplus total Germany 2 051 31 510 1 880 74 151 774 -2 -4 102 9 18 893 16 32 816 -3 -6 153 Poland 14 247 229 990 18 760 57 812 079 19 270 693 15 216 825 32 455 904 -2 -28 494 Est/Lat/Lith 7 513 122 620 4 070 19 141 034 3 21 379 16 117 349 41 308 033 -0,5 -3 757 Finland 2 387 146 560 6 370 75 179 025 7 16 709 14 32 504 38 90 706 3 7 161 Sweden 2 698 175 610 7 320 79 184 000 3 8 094 22 58 350 36 97 128 -2 -5 396 Denmark 2 077 62 240 1 180 129 267 933 8 16 616 54 112 298 68 141 236 -2 -4 154 Total 30 973 768 530 39 580 56 1 735 845 11 329 389 556 219 36 1 125 823 -1 -40 793 Ammoniak loss 18 556 219 17 524 207 Field losses 38 1 179 626 1 16 469 19 601 616 0 -2 040

60 56 Farm gate balance

50 Field balance

38 40 36

19 20 11 10

Nutrient balance (kg / ha year) 1 0 0 -1 -10 AA - N AA - P BERAS - N BERAS - P

Figure 2-25. The average nutrient farm-gate and field balance in the agriculture of the BERAS-countries (AA) and on the BERAS-farms (BERAS). N - nitrogen, P - phosphorus. Average nitrogen fixation per ha for all BERAS-farms is 40 kg N per ha and year. Should nitrogen fixation be underestimated and nitrogen fixation 20 % higher should the calculated field surplus be 30 % lower and if nitrogen fixation should be overestimated with 20 % should the field surplus be 70 % lower compared to the average agriculture.

Based on these results, we have calculated one fully realistic scenario where Poland and the Baltic countries introduce an agriculture like the one we have in Sweden today and one scenario working in the opposite way with converting the whole Baltic Sea drainage area to ERA-type agriculture (Figure 2-26), where also the Baltic countries’ ERA-farms produce similar to the Swedish ERA-farms.

28 The first conventional scenario would give an 50 % increase of the field surplus of nitrogen and corresponding increase load to the Baltic Sea and probably also a similar increase of the phosphorus load. The consequences would be the opposite with a decrease of the nitrogen surplus with 50 % if the whole agriculture in the Baltic Sea area should convert to an ecological recycling agriculture (ERA scenario). This scenario would also result in a null- surplus of phosphorus which would result in a significant decrease of the phosphorus load from the agriculture to the Baltic Sea. 90 79 80 Farm gate balance

70 Field balance 60 60 56

50 38 40 36

30 19 20 11 13 Nutrient balance (kg / ha year) 10 1 1 0 0 -1 -10 N P N P N P

Todays situation Conventional scenario ERA scenario

Figure 2-26. Surplus of nitrogen an phosphorus in farm-gate and field balances calculated for three alternatives: The present agriculture (Today’s situation), if also Poland and the Baltic countries introduce an agriculture like the Swedish (Conventional scenario) and with converting the whole Baltic Sea drainage area agriculture to Ecological Recycling Agriculture (ERA scenario).

600 550

497 Crop production 500 447 426 Animal production

400 336 304 300

200 Nitrogen production (kt /year)

100

0 Todays situation Conventional scenario ERA scenario

Figure 2-27. Vegetables and animal production in the conventional Scenario and ERA scenario should give an increased and reduced agricultural production in the BERAS-countries respectively.

29 The output of vegetables and animal production in total kt nitrogen is calculated for the conventional Scenario and ERA scenario. The conventional scenario should give an increased vegetal production of about 30 % and increased animal production of abot 10 % (average + 15 %) and reduced. The alternative BERAS Scenario with a reduced nutrient leakage should result in a decreased vegetable production of about 20 % and an decreased animal production of about 30 % the BERAS-countries. In the second scenario is an higher part of the animal production based on grass fodder compared to the conventional production with a high part meat produced with grain and concentrate as fodder. The lower production in the BERAS scenario should reduce the overproduction in Europe to day but could also be compensated through using to day not utilized agriculture land in the Baltic countries.

References Brand and Ejhed, 2002 Brink, 1990. Granstedt et al., 2004 HELCOM, 2004. Myrbeck, 1999 SCB 2000.

Evaluation of nitrogen utilization by means of the concept of primary production balance Pentti Seuri, Helena Kahiluoto MTT Agrifood Research Finland, Huttulantie 1, FI-51900 Juva, Finland, Tel. +358 15 321230, [email protected], [email protected], www.mtt.fi Key Words: nitrogen utilization, nutrient balance, primary nutrient, secondary nutrient, nutrient loading

Abstract Primary production balance (PPB) indicates the ratio (*) between harvested nutrients and nutrients put into crop production from outside the system (=primary nutrients). The PPB is independent of the final output (crop vs. animal products) and can be used to evaluate different systems (farms). The utilization of nitrogen was evaluated by means of the PPB on nine organic farms, stockless and mixed ones, in eastern Finland. The mixed farms were able to reach a remarkably higher primary production balance (1.0 – 1.2) compared with the stockless ones (0.5). Values higher than 1.0 indicate recirculation of nutrients. Other components of the high PPB were legumes as nitrogen source and optimum livestock density.

Introduction Nutrient balances (farm-gate balance, surface balance or cattle balance) only indicate an absolute load of nutrients as a difference between input nutrients and output nutrients (kg or kg/ha); basically they do not tell anything about the efficiency of nutrient utilization.

It is also possible to count a ratio between output and input. Sometimes this type of ratio has been used as a measure of efficiency of nutrient utilization. As long as the system is simple enough, i.e. stockless farm without any recycling nutrients, the output/input ratio indicates the efficiency of nutrient utilization. However, as soon as a system involves recycled nutrients, the output/input ratios are difficult to interpret (Myrbeck 1999).

From an ecological point of view there is only one production process in the agricultural system, i.e. crop production or primary production. Primary production can be used either directly as human food or fed to animals. Nutrient load and nutrient utilization, i.e. efficiency to utilize nutrients, are two separate dimensions. If only crop products are produced, the nutrient load is less than an equal amount (kg nitrogen) of animal products, even the efficiency to utilize nutrients is equal. This is because more crop products are needed to produce an equal amount of animal products.

In order to reduce the nutrient load there are two choices: either to produce less or to improve the efficiency of nutrient utilization. Since the amount of primary production is highly dependent on the priorities in the human diet, it can be taken as a given constant. According to this assumption, the harvested yield (Y) to external nutrient input (=primary nutrients, P) ratio alone indicates the nutrient utilization in any system. The concept of primary production balance (PPB) is based on this fact (Seuri 2002).

The aims of this study were:

To introduce a new method, primary production balance, for the evaluation of nutrient utilization To demonstrate and find the key factors to reach a high utilization rate of nutrients

(*) Note: the term “balance” is used to indicate a difference (input – output) or ratio (output/input); the unit of difference is kg or kg/ha, the unit of ratio is either per cent or no unit.

Material and methods A more profound analysis was made on nitrogen utilization on nine organic farms in eastern Finland. Data was collected by personally interviewing farmers in 2004. An overall picture was drawn on how the farms were functioning and, to ensure the validity of data, the results were discussed personally with the farmers. The estimations of harvested yield (dry matter & nitrogen) were adjusted with the number of animals and total animal production. The nitrogen contents of all organic materials within the system (crops, fodder, bedding materials, seeds, animal products, purchased manure) were estimated by means of standard figures, unless measured values were available. Atmospheric deposition, 5 kg nitrogen/ha, was included as input.

All the main nutrient flows were identified. However, because of the steady-state assumption (i.e. balanced systems, no change in reserve nutrients in soil) and estimation of biologically fixed nitrogen the results may include some error.

Biological nitrogen fixation was estimated using harvested legume yield: the assumption was 50 kg nitrogen per 1000 kg harvested dry matter of legume. That means that roughly 70 % of the total nitrogen content in the legume biomass originated from BNF. The assumption was derived from the Swedish STANK model (STANK 1998), the Danish model by Kristensen et al. (1995) and the Finnish model by Väisänen (2000). On all farms the most important legume was red clover. However, some white clover and alsike clover were grown in

31 perennial mixture lays, too. Besides pea, which was the most important annual legume crop, some annual vetch was grown.

Farm-gate balance, surface balance and primary nutrient balance were calculated for each individual farm (Table 2-2). Primary nutrient balance can be calculated from the following two equations (Seuri 2002):

(I) PPB = Y/P where Y = total harvested yield P = primary nutrients (=external nutrients)

(II) PPB = U x C where U = utilization rate (=surface balance) C = circulation factor = (P + S)/P S = secondary nutrients (=recirculated nutrients)

Equation (I) follows the definition of PPB. Equation (II) illustrates two components of PPB: utilization rate, which is equal to surface balance, and circulation factor, which indicates the importance of recirculated nutrients in the system. There is a major difference between stockless farms and livestock farms. Since there are no recirculated nutrients (S) on stockless farms, the circulation factor is always 1.0. On livestock farms the circulation factor is always higher than 1.0.

To illustrate the difference between primary and secondary nutrients and to point out the role of recirculation in order to improve the nutrient utilization, some simple simulation was made on two stockless farms. The farms produce some fodder and receive some farmyard manure (FYM) from the neighbouring farm. The initial balance (A) indicates utilization in case manure from the neighbouring farm is external nutrient input (primary nutrient). The simulated balance (B) indicates the utilization in case all the harvested fodder yield is used on the farm for dairy cattle. It is assumed that 25 % of the nitrogen in the fodder is sold out from the farm in the form of milk and beef and 25 % is lost in the gaseous form before the manure is spread on the field. The rest of the nitrogen (50%) remains in the manure.

The average utilization rate of the primary nitrogen in the agriculture in Finland was calculated from statistics. Rough estimations and comparisons were made between the farms in this study and national average utilization rates.

Results and discussion Table 2-2. Comparison between primary production balance (PPB), surface balance (SB) and farm-gate balance (FGB) of nitrogen on nine organic farms in eastern Finland. Farms 8B and 9B are simulated from 8A and 9A, respectively.

Farm Production Primary Total N Harvested PPB SB FGB Circulation N surplus type N input on field N yield factor (kg/ha) (kg/ha) (kg/ha) (kg/ha)

1 dairy 60 92 69 1.15 0.75 0.34 1.53 40 2 dairy 68 108 75 1.11 0.69 0.3 1.60 49 3 dairy 53 83 56 1.06 0.68 0.3 1.56 44 4 beef 69 113 84 1.22 0.74 0.18 1.64 60

32 5 beef 65 113 73 1.13 0.65 0.20 1.74 53 6 beef(+crop) 52 89 50 1.06 0.62 0.17 1.70 48 7 goat(+crop) 56 73 45 0.80 0.62 0.30 1.30 55 8A crop 87 87 49 0.56 0.56 0.56 1.0 39 8B ”dairy” 63 87 49 0.77 0.56 0.19 1.39 51 9A crop 66 66 34 0.51 0.51 0.51 1.0 33 9B ”dairy”(+crop) 49 66 41 0.84 0.62 0.3 1.36 34

The PPB of nitrogen fell in the range 1.0 – 1.2 on all mixed farms except for farm 7, i.e. the farms were able to harvest more nitrogen than they received as an input into the crop production from outside the farm. Both stockless farms reached a PPB as low as around 0.5; the dairy farm simulation increased the PPB up to 0.8.

The SB of nitrogen fell in the range 0.6 – 0.75 on all mixed farms and by definition PPB and SB are identical in a stockless system (around 0.5), i.e. in any system without recirculated nutrients. The FGB of nitrogen correlated strongly with production type, being around 0.3 on dairy farms and around 0.2 on beef farms. Analogously to PPB and SB, also FGB was identical on stockless farms (around 0.5). The dairy farm simulation decreased the FGP down to 0.19 on farm 8 and down to 0.3 on farm 9.

Simulation on farm 8 shows clearly the role of recirculation and the difference between PPB and SB. On farm 8, the only difference between the real farm and the simulated farm is the method of definition of the origin of input nitrogen, i.e. the initial yield harvested and the initial amount of nitrogen available in the field are exactly the same. On farm 8A, all the nitrogen in FYM from the neighbouring farm is considered as primary nitrogen analogous to the nitrogen in artificial fertilizers or the nitrogen from BNF. This is analogous to any nitrogen input which increases the total amount of nitrogen in the system. On farm 8B, the nitrogen in the FYM from the neighbouring farm is considered as secondary nitrogen analogous to the nitrogen in FYM originating from the farm. This is analogous to any recycled nitrogen which does not increase the total amount of nitrogen in the system. However, the SB method does not identify the origin of the nutrients in the field, i.e. unlike PPB, SB remains constant on farm 8. The higher PPB value on the simulated farm 8B indicates higher efficiency of primary nitrogen utilization, thereby a lower nitrogen load potential.

On farm 9B there are some green manure fields, from where yield is harvested instead of ploughing directly. Therefore also the SB is influenced by simulation on farm 9, but otherwise it is analogous to farm 8.

In Finland (1995 – 1999), calculation of nitrogen balance in agriculture shows that the annual total primary nitrogen input (artificial fertilizers, atmospheric deposition and symbiotically fixed nitrogen) is about 100 kg/ha. The total harvested nitrogen yield is about 74 kg/ha, respectively (Lemola & Esala 2004). Thus, the PPB in agriculture averages 74 kg/ha / 100 kg/ha = 0.74, indicating serious lack of nutrient re-cycling. However, there is huge potential to recycle nutrients in agriculture, because 80 % of the total crop yield is used as animal fodder.

In this study, all the livestock farms exceeded the value 0.74, range 0.8 – 1.2, average around 1.0. The high PPB on nitrogen was not only due to recycling but also to biological nitrogen fixation. The main source of primary nitrogen input was symbiotic fixed nitrogen by legumes.

33 The utilization rate of nitrogen by legumes is clearly higher than for any other source of nitrogen into a system. In most cases about the same amount of nitrogen was harvested than was symbiotically fixed, i.e. the utilization rate is approx. 100%.

In addition, the balance between livestock and field area (fodder production) was of major importance in reaching a high PPB. Whenever the livestock density was increased by means of purchased fodder, the utilization of farmyard manure was poor and resulted in lower PPB (farms 3, 6 and 7). Self-sufficient fodder production was the optimum. The farms with high PPB had also a slightly higher yield level than farms with lower PPB.

On the other hand, two stockless organic farms indicated that without recirculation an organic system cannot utilize nitrogen very efficiently. On these farms the primary source of nitrogen consisted of legumes, but because the legume crop was partly used as green manure, there were heavy losses of nitrogen with resultant lower total PPB.

Conclusions It was fairly easy to calculate the primary production balance for each of the nine farms included in this study. The estimation of biological nitrogen fixation and harvested nitrogen yield are, however, obvious sources of error. The assumption of steady state is not necessarily valid in all cases. Even though crop production causes only minor nutrient load compared with animal production, it does not necessarily mean that crop farms utilize nutrients effectively. Using the PPB it is easy to compare different farms. The results of this study show clearly that livestock farms are able to reach a remarkably higher PPB compared with crop farms despite the very low farm-gate balance on livestock farms.

References Kristensen, ES., Högh-Jensen, H. & Halberg, N. (1995) A simple model for estimation of atmospherically-derived nitrogen in grass/clover systems. Biol Agric Hortic 12: 263-276. Lemola, R. & Esala, M. (2004) Typen ja fosforin virrat kasvintuotannossa. AESOPUS- hankkeen loppuseminaari 4.11.2004. SYKE, Helsinki. (Nitrogen and phosphorus flows in agriculture. Seminar on the AESOPUS project 4.11.2004. Finnish Environment Institute, Helsinki.) Myrbeck, Å., (1999) Nutrient flows and balances in different farming systems – A study of 1300 Swedish farms. Bulletins from the Division of Soil Management, Department of Soil Sciences, Swedish University of Agricultural Sciences. 30: 1-47, 11 app. Seuri, P. (2002) Nutrient utilization with and without recycling within farming systems. In: eds. Jakob Magid et al.. Urban Areas - Rural Areas and Recycling - the Organic Way Forward?. DARCOF Report 3: p. 175-181. http://www.agsci.kvl.dk/njf327/papers/NJF- Co-development.pdf STANK. 1998. Statens Jordbruksverks dataprogram för växtnäringsbalansberäkning. Stallgödsel och växtnäring i kretslopp (STANK) Jordbruksverket, Jönköping. (Plant nutrient balance calculation program) Väisänen, J. (2000) Biological nitrogen fixation in organic and conventional grass-clover swards and a model for its estimation. Licentiate’s thesis. University of Helsinki Department of Plant Production Section of Crop Husbandry. 42 p. + 2 app.

34 3. Effects of 100 % organic production on Funen, Denmark Ib Sillebak Kristensen1 , Ole Tyrsted Jørgensen2 and Inge T. Kristensen1 1 Danish Institute of Agricultural Sciences, Department of Agroecological, Research Group of Farming Systems, Post-box 50. DK-8830 Tjele, DENMARK. 2 Fyns Amt (the Funen County) Contact: Ib Sillebak Kristensen, Tel. +45 89 99 12 05, E-mail: [email protected]

Introduction Organic production has been found to lover the environmental impact compared to present conventional production. The effect is partly because of lower stocking rate and partly because of higher utilization (Nielsen and Kristensen, 2005). The present organic production in Denmark is mainly producing organic milk on specialized organic dairy farms, with a high stocking rate of 1.4 Livestock Unit (LSU) per ha cultivated land and only 5 % N-export in crop production, found on representative farms, Kristensen et al. (2005b). The milk production both use external inputs in feed, straw and animal manure. The inputs is partly conventional and partly organic from mainly organic arable farms, who also have considerable external inputs from manure (32 % of total N-inputs, Kristensen, 2005). A closer integration between animal and arable production within the same farm area producing both animal and vegetable products is supposed to contribute to a better resource utilization and a lower environmental impact.

In order to quantify the consequences of changing the present Danish Agriculture to organic production a modelling approach has been developed on the background of the present Danish agricultural production and the hereby estimated environmental impact of representative farm types (Knudsen et al (2005) and Kristensen et al (2005b). In the organic model approach no external inputs were used in order to quantify the maximal effect of Ecological Recycling Agriculture, ERA-farms, see chapter 7, Granstedt (2005). The assumptions for conversion of present agricultural situation into 100 % organic agriculture has been taken from the Danish “Bichel”-work, Anon. (2001).

The aim is to analyse possible impact of organic production on the nutrient loss in the Baltic Sea compared to the present conventional situation. The Funen county has been chosen because it represent average stocking rate of 0.8 LSU/ha (table 1) for the Danish agricultural production influencing the Baltic Sea. Also considerable previous work has been done on the Funen county in order to quantify the present environmental impact on the Baltic Sea (Madsen et al., 2003 and Terlikowska et al (2000).

Methods Average farm-type based on main production and stock density is established from a set of representative data sets, briefly described in Kristensen and Kristensen (2004) and in detail in Larsen (2003). The production and nutrient flows for each farm type were based partly on farm accounts and partly on detailed information from case studies. The individual farm in the Danish register data are classified according to the farm-type and the total N-surplus is estimated based on the areas within each farm type. The method has shown good results in a small vulnerable watershed of 36,000 ha, 1.5% of DK agricultural area, Kristensen and Kristensen (2004). This preliminary results look promising and the work is with this paper extended to a model of the present situation on Funen (223,000 ha = 9% of DK agricultural area).

All Danish farms are obliged to keep detailed records of purchases and sales for tax purposes and the yearly accounts are made with professional help. A representative set of these accounts, 2239, are reported by the advisors to the Danish Research Institute of Food Economics to the FADN-data (The Farm Accountancy Data Network, see Anon. (2005) and http://www.foi.dk/. The account data constitute the basic empirical input to the representative farm types presented here. Besides the economical data, information on land use, livestock numbers and amounts of produce are included in the data set compiled by the advisors. The modelled representative farm types were based on farm accounts in year 2002, sampled as to represent the total Danish agricultural sector for the main livestock and crop production. The data are described in detail on the homepage www.lcafood.dk. The conventional and organic dairy farms can be seen at Kristensen (2005b), pig farms at Kristensen (2005c) and organic and conventional cash crop farms at Kristensen (2005a). The same overall method was used in the years 1995-1996 (Halberg et al., 1999), year 1999 for dairy and arable farms (Knudsen, Kristensen et al (2005) and Kristensen, Halberg et al (2005b).

The representative FADN-data sample represents 5% of the conventional farms and 18% of the organic farms in Denmark. Data can be looked up at the above homepage. Seventy-five percent of the dairy cows are on sandy soil, mostly in western Jutland. The dairy farm types cover 23% of the agricultural land and include 75% of all dairy cows. The average herd size on conventional farms is 61 cows and 82 cows on organic farms. The average stocking rate on conventional farms is 1.46 LSU ha-1 and 1.28 on organic farms. Eighty-five percent of the farm area is part of a crop rotation. Twenty-six percent of the farm area of conventional farms is grass/clover. This percentage on organic farms is twice as large. The remainder of the rotating area is mainly under cereals, partly for grain and partly for whole crop silage harvested 2-3 weeks before full ripe. Maize for silage has become important, mainly on conventional farms. Cereal grain yield is 4.9 t DM ha-1 on conventional farms and 41% lower on organic farms. Roughage yields are assumed to be on the average level of pilot farms, and the average yield of rotating crops is 5,700 SFU ha-1 = 6,300 kg DM ha-1 on conventional dairy farms and 21% lower on organic farms.

The milk yield level in 2002 was 7,740 kg ECM milk cow-1 year-1 on conventional farms and 10% lower on organic farms. Of the total SFU-intake by cows, 58% is homegrown roughage on conventional farms and 75% on organic farms. An average level of protein was 18.4% of SFU ˜ 20% of DM.

In the Danish dairy sector, biological fixation is an important N input, especially on organic farms. On pilot farms, fixation in grass/clover was calculated form visual soil legume cover, based on the method presented by Kristensen et al (1995). For the representative farms, the fixation and mineral fertilizer for grass/clover was assumed to be the same level as on the pilot farms. Details of the level of fixation in grass/clover and information regarding fodder, N demand, crop yield estimates and fertilizer use are given in the Appendix of Kristensen, Halberg et al (2005b). The ammonia losses shown were calculated using the emission coefficients of stall and houses mainly after Poulsen and Kristensen (1998), and updated by Hutchings et al (2001) and Illerup et al (2002) and the denitrification losses after Vinther and Hansen (2004) respectively.

Farm balances were calculated following the principles described in Halberg et al (1995) and further developed by Kristensen (2002).Coherent N balances at farm level were calculated, incorporating imports of fodder (cereal and concentrates) and mineral fertilizers and exports

36 of milk, meat, cash crops and surplus manure. Summarizing and upscaling key in- and outputs across representative farm types showed good agreement with national statistics of land use, livestock numbers, average yield per crop, input of fodder concentrate and fertilizer per crop.

In order to describe the present situation on Funen it is assumed that the representative types for the entire Danish agriculture is representative for the Funen agriculture as well. These data are combined with the individual farm data of the Funen farmes. The register data for all farms is present in 3 central registers: The fertilizer accounts, land use (General Agricultural Register) and animal numbers (Central Husbandry Register). The farms were classified into the before mentioned types and the estimated nutrient load was used for calculating the present nutrient loads per type of the actual agriculture of Funen in year 2002.

Results

Present agricultural production of Funen in year 2002. In table 1 the agricultural production in year 2002 is presented. At the right column the entire agricultural production is shown and different categories of full time farmers (above 1.665 man-hours per year, see Larsen (2003)). In “Total Funen” specialized horticulture is not included in the dataset because the horticulture land use are not updated in the Central Registers. The area use is a direct output from the central register (CLR) and 60 % of the agriculture is on sandy loam with more than 10 % clay of texture. The Funen dairy farms only grow 13 % of the arable land with grass/clover, the entire conventional dairy sector grow double area with grass/clover, see Kristensen et al. (2005). Only 8 organic dairy farms are situated on Funen. These few farms has 20 % of the area with low productive permanent grassland, making the per ha figures lower than average organic dairy farms with only 9 % area with permanent grass. Pig and arable farms grow 44-53 % of the area with winter cereals. No dairy organic farmers are mainly growing spring cereals. The yields on Funen are calculated. Roughage yield is assumed to be the same as studies on DK pilot farms, and the cereal yield are takes from Danish Statistics from Funen County. Calculating the yield per type-average the same yield is assumed for all Funens farmers and only the different area proportion per type is included in the calculated average yield per average farm-type. The average grain yield for organic farms is set to 70 % of conventional grain yields after Halberg and Kristensen (1997). The low organic grain yield of around 4000 kg grain (85 % DM) per ha is 35 % lower than conventional yield because of 30 % lower yield per ha and a higher area proportion of low yielding spring grain compared to high yielding winter cereals.

I table 2 the N-balance in year 2002 is calculated assuming that the Funen farmers has the same feed and product N- and P-turnover as the Danish average farmers, see Kristensen and Kristensen (2004) and methods for assumptions. Artificial fertilizer and exchange of animal manure is measured values from individual farms, these amounts are included directly from the central register of “Fertilizer control. Assuming the same percentage losses per farm-type the aerial losses and soil-N changes can be deducted the farm gate N-surplus and the N- leaching calculated as the rest is shown in the bottom of the table. I total the leaching from Funen is 63 kg N/ha, which is only 1 kg N/ha lower than the independent calculation made by Schrøder (2004). As previous found the N-leaching calculated by difference is in good agreement of direct modelling of N-leaching (Kristensen et al (2005a). In table 3 the P- balance is 12 kg P-surplus/ha, which is 40 % higher than Schrøder (2004) and Nielsen et al (2004).

37 Modelled agricultural production and emissions with 100 % organic production In table 4 the assumptions for making Funen 100 % organic is presented. The details for the assumptions can be seen in Hermansen (1998), Alrøe og Kristensen (2001), Anon. (1999) and Anon. (2001). In order to illustrate the maximum benefit of organic production the scenario of 100 % organic production with no external import has been calculated. In short the main assumptions are: o The same area as present situation. o 40 % of the area is used for production of grass/clover in order to keep up the soil fertility at the minimum level of actual organic dairy farms with low sticking rate. At lover soil fertility the grain yield becomes unstable and sometimes low due to low N-level and infestation with weeds, especially rood weeds. o The dairy production is produced 84 % on roughage feed from grass/clover and 16 % grain feed, see table 4. o 12 % of the area is used for grain and vegetables for human consumptions. o The rest of the area in crop rotation (48 % = 100 – 40 – 12) is used for organic pig production, where 8 % area is grown with rape and peas in order to make a balanced feed ration for pig production. o The organic yields are assumed to be average from pilot farms studies, mainly after Halberg and Kristensen (1997).

In table 5 the consequences of the assumptions in table 3 are shown. The main results found are: o With high roughage uptake the milk production per dairy cow is depressed 25 % compared to present level. o In order to use the fodder from 40 % area grown with grass/clover the cattle herd shall be increased by 40 % extra cattle livestock units. o The entire area used for cattle production is 65 % of the area, leaving 23 % of the area for pig feed production (23 = 100 – 65 – 12 (= crops for human consumption)). o The pig production is assumed to rely on 10 % from roughage and 60 % grain. With this feed use around 30 % of the present pig production can be produced on Funen. o It is assumed that production cattle, human and pig product is produced within the same crop rotation in order to keep up the grain production to the same level as measured on present private cattle farms. So it is assumed that all farms produce both cattle and pig products.

In table 6 the consequences on farm gate balances and emissions is calculated when same coefficients of emissions as present agriculture is assumed. The main findings are: o The milk production per cow is reduced by approximately 25 % less milk/cow and around 15 % less milk from Funen County. o All male cattle were raised for steers and the cattle meat production was doubled compared to present situation. o The fertilizer and feed import is zero o The fixation input is increased mainly from the 40 % grass/clover area with a fixation of 150 kg N/ha. o The total inputs (only from fixation and deposition) was reduced to 50 % of present inputs. o The pig production is reduced by 70 %, mainly because of lower pig number and partly because of lover productivity of only 19 slaughter pigs per sow compared to present 22.

38 o The farm gate balance is reduced by 50 %, bringing the farm gate surplus down to 59 kg N-surplus/ha. o With same coefficient of emissions and soil-N changes the balance will contribute to only 35 kg N-leaching per ha, around a 50 % reduction compared to the present situation in year 2002. o Without any external P-inputs the P-balance is about – 6 kg P/ha.

Table 3-1. Farm characteristics on Funen in 2002.

Table 1. Farm characteristics on Fyen in 2002. Sandy loam Sandy Dairy Dairy Pig Arable, Not dairy Total

conv. org conv. conv. org Fyen Number of farms No 2.335 1.388 462 8 653 844 186 6.328 Area, incl. set aside & perm. grassha 136.925 74.574 33.236 895 59.700 78.594 3.377 223.376 Area, incl. set aside & perm.ha/farm grass 58,6 53,7 71,9 111,9 91,4 93,1 18,2 35,3 Set aside ha/farm 4,0 3,8 4,8 3,7 6,4 7,5 0,7 2,4 Permanent grass ha/farm 2,0 2,4 5,0 22,4 0,8 2,1 3,0 1,3 Area use of ploughed ha Grass/clover % of ha 4% 5% 13% 40% 1% 1% 18% 4% Maize & Whole crops% of ha 5% 6% 30% 17% 0% 1% 4% 5% Winter cereals % of ha 43% 42% 26% 6% 53% 44% 17% 42% Spring cereals % of ha 31% 31% 25% 31% 26% 31% 51% 31% Beets % of ha 6% 8% 6% 4% 7% 7% 0% 7% Rape % of ha 4% 4% 2% 0% 7% 4% 1% 4% Other % of ha 7% 4% -3% 2% 7% 13% 8% 7% Net yield Grass/clover SFU/ha 6000 6000 6000 5520 6000 6000 5520 6000 Permanent SFU/ha 2320 2320 2320 2018 2320 2320 2018 2320 Maize & whole crops SFU/ha 8521 8649 8565 5095 8264 8805 4512 8557 Undersown grass/cloverSFU/ha incl. G/C incl. G/C incl. G/C incl. G/C incl. G/C incl. G/C incl. G/C incl. G/C Cereals SFU/ha 6339 6170 6167 4065 6422 6385 4001 6246 Peas kg/ha 3940 3940 3940 3940 3940 3940 Winter rape kg/ha 2820 2820 2820 2820 2820 2820 Total SFU/ha 6604 6659 7080 4492 6662 6644 3775 6579

39 Table 3-2. Farm-gate balances of N and P on Funen in 2002.

Table 2. Farm gate N-and P-balances of Fyen in 2002.

Sandy loam Sandy Dairy Dairy Pig Arable, Not dairy Total Units conv. org conv. conv. org Fyen Farm characteristics LSU 121.483 70.851 55.532 778 92.560 9.706 1.119 200.284 ECM milk/cow 7523 Pigs/sow 21,5 Farm gate balance Inputs Artificial fertilizer N/ha 84 77 59 0 68 105 6 83 Animal manure N/ha 17 21 14 29 15 22 43 18 Net feed import N/ha 52 68 153 49 153 -49 -3 56 Fixation N/ha 5 6 12 52 2 3 27 5 Deposition N/ha 18 18 18 18 18 18 18 18 Total inputs N/ha 177 189 256 147 256 99 92 180 Outputs Milk N/ha -10 -12 -53 -27 0 -1 -3 -10 Meat, cattle N/ha -1 -1 0 0 0 0 -6 -2 Meat, pigs N/ha -28 -31 -1 0 -82 -5 0 -28 Animal manure N/ha -18 -19 -16 -7 -37 -2 -3 -18 Vegetable products N/ha -13 -15 -7 -12 -15 -16 -15 -13 Total outputs N/ha -69 -77 -78 -46 -135 -24 -28 -71 Balance N/ha 107 112 178 101 121 75 64 109 N-efficiency % 39% 41% 30% 31% 53% 24% 30% 40%

Ammonia losses % of N-surpluss 28% 26% 26% 25% 37% 21% 17% 27% Ammonia losses N/ha 30 29 46 25 45 16 11 29 Denitrication % of N-surpluss 21% 8% 18% 28% 15% 18% 24% 15% Denitrication N/ha 23 9 33 28 18 13 15 17 Nitrate leaching & soil-N changesN/ha 54 73 100 48 58 46 38 63

Soil-N change (- breaking down,N/ha + built up) 3 4 6 24 1 -2 0 3 Nitrate leaching N/ha 52 69 94 23 57 48 38 60 P-balance P/ha 11 13 17 2 20 6 0 11 Soil-N changes are calculated with C-tool, see www.agrsci.dk/c-tool/

Table 3-3. Farm-gate balances of P on Funen in 2002. Table 3. Farm gate P-balances of Fyen in 2002. Dairy Dairy Pig Arable, Not dairy Total conv. org conv. conv. org Fyen Inputs Soya feed P/ha 9 2 10 1 8 5 Grain P/ha 5 3 8 0 8 5 Feed minerals P/ha 3 1 23 2 8 8 Artificial fertilizer P/ha 13 0 5 11 8 9 Animal manure P/ha 4 8 4 6 12 5 Total inputs P/ha 34 14 50 20 43 33 Outputs Vegetable products P/ha 2 0 7 11 8 8 Meat P/ha 2 1 14 1 8 6 Milk P/ha 8 5 0 0 8 2 Animal manure P/ha 3 1 10 1 1 5 Total outputs P/ha 15 8 31 14 24 21 Balance P/ha 19 3 21 6 11 12 P-efficiency % 44% 61% 63% 66% 56% 63%

Feed minerals P/LSU1) 2 1 16 10 5 8 Balance P/LSU1) 11 2 14 29 34 13 1) 1 LSU = 0.85 frisian dairy cow or 0.34 heifer or steer or 36 slaughter pigs

40 Table 3-4 Farm characteristics on Funen in 2002 and in Scenario Funen 100% organic. Funen 2002 Scenario unit % of conv. (conv.) 100% org. Number of farms no 6 328 Area, incl. set aside & perm. grass ha 223 376 223 376 Area, incl. set aside & perm. grass ha/farm 35.3 35.3 Set aside ha/farm 2.4 0 Permanent grass ha/farm 1.3 1.3 Area use of ploughed ha Grass/clover % of ha 4 40 Conv. maize or Org. whole crop % of ha 5 12 Winter cereals % of ha 41 40 Spring cereals % of ha 31 Beets % of ha 6 0 Rape % of ha 4 3.2 Other (peas in org. scen.) % of ha 9 5.2 Net yields assumed Grass/clover SFU1/ha 6 000 5 200 87 Permanent SFU/ha 3 000 1 500 50 Conv. maize & org. whole crop SFU/ha 8 000 3 000 38 Undersown grass/clover SFU/ha incl. in G/C 400 Cereals SFU/ha 6 233 3 400 55 Peas kg/ha 3 940 3 128 79 Winter rape seed kg/ha 2 820 1 500 53 Total SFU/ha 6 302 3 599 57

Table 3-5 Assumptions for Scenario Funen 100 % organic with no import. unit assumed value % of conv. Feed and area use for cattle Grass/clover & whole crop silage SFU/MPU 2 6 975 Cereals SFU/MPU 1 302

No of cattle in herds MPU 66 269 No of cattle in herds LSU 3 102 067 132 Cereals for cattle SFU 86 281 783 Grass/clover, whole crop & perm. ha 116 805 Area use for cattle production ha 142 182 Livestock density on cattle farms on Funen LSU/ha 0.72 Feed and area use for pigs Grass/clover & whole crop silage SFU/PU 4 600 Cereals SFU/PU 3 971 Peas SFU/PU 1 346 Rape-cake SFU/PU 382 Cereals for pigs etc. ha 33 565 Cereals for pigs etc. SFU 114 119 854 No of pigs PU 28 738 No of pigs LSU 33 911 30

1 1 SFU (Standard Feed Unit) = ??? 2 1 MPU (M??? P??? Unit) =1 frisian dairy cow + 1.03 frisian heifer + 1 steer 3 1 LSU (LiveStock Unit) = 0.85 frisian dairy cow/ 0.34 heifer or steer = 100 kg N from manure storage 4 1 PU (Pig Unit) = 1 year sow + 18,7 slaughter pigs

41 Grass/clover & whole crop silage ha 3 316 Peas ha 11 377 Rape seed cakes ha 7 075 Area use for pigs ha 55 333 Livestock density of pig area on cattle farms LSU/ha 0.61 Manure production N-efficiency, cattle herd % 15.7 76 N-production, cattle kg N/ha 104 N-production, cattle on grass kg N/ha 36 N-production, cattle in stable kg N/ha 68 N-efficiency, pig herd % 365 100 N-production, pigs kg N/ha 61

Table 3-6 Farm gate N-balances of Funen in 2002 and in Scenario Funen 100% organic. Funen 2002 Scenario unit % of conv. (conventional) 100% organic Farm characteristics LSU 200 284 135 978 68 LSU/ha 0.90 0.61 68 kg ECM6/cow 7 672 5 540 72 piglets/sow 21.5 18.7 87 Farm gate balance Inputs Artificial fertilizer kg N/ha 83 0 Net feed import kg N/ha 56 0 Fixation kg N/ha 5 66 Deposition kg N/ha 16 16 Total inputs kg N/ha 160 82 51 Outputs Milk kg N/ha -10 -8 Meat, cattle kg N/ha -2 -4 Meat, pigs kg N/ha -28 -5 Vegetable products kg N/ha -13 -7 Total outputs kg N/ha -53 -25 47 Balance kg N/ha 107 57 54 N-efficiency % 33 30 91 Ammonia losses % of N-surplus 22 22 Ammonia losses kg N/ha 24 13 54 Denitrification % of N-surplus 14 14 Denitrification kg N/ha 14 8 54 Nitrate leaching & soil-N changes kg N/ha 69 37 54 Soil-N change (- breaking down, + built kg N/ha 3 3 100 up) Nitrate leaching kg N/ha 65 34 51 Soil-N changes are calculated with C-tool, see www.agrsci.dk/c-tool/

5 including bone meal and synthetic amino acids. Not allowed in 2002, expected 35%. 6 Energy Corrected Milk

42 References

Alrøe, H. F. and Kristensen, E. S. Matthies, M., Malchow, H., and Kriz, J. 2001: Researching alternative, sustainable agriculture systems. A modelling approach by examples from Denmark.Integrative Systems Approaches to Natural and Social Dynamics. 113-140. Anon. 1999: Organic Scenarios for Denmark. The Bichel Committee, 1999. http://www.mst.dk/udgiv/publications/2001/87-7944-622-1/pdf/87-7944-624-8.pdf, 1- 118. MILJØstyrelsen. Sekretariat for pesticidudvalget. Anon. 2001: The Bichel Committee 1999. Report from the Bichel Committee - Organic Scenarios for Denmark Report from the Interdisciplinary Group of the Bichel Committee. http://www.mst.dk/udgiv/Publications/2001/87-7944-622-1/html/default_eng.htm., 1- 118. Anon. 2005: The Farm Accountancy Data Network (FADN) is an instrument for evaluating the income of agricultural holdings and the impacts of the Common Agricultural Policy. http://europa.eu.int/comm/agriculture/rica/index_en.cfm. Granstedt, A. 2005: Plant nutrient balance studies. In Eds. Gransted,A. and Thomasson,O.: Environmental impact of eco-local food systems. Baltic Ecological Recycling Agriculture and Society (BERAS) [3], 1-1?? Halberg, N., Kristensen, E. S., and Kristensen, I. S. 1995: Nitrogen turnover on organic and conventional mixed farms. Journal of Agricultural and Environmental Ethics 8, 30-51. Hermansen, J. E. 1998: Samhørende værdier for foderforbrug og produktion i økologiske husdyrproduktionssystemer., 1-8. Hutchings, N., Sommer, S. G., Andersen, J. M., and Asman, W. A. H. 2001: A detailed ammonia emmision inventory for Denmark. Atmospheric Environment 35, 1959- 1968. Illerup, J. B., Birr-Pedersen, K., Mikkelsen, M. H., Winther, M., Gyldenkærne, S., Bruun, H. G., and Fenhann, J. 2002: Projektion Models 2010. Danish emissions of SO2, NOX, NMVOC and NH3. NERI Technical Report 414, 192-200. Knudsen, M. T., Kristensen, I. S., Berntsen, J., Petersen, B. M., and Kristensen, E. S. 2005: The effect of organic farming on N leaching. Submitted. Journal of Agricultural Science , 1-35. Kristensen, E. S., Høgh-Jensen, H., and Kristensen, I. S. 1995: A simple model for estimation of atmospherically-derived nitrogen in grass-clover systems. Biological Agriculture and Horticulture 12[3], 263-276. Kristensen, I. S. 2002: Principles and methods for collecting and evaluation nutrient balances, 29-40. http://www.agrsci.dk/jbs/demolit/Principles%20and%20methods.pdf, Danish Institute of Agricultural Science. Lithuanian Dairy Farms Demonstration Project. Kristensen, I. S. 2005a: Nitrogen balance from cash crop farms (2002). http://www.lcafood.dk/processes/agriculture/N_balance_cash_crop_farms_2002.htm. Kristensen, I. S. 2005b: Nitrogen balance from dairy farms (2002). http://www.lcafood.dk/processes/agriculture/N_balance_dairyfarms_2002.htm#table_ 1. Kristensen, I. S. 2005c: Nitrogen balance from pig farms (2002). http://www.lcafood.dk/processes/agriculture/N_balance_pigfarms_2002.htm. Kristensen, I. S., Dalgaard, R., Halberg, N., and Petersen, B. M. 2005a: Kvælstofbalance og - tab fra forskellige bedriftstyper. Plantekongres 2005. http://www.lr.dk/planteavl/diverse/PLK05_11_1_3_I_S_Kristensen.pdf. og http://www.lr.dk/planteavl/diverse/plk05_11_1_3_i_s_kristensen.ppt. Planteproduktion i landbruget , 1-2.

43 Kristensen, I. S., Halberg, N., Nielsen, A. H., and Dalgaard, R. 2005b: N-turnover on danish mixed dairy farms. Part II. In: Bos,J.; Pflimlin,A., Aarts,F. and Vertés,F.(Eds): "Nutrient management on farm scale. How to attain policy objectives in regions with intensive dairy farming. Report of the first workshop of the EGF Workshop. http://www.agrsci.dk/var/agrsci/storage/original/application/57b9a70804960c2c9f54a d255b11f22d.pdf. Plant Research International. 83, 91-109. Kristensen, I. T. and Kristensen, I. S. 2004: Farm types as an alternative to detailed models in evaluation of agricultural practise in a certain area. Oral precentation on MIS 2004: "Fourth International Conference on Management Information Systems, incorporating GIS and Remote Sensing. 13 - 15 September 2004. Malaga, Spain., 1-10.

Larsen, I. 2003: LCA project. Method description.

http://www.lcafood.dk/processes/agriculture/MethodDescription.pdf , 1- 16. www.lcafood.dk. Goto Processes; Agriculture; Farm types; Larsen et al., 2003. Madsen Nielsen, K., Andersen, H. E., Larsen, S. E., Kronvang, B., Stjernholm, M., Styczen, M., Poulsen, R. N., Villholt, K., Krogsgaard, J., Dahl-Madsen, K. I., Friis-Christensen, A, Uhrenholdt, T., Hansen, I. S, Pedersen, S. E., Jørgensen, O., Windolf, J., Jensen, M. H., Refsgaard, J. C., Hansen, J. R., Ernstsen, V., Børgesen, C. D., and Wiggers, L. 2004: Odense Fjord - Scenarier for reduktion af næringsstoffer485, 1-246. København, Danmarks Miljøundersøgelser, Miljøministeriet. Poulsen, H. D. and Kristensen, V. F. 1998: Standard values for farm manure. A revaluation of the Danish standard values concerning the nitrogen, phosphorus and potassium content of manure. Danish Institute of Agricultural Science report Animal Husbandry. 7, 1-160. Schrøder, H. 2004: N- og P-regnskaber for landbruget i Fyns Amt og Danmark 1984/85- 2001/02., 1-94. Danedi for Fyns Amt. Terlikowska, K., Muus, K., Antsmäe, H., Effenberg, G., Joelsson, A., Kolk, M., Koppe, A., Maximova, L., Nilsson, B., Pedersen, S. E., Pevnev, E. N., Steinmann, F., Westberg, V., and Smeds, L. Muus, K. and Pedersen, S. E. 2000: BERNET Theme Report: Sustainable Agriculture and Forestry, 1-143. BERNET. Vinther, F. P. and Hansen, S. Vinther, F. P. and Hansen, S. 2004: SimDen - en simpel model til kvantificering af N2O-emission og denitrifikation. see www.agrsci.dk/simden .104, 1-47. DJF rapport Markbrug.

44 4. Nitrogen and phosphorus leakage in ecological recycling agriculture Thomas Schneider1), Miia Kuisma2) and Pentti Seuri2)

1)Department of Physical Geography and Quaternary Geology, Stockholm University, SE -106 91 Stockholm, [email protected] 2)MTT Agrifood Research Finland, Ecological Production, FI-51 900 Juva, Finland

Introduction Organic farming is often considered as one solution to reduce eutrophication of the Baltic Sea. However, the efficiency of organic agriculture in reducing nutrient leakage from primary food production to the aquatic environment is still questioned. The most crucial part of organic agriculture for reducing nutrient leakage is nutrient management in the crop rotation. Nutrient balance studies showed that ecological recycling agriculture (ERA) had lower nutrient surplus and thus, lower potential of leakage (this report and Granstedt and others, 2004). However, direct measurements of nutrient leakage in ecological agriculture are scarce. Bergström and Kirchmann, 2000 reviewed the available literature on nitrogen leakage in organic agriculture and concluded that organic agriculture seems to have lower nitrogen leaching per hectare than conventional systems but differences were small and they were sensitive for small changes in either production system. The nitrogen leakage per mass of produced crops was higher in organic agriculture. They summarized, that nitrogen leakage is more a question of nitrogen management, e.g. crops and crop rotation, than of production system. They believed that a decrease of nitrogen leakage can be achieved by optimizing the conventional system. However, Bergström and Kirchmann, 2000 did not consider differences of different organic systems. ERA is a nutrient extensive system based on an animal density adjusted to the own fodder production on each farm unit, which has a potential of low N- leakage (Granstedt and others, 2004).

In this report we quantify nitrogen and phosphorus leakage on three ERA farms in the BERAS project. The study is based on direct measurements of nitrogen and phosphorus concentration in drainage water and water flow measurements. The results are compared with calculated standard leakage as reported to the HELCOM PLC-4 report (Brandt and Ejhed, 2002, HELCOM, 2004). Methods

Physical settings of the test fields The test sites in Sweden are located at Skilleby farm in Järna, 50 km south of Stockholm, and on Solmarka farm, 20 km south of Kalmar (Fig. 4-1 and Figure 2-3). Both Skilleby and Solmarka farm are managed according to the biodynamic farming practice since the 1960:s and 1970:s, respectively. The five-year crop rotation consists of three years of ley followed by winter cereals and spring cereals with insown clover grass. Skilleby farm is managed by the nearby Yttereneby farm and the animal density corresponds to 0.6 au/ha. Solmarka has its own cows and cattle and an animal density of 0,7 au/ha.

The Finnish test site is located in Juva 270 km north of Helsinki (Figure 4-1 and Figure 2-3). Organic farming practices have been applied since 1985. The six year mixed crop rotation consists of one year of spring cereal and grass seeds, two years of grassland followed by

45 winter cereal, green manure, and spring cereal. Approximately 30 t/ha cattle sludge is spread continuously on the 1st year plot which equals about 0.3 au/ha for the whole crop rotation per year. Harvested grain yields as net yields amounted to about 2 t DM/ha and grass yields to about 6 t DM/ha.

The physical characteristics of the test fields are summarized in (Table 4-1). For more background information on the three investigation farms see Seppanen, 2004 and. Granstedt and others, 2004

Figure 4-1 Map of investigation fields in Skilleby (6), Solmarka (10) and Partala (14) with sampling site, drainage area and drainage system. Number in brackets according to location map (Figure 2-3). Farm characteristics are shown in Table 4-1.

Table 4-1. Characteristics of test fields at Skilleby, Solmarka and Partala farm. Skilleby farm Solmarka farm Partala farm Available data 030701 - 050630 040701 - 050630 01/2005-04/2005 Soil type Clay Sandy loam - Moraine silty loam Mean air temperature (°C) 6.6 7,3 4 1) Mean precipitation (mm/a) 518 566 620 2) Total drainage area (ha) 22.7 11.0 4.87 1)Juva 1997-2003 2)Mikkeli 1997-2004

Data sampling Skilleby farm. Water samples were sampled manually every second week at the outlet of a drainage pipe from the test field. The samples were analysed on N-and P-concentration in accredited laboratories. Nutrient concentration was interpolated linearly between two sampling events. Water stage was measured continuously at a V-notch thin plate weir (90°) by a pressure transducer. Stage values were transformed into discharge data by applying standard hydraulic equations (e.g. Shaw, 1993). The product of daily mean values of water discharge and interpolated nutrient concentration data yielded nutrient load from the test site. Air temperature and precipitation were measured continuously by an automatic climate station located in the test field.

Solmarka farm. Water samples were sampled manually every week at the outlet of a drainage pipe from the test field. The samples were analysed on N-and P-concentration

46 according to accredited methods. Nutrient concentration was interpolated linearly between two sampling events. Water discharge was measured directly during the sampling events by means of a calibrated bucket and was interpolated linearly between sampling events. Discharge data was multiplied with measured nutrient concentration to obtain nutrient load. Air temperature and precipitation were measured continuously by an automatic climate station located in the test field.

Partala farm. The test field at Partala consists of five plots with five of six crops in an entire crop rotation cycle. The drainage water from the plots is directed to a V-notch weir where water stage is measured continuously with a pressure transducer. Water samples were taken manually in proportion to water flow (once a day to once a week). The water samples were analysed for both total and soluble nitrogen and phosphorus, total solids, pH and conductivity. Nutrient concentration will be interpolated linearly between two sampling events. An automatic climate station located nearby the test field measured air temperature continuously and precipitation is measured manually every day. However, measurements at Partala started in April 2005, thus, no results on nutrient leakage are available yet.

Comparison with TRK The obtained results at Skilleby and Solmarka on annual nutrient leakage were compared with official Swedish leakage data published as the TRK-report (Brandt and Ejhed, 2002) and reported to the Helcom pollution load compilation (HELCOM, 2004).

The TRK data was based on modelling nitrogen leakage with the SOIL-N (Johnsson and others, 1987) and HBV-N model (Arheimer and Brandt, 1998). The models simulate nitrogen leakage depending on soil type and crops. The TRK-dataset was normalized for long-term climatic fluctuations. In order to obtain comparable values of nutrient leakage, the standard values for the test fields were calculated from the standard leakage data in the TRK area 22 (Solmarka) and 60 (Skilleby) as presented in Table 4 in the TRK-report (Brandt and Ejhed, 2002). The standard nitrogen leakage was calculated for soil type and grown crops.

2 Standard phosphorus leakage, PL, (kg km /year) was calculated according to Brandt and Ejhed, 2002 and Ulén and others, 2001 by the following equation:

PL = (-0.0803 + 0.1 · LD + 0.003 · S + 0.0025 · PHCl) · Q, (1)

where S = (8.0 · xclay + 2.2 · xsilt + 0.3 · xsand) · r · 0.001 (2)

LD Livestock density (livestock unit ha-1) S Soil specific area (m2 m-3 10-6) PHCl HCl extractable phosphorus (mg/100 g dry soil) r Bulk density of soil = 1250 kg m-3 xclay Clay fraction in top soil (0-30 cm), < 2 mm xsilt Silt fraction in top soil (0-30 cm), 2 mm – 60 mm xsand Sand fraction in top soil (0-30 cm), 60 mm – 200 mm Q Runoff (mm)

47 Results and discussion

Standard leakage - TRK Nitrogen leakage from the test fields calculated according to the method described in the TRK-report (Brandt and Ejhed, 2002) was 5.3 kg/ha year in Skilleby and 9.2 kg/ha year in Solmarka (Table 4-2 and 4-3). Table 4-2. Nitrogen leakage at Skilleby calculated from standard leakage defined in TRK report (Brandt and Ejhed, 2002). NTRK - standard leakage depending on soiltype, crops and climate zone according to TRK. NTRK-Skilleby is calculated as the product of NTRK and the share of the respective soiltype and crop.

Area Share Soiltype Crops NTRK NTRK-Skilleby ha kg/ha year kg/ha year 8,08 35% Clay Oats 12 4.2 9,73 43% Clay Ley 2 0.9

2003 5,07 22% Forest 1 0.2 22.9 5.3 8,08 35% Clay Ley 2 0.7 9,73 43% Clay Winterwheat 10 4.3

2004 5,07 22% Forest 1 0.2 22.9 5.2

Table 4-3. Nitrogen leakage at Solmarka calculated from standard leakage defined in TRK report

(Brandt and Ejhed, 2002). NTRK - standard leakage depending on soiltype, crops and climate zone according to TRK. NTRK-Solmarka is calculated as the product of NTRK and the share of the respective soiltype and crop.

Area Share Soiltype Crops NTRK NTRK-Solmarka ha kg/ha year kg/ha year 0,93 8% Sandy loam Ley 6 0,5 7,47 68% Silty loam Ley 3 2,0

0,74 7% Silty loam Potatoes 31 2,1 0,74 7% Silty loam Winter wheat 24 1,6 2004 0,37 3% Silty loam Broccoli 27 0,9 0,75 7% Silty loam Oats 30 2,0 11. 0 9.2

Standard phosphorus leakage was calculated according to equation (1) to 0.13 kg/ha year in Skilleby and to 0.14 kg/ha year in Solmarka. Input data are summarized in Table 4-4 and 4-5.

Table 4-4. Input data from Skilleby farm to calculate phosphorus leakage according to TRK (Brandt and

Ejhed, 2002). Pforest is standard leakage for forest, other parameters are defined in the text. Parameter Units References Area 22.6 ha Forest 20% Arable land 80% LD 0.6 LU/ha Granstedt and others, 2004 S 5.93 10-6 m2/m3

PHCL 55 mg/100 g dry soil SBFI, 2002 r 1250 kg/m3 Ulén and others, 2001 xclay 0.43 Granstedt, 1990 xsilt 0.57 Granstedt, 1990 xsand 0.24 Granstedt, 1990 Q2003/04 121 mm

48 Q2004/05 185 mm Pforest 0.045 kg/ha a Ulén and others, 2001

Table 4-5. Input data from Solmarka farm to calculate phosphorus leakage according to TRK (Brandt and Ejhed, 2002). All parameters are defined in the text. Parameter Units References Area 11.0 ha LD 0.7 LU/ha Granstedt and others, 2004 S 3.13 10-6 m2/m3 PHCL 50 mg/100 g dry soil Eriksson, 1997 r 1250 kg/m3 Ulén and others, 2001 xclay 0.15 Bernhard, 2005 xsilt 0.55 Bernhard, 2005 xsand 0.30 Bernhard, 2005 Q2003/04 163 mm

Climatic and hydrologic conditions Measured precipitation is often lower than real precipitation because of losses due to evaporation from the rain gauge and due to wind effects. Results from discharge and precipitation measurements were compared with official data from the Swedish Meteorological and Hydrological Institute (SMHI). Precipitation at Skilleby was ~20% lower and runoff was ~30% lower than at the surrounding SMHI stations (Table 4-6). The difference in precipitation fits into the general precipitation pattern. However, the difference in runoff might have been caused by measuring errors e.g. ice damming in winter and during the spring flood event or leakage of water besides the gauging station. Accordingly, measured runoff was assumed to underestimate real runoff by 10%.

Differences were larger at Solmarka. Runoff at Ljungbyån was three times as large as at Solmarka whereas precipitation was similar to that at the SMHI station. The lower runoff at Solmarka is most probably an underestimation of real runoff due to the lack of continuous measurements and difficulties of measuring high discharges at the outlet of the drain pipe. Additionally, there is only one year of data available. More data is needed to draw reasonable conclusions. All results from Solmarka are based on corrected runoff, where measured runoff was multiplied by 3.3 according to the relationship between runoff at Solmarka and runoff at Ljungbyån (Table 4-6). Table 4-6. Runoff and precipitation for the period 030701-040630 at Skilleby and Solmarka and at different gauging stations from the Swedish Meteorological and Hydrological Institute (SMHI). The SMHI stations are located ~16 km from the investigation sites. Coordinates are given in local Swedish Grid (RT90, 2.5g W) Location Runoff (mm) Location Precipitation (mm) Skilleby 6548195 N 110 Skilleby 6548195 N 495 1602410 E 1602410 E Trosaån 6554410 N 165 Gnesta 6553550 N 634 1589910 E 1586690 E Saxbroån 6556690 N 172 Södertälje 6563670 N 634 1614570 E 1603480 E

Solmarka 6270230 N 50 Solmarka 6270230 N 566 1520380 E 1520380 E Ljungbyån 6285510 N 163 Kalmar 6283560 N 587 1520430 E 1529660 E

49 The regions of Skilleby and Solmarka had similar climatic conditions during the normal period of the TRK project (Precipitation, P = 650 mm, Mean air temperature, T = 7°C) (Johnsson and Mårtensson, 2002). Compared to the investigation period in the BERAS project, the Skilleby region had similar conditions (Södertälje: P = 634 mm, T = 6,8°C). The Solmarka region, however, was dryer (Solmarka: P = 566, T = 7,3°C). Thus, the climatic influence on nutrient leakage is similar at Skilleby during both the TRK period and the BERAS investigation period which implies that differences in nutrient leakage solely depend on crops and the production system. At Solmarka, however, lower precipitation might cause lower nutrient leakage than during the TRK period.

Measured leakage from ERA Water discharge, nitrogen and phosphorus concentration are shown in Figure 4-2 and 4-3. Discharge follows the usual pattern with flood events in fall and during the snow melt in spring and low discharge during summer. This results in high variability of nutrient load where large amounts can be leached during some few flood events. The difference between nutrient concentration in summer 2003 and 2004 at Skilleby were due to dry conditions in 2003, where no samples were taken. However, nutrient load is not affected significantly as discharge is low during summer.

0.4 0.4 ) ) l 2003-04 l 2004-05 / / g g m m ( (

0.2 0.2 c c n n o o c c P P 0 0 20 20 ) ) l l / / g g m m ( (

10 10 c c n n o o c c N N 0 0 0.06 0.06 ) ) s s / / l l ( ( 0.04 e e g g r r 0.03 a a h h

c 0.02 c s s i i D D 0 0 030701 030909 031118 040127 040406 040615 040701 040909 041118 050127 050407 050616

Figure 4-2. Discharge, nitrogen and phosphorus concentration, Skilleby 2003-05.

0.4 ) l 2004-05 / g m (

0.2 c n o c P 0 40 ) l / g m (

20 c n o c N 0 2 ) s / l

( 1.5

e g r 1 a h c s

i 0.5 D 0 040701 040909 041118 050127 050407 050616 Figure 4-3. Discharge, nitrogen and phosphorus concentration, Solmarka 2004-05.

Annual nutrient load is summarized in Table 4-7. In Skilleby, nitrogen leakage in 2003/04 was of the same magnitude as standard leakage for this area in the TRK project, but for 2004/05 it was the double. Nitrogen load differs significantly between the two years, whereas differences in phosphorus leakage were smaller. The large N-leakage 2004/05 most probably

50 can be explained by releasing fixed nitrogen due to ley ploughing of half of the drainage area and spreading of manure on the same area during fall 2004. Further analyses and measurements are necessary to study whether mineralization of organically fixed nitrogen in the manure can produce enough movable nitrogen during such a short time. Even in Solmarka a large part of the field area was grassland which was ploughed in fall 2004 causing large nitrogen pulses. However, results from Solmarka are uncertain due to difficulties in discharge measurements. Table 4-7. Nitrogen and phosphorus leakage at Skilleby and Solmarka in comparison with the standard leakage according to the TRK-project (Brandt and Ejhed, 2002). NTRK was calculated according to data shown in Table 4-2 and Table 4-3. PTRK was calculated according to Equation 1.

N NTRK P PTRK kg/ha year kg/ha year kg/ha year kg/ha year Skilleby 2003/04 5.7 5.3 0.18 0.14 Skilleby 2004/05 11.8 5.2 0.25 0.22

Solmarka 2004/05 21.6 1) 9.2 1) 0.14 1) 0.19 1) 1) Results are based on corrected runoff, see text for more details.

It is always difficult to draw conclusions about general nutrient leakage from a two year data series in a five-year crop rotation system. Nevertheless, we can try to generalise our two-year- results to a larger scale. The test fields at Skilleby farm consist of two lots, on which ley was ploughed in 2004 on one lot and in 2005 on the other one. In a five-year crop rotation on two fields there are two years of ley-ploughing and three years of non-ley-ploughing. Assuming, the 2003/04 results being representative for a non-ley-ploughing season (Nnlp) and the 2004/05 results for a ley-ploughing season (Nlp), mean nitrogen leakage (Nmean) can be estimated as: 2× N + 3× N N = nlp lp (3) mean 5 By applying this relationship, mean nitrogen leakage from the test fields at Skilleby farm was calculated to 8.14 kg/ha N.

The TRK results are calculated for respective area with its characteristic climate, soiltypes and livestock density. In the TRK area 60 (Skilleby) livestock density is 0.2 – 0.4 au/ha (Granstedt, 2000) whereas Skilleby farm has a livestock density of 0.6 au/ha. Nutrient surplus is strongly depending on livestock density. As shown earlier in this report (see Chapter ???) nitrogen surplus can be decreased by 40% by halving livestock density. The ERA farm produces ~40% more nitrogen leakage than farms with a 50% lower livestock density in the same area. Taking livestock density into account nitrogen leakage from ERA is of the same magnitude as the calculated standard leakage in the respective TRK area.

In the TRK area 22 (Solmarka) mean livestock density is 0.8-1 au/ha. Solmarka farm has a livestock density of 0.7 au/ha and a nitrogen leakage twice the leakage calculated in the TRK project. This is similar to the results at Skilleby during the ley-ploughing season. However, the results are limited and crop rotation is more complicated at Solmarka with several fields and crops than at Skilleby. More data is needed to draw reliable conclusions.

Phosphorus leakage from ERA is ~0.2 kg/ha year which confirms the calculated P leakage in the TRK project. The calculation of phosphorus in the TRK project depends mainly on livestock density (see Equation 1).

51 Comparison with nutrient balances Mean nitrogen surplus of all 36 BERAS farms was 36 kg/ha year. With ammonium losses of 30% respective 40%, the theoretical nitrogen leakage from the ERA farms in the project were 10 respective 7 kg N/ha year, assuming that leakage and denitrification in the soil contribute with equal parts. In addition, Skilleby farm has a livestock density of 0.6 au/ha which is similar to the mean of all BERAS farms. Mean leakage at Skilleby was calculated above to 8.1 kg/ha year. These results are, thus, in good agreement with results from the nutrient balances.

Phosphorus surplus for all BERAS farms was calculated in Chapter 2 to -1 kg/ha year. Together with a measured leakage of 0.2 kg/ha year this indicates a constant loss of phosphorus from the fields, which most probably is fed by weathering of bedrock material in the soil matrix.

Conclusions Within the BERAS project direct measurements of nitrogen and phosphorus leakage from fields were carried out on two ERA farms in Sweden and one in Finland. The data series which is available contains two years of measurements on Skilleby farm (Stockholm County). The data series from Solmarka farm (Kalmar County) and from the Finnish test farm in Partala were too short and are not reported here. The results from the measurements from Skilleby farm lead to the following conclusions: · Nitrogen and phosphorus leakage from the ERA farm in Stockholm County were 8 kg N/ha year and 0.2 kg P/ha year, respectively. These results are in good agreement with the official nutrient leakage calculated in the TRK project for the same area, when taking into account differences in livestock density. · The measured nitrogen leakage supports the results from the nutrient balances in the previous chapter of this report, where nitrogen leakage is 7-10 kg/ha year depending on the magnitude of ammoniac loses. · The measured phosphorus leakage together with the deficit in the nutrient balances indicate a constant loss of phosphorus from the soil which most probably is fed by weathering of bedrock material in the soil matrix.

References Arheimer, B. and M. Brandt. 1998. Modelling Nitrogen Transport and Retention in the Catchments of Southern Sweden. Ambio, 27(6), 471-480. Bergström, L. and H. Kirchmann. 2000. Är övergång till ekologisk odling ett sätt att minska utlakningen av kväve från åkermark? In: G. Torstensson, A. Gustafson, L. Bergström and B. Ulén (Editors), Utredning om effekterna på kväveutlakning vid övergång till ekologisk odling (Investigation of the effects of conversion to ecological (organic) agriculture on nitrogen leaching - English summary). Ekohydrologi 56. Swedish University of Agricultural Sciences, Division of Water Quality Management, Uppsala, 49-73. Bernhard, B. 2005. Personal communication, May 2005. Brandt, M. and H. Ejhed. 2002. TRK Transport – Retention – Källfördelning. Belastning på havet. Naturvårdsverket, SE-106 48 Stockholm, Rapport 5247, 120 p. Granstedt. 1990. Fallstudier av kväveförsörjning i alternativ odling. (Case studies on the flow and supply of nitrogen in alternative farming. In Swedish with English summary.). Swedish University of Agricultural Sciences, Uppsala, PhD Dissertation. Alternativ odling 4. Granstedt, A. 2000. Increasing the efficiency of plant nutrient recycling within the agricultural system as a way of reducing the load to the environment - Experience from Sweden and Finland. Agriculture Ecosystems and Environment, 80(1-2), 169-185. Granstedt, A., P. Seuri and O. Thomsson. 2004. Effective recycling agriculture around the Baltic Sea - Background report. BERAS report 2. CUL - Sveriges Lantbruksuniversitet, Uppsala, Ekologiskt Lantbruk 41.

52 HELCOM. 2004. The Fourth Baltic Sea Pollution Load Compilation (PLC-4). Helsinki Commission, Baltic Marine Environment Protection Commission, Baltic Sea Environment Proceedings 93. Johnsson, H., L. Bergström, P.-E. Jansson and K. Paustian. 1987. Simulated nitrogen dynamics and losses in a layered agricultural soil. Agric. Ecosystems Environ, 18, 333-356. Johnsson, H. and K. Mårtensson. 2002. Kväveläckage från svensk åkermark. Beräkningar av normalutlakning för 1995 och 1999. Underlagsrapport till TRK. Naturvårdsverket, SE-106 48 Stockholm, Rapport 5248, 89 p. SBFI. 2002. Measurements of phosphorus in soils on Skilleby research farm. Swedish Biodynamic Research Institute. Unpublished report. Seppanen, L. (Editor). 2004. Local and organic food and farming around the Baltic Sea. Ekologiskt jordbruk, 40. Centre for Sustainable Agriculture (CUL), Swedish University of Agricultural Sciences, Uppsala, 98 p. Shaw, E.M. 1993. Hydrology in practice. Chapman & Hall, London, 539 p. Ulén, B., G. Johansson and K. Kyllmar. 2001. Model predictions and long-term trends in phosphorus transport from arable lands in Sweden. Agricultural Water Management, 49, 197-210.

53 5. Global warming and fossil energy use Olof Thomsson, Swedish Biodynamic Research Institute, Järna Sweden Christine Wallgren, fms, Stockholm University, Sweden

Global warming and energy use are two environmental impact categories that are closely connected to each other since much of the greenhouse gas emission is caused by burning of fossil fuels. However, also emission of methane and nitrous oxide in agriculture and in industry are contributing to global warming.

The aim of this part of BERAS was to compare the environmental impacts of conventional and alternative system settings in the three most important parts of the food chain: the agriculture, the transports and the food industry. The agriculture and food industry were assessed by Olof Thomsson while the transportation energy use assessment was performed by Christine Wallgren, Center for Environmental Strategies Research (fms), Royal Institute of Technology, Stockholm. At the end of the chapter, a section written by Olof Thomsson presents the global warming impact and consumption of primary energy resources in the transportation and processing parts of the food chain.

For the agriculture, the assessment concerns comparison of conventional and ecological recycling agriculture, the latter earlier in this report defined as ERA. In this section, both the inventory results and the environmental impact results are reported together.

For the transportation and food processing assessments, the comparison concern conventional large-scale systems vs. local more small-scale systems. In the transports section, existing local distribution of locally produced food in Järna and the conventional long-distance transportation of the same product groups found in Järna stores were compared. For the food industry, the same small-scale food processors operating in Järna were assessed and compared to large-scale food processing industries. For both these system parts, the inventory results of direct and in-direct energy and resource use first are reported separately. Then, the environmental impact results are reported together for the two parts in a separate section at the end of the chapter.

Methodology The methodological aspects common for the three sub-studies are presented here. Specific methodological issues are described for each sub-study.

Data inventory and impact assessment The environmental impact assessment used in this study follows the principle of the life cycle assessment (LCA) methodology, although it is not complete LCAs. First data concerning direct and indirect energy use and resource consumption were inventoried, in LCA terminology this is called LCI (life cycle inventory). Then, these data were characterised into impact groups. One emission may contribute to several impact categories. There is a list of 15 impact categories given in the Nordic Guidelines for LCA (Lindfors et al. 1995). In this part of the study we have paid attention to two of them; Global warming impact and Use of resources, fossil energy. The global warming is measured in global warming potentials

(GWP) where all emissions are transformed into CO2 equivalents. There are three given time- spans that could be taken into account; 20, 100 and 500 years. We have chosen the 100 year perspective. Direct impacting gases only, have been counted. They are carbon dioxide (CO2),

54 methane (CH4) and nitrous oxide (N2O). The GWP of CH4 and N2O correspond to 23 and 296 CO2 equivalents respectively (IPCC, 2001). This means that every kg of methane gives as much global warming impact as 23 kg of carbon dioxide, and is thus multiplied by 23 in order to get the GWP.

The use of the resource fossil energy inventory separated the energy use in two categories of energy carriers; electricity and fossil fuels. These were then re-calculated to primary energy, i.e. the use of energy carriers were traced back to the consumption of primary energy resources. By that, it is possible to make comparisons between scenarios and activities which use much electricity with those mainly using fuels. This measure considers the consumption of resources in the lifecycle of the energy carriers. The primary energy consumption in production of average Swedish electricity demand is shown in Table 5-1. Transmission losses in the distribution net (7 %), pre-combustion energy consumption for fuels and efficiency in e.g. hydropower and nuclear power are included, making 2.35 MJ primary energy per MJ electricity used. The equivalent value for electricity produced in oil-fired power plants is 2.69 (Habersatter et al., 1998). For fuels, an average for different fuels has been used: 1.25 MJ primary energy per MJ fuel (calc. from Tables 16.4 and 16.9 in Habersatter et al., 1998). Fuel oil, fossil gas, petrol, and biofuels all have values ranging between 1.09 and 1.35 MJ.

Table 5-1. Energy resources used for average Swedish electricity use (Lundgren, 1992) Energy resource MJ energy resource per MJ electricity Fossil oil 0.064 Fossil gas 0.0093 Coal 0.040 Peat 0.0045 Biofuels 0.045 1 Uranium 1.60 2 Hydropower 0.588 Sum 2.35 1 Calculated as MJ in uranium. 35 % efficiency is used in the conversion of MJ nuclear electricity to MJ in uranium, i.e. 35 % of the theoretical heat obtained in the fission process can be utilised as electricity. 2 Calculated as MJ potential energy. 80 % efficiency is used in the conversion of MJ electricity to MJ potential energy.

Ecological Recycling Agriculture Olof Thomsson, Swedish Biodynamic Research Institute, Järna Sweden

This sub-study has investigated consumption of energy and resources and emission of gases that contribute to global warming on the 12 Swedish BERAS-farms (Table A1-1) and in average Swedish agriculture. The results are presented as consumption of primary energy resources and global warming impact.

Method Fossil energy use and global warming impact from both direct and in-direct sources were included in the study. The direct use of fossil vehicle fuels, heating oil, electricity, and lubricants were included. In-direct impacts from production of fertilisers, fodder imported to the farm, packaging materials (primarily plastics for silage wrapping), and machinery (only primary energy consumption) were also investigated.

Data for energy use, amounts of imported necessities, and production exported from the BERAS-farms were obtained from the farm accountings and interviews. For the average

55 Swedish agriculture, these data were obtained from Swedish official statistical data. Following Dalgaard et al. (2001), the energy use was completed with norm values for lubricants and machinery. Literature data were used to obtain data for energy use and emissions in the production of all the energy carriers and necessities. For the calculation of the global warming impact, literature data on direct methane emission from animals (ref) and nitrous oxide emission from soil (IPCC, 2001) were used. Methane emissions from manure storages were neglected. Energy use and greenhouse gas emissions for the production of the different inputs are shown in Table 5-2. Impacts from the production and use of pesticides were omitted since their contribution was assumed to be small (ref?). Table 5-2. Norms for the energy use in production of different agricultural inputs (with references) value unit reference vehicle fuels (diesel) MJ/l heating oil MJ/l electricity MJ/kWh lubricants 3.6 MJ/l diesel fertilisers MJ/kg N fodder MJ/kg plastics (LLDPE) MJ/kg machinery 12 MJ/l diesel

The calculations were performed for each BERAS-farm (not shown here) and for four production type groups of the BERAS-farms. All farms have more or less diversified production but they were group according to their main production in: 1) Mixed production with vegetables (2 farms), 2) Milk, beef and cereals (6 farms) 3) Pork, poultry, egg and cereals (2 farms) 4) Cereals and beef (2 farms)

When grouped, the data for the included farms were aggregated and the group treated as one farm unit. The results are reported for the production groups of the BERAS-farms, for an average of the BERAS-farms and for average Swedish agriculture. Table 5-3 shows the number of different animals per hectare for the different groups .

Table 5-3. Number of animals of different kind per hectare for the different BERAS-farm groups and for Swedish average agriculture, number per hectare BERAS type-farms Cattle Sheep and goats Pigs Poultry BERAS farms - mixed prod. with vegetables 1.1 0.0 0.0 0.4 BERAS farms - milk, beef and cereals 1.1 0.3 0.0 0.0 BERAS farms - pork, poultry, egg and cereals 0.7 0.2 1.1 2.9 BERAS farms - cereal and cattle prod. 0.4 0.0 0.0 0.0 BERAS farms – average 0.9 0.2 0.3 0.9 Swedish average agriculture 0.7 0.2 0.8 2.7

Results and discussion

The global warming impact, measured in Global Warming Potentials (GWP) as CO2 equivalents, is reported per kg products exported from the farm (Figure 5-1) and per hectare (Figure 5-2). Please observe that the per-hectare figures are calculated per hectare on the actual farms, i.e. the acreage used for imported fodder is not included. In both cases the GWP were lower for the average BERAS-farm than for the average Swedish agriculture. The difference was reduced by the larger emission of methane from animals. That is explained by

56 the larger share of ruminant animals on the BERAS-farms, compared to average Swedish agriculture having a larger part of monogastric animals which emit very little methane. Secondly, it is assumed that these cattle are fed on more roughage making them emit more methane compared to conventional cattle. The very large GWP from animals in the “Pork, poultry, egg”-group originates mainly from cattle (beef production) that also occur on these farms. The global warming impact from other factors investigated was substantially lower on the BERAS-farms compared to average Swedish agriculture. 1,6

1,4

1,2

1,0

0,8

0,6 equiv. per kg products exported

2 0,4

kg CO 0,2

0,0 BERAS farms - BERAS farms - BERAS farms - BERAS farms - BERAS farms - Swedish average mixed prod. with milk, beef and pork, poultry, egg cereal and cattle average agriculture vegetables cereals and cereals prod. Soil 0,05 0,04 0,20 0,08 0,06 0,09 Animals 0,40 0,49 0,83 0,29 0,49 0,33 Plastics 0,00 0,00 0,01 0,01 0,00 0,00 Imported fertiliser/manure 0,04 0,00 0,00 0,06 0,01 0,21 Imported fodder 0,01 0,01 0,04 0,00 0,01 0,03 Fossil energy 0,14 0,08 0,34 0,16 0,11 0,19

Figure 5-1. Global warming potentials for different BERAS-farm groups and for Swedish average agriculture, kg CO2 equivalents per kg products exported

3 000

2 500

2 000

1 500 equiv. per hectare 2 1 000 kg CO

500

0 BERAS farms - BERAS farms - BERAS farms - BERAS farms - BERAS farms - Swedish average mixed prod. with milk, beef and pork, poultry, egg cereal and cattle average agriculture vegetables cereals and cereals prod. Soil 216 157 163 121 160 250 Animals 1 725 1 867 681 458 1 385 884 Plastics 3 5 7 8 6 10 Imported fertiliser/manure 166 0 1 98 22 562 Imported fodder 29 35 32 4 30 69 Fossil energy 582 321 280 249 321 511

Figure 5-2. Global warming potentials for different BERAS-farm groups and for Swedish average agriculture, kg CO2 equivalents per hectare

The consumption of primary energy resources are shown in Figure 5-3 and Figure 5-4. Counted both per kg products and per hectare, the consumption is substantially lower on the BERAS-farms in average compared to average Swedish agriculture. The most important reason is the lesser use of fire oil (for drying of grain), fertilisers and electricity. Concerning the fire oil, this might be due to a lower rate of on-farm drying but this was not investigated. If that was the case, that oil would be used in the industry instead and therefore, in a systems perspective, that part should be neglected. The differences is, however, still obvious. The difference in electricity consumption was not possible to explain within the study.

Notable is also the very large diesel consumption in the “Pork, poultry, egg”-group, making this production more energy resources demanding than other production. This contradicts the common opinion that meat production from monogastric animals like pigs and poultry is more energy efficient. Whether this is a result of chance, there are only two farms in the group, or some other factors have not been possible to sort out. 14

4 MJ per kg products exported

0 BERAS farms - BERAS farms - pork, BERAS farms - BERAS farms - milk, BERAS farms - Swedish average mixed prod. with poultry, egg and cereal and cattle beef and cereals average agriculture vegetables cereals prod. machinery 0,90 0,58 2,46 1,03 0,79 0,77 plastics 0,04 0,08 0,55 0,30 0,13 0,24 fertilisers 0,27 0,00 0,01 0,02 0,03 1,49 imported fodder 0,13 0,26 0,73 0,06 0,26 0,54 electricity 1,40 2,22 2,26 2,06 2,12 3,77 fireoil 0,00 0,00 0,12 0,24 0,02 1,25 vehicle fuel 2,26 1,37 5,70 2,39 1,88 1,75 Figure 5-3. Consumption of primary energy resources for different BERAS-farm groups and for Swedish average agriculture, MJ primary energy resources per kg products exported

58 30 000

25 000

20 000

15 000 MJ per ha

10 000

5 000

0 BERAS farms - BERAS farms - BERAS farms - BERAS farms - BERAS farms - Swedish average mixed prod. with milk, beef and pork, poultry, egg cereal and cattle average agriculture vegetables cereals and cereals prod. machinery 3 873 2 213 2 010 1 619 2 219 2 032 plastics 178 309 445 477 354 630 fertilisers 1 167 0 7 37 91 3 939 imported fodder 572 980 593 99 736 1 438 electricity 6 021 8 515 1 843 3 227 5 934 9 986 fireoil 0 0 101 370 64 3 305 vehicle fuel 9 721 5 266 4 651 3 757 5 272 4 631

Figure 5-4. Consumption of primary energy resources for different BERAS-farm groups and for Swedish average agriculture, MJ primary energy resources per hectare

Discussion To use Swedish average agriculture as representative for the conventional food producer for Swedish consumption in the global warming and energy use impact assessment is a weak point in the comparison. The data available are not fully comparable to the data obtained at farm-level recording….

Comparison to other studies The assessment of the agriculture was compared to results obtained by Engström (2004), which assessed the environmental impacts from Swedish food production and consumption. The results concerning Swedish average agriculture… A comparison between the studies concerning the environmental impacts from Swedish food consumption is presented in Chapter Food basket scenario, Järna Sweden.

Transports of locally produced food in Järna Christine Wallgren, fms, Stockholm University, Sweden

This sub-study describes how locally produced and consumed food in Järna Sweden, are transported, and how much fossil energy that transportation has used. The study comprises vegetables, potatoes and root crops from four local growers; milk and dairy products from Järna Dairy; bread from Saltå Mill & Bakery; and meat from animal keepers in the vicinity of Järna. The products are collected at the different producers and delivered to stores, schools and other large kitchens in both the Järna area and in Stockholm (about 60 km away).

Method The transports are described in terms of which vehicles that have been used, which routes that have been driven, distances, amount of products transported, and energy use for the

59 transports. The energy use has been calculated as MJ/kg product transported. Only direct use was counted and transports for necessities and other ingredients than the main raw products were usually omitted. If included that is commented in the text. The results have been compared to the transportation of equivalent products in today’s conventional (large-scale) food system. Data used for comparisons to the conventional system are in most cases earlier reported in Carlsson-Kanyama et al. (2004).

All calculations are based on data for year 2004 and/or measurements performed during spring 2005. For some not accessible basic data, estimations have been done based on assumptions presented in the report.

Three case studies of small-scale, ecological and local production and distribution are included: 1. Saltå Kvarn (mill and bakery); buying grain from ecological farms in Järna and in the south-central part of Sweden. Producing cereal products and bread sold all over Sweden. Included here are the transports to Järna and Stockholm consumers. 2. Järna Odlarring (a local farmers cooperation); buying and selling both newly harvested vegetables and root crops from four farms in Järna (during season), and meat from ecological beef producers in Järna (all year around). 3. Järna Mejeri (a farm-size dairy); producing and selling milk, yoghurt and cheese from cow milk produced on two farms in Järna.

Data collection Data for the transportation of vegetables have been collected from the bookkeeping for Järna Odlarring 2004, which record sold amounts of vegetables per every single day. The amount delivered is equivalent to amount sold since there are no intermediate storage and no losses (since the products are harvested the same day). Possible losses on the farms are ploughed into the soil. Weekly averages of deliverances were calculated from the basic data. Transport data was calculated for two different deliverance routes during three average seasonal routes; early, mid and late seasons. When vegetables and meat were co-transported that was taken into consideration. Two different cars were used. The average fuel consumptions for the two cars were calculated separately per yearly basis. The route distances were measured by the trip-km-meter in the cars.

For the meat transports, data were obtained both from the slaughterhouse company (that transported the animals from the farms to the slaughterhouse) and from Järna Odlarring (that performed the deliverances). Data for the animal transports were average data for distances and fuel consumption, and a combination of detailed and yearly average data concerning the number of animals. The volumes and distances for the deliverance routes were recorded during two weeks in April 2005. The measured routes were two transports from the slaughterhouse combined with two routes customers in the vicinity of Järna, and two routes to Stockholm. Delivered amounts were recorded at the loading of the cars and the distances driven were recorded for each trip. The cars used for vegetable transports were used also for the meat transport and the fuel consumption and deliverance route distances were assumed equal in both cases.

During the winter season, only meat is transported in the cars. During the early and late vegetable seasons, vegetables and meat are co-transported as often as possible. Strict rules for this, concerning packaging and temperatures, have to be followed. However, the energy use

60 calculations were performed separately for vegetables and meat. Allocation per weight was used in order to separate the energy use proportionally to vegetables and meat respectively.

For the milk and dairy products, the data collection was performed during April 2005. The routes are one collecting milk on two Järna farms (the dairy is situated at one of the farms), and three deliverance routes; two local routes to customers in Järna and one to Stockholm. The three deliverance routes are about the same every week all year around. The distances were recorded for each route. The fuel consumption was measured by filling up the car before and after each route. Two different cars were used. The amounts of products loaded were recorded at each measuring occasion.

The bread transports calculations were based on average deliverances during one year. This was assumed appropriate since the bread deliverances to Järna, and thus the consumption, are relatively constant over the year. The transports included here are performed by the company’s own car to grocery stores in Järna and Stockholm. Bread is also delivered to other distributing channels, primarily to COOP, but they are omitted in this study. The transports in these cases are performed by the distributors themselves.

Energy use calculations For each transport, data for fuel use, distance and amount products were collected as described above. All transports were performed by diesel-fuelled vehicles. The energy use per kg product was calculated in MJ/kg product. First, the fuel use per km was multiplied by the number of km per route – giving the fuel use per route in litre/route. This figure was then multiplied by 35.86, which is the energy content of 1 litre diesel – giving energy use per route in MJ/route. Lastly, the energy use per route was divided by the number of kg products transported – giving energy use per kg product transported. These calculations were performed for meat, vegetables, milk and dairy products, and bread separately.

Comparing data for transports of similar conventional products are collected from several sources but all are referred to in Carlsson-Kanyama et al. (2004).

Results and discussion The results are presented in text here and included in diagram form together with the processing industry in the following sub-chapter.

Saltå Kvarn Saltå Kvarn produced 11 000 loafs of bread per week (production on 5 days) during 2004 out of flour produced in the own mill. The grain was bought from about 20 farms, some in Järna but also farms in the south and central part of Sweden. The largest farm is situated outside Lidköping in the west of Sweden, some 330 km away. Grain is also bought from a wholesaler. The farms around Järna delivers about 500 ton. In total, Saltå Kvarn buys about 4000 ton per year. Everything is transported by truck or tractor and wagon.

The weight of the bread produced in the bakery was 6 600 kg per week or 343 ton per year. To produce one kg bread, about 0.6 kg flour was used and to produce one kg of flour about 1.3 kg grain was used making 0.78 kg grain per kg bread. So, in order to produce the 343 ton bread per year 267 ton grain was used. That amount is well within what the Järna farms delivers, thus we have only accounted for the local grain in-transports.

61 Transportation of ingredients The transport of grain produced locally is transported by the farmers with tractor and wagon. The fuel consumption is 10-12 litre diesel per hour. The grain is transported direct from the field to a silo at Järna Kvarn. This transport is between 3 and 5 km and take scarcely an quarter of an hour one way. Assuming the average load was 13 ton and each trip took 26 minutes give 4.8 litre diesel per 13 ton. This makes 0.013 MJ per kg grain or about 0.01 MJ per kg bread. This grain was, after drying, transported by truck to storage in Tystberga 42 km away. This transport used 0.026 MJ/kg bread. Together the grain transports used 0.036 MJ per kg bread.

Other ingredients as dried fruit and seeds stand for about 7 % of the bread weight, some of them imported from e.g. Turkey. A rough estimation, assuming that most of them originate in Europe and use 5 MJ/kg for transports, give 0.35 MJ per kg bread. Hypothetically, it could of course be argued that the ingredients could be exchanged for locally produced berries and seeds but we have chosen to include the energy use for the imported ingredients. The sum for transportation of ingredients in the actual case was about 0.4 MJ per kg bread.

Deliverances of bread Of the bread produced, about 16 ton was delivered locally in Järna with vicinity. 187 ton was delivered with own cars to Stockholm and 140 ton was collected by distributors at the bakery. Saltå Kvarn also sell flour, groats, muesli and imported colonial products like seeds and dried fruit but these are also transported by distributors and not included in this study.

The local deliverances were made with a small lorry consuming 15 litre diesel per 100 km, to two grocery stores in Järna (ICA and COOP). The distance is 8 km round trip. About 100 loafs (55 kg) were delivered each time. No other transports were done at this transport and the car went empty back to the bakery. The fuel consumption was thus 1.2 l/route, making the energy use 43 MJ/route or 0.78 MJ per kg bread delivered.

Deliverances to Stockholm were made five days a week with the same car used for the local transports. About 1200 loafs weighing 720 kg were delivered each time. The average fuel use was the same as for the local transports (15 l/100 km) and the route was 150 km. The fuel consumption was 22.5 litre diesel, making 807 MJ per route or 1.12 MJ per kg bread delivered. Also here, no other transports were done and the car went empty back to the bakery. The car was, however, full when starting from the bakery, meaning that no extra space was available.

Energy use per kg bread The conclusion is that the total energy use for transports of bread was 1.2 MJ per kg bread delivered locally in Järna. Using 100 % locally produced ingredients and local storage would lower the energy use by about 0.3 M//kg bread (a part of 0.35 and 0.026), resulting in a hypothetical energy use of 0.9 M//kg bread. However, the most influencing factor was how much that was transported each time. If the volume would have been 300 loafs instead of 100, the energy use would be a third. The second most influencing factor is which car that is used. Using a smaller car with fuel consumption of 7 l/100 km would halve the energy use.

Comparison to LCA-studies of conventional bread transports show 1.0 MJ/kg bread (Carlsson-Kanyama and Faist, 2000) and 3.8 MJ/kg bread (Johannisson and Olsson, 1998).

62 Järna Odlarring Järna Odlarring has fixed deliverance routes of meat and vegetables but delivers also at other times and places in order to be flexible and to fulfil wishes of their costumers. Not all routes are, therefore, optimised from an energy efficiency point of view. Two vehicles are used; a Citroën Berlingo model 1999 with a fuel consumption of 7 litre diesel/100 km (cooling cabin) and a Fiat Ducato model 2000 with a fuel consumption of 12 litre diesel/100 km. Both cars are run without trailer.

Animal transports Animals were collected on six farms around Järna and transported to a small slaughterhouse in Stigtomta outside Nyköping some 60-70 km away. These transports were performed by the slaughterhouse company. Animals were collected once every other week, usually 2-4 cattle, and during October-November also ca 10 lambs, each time. Besides these, also 10 calves from these farms are killed each year. Per year it makes about 75 cattle, 40 lambs and 10 calves.

The vehicle was a Scania animal transport lorry of the smallest size, consuming 30-50 litre diesel/100 km. Max load is 13 cattle or 50 lambs. It is not allowed to mix animals from ecological and conventional farms, so the car was not always full. The route distance varied depending on which farms that delivered animals. An average round trip was assumed to be 156 km.

Meat transports All deliverances of meat from the slaughterhouse to the Järna consumers were done by Järna Odlarring. The meat was delivered in closed packages, with the same cars that transported vegetables. Beef was cut in ready-to-eat pieces and vacuum-sealed, which is allowed to be co- transported with vegetables. Lambs were cut with bones and therefore not vacuum-sealed but packed in plastic bags that were put in cardboard boxes, which are not allowed to be co- transported. A thorough washing has to be performed between the transports. When possible the small van is used. The larger lorry is used about half of the tours to Stockholm.

On Wednesday, even weeks, meat was fetched at the slaughterhouse and delivered to the two grocery stores in Järna. Some, that was not possible to sell immediately, was put in cool and cold storages in Järna. On Wednesdays, odd weeks, meat from the own storages were delivered to the two stores in Järna and to stores in Stockholm. The routes are about the same all year around. Deliverances to large kitchens were performed on Tuesdays each week. The orders were more varied, thus no average routes are possible to set. About 23 routes from the slaughterhouse and to Stockholm are performed per year.

Each cattle give about 140 kg bone-free beef. A calf gives about 60 kg. An average lamb gives 17.5 kg meat including bones. In one year that makes about 10.5 ton beef, 600 kg calf meat and 700 kg lambs meat. Per deliverance that makes about 480 kg beef and calf meat. Adding the 700 kg lambs meat produced during October-November, 510 kg meat is transported per deliverance in average. During autumn 2004, a test was performed with slaughter and of deliverance pig meat. This was not included in the calculation. If there are enough sales it is assumed that 3 pigs every other week could be added to the deliverances. This would probably make the energy efficiency for the transports better.

Calculations of energy use for animal and meat transportations are performed in three steps: 1. The energy use for transporting the animals to the slaughterhouse was 156 km per route * 0.4 l diesel per km * 35.86 MJ per l diesel / 510 kg meat per route = 4.4 MJ per kg meat

63 2. The energy use for fetching the meat at the slaughterhouse was 156 km per route * 0.103 l diesel per km * 35.86 MJ per l diesel / 510 kg meat per route = 1.1 MJ per kg meat 3. The energy use for deliverances to the Järna customers was …. = 0.45 MJ per kg meat. The value is a weighted average of different routes and seasons where the energy use has been allocated between meat and vegetable transports. (The calculations can be obtained in an Excel-sheet from the author.)

Energy use per kg meat Together this makes 6 MJ per kg meat delivered in Järna. For the meat delivered to Stockholm, the deliverance route used 5.6 MJ per kg meat, making the total energy use for transports around 11 MJ per kg meat.

The addition of three pigs every other week would lower the energy use to X.Y MJ per kg meat. If the volumes increased to a level where the animal transport lorry was fully used, that energy use would be lowered from 3.9 MJ per kg meat to about 1 MJ per kg meat. The total energy use should then be around 4 MJ per kg meat.

Energy use for conventional animal and meat transports, collected from different LCA- studies, are presented in Table 5-4. Table 5-4. Energy use for conventional meat transports, from other studies. Examples Reference MJ per kg meat Meat from Sweden Johanisson and Olsson (1998) 0.45 Meat from Sweden Edsjö (1995) 0.97 Meat from Linköping to Sunnerstedt (1996) 0.46 Stockholm Meat in Europe Carlsson-Kanyama and Faist (2001) 5.7 Meat from Ireland to Stockholm Sunnerstedt (1996) 1.2 Meat from Latin America Wallgren (2005) 12

Vegetable transports Collection of the vegetables and root crops at the growers was made at the same route as the deliverance. No intermediate storage was used. The deliverances 2004 started the last week in May. During the first period, here called early season, which lasted to 1 July, one deliverance route was made per week in Järna and its vicinity. This route was 48 km.

The second period, here called main season, three deliverance routes per week were made. Two of these went to the Järna area, were performed on Tuesdays and Thursdays and were 33 km long. The third route, on Wednesdays, delivered vegetables to Stockholm. This route was 123 km long. Together this makes 189 km per week for the three routes.

During the late season, 1 October to 30 November, one local route in Järna was carried out each week and one route to Stockholm every other week. Distances were as before. A few after-season deliverances were performed during November-December of root crops and vegetables stored at one of the producers, Skilleby farm.

The harvest, and thereby the deliverances, varied much between the periods. The deliverances per week are shown in Figure 5-5. In total, 21 500 kg was delivered during 2004. The average weekly deliverances for the three seasons are shown in Table 5-5.

64 2500

2000

1500

1000 kg per week 500

0 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 week

Figure 5-5. Weekly deliverances of vegetables and root crops by Järna Odlarring from week 20 to week 44 during 2004.

Table 5-5. Average weekly deliverances for the three seasons. Period number of weeks weight per period weight per week Early season 9 1653 184 Main season 12 14561 1213 Late season 5 4579 925 2004 26 21498 827

Energy use per kg vegetable The energy use was calculated as a weighted average of the deliverances at the three seasons given in Table 5-5 and the estimated fuel consumption for the different routes. Also the co- transportation of vegetables and meat was considered. The results are that the energy use for the local transportation of vegetables and root crops in the Järna area consumed 0.3 MJ per kg product. Equivalent value for the deliverances to Stockholm was 5.9 MJ per kg product.

Comparisons to LCA-studies show some results in the same range but most of them show a larger energy use per kg product (Table 5-6). Table 5-6. Energy use for vegetable transportation, results presented in other studies. Carlsson-Kanyama et al. (2004) Sunnerstedt (1996) Sweden/Denmark Potatoes 0.8 0.6 (locally 0.06-0.15) Tomatoes 0.8-0.9 0.6 Carrots 0.8-1.0 0.17 (Gotland) Apples 0.9-1.5 0.6 Europe Tomatoes 1.8-3.9 3.0 Tomatoes, by air 50 Carrots 1.2-3-2 0.9-2.3 Lettuce 1.5* Apples 2.8 1.8-2.1 Other continents Apples 5.6-7.7 *Average for conditions in Europe, not for Stockholm

65 Järna Mejeri

Milk in-transports Transport of milk to the dairy is carried out every second day all year around. The vehicle used was a Peugeot 1999 with trailer. Fuel consumption was 8 litres per 100 km. The milk was fetched at two farms at 0.5 and 4 km distance respectively. The collection route was 9 km round trip. In average, 1000 litres were collected at Nibble and 800 litres at Yttereneby each time. The milk was transported in a cooler tank. Each collection trip used 0.72 litre diesel, making the energy use 25.8 MJ per 1800 litres. It is assumed that one litre milk is equivalent to one kg milk, resulting in 0.014 MJ per kg milk.

Deliverances of dairy products Deliverances of milk and dairy products to costumers are done with two different cars; the above mentioned Peugeot and a diesel fuelled Daewoo 2000 with a fuel consumption of 12 litre diesel per 100 km. Three different deliverance routes are used. Route 1 – Järna with vicinity. The route was 39 km and the Peugeot (8 l/100 km) was used. It starts at the dairy and stops at Nibble garden – Saltå mill – Mossvägens Livs – Konsum Järna – ICA Järna – Konsum Hölö and then back to the dairy. The tour went once a week and the deliverance at the measuring week (w. 11) was 350 kg drinking milk, 108 kg sour-milk, 78 kg yoghurt and 10 kg other products (but no cheese). In total 545 kg delivered products. The fuel consumption was 3.1 litre diesel. Thus, the energy use was 0.204 MJ per kg product. When adding the energy use for the in-transport to the dairy (0.014 MJ/kg) the total energy use was 0.22 MJ/kg. Route 2 – Järna with vicinity. This route goes from the dairy – Nibble garden – ICA Järna – Konsum Järna – Konsum Gnesta – Konsum Hölö– The Culturehouse in Ytterjärna – the dairy. The route was 57 km and the Daewoo (12 l/100 km) was used. This route was driven twice a week and the fuel consumption was 6.8 litre diesel per route. At the measuring week 620 kg products were delivered, making the energy use 0.39 MJ per kg product. When adding the in- transport (0.014 MJ/kg) we get 0.40 MJ/kg. Route 3 – to Stockholm delivered to about 20 costumers using the Daewoo. The route was 195 km 23.4 litre diesel was used. The deliverances were; 268 litre milk, 70 litre sour-milk, 136 litre yoghurt, 72 kg cheese and 29 kg other products, in total 585 kg. That makes the energy use 1.43 MJ per kg product delivered. When adding the energy use for in-transport (0.014 MJ/kg) the total was 1.44 MJ/kg.

Energy use per kg dairy product The weighted average of the energy use for the locally delivered dairy products (Route 1 and 2) was 0.32 MJ per kg.

Results from LCA-studies of conventional production are shown in Table 5-7 Table 5-7. Energy use for transports, results from other studies. Carlsson-Kanyama et al. (2004) Sunnerstedt (1996) Sweden/Denmark/Finland Drinking milk Yoghurt 0.30-0.38 Cheese 1.7 1.2 Europe Yoghurt 2.1 Cheese 9* 1.9 *Average for conditions in Europe, not for Stockholm

66 Small-scale food processing industries in Järna Olof Thomsson, Swedish Biodynamic Research Institute, Järna Sweden

This sub-study has investigated the use of energy and packages in small-scale food processing industries for bread, vegetables and root crops, dairy products, and meat. The industries studied are the same companies studied in the transports sub-study; a mill and bakery company (Saltå Kvarn), a farmer’s cooperative meat and vegetables wholesaler (Järna Odlarring), a farm-size dairy (Järna Mejeri), all situated in the vicinity of Järna, and a small slaughterhouse (Stigtomta Slakteri) situated some 65 km away from Järna. These are compared to “conventional” large-scale food processors.

Method Data for direct energy use, use of packaging materials (in-direct energy use), and production in the small-scale industries were collected through bookkeeping records and interviews with responsible persons on each company for the year 2004.

Direct energy use, electricity, heating oil and fossil gas were inventoried separately in order to be able to calculate the consumption of primary energy resources reported in the next section. For the indirect energy use, the production and inherent energy of packaging materials were included (Audsley et al. 1997; Habersatter et al. 1998), also here electricity and fossil fuels were recorded separately. Energy use and recovery from packaging waste management, from the import of other raw materials and from water production and wastewater treatment were not included.

Data for similar products produced in large-scale industries were mainly obtained from LCA Livsmedel (2002). That was a project within the Swedish cooperative food industry performing seven LCAs on food products. The complete studies are not published but the aggregated results were presented in Swedish in the popular publication referred to and on some of the companies’ websites. The products studied were: drinking milk (1.5 % fat), beef (from dairy production animals, which represent 70 % of the Swedish beef production), pork, chicken, hamburger bread, potatoes, and iceberg lettuce. Here the waste management of packaging material was included.

The results are presented as MJ per kg product delivered.

Results and discussion The direct energy use and indirect energy in MJ per kg product for bread, vegetables and root crops, meat, and dairy products in the Järna food industries and in examples or averages of large-scale food industries are shown in Table 5-8. The results show no trend whether the small-scale or large-scale in general use the least energy. One could expect that the large-scale industries should use less energy per kg product due to the large-scale efficiency factor. For the meat and dairy products examples that is true but for the bread and vegetable examples the contrary is shown. One should, however, remember that there are large differences concerning energy use in the literature and probably also in practise. The energy use in the different productions is further discussed in separate sections below.

67 Table 5-8. Energy use in processing of four different product groups for conventional food systems and small-scale local food systems, MJ per kg product direct use direct use packages packages Total electricity fossil fuel electricity fossil fuel conventional7 3.8 0.5 0.4 1.7 6.4 bread Järna8 1.3 2.2 0.2 1.6 5.3 vegetables, root crops conventional9 0.2 0.0 0.7 1.1 2.0 and potatoes Järna 0.0 0.0 0.0 0.0 0.0 conventional10 2.6 1.4 0.1 2.9 7.0 meat Järna 5.2 0.0 0.3 3.3 8.8 conventional11 0.2 0.2 0.2 1.0 1.6 dairy products Järna 0.9 0.0 0.0 1.3 2.2

Bread The bakery on Saltå Kvarn used less fossil energy per kg bread than the conventional example but that was production of hamburger bread. In Saltå many different kinds of bread are baked. Saltå also used some fire-wood (some 500 m3 piled paper wood annually) but that was assumed to be CO2-neutral and thus not included in the assessment. In scenario simulations based on other sources, presented in Thomsson (1999), the direct energy use in bakeries ranged from 1.6 to 3.8 MJ per kg bread and indirect energy use in packaging from 1.2 to 6.3 MJ per kg bread (0.5 – 2.5 when the packaging material was recycled and incinerated). Together the range in that study was 2.8 to 10.1 MJ per kg bread (or 2.1 to 6.3 when the recycling was included) indicating that this comparison might be a bit unfair. On the other hand, if recycling of the packaging material would have been included in the Järna study the total would be in the range of 4.5 MJ per kg bread.

Vegetables and root crops The processing of vegetables and root crops were not using any large amounts of energy. In Järna, the most products were delivered the same day as it was harvested without much treatment, and the handling that was carried out during the growing season implying that no heated locals were needed. The office was not included, which also seem to be the case in the conventional study. Concerning packages, the major energy use was used for the lettuce in the conventional example. The lettuce was put into plastic bags (primary package) and then transported in single-use cardboard boxes (secondary package). In the Järna case, the most of the vegetables were transported without primary package in re-usable plastic boxes.

Meat As expected the conventional meat processing use less energy in the plant (25 % less). Since Stigtomta is a small slaughterhouse, and since it is specialised on “individual slaughter”, i.e. farmers are guaranteed to get their own animals if they want, that difference is almost surprisingly small – a difference that very well can be compensated in other parts of the food system. Concerning the packages, the difference is around 15 % and that probably is within the margins of error in the estimation.

7 hamburger bread (LCA Livsmedel 2002) 8 all kinds of bread 9 average of potatoes and lettuce (LCA Livsmedel 2002) 10 average of beef and pork, since that is what is slaughtered in Stigtomta (LCA Livsmedel 2002) 11 drinking milk (LCA Livsmedel 2002)

68 Dairy products The energy use in the small-scale Järna dairy was more than double than the conventional example. However, they also produced a lot of cheese which use about 10 kg milk per kg cheese, which probably has an effect of the energy use since it makes less kg products to divide the energy use with.

Global warming impact and primary energy consumption in a local food chain in Järna Olof Thomsson, Swedish Biodynamic Research Institute, Järna Sweden

This section describe emission of gases that contribute to global warming and consumption of primary energy resources in the transportation and processing parts of the food chain described in previous sections.

Method The environmental impacts were assessed as described above in the beginning of this chapter. The energy and packaging use data collected in the inventories of the transportation and processing were the starting point for the calculations. It was assumed that the electricity used was produced like the Swedish average, which roughly is produced by half hydropower and half nuclear power, and minor parts of fossil fuels. Concerning the emission of global warming gases in the production of packaging material, data from Audsley et al. (1997) and Habersatter et al. (1998) were used. The results are presented as CO2-equivalents per kg product and MJ primary energy resources per kg product. Please note that this MJ is not the same unit as the MJ presented in the previous sections. The unit used here represents the consumption of resources while the ordinary MJ-unit represents the use of an un-defined energy quality in a process where the energy quality is degraded, e.g. chemical energy in wood degraded to diffuse heat.

Results and discussion When looking at the GWPs for the processing and transports together, the difference between conventional and the Järna systems is substantial for bread, vegetables & potatoes and dairy products – especially for the latter two (Figure 5-6). The very inefficient transports for the meat produced in Järna and slaughtered in Stigtomta makes the GWP equal to the conventional. Somewhat larger volumes, as discussed in the Transports section, would make the result for the Järna alternative better.

It should, however, be remembered that the data for the conventional processing is not entirely comparable to the Järna cases, and the literature data for transports from different references vary a lot. In other words, this comparison has to be taken not as a truth but as an example on differences that may occur. Both the conventional and the local small-scale can be better or worse concerning global warming impact. For example is the choice of energy carrier very important for the result. Comparing the results for meat in Table 5-8 and Figure 5-6 may serve as a good example. The energy use in the conventional plant is 2.6 MJ el +1.4 MJ fossil fuels, equals 4.0 MJ per kg meat, while the Järna plant use 5.2 MJ electricity. Thus 30 percent more energy is used in the Järna plant but when it comes to GWPs the Järna plant show 90 percent less impact. Looking at the bread processing alternatives, the comparison between large-scale and small-scale is the opposite. If the electricity had been produced from fossil fuels, the proportional difference between the systems for global warming impact from

69 direct energy use would have been equal to that in Table 5-8. In order to show a broader picture of the environmental impacts more impact categories should have been included but that was not possible within this project.

Concerning the GWPs from packages and transports, these represent a “fair” difference since both packaging materials and fuels in principle are of equal origin.

0,7

0,6

0,5

0,4

0,3

0,2 kg CO2 equiv. per product

0,1

0,0 conventional Järna conventional Järna conventional Järna conventional Järna transports 0,19 0,09 0,48 0,02 0,26 0,47 0,16 0,03 packages 0,15 0,03 0,09 0,00 0,15 0,07 0,04 0,03 direct energy use 0,09 0,18 0,00 0,00 0,18 0,02 0,02 0,00 bread vegetables and potatoes meat dairy products Figure 5-6. Global warming potentials for conventional processors and Järna small-scale processors for different product groups, kg CO2 equivalents per kg products exported The general results for consumption of primary energy resources show similar patterns as the GWP (Figure 5-7). The exception is for the meat processing, where the Järna case show a substantially larger consumption. This is due to that the slaughterhouse only uses electricity which, as also discussed above, in Sweden is very “global warming friendly”, but that use a lot of primary energy resources. Also here, the contrary phenomena can be observed for the bread processing cases. A sensitivity analysis, where the Järna cases use the same proportion of electricity and fossil fuels as the conventional, show …

70 30

5 MJ primary energy resources per kg product

0 conventional Järna conventional Järna conventional Järna conventional Järna transports 3,0 1,5 7,6 0,4 4,2 7,5 2,6 0,4 packages 3,1 2,4 3,1 0,0 3,9 4,8 1,8 1,7 direct energy use 9,5 5,7 0,4 0,0 7,9 12,2 0,8 2,1 bread vegetables and potatoes meat dairy products Figure 5-7. Consumption of primary energy resources for conventional processors and Järna small-scale processors for different product groups, MJ primary energy resources per kg product

References Audsley, Eric. Alber, Sebastian. Clift, Roland. Cowell, Sarah. Crettaz, Pierre. Gaillard, Gérard. Hausheer, Judith. Jolliett, Olivier. Kleijn, Rene. Mortensen, Bente. Pearce, David. Roger, Ettiene. Teulon, Hélène. Weidema, Bo. van Zeijts, Henk. 1997. Harmonisation of environmental life cycle assessment for agriculture. Final Report Concerted Action AIR3- CT94-2028. European Commission DG VI Agriculture (Silsoe, UK) Berlin, J. (2001), LCI of Semi-Hard Cheese Mat21, SIK-rapport 692. Carlsson-Kanyama & Faist (2000), Energy Use in the Food Sector - a data survey. Carlsson-Kanyama, Sundkvist & Wallgren. 2004. Lokala livsmedelsmarknader – en fallstudie. Miljöaspekter på transporter och funktion för ökat medvetande om miljövänlig matproduktion. (In Swedish: Local food markets – a case study.) Centrum för miljöstrategisk forskning – fms (Center for Environmental Strategies Research). KTH. Stockholm Dalgaard et al. 2001. A model for fossil energy use in Danish agriculture used to compare organic and conventional farming. Agriculture, Ecosystems and Environment 87 (2001). pp 51-65. Fernlund, T. m.fl. (1997) Baktankar – En liten bok med brödrecept, Saltå kvarn, Balders förlag. Johannisson och Olsson, (1998) Miljöanalys ur livscykelpersepktiv av fläskkött och vitt bröd. Förstudie. SIK-rapport 640. Habersatter et al. 1998. Life Cycle Inventories for Packagings. Volume 1 and 2. Environmental Series Waste (250). SAEFL. Berne Husus-rapport 2005; Wallgren & Höjer (forthcoming). IPCC. 2001. (Authors: Boucher, O. Haigh, J. Hauglustaine, D. Haywood, J. Myhre, G. Nakajima, T. Shi, G Y. Solomon, S.) Radiative Forcing of Climate Change. In: Climate Change 2001 (6): The Scientific Basis. pp 349-416. IPCC. Cambridge LCA Livsmedel. 2002. Maten och miljön. Livscykelanalys av sju livsmedel. (In Swedish: The food and the environment. Life cycle assessment of seven foodstuffs). LRF. Stockholm.

71 Lindfors et al. 1995. Nordic Guidelines on Life-Cycle Assessment. NORD 20. Nordic Council of Ministers. Lundgren, Lars. 1992. Hur mycket påverkas miljön av en kilowattimme el i Sverige? (In Swedish: How much is the environment affected by a kWh used in Sweden?). Vattenfall Research. Sortimentlista ??? Saltå kvarn Sunnerstedt 1996 Stockholms Miljöförvaltning …. Thomsson, Olof. 1999. Systems Analysis of Small-Scale Systems for Food Supply and Organic Waste Management. Acta Universitatis Agriculturae Sueciae, Agraria 185. Uppsala

Personal communications Jan Gustafsson, Saltå kvarn, market director, +46 8 55150803 Henrik Almhagen, Saltå Kvarn, +46 8 55150800 Mats Arman, Saltå kvarn, mill director, +46 8 55150814 Dagfinn Reeder, Yttereneby farm, +46 8 55150388 Hans-Petter Sveen, Järna Odlarring, executive director, +46 8 55151270 Simon Vidermark, Järna Odlarring, +46 8 55151270 Kristina Åström, Stigtomta slakteri, +46 155 227023 Rainer Höglund, Järna Mejeri, executive director, +46 8 55150588 Job Michielsen, Järna Mejeri, +46 8 55150588

72 6. Organic waste management studies Inventory in Juva Tiina Lehto, South Savo Environment Centre, Finland

The objective in the inventory conducted at Juva is to determine the possibilities for recycling biowaste (organic waste) produced by local food actors back into agriculture. The examined food actors belong to the food chain after primary production (agriculture): food processors, grocery stores, schools, communal kitchens, and private consumers. The research methods used are waste flow and substance flow study.

The study first looked into the current waste management system within the Juva community and among specified food actors. Special attention was focused on biowaste and nutrient flows. Because biowaste recycling into fields in Juva was scarce in 2002, an assessment was made of the possibilities to recycle biowaste back into agriculture.

Recycling of biowaste can recycle nutrients back into agriculture and thereby add humus to agricultural land. In organic agriculture, nitrogen is fixated by plants, but it is estimated that there will be lack of phosphorus in the future. Biowaste could compensate nitrogen fixation and it could be a good source of phosphorus. Legislation sets demands on biowaste recycling and sewage sludge, for example, is not permitted to be used in organic agriculture in Finland. Considering the recycling of food waste and animal-based by-products, it is necessary to take into account regulation (EC) No. 1774/2002, which lays down health rules concerning animal by-products not intended for human consumption. This regulation was enacted mainly because of the BSC epidemic and foot and mouth disease, and it affects and limits the recycling of animal-based by-products, including food waste from kitchens and animal-based biowaste from grocery stores.

Materials and methods The system under study is aimed at food actors in the Juva population centre; a few food processors are situated in urban area. Food actors are food processors processing meat, milk, vegetables and cereals, three communal kitchens including schools, three grocery stores, and private households. Also included is the wastewater led to the communal wastewater treatment plant and the sewage sludge thus formed. The research methods used were waste flow and substance flow study. The substance flow study considered especially N and P flows included in the biowaste. Also, the study covered assessment of greenhouse gas emissions caused by biowaste treatment.

Data for the study was mainly obtained through interviews, enquiries and analyses results (Savolab Oy 2003, Jyväskylän yliopisto, 2005). Also included was an assessment of the amount of biowaste produced by households because is assumed that part of biowaste ends up as landfill waste. The calculated amounts of biowaste produced by households are based on the following assumptions: one person in the region of South Savo produced 221 kg of waste in 2000 (Angervuori, 2002); about 33% of all household waste in population centres and about 40% of the household waste produced in rural areas is biowaste (Statistics Finland 1994); in the case of the Juva population centre, about 40% of the inhabitants are included in separate biowaste collection, and the rest compost their biowaste on their property (YTI- tutkimuskeskus, 2004); the inhabitants of the rural area of Juva are obligated to compost their biowaste; the yield of biowaste via separate collection is 70% (Rejlers Oy 2000); Juva had a

73 population of 7449 in 2002, and 3628 of them are estimated to live within the population centre (48.7%).

The estimates of the nutrient flows resulting from the biowaste (Table 6-1) are based on the Fineli-Finnish Food consumption Database of the National Public Health Institute (National Public Health Institute, 2005) and the estimates of the nutrients included in household wastewater are based on the estimates presented by Eilersen et al. (Magid et al., 2002). The dry matter content of food products have been estimated based on the feed tables produced by Agronet (2005). Separately collected biowaste is assumed to contain 20 g/kg N and 4 g/kg P in dry matter, which is assumed to correspond to 35 mass-% (Sokka et al., 2004). Laboratory analyses were made by studying sewage sludge samples from Juva wastewater treatment plant and from compost samples taken from the Juva stack-composting area (Jyväskylän yliopisto, 2005) (

74 Table 6-2) to determine the nutrients that could be recycled at present and to determine whether the heavy metal concentrations are below the limits set by regulations. The analysis results for the incoming wastewater at the treatment plant were already available (Savolab Oy, 2003) (Table 6-1). These data are based on the average daily nutrient concentration of five samples analysed in laboratory conditions. The conversion factor used for biowaste is 0.3 tonnes/m3. Table 6-1. The nutrient values used for different kinds of foodstuffs and biowaste. Waste component N P Dry matter Lettuce (g/kg) 1,76 0,4 10,3 %(1 Oat husk (g/kg) 28 2 4,3 2 88 % Turkey offal (g/kg) 35,2 1,5 35 % Raw milk (g/kg) (Tuhkanen 2005) 5,2 1 12 % Organic milk (g/kg) 4,8 0,9 12 % Pig carcass (g/kg) 27,5 1,8 35 % Separately collected biowaste (g/kg dry matter) 20 4 35 % Household wastewater 14 2,2 135 - faeces g/cap/day 1 0,5 35 - urine g/cap /d 11 1,5 60 - kitchen liquid waste g/cap/d 0,5 0,1 20 - bathroom, grey water g/cap/d 1 0,3 20 Nutrients to treatment plant, kg/d (Savolab Oy 2003) 60,4 12,8 Nutrients into water systems, kg/d (Savolab Oy 2003) 49,2 0,46 1 Dry matter of cabbage 2 Nutrients in oat bran

Waste system in Juva Figure 6-1 shows the biowaste flows produced by the various food actors in 2002 and Figure 6-2 presents an assessment of the nutrient flows included in biowaste. The total solid biowaste amount produced by the various food actors was about 1 150 tonnes, containing 27.6 tonnes of N and 1.6 tonnes of P (excluding quill waste). About 214 tonnes were treated in the communal waste management system by stack-composting the waste. The biowaste in the stack-composting area contained about 0.71 tonnes of N and 0.15 tonnes of P. The stack- composted biowaste originated mainly from households (about 45 tonnes; 0.32 tonnes of N and 0.06 tonnes of P) and from one vegetable processor (about 150 tonnes; 0.26 tonnes of N and 0.06 tonnes of P); the remainder came from grocery stores in the population centre and from communal kitchens. About 10% of the biowaste originating from grocery stores was estimated to be former animal-based foodstuffs. About 70% of the biowaste produced was transported outside Juva; a large amount went to animal fodder industry, containing as much as 89% N (24.6 tonnes of N) and 66% of P (1.05 tonnes of P) of the total nutrient amount of the biowaste produced by the food actors covered by this study. A smaller amount appears to have ended up in a landfill in Mikkeli (50 km away) along with mixed waste. At maximum 12% of the biowaste (5% of N and 15% of P) was treated at the food actors’ own systems in Juva, but the said proportion appears not to have been utilised on fields to benefit food production.

Wastewater led to the Juva waste-water treatment plant is mainly from households. Most significant industrial plants in the area are a diary and a slaughterhouse, adding BOD and nutrient load to the incoming waste water. In addition waste waters of a vegetable processor, a printer house and a fur refinery are led to the waste water treatment plant. The wastewater treatment plant in Juva is a BIOLAK treatment plant built in 1982, which has later been

75 provided with a facility for phosphorus precipitation by iron(II)sulphate. The resultant sewage sludge is subjected to filter-pressing; with polymers being used as aiding agents. In 2002, the incoming wastewater to the treatment plant amounted to 421 000 m3 with a solid matter content of 134 tonnes and containing 22.0 tonnes of N and 4.7 tonnes of P. The amount of iron(II)sulphate used was 80.7 tonnes (Savolab Oy 2003) and the removal efficiency of phosphorous was 96% and that of nitrogen was 19%.

The sewage sludge amount formed in 2002 was 512 m3 (about 364 tonnes12) with a dry matter content of 18%. According to analyses results obtained in 2005 (

12 According to analysis made in 2005 volume weight was 771 g/l ( Table 6-2).

76 Table 6-2) , the sewage sludge contained 4.4% of N and 2.9% of P (in terms of dry matter). According to the removal efficiency in 2002, the sewage sludge contains at maximum 4.1 tonnes of N and 4.5 tonnes of P. Nitrogen can escape from the process through denitrification and evaporation, but phosphorus is believed to be slightly soluble because of phosphorus precipitation. The heavy metal concentrations of the sewage sludge are below the requirements of the relevant regulations imposed on sewage sludge for agricultural use; the exception is chromium. The chromium concentration was much higher than the permitted amount.

Figure 6-1. Biowaste flows involving Juva food actors in 2002 (DM = dry matter).

Figure 6-2. Nutrient flows of biowaste involving Juva food actors in 2002.

The sewage sludge formed was composted in the communal stack-composting area along with the separately collected biowaste. Tree bark and leaves and gravel were mixed in with the sludge and bio waste being composted to promote the composting process and to produce a final product of better the quality. The nutrient concentrations of the compost as analysed in 2005 were low; 0.28% of N and 0.32% of –P, but the heavy metal concentrations fulfilled the requirements of the regulations. The nutrient concentrations of the biowaste can be assumed to be the same as in 2002, because no noteworthy changes in biowaste flows or treatment have occurred. The organic matter (33%) and dry matter content (74%) of the compost samples indicate that the mineral content is relatively high (Jyväskylän yliopisto, 2005). According to a decision of the Ministry of Agriculture and Forestry (46/1994), the humus content has to be at least 20% of the dry matter content. The compost product has previously been used in green areas and in road construction, and now it is being stored for the purpose of landscaping an old landfill; local farmers have not accepted it.

78 Table 6-2. Analyses results for the compost product from the stack-composting area and the sewage sludge from the wastewater treatment plant in Juva (Jyväskylän yliopisto, 2005). Analysed property Unit Sewage sludge Vnp1 MMMp2 Compost Sewage sludge from Juva 282/1994 46/1994 from from Juva Juva Volume weight g/l 771 696 771 Conductivity mS/m 30.6 2.70 30.6 pH-value pH 6.5 6.6 6.5 Ignition loss % DM 65 33 65 Dry matter (DM) % 18 74 18 Total nitrogen % DM 4.4 0.28 4.4 Total phosphorus % DM 2.9 0.32 2.9 Total potassium % DM 0.20 0.18 0.20 Total magnesium % DM 0.16 0.21 0.16 Total calcium % DM 1.1 0.28 1.1 Mercury mg/kg DM 0.24 1.0 2.0 0.07 0.24 Cadmium mg/kg DM 0.6 1.5 3.0 < 0.5 0.6 Total chromium mg/kg DM 530 300 - 21 530 Copper mg/kg DM 240 600 600 39 240 Lead mg/kg DM 31 100 150 6 31 Molybdenum mg/kg DM 3 - - < 1 3 Nickel mg/kg DM 20 100 100 7 20 Sulphur mg/kg DM 6 800 - - 520 6 800 Zinc mg/kg DM 480 1 500 1 500 88 480 DM = dry matter 1 Decision of the Council of State (Vnp) No 282/1994 2 Decision of Ministy of Agriculture and Forestry (MMMp) No 46/1994

Possibilities for developed recycling of biowaste for use in agriculture According to the waste flow study conducted, the recyclable biowaste in the Juva area amounts to about 1 150 ton per year and contains 27.6 tonnes of N and 1.6 tonnes of P. Figure 6-3 shows the total nutrient and dry matter percentages of the biowaste produced by food actors, including wastewater to the treatment plant, in Juva in 2002. Most of the nutrient- containing components originated from food producers and from urine. Slaughterhouse waste and food-processing waste contained 52% of N and 19% of P of the total nutrient amount and 93% of N and 75% of P of the solid biowaste. Urine contained 29% of N and 32% of P of the total nutrient amount and 66% of N and 43% of P of wastewater nutrients. Slaughterhouse waste could be a good nutrient source, but the current regulations pertaining to former animal- based by-products may prevent the recycling of these nutrients back to agricultural land. Food waste from households and communal kitchens and the former foodstuffs from grocery stores are estimated to contain only 4% of N and 6% of P of the total nutrient amount, but 7% of N and as much as 25% of P in solid biowaste.

Figure 6-3. N, P and dry matter flows of biowaste and wastewater to the treatment plant in 2002, Juva. (Dry matter of wastewater is assumed to be double that of solid matter content.) Alternative treatment processes to recycling of nutrients and humus of biowaste are composting and biogas treatment (centralized or small-scale treatment or co-digestion plant). The treatment of biowaste and wastewater cause nutrient losses and all nutrients in the treated biowaste do not become available to plants. The composting process mainly affects the amount of nitrogen. The loss of nitrogen in composting can be as much as 50% (Pipatti et al. 1996) due to evaporation and leakage. Phosphorus can be assumed to remain in the compost; a little phosphorus loss may occur as leachates. If biowaste is anaerobically treated, the losses during treatment are minor. In the biogas process, most of the nutrients turn into soluble form (NH4-N, PO4-P, K2O-K) in the same ratio as organic matter is reduced (Taavitsainen et al. 2002). The biowaste’s organic matter content reduction is about 50 mass-% and that of sewage sludge 30 mass-% (Isoaho et al. 2003). If the hydrolysis residual is separated into liquid and solid fractions, losses will take place. How the hydrolysis residual is stored and spread onto the fields affects also the nutrient loss13 and possible post-composting requirements as per Finnish regulations.

It should be noted that urine contains about 1/3 of the nutrients formed within the studied food system and even a greater share of the nutrients in incoming wastewater flow to wastewater treatment plant in Juva in 2002. From the environmental viewpoint, separate urine collection would be preferable to the conventional system. It would increase the recyclable nutrient amounts to fields and at the same time decrease the emissions into water systems. In addition, phosphorous would be in more usable form for plants and risk of heavy metals would be minor compared to sewage sludge. In conventional wastewater treatment system phosphorus is precipitated from wastewater and is in slightly soluble form in sewage sludge. About 30%

13 For storage losses (of ammonia) the figures range from 1% (tightly covered) to 10-25% (uncovered). The differences are probably as big in practice, too. In regard to spreading the mulching of hydrolysis residual, it resulted in only 3-4% losses of total-N; mulching within 10 hours resulted in 10% losses and mulching after 6 days resulted in 40% losses of total-N. (Thomsson 2004)

80 of the total phosphorus in sewage sludge appears to be usable to plants (Tritonet Oy, 2005, Luomat at al., 1986). The usable nitrogen for plants contained in undigested and dried sewage sludge is about 40% (Chaussod et al., 1985). In addition, sewage sludge has to be stabilized (e.g. composting, anaerobic treatment) before it is used on fields, and this further diminishes its nutrient content. The biological removal of phosphorus and nitrogen from wastewater in Finland is being studied by the Finnish Environment Institute (Vesitalous, 2001). The advantages of biological removal of nutrients include that amount of chemicals in sewage sludge entering the environment would be reduced and that the nutrients contained in sewage sludge would be in form more usable to plants than is the case following chemical-biological treatment.

As regards greenhouse gas emissions in connection with biowaste treatment, it is known that the treatment does produce greenhouse gas emissions through composting, biowaste crushing, and as energy is needed in treatment plants and also in collecting and transporting biowaste. The effect of biowaste treatment on greenhouse gas (GHG) emissions is examined in Figure 6-4 and Figure 6-5 (Biokaasulaitos ry, Magid, 2002, Pelkonen et al., 2000, Tuhkanen 2002, YTI-tutkimuskeskus, 2005). The GHG emissions involved in biowaste transportation are not taken into account because the location of the treatment place is not considered here. The GHG emissions in 2002 are based on the biowaste amounts, which are separately collected, composted in small-scale operations or taken to a landfill. In the evaluation of the effects of the different treatment options on GHG emissions we have taken into a consideration composting and biogas treatments and as a comparison the landfill option, and covering the total solid biowaste amount produced by the food actors included in the study.

Figure 6-4. Effects of biowaste treatment on greenhouse gas emissions. Conversion factor used for biogas treatment is 77 kg of CO2 released per GJ of energy produced from oil. (Biokaasulaitos ry, Magid 2002, Pelkonen et al. 2000, Tuhkanen 2002, YTI-tutkimuskeskus 2005)

Figure 6-5. Effects of biowaste treatment on greenhouse gas emissions estimated by biowaste amounts produced by the food actors studied in 2002, Juva.

Anaerobic treatment would be the best solution when considering greenhouse gas emissions, because that treatment produces energy and does not cause any process emissions. Composting is an effective solution when compared to the landfill option. For example, if the input of biowaste in a co-digestion plant were 300 tonnes/year and the input of manure 3000 tonnes/year, it would produced approx. 400 MWh/heat energy and approx. 280 MWh/electrical energy. The solid content would be about 12% and no water would need to be added. If the energy consumption of a single-family house is estimated to be 20 MWh/year, then the heat generated by a co-digestion plant would be enough to heat ten houses and in addition 260 MWh/year of electricity would be produced to be consumed outside the plant.

An obstacle related to the recycling of biowaste back into agriculture at present is in the requirements imposed by regulations on existing treatment systems. In 2002, biowaste composted in stacks could not be recycled back into agriculture at Juva. The treatment process did not fulfil the requirements; especially the by-product directive 1774/2002 as the compost included former animal-based by-products in the form of food waste and former animal-based foodstuffs from grocery stores. The final product could then only be used in landfills (as covering or filling material). If final product is used in agriculture or in greening areas in the future, then the composting treatment has to fulfil the requirements of the by-product directive. Otherwise, the product can be taken only to landfills. Alternatively, it is possible to compost only the biowaste of vegetables and other non-animal-based by-products and sewage sludge.

Also, the nutrient availability from the final product and some biowaste applicability to treatment, their pre-processing requirements, and the cost of treatment plants may become obstacles. Attitudes towards biowaste products may prevent biowaste recycling for

82 agricultural purposes too. The quality of the final product depends on the quality of the treated biowaste. The use of biowaste as fertilizer may involve the risk of pathogens, especially in the case of human and animal-based wastes. If the treatment process fulfils the requirements of legislation, then the final product should be safe and hygienic as a soil conditioner. Beside agriculture, compost can be used also in landscaping and green areas, or for raising energy plants and for preventing erosion.

References Agronet. Agrifood Research Finland. Feed tables and recommendations of feeding. Available: http://www.agronet.fi/rehutaulukot. Updated: 2005. Angervuori. Etelä-Savon ympäristökeskuksen alueellinen jätesuunnitelma. Seuranta ja tarkastaminen 2001. Etelä-Savon ympäristökeskuksen moniste 40. Mikkeli 2002. Biokaasulaitos ry. Perustietoja biokaasusta. Available: http://www.kolumbus.fi/suomen.biokaasukeskus/. Referred: 2004. Chaussod R., Gupta S.K., Hall J.E., Pommel B., Williams J.H. Nitrogen and phosphorus value of sewage sludges. Concerted action treatment and use of sewage sludge. Revision of Document Nr. SL/82/82-XII/ENV/35/82. Commission of the European Communities. Directorate-General Science, Research and Development; Environment and Raw Materials Research Programmes. 1985. Isoaho, Vinnari. Pirkanmaan jätehuollon järjestelmä- ja kustannustarkastelu. Tampereen teknillinen yliopisto, Ympäristötekniikan laitos. Tampere 2003. Jyväskylän yliopisto. Ympäristötutkimuskeskus. Tutkimustodistus 2005-1648. 07.06.2005. Luoma Tarmo, Sipilä Ilkka. Jätevesilietteen käyttö maataloudessa. Helsingin yliopiston maatalousteknologian laitos. Tutkimustiedote N:o 49. Helsinki 1986. ISBN 951-45-4063- 8. Magid Jacob. Granstedt Artur, Dýrmudsson Ólafur, Kahiluoto Helena & Ruissen Theo (eds.). Urban Areas - Rural Areas and Recycling - The organic way forward? Proceedings from NJF-seminar No. 327. Copenhagen, Denmark 20-21 August 2001. Danish Research Centre for Organic Farming 2002. National Public Health Institute. Fineli -Finnish Food Consumption Database. Available: http://www.fineli.fi/. Updated: 2005. Pelkonen, Rauta, Tanskanen. Yhdyskuntajätehuollon päästöjen järjestelmätarkastelu. Teknillinen korkeakoulu Pipatti, Hänninen, Vesterinen, Wihersaari, Savolainen. Jätteiden käsittelyvaihtoehtojen vaikutus kasvihuonekaasupäästöihin. VTT julkaisuja. Espoo 1996. ISBN 951-38-4520-6. Rejlers Oy. Ympäristötekniikan asiantuntijapalvelut. Pieksämäen seudun jätehuollon kehittämissuunnitelma. Loppuraportti. Mikkeli 2000. Savolab Oy. Juvan kunta. Juvan kunnan jätevedenpuhdistamo. Vuosiyhteenveto 2002. 22.9.2003. Sokka, Antikainen, Kauppi. Flows of nitrogen and phosphorus in municipal waste: a substance flow analysis in Finland. Progress in Industrial Ecology, Vol. 1, Nos. 1/2/3, 2004. Statistics Finland. Waste management in households in 1994. Available: http//:www.tilastokeskus.fi/tk/yr/ye16.html. Referred 2.8.2004 Taavitsainen, Kapunen, Survo. MaLLa-hankkeen loppuraportti: Maatalouden lietteiden ja lantojen keskitetyn käsittelyn mallinnus. Pohjois-Savon ammattikorkeakoulu. 18.6.2002. Thomsson. Swedish University of Agricultural Science. Information of losses concerning biogas treatment. 13.12.2004 and 6.12.2004. Tritonet Oy. Alueellinen ympäristötase –tutkimushanke. Alueellinen ympäristötasetutkimus. Työ 03106. Tampere 31.1.2005. 162 s.

83 Tuhkanen. Agrifood Research Finland. Information of nutrient content. 30.5.2005. Tuhkanen Sami. Jätehuollon merkitys Suomen kasvihuonekaasupäästöjen vähentämisessä. Kaatopaikkojen metaanipäästöt ja niiden talteenotto. Climtech. VTT tiedotteita 2142. Espoo 2002. ISBN 951-38-5895-2. Vesihuoltotekniikan laboratorio. Espoo 2000. ISBN 951-22-5304-6. Vesitalous 4/2001. Jätevesien ravinteiden poisto tehostuu jälkisuodatuksella. 16-21 pp. YTI-tutkimuskeskus. Biojätteiden ja jätevesilietteiden nykytila ja tavoitteet. Biojätteen ja jätevesilietteen käsittelymenetelmät ja niiden soveltuvuus Etelä-Savossa –hanke. Raportti 20.9.2004. YTI-tutkimuskeskus. Biokaasun toiminta- ja käyttöedellytykset Etelä-Savossa. Loppuraportti. 2004.

Biogas plant in Järna Winfried Schäfer, Agricultural Engineering Research (VAKOLA) MTT Agrifood Research Finland Artur Granstedt and Lars Evers, Swedish Biodynamic Research Institute, Järna Sweden

Introduction The Biodynamic Research Institute in Järna developed an on-farm biogas plant integrated within the self-contained farm organism, BERAS-farm Skilleby-Yttereneby. The plant digests manure of dairy cattle and organic residues originating from the farm and the surrounding food processing units containing 17.7-19.6 % total solids. The developed and implemented technique is based on stable manure and food residues on a two-phase process and new technology for continuously filling and discharging the hydrolysis reactor. The output of the hydrolysis reactor is separated into a solid and liquid fraction. The solid fraction is composted. The liquid fraction is further digested in a methane reactor and the effluent used as liquid fertiliser. Initial results show that anaerobic digestion followed by aerobic composting of the solid fraction improves the nutrient balance of the farm compared to mere aerobic composting.

Methodology Manure of a dairy stanchion barn with 65 adult bovine units is shifted by a hydraulic powered scraper into the feeder channel of the hydrolysis reactor. The urine is separated in the barn via a perforated scraper floor. The manure is a mixture of faeces, straw and oat husks. From the feeder channel the manure is pressed via a 400 mm wide feeder pipe to the top of the 30° inclined hydrolysis reactor of 53 m3 capacity. The manure mixes itself with the substrate sinking down by gravity force. After a hydraulic retention time of about 22-25 days at 38°C, the substrate is discharged by a bottomless drawer from the lower part of the reactor. Every drawer cycle removes about 100 l substrate from the hydrolysis reactor to be discharged into the transport screw underneath. From the transport screw the substrate partly drops into a down crossing extruder screw where it is separated into solid and liquid fractions. The remaining material is conveyed back to the feeder channel and inoculated into the fresh manure. The solid fraction from the extruder screw is stored at the dung yard for composting. The liquid fraction is collected into a buffer and from there pumped into the methane reactor with 17.6 m3 capacity. Liquid from the buffer and from the methane reactor partly returns into the feeder pipe to improve the flow ability. After a hydraulic retention time of 19-21 days at 38°C the effluent is pumped into a slurry store covered by a floating canvas. The gas generated by both reactors is stored in a sack and fed by a compressor to the process heater and the furnace of the estate for heating purposes (Schäfer, 2003).

84 Results The results concerning the nutrient contents are presented in more detail in a separate on BERAS homepage available manuscript (Schäfer et al, 2005). During the anaerobic digestion in process A, 14.6-15.4 % of carbon was found in the biogas. During aerobic composting escaped 26-31 % of the input carbon of the solid fraction. In process B 58-60 % of the carbon escaped during aerobic composting. Even if the biogas yield would be threefold more, there would still be 41-42.5 % carbon available for composting of the solid fraction. This confirms the hypothesis that biogas production before composting hardly has a negative impact on the humus balance (Möller, 2002) compared to mere aerobic composting. Total nitrogen losses ranged between 30% and 48% in process B and between 19% and 29% in process A. Similar values we found for NH4: up to 6% losses in process A versus 96% in process B. Potassium and phosphorus losses were higher in process A than process B. The results confirm the calculations of Möller (2002) that biogas production increases recycling of NH4 and reduces overall nitrogen losses compared to mere aerobic composting.

Conclusions The two phase prototype biogas plant in Järna is suitable for digestion of organic residues of the farm and the surrounding food processing units. The prototype put many recent research results into practice. But there is still a lack of appropriate technical solutions in terms of handling organic material of high dry matter content, and process optimisation. The innovative continuously feeding and discharging technique is appropriate for the consistency and the dry matter content of the organic residues of the farm. It is probably not suitable for larger quantities of unchopped straw or green cut. Anaerobic digestion of manure and organic residues followed by composting the dry fraction of the hydrolysis reactor improves the energy and nutrient balance on-farm compared to mere aerobic composting. Appropriate new technical design as used at the prototype biogas plant in Järna is a key factor. To confirm the present results more measurements are necessary. The optimisation of the plant in respect of hydraulic retention time and load rate may lead to higher gas generation but requires an improved measuring technique. Thereafter an economic evaluation is necessary to assess the competitiveness of the new technology. The benefit of an on-farm biogas plant may be considered in a larger extent if we include into the nutrient balance not only the biogas plant but also the nutrient circle of a whole crop rotation period of the farm organism. The quality of the nutrients is finally related to soil fertility, fodder quality and animal health.

References Schäfer W. 2003. Biogas on-farm: energy and material flow. Paper presented at Nordic Association of Agricultural Scientists 22nd Congress, "Nordic Agriculture in Global Perspective", Turku, Finland, July 1-4 2003. Schäfer W., Evers L., Lehto M., Sorvala S., Teys F., Granstedt, A. 2005. Nutrient balance of a two-phase solid manure biogas plant. Manuscript: http://www.jdb.se/beras/

85 Possibilities for developed recycling and renewable energy production in Juva and Järna Tiina Lehto, South Savo Environment Centre, Finland Artur Granstedt, Swedish Biodynamic Research Institute, Järna Sweden

The plant nutrients food stuffs from agriculture end up in slaughterhouse wastes, domestic wastes (wastes from household and food industry) and sewage wastes. This three fractions contains 4, 3 and 2 kg N per capita and year and 2, 0,5 and 1 kg P calculated per capita an year calculated from Magid et al (2002). About 60 percent of the nitrogen and 45 % of the phosphorus belonged to the liquid wastes residues manly in the human urine fraction. Of the total phosphorus taken up by plants (20 kg P per ha) about 75 % can be recycled within the farming system on ERA farms with an optimal utilization of the recycled nutrients in manure. But the 15 % is found in the sewage fraction from the human consumption and could be recirculated to the agriculture trough urine separation if also the hygienisation can be realised on a secure way. Another 10 is found in the slaughter wastes which also in an important resource for the sustainable agriculture.

Two ways of local recycling of the solid fraction of biomass combined with producing biogas are presented within the BERAS-project. The goal is safe recycling of nutrients, reduced emissions of greenhouse gases and reduced emissions of reactive nitrogen. One way is the central recycling on commune level with production of biogas according the above description for Juva . But the central biowaste treatment raise problem with the quality control and with high risk with contamination of heavy metals, medicaments and animal and human pathogens which not can be used on soil for food production.

Recycling of food residues is introduce in the small scale biogas plant on Skilleby in Järna. This is established as an essential link in the local ecological recycling system. Only food residues from the ecological farms, ecological small scale food processor and ecological consumer are used. Permission for this local recycling is become from the commune. The biowaste is treated with one hygienisation stage and consist of one our heating under 70 0C with help of the energy from the biogas plant. For the farmer is this recycling of the nutrients from the food chain combined with the manure from the farm a valuable resource. A further step is to also use slaughter wastes if also a more local slaughter will be build according the plans in Järna. This local small scale biogas plant on farm level seemed to be the best solution for recycling of nutrients from the human food (local processors, ecological public kitchen and consumer), to reduce emissions of greenhouse gases and also have a good control against contaminations of pathogens and harmful substances.

86 7. Consumer surveys in Juva and Järna for identification of eco-local food baskets Annamari Hannula, Savonlinna Department of Education, University of Joensuu, Finland Olof Thomsson, Swedish Biodynamical Research Institute, Järna, Sweden

Introduction Our food habits are, unquestionably, important both for the health and for the environment. That is recognised as one of the starting points in BERAS. That is also one of the key issues of the S.M.A.R.T. recommendations (CTN 2001). The consumer survey presented here was aimed to give knowledge about the food habits of environmentally concerned inhabitants in two of the BERAS case study sites. The aim was that this knowledge should be used as input to other BERAS studies for comparisons to Swedish and Finnish average food baskets concerning environmental impacts in the whole food system, reported in other chapters in this report.

The main objective of the consumer surveys was to put together realistic food baskets (consumption profiles) for a Swedish and a Finnish case, containing mainly locally and ecologically produced foodstuffs. The intended unit was kg food in different product groups per capita and year. Economic results as € per capita for different household types and year are shortly reported here, and in more detail in other BERAS reports.

Case study sites The case studies were performed in Juva, Finland and Järna, Sweden – the same sites used for many other studies in the BERAS project. Both the municipality of Juva and the community of Järna have around 7500 inhabitants. For a more extended presentation of the sites see Seppänen (ed. 2004).

Juva is a rural municipality in South-Savo region, about 270 km northeast of Helsinki. In South-Savo there is about xxx inhabitants. Compared to neighbouring areas, Juva has a strong tradition of organic farming. In Juva 15,8 % of the cultivated land is organic, compared to Finnish average 7,6 %. Juva has a strong food processing industry, compared to many other rural municipalities, comprising a dairy, a flour-mill, vegetable processing enterprises, meat processing enterprises and bakeries. The dairy is processing organic milk while the other enterprises are using conventional, non-organic, raw-products.

The small town of Järna is part of Södertälje municipality, located in Stockholm County, 60 km south of Stockholm. The hart of the case study lies in the outskirts of Järna and is connected to an antroposophic community with a high concentration of antroposophic orientated initiatives and small businesses which prefer biodynamic and organic produced products. There are several biodynamic farms and market gardens in the area that serve the local market and a well-developed consumer network linked to these farms. There are also several food processing industries like a mill and bakery (working on the local as well as on the national market), a farm-size dairy and a farmer cooperative selling vegetables and meat.

Subjects Most of the research subjects in the study are persons or families devoted for environment and health, living in Juva County in Finland and in Järna community in Sweden (Table 7-1). The

87 families were invited to take part in the survey through local food and environment organisations and through staff in the local (ecological farming) research institutes.

Table 7-1. Composition of research subjects. Juva Järna period April 2004 October 2004 February 2004 September 2004 no. households 9 9 15 13 no. adults 15 15 29.5 25 no. children 0-19 years 13 12 19.5 18.5

Methodology The methodology used for the data collection differed slightly between the two case studies but, basically, the families recorded their food purchases for two two-week periods; one period during winter/spring (when local products are scarce) and one period in late summer/early autumn (when local products are easy available). The periods were chosen in order to get representative results for the yearly consumption. In Finland, the first period was performed during April 2004 and the second in September and November 2004. In Sweden, the survey started in February 2004. The second period was performed in September and early October.

Both in Finland and in Sweden a family member collected the receipts or filled in purchase diaries for all food entering the household for human consumption during 14 days. Amount, price, origin and environmental branding of all food products were recorded either on the (detailed) receipts or on the specified lists (supplied by the project).

After the recording period, the families were interviewed about their food choices, food consumption and food purchasing habits. In the interviews, we also tried to get a picture of the quantities of different kinds of food that was put into or taken out of the household stores and the quantities home-produced food, to get representative values for the consumption during a two-week period at that time of the year.

The amounts of different products purchased during the measured four weeks were then extrapolated to get values for the whole-year consumption. In Juva, meals eaten outside was not reported in this paper because in the Finnish statistics there are only the expenses for the food which is bought and eaten at home.

For some comparisons to Swedish average figures, the measured results for the Järna consumers were extrapolated to cover also meals eaten outside home by an estimated factor. The factor was obtained through an estimation made by each household of how many meals they were eating outside home at an average week during the measuring periods, and an assumption that each person eats three meals a day. In average, the Järna households ate 16 % of there meals outside home. Thus, when measured consumption was compensated for “out- eating” the original figures were multiplied by 1.16, implying an assumption that food eaten outside home had the same proportions of different product groups and energy content as that purchased for home-consumption.

The method used in this consumer survey has some limitations which should be taken into account when interpreting the results. The purchase diary used in these studies record food availability in households, not the food consumption of single persons. In other words, the

88 results represented here per capita per year are estimations about purchased food, not real food consumption. Also, purchasing patterns may be distorted and no information on the distribution of foods within households is normally obtained (Cameron, 1988). One problem is the possible lack of information about whether non-buyers never use a product or whether they simply not were buying it the recorded weeks (Irish, 1982). When estimating the annual food expenditures, the bulk purchases would make it more difficult than if the consumers acquire all or part of their food in relatively small quantities once or several times per week (Pena, 1998). However, we tried to take these things in to account when interviewing the families and checking their purchase diaries and collected receipts. Results and discussion The results for amounts of different food products consumed are presented and discussed separately for the two surveys and compared to the national averages respectively. The shares of ecologically and locally produced foods and the expenditures for food are presented and discussed in the following sub-chapters.

Amounts of food consumed in the Juva households The largest differences between the consumption patterns in the investigated households in Juva and national average households in Finland are the lower consumption of potatoes and meat and the higher consumption of garden products (Figure 7-1). However, concerning the potatoes and garden products we can only say that the purchased amount is lower or higher. Some families grow their own potatoes and vegetables, which is not accounted for in the result, so the consumption might be substantially higher than the result indicates. One vegetarian and one meat producer may account for lower meat consumption. On the other hand, there were some bulk purchases of carrots and the difference may be partly explained by methodological reasons. Other differences are small when looking at whole product groups.

Juva survey 2004 52 MEAT, FISH AND EGGS Fin. average 1998 74

19 FAT PRODUCTS 11

166 MILK PRODUCTS 175

FRUIT, BERRIES, NUTS 45 AND SEEDS 52

GARDEN PRODUCTS, 71 EXCL. LEGUMES 39

31 POTATO PRODUCTS 44

CEREAL PRODUCTS 85 INCL. LEGUMES 72

0 50 100 150 200 kg per person and year Figure 7-1. Food consumption of product groups in Finland 1998 and in the Juva survey 2004. Meals outside home excluded. kg per person and year

89 When dividing the product groups in smaller groups (Figure 7-2) some more differences but no striking new patterns appear.

Egg 7 7 Juva survey 2004 14 Fish, fish products and caviar 11 Fin. average 1998 20 Sausage, paté etc. 33 0,0 Game 0 0,0 Deer and wild boar (pig) 0 3 Poultry 0 0,2 Lamb 0 6 Pork 15 2 Beef 9 5 Vegetable oil 0 8 Margarine 6 6 Butter (incl. Bregott) 5 4 Cream and milkpowder 4 22 Cheese 13 25 Yoghurt, quark etc. 37 Milk 115 121 1 Jam and marmelade 0 17 Fruit juices and fruit-syrup 25 3 Nuts and seeds 0 42 Fruit and berries 52 9 Tinned and preserved vegetables etc. 8 47 Fresh vegetables and legumes 23 21 Root vegetables 9 6 Potato chips etc. 5 25 Potatoes 39 3 Rice 0 4 Pasta 3 18 Groats, flakes and other grains 20 10 Flour 0 7 Sweet bread, cookies etc. 12 37 Bread, dry and soft and rusks 37

0 20 40 60 80 100 120 kg per person and year Figure 7-2. Food consumption of detailed product groups in Finland 1998 and in the Juva survey 2004. kg per person and year

90 Amounts of food consumed in the Järna households When studying the results from the study in Järna there are some more evident differences between the consumption patterns in the investigated households and the Swedish average households (Figure 7-3). The most obvious are the lower consumption of meat and potatoes and the higher vegetable consumption. The smaller meat consumption and larger vegetable consumption was expected but that potatoes seem to be less favoured by these households was somewhat surprising. This fact might partly be explained by antroposophic nutritional concepts recommending limited intake of solanin producing products, like potatoes and tomatoes. Thus, it is likely that the result reflect an actual lower consumption of potatoes.

Järna survey 2004 26 MEAT, FISH AND EGGS Sw. average 2002 105

15 FAT PRODUCTS 13

199 MILK PRODUCTS 168

FRUIT, BERRIES, NUTS 80 AND SEEDS 63

GARDEN PRODUCTS, 130 EXCL. LEGUMES 67

23 POTATO PRODUCTS 54

CEREAL PRODUCTS 112 INCL. LEGUMES 103

0 50 100 150 200 kg per person and year Figure 7-3. Food consumption of product groups in the Järna survey 2004 and in Sweden 2002. Swedish averages include all meals. Järna figures are corrected for meals outside home – purchased amounts added by the part of meals eaten outside home (16 %). kg per person and year When looking at more detailed products or product groups there are some more interesting differences (Figure 7-4). Although there is no big difference in cereal products as a group it can easily be seen that these households seem to bake more of their bread at home. They also eat more groats and flakes, which is in accordance with the higher consumption of yoghurt and other fermented dairy products. Concerning the fat products consumed, it is obvious that these households prefer butter in favour of the more processed margarine. Figure 7-4 also shows that when these households eat meat, they seem to choose meat from animals that have been kept outside (e.g. lamb and wild boar) in an assumable animal friendly production.

91 7 Egg 9 Järna survey 2004 5 Fish, fish products and caviar 18 Swe. average 4 Sausage, paté etc. 35 0,2 Game 2 1 Deer and wild boar (pig) 0 1 Poultry 14 4 Lamb 1 1 Pork 15 4 Beef 11 3 Vegetable oil 1 0,5 Margarine 11 11 Butter (incl. Bregott) 1 8 Cream and milkpowder 8 20 Cheese 17 73 Yoghurt, quark etc. 31 98 Milk 112 4 Jam and marmelade 7 25 Fruit juices and fruit-syrup 23 4 Nuts and seeds 1 76 Fruit and berries 62 17 Tinned and preserved vegetables etc. 19 80 Fresh vegetables and legumes 39 42 Root vegetables 9 1 Potato chips etc. 2 22 Potatoes 52 3 Rice 5 6 Pasta 8 18 Groats, flakes and other grains 7 36 Flour 13 5 Sweet bread, cookies etc. 19 35 Bread, dry and soft and rusks 53

0 20 40 60 80 100 120 kg per person and year Figure 7-4. Food consumption of detailed product groups in the Järna survey 2004 and in Sweden 2002. Swedish averages include all meals. Järna figures are corrected for meals outside home – purchased amounts added by the part of meals eaten outside home (16%). kg per person and year

92 Share of ecological and local food The main objective of this study was to present data for an “eco-local” food basket; i.e. a food basket mainly consisting of ecologically and locally produced food. The shares of ecologically and locally produced food reported in the surveys and for national averages are presented in

93 Table 7-2. The households in both Järna and Juva buy a much larger share ecological food compared to the national averages. The part of ecologically produced food purchased is substantial for some product groups, especially in Järna. In Juva the share of ecological food in general is smaller but the availability of such products most certain is a large constraint. For example, no ecological meat or fish was bought in Juva. The reason was said to be that there very rarely are ecological alternatives on sale. However, the Juvan households bought much more ecological milk more than average Finns do. Also the share of ecological fresh garden products, cereal products and eggs were larger in the Juva food basket than in the typical Finish food basket.

It is worth noting that the share of ecological food in the Järna households is very large, 73 % of the weight for what we here have called “real food” (sugar, candy, beverages etc. not counted). That might be a result of availability, but the availability, on the other hand, is probably a result of the long-lasting demand for ecologically produced food in Järna. Some of the Järna consumers even mentioned that they would have bought more eco-food if it was available, and not too expensive.

In the food basked of the average Finnish consumer, the share of organic food is 1 %. In a survey only 4 % of the Finnish households have estimated that the share of organic products in their food basket is 6 % or more (Nielsen, 2004). In Finland about 20 % of the consumers answered in interviews that they buy organic products continuously. Half on them have estimated that the share of organic products in their food basket is less than 20 %. (http://www.finfood.fi)

Concerning the second important issue investigated in BERAS, the locally produced food, the part purchased by the investigated households was found to be substantial for some product groups (Table 7-2). Also here the shares in general are larger in Järna, probably depending on the good availability. In Juva the share of local food varied very much between the food groups. For example the share of local milk was 37 % of the weight. That is possible because in Juva there is a local dairy. The share of local cereal products was 10 % of weight. In Juva there is a local mill, but according to these data people didn’t buy very much of their products. The people in the study seem to have preferred the ecological cereal products more than the local ones. However, about one fifth of the bread is produced by the local bakery.

For local food, no comparisons with national averages have been possible to do, but one can assume that the average share is very low since retailer chains tend to uniform their assortment all over the countries.

94 Table 7-2. The share of ecological (organic) and local food purchases in Sweden, in 15 Järna households, in Finland and in 10 Juva households. kg per capita and year, and % of weight Swe. Fin. Product group Järna survey 20041 Juva survey 20042 average average eco- eco- total3 eco4 total eco total6 eco total eco local7 local5 local kg %8 kg kg % kg % kg % kg kg % kg % kg % Cereal products 103 1.6 112 89 79 61 54 72 3.4 85 14 17 8 9 4 4 incl. legumes Potatoes 54 3.3 23 22 96 9 38 44 2.7 31 7 24 8 25 4 13 Root crops 9 9.9 42 39 92 17 40 9 3.59 21 18 87 16 79 16 75 Vegetables 58 2.0 88 56 64 25 29 30 3.9 51 10 19 12 24 5 10 Milk products 168 5.1 199 162 81 72 36 175 1.8 166 78 47 50 30 50 30 Meat ruminants 12 7 5 70 4 49 9 * 2 0 0 2 100 0 0 (beef and lamb) Meat monogastrics 28 0.810 2 1 48 1 28 15 * 9 0 0 3 32 0 0 (pork and poultry) Other meat and mixed 37 5 3 62 2 41 33 * 20 1 4 1 7 0 0 meat products Egg 9 9.7 7 6 88 2 22 7 2.1 7 1 9 2 28 0 0 Fish and fish products 18 011 5 0 3 0 0 11 011 14 0 0 1 5 0 0 Fat 13 2.7 15 6 42 0 0 11 3.312 13 1 5 0 0 0 0 Fruit, berries, nuts 63 2.6 80 39 48 2 3 5213 * 45 1 3 0 0 0 0 and seeds Total "real food", excl. sugar, candy, 572 2.2 584 428 73 194 33 466 114 462 131 28 103 22 79 17 beverages etc.

1 compensated for meals eaten outside home 2 not compensated for meals eaten outside home 3 Swedish average 2002 (Jordbruksverket, 2004) 4 certified KRAV, Luomu and/or Demeter 5 produced in Järna district and, since all local is eco in Järna, certified KRAV and/or Demeter 6 Finnish average 1998 (Tennilä 2000) 7 produced in Juva district 8 % of expenditures per product group 9 carrots (Finfood Luomu / A.C.Nielsen ScanTrack) 10 % of all meat and meat products 11 not possible to certify at that time 12 oil (Finfood Luomu / A.C.Nielsen ScanTrack) 13 fruit and berries only 14 % for all foods

Households’ expenditure on food The expenditures on food are summarised in Table 7-3 and somewhat discussed below. For more detailed results, see BERAS report 4. Table 7-3. Expenditures on food. €/CU15 €/person/year €/household/year Juva 2213 1642 4334 Finnish average no reference 1580 3397 Järna 2584 1800 5833 Swedish average 2084 1600 3376

15 CU = Consumption Unit, a measure that compensates for household structure and the ages of the household members to make more relevant comparisons of consumption between different household types

95 In Juva households’ expenditure for food was between 1622 and 6815 €/household/year and mean was 4334 €/household/year. The households in Juva consisted of 2.9 persons in average. Consumption expenditure of households in Finland at year 2001 was 3397 EUR/year per household for food and non-alcoholic beverages. The value of home grown products is not taken into account in these numbers. Mean Finnish household had 2.15 persons in year 2001 (Statistical Yearbook, 2004).

In Juva households’ expenditure for food per consumption unit (CU) was 908-4803 €/CU/year, the mean was3013 €/CU/year. The consumption unit compensates for household structure and age to make more relevant comparisons between different household types. There is no reference for € per CU in Finland. In Juva average expenditures for food was a little bit higher than the Finnish averages. One reason for that may be the second purchase diary period was near Christmas and according to the diaries families have bought dried fruits etc. for baking, Christmas cakes and ginger biscuits in advance.

In Järna the investigated households seem to spend more money on food than average Swedish households. The mean value for food expenditures per household was 5833 €/household/year in the monitored households, while the Swedish average household expenditures was 3376 €, alcoholic beverages and restaurant meals not counted (Statistics Sweden, 2004). However, when calculated per consumption unit (CU) the difference is smaller. The results was 2600 €/CU/year in Järna and 2100 € for the Swedish average CU; i.e. 24 % larger food expenditures for the monitored households compared to the Swedish average. Whether this is a result of that these families really prioritise the food higher, or something else, is, however, hard to say. Of course the ecological food generally is more expensive but the difference could as well be a result of socio-economic status of the included households, which was not investigated in Järna. Another impact, not investigated, affecting the result could be the proportion of meals eaten at home.

Validity of the method The amount of energy according to the purchase diary was 11.5 MJ/person/day in the Juva district. According to the FINDIET 2002, dietary energy intake was 9.2 MJ/day among men and 6.6 MJ/day among women. For the Järna case, the energy content of consumed (purchased + restaurant meals) “real” food (excl. sugar, sweets, beverages etc.) was 10.7 MJ/person/day, while the Swedish average 2002 was 10.2 MJ/person/day. Thus, we can conclude that our results are in a reasonable range concerning energy content of the purchased food. However, the results are not easily comparable to statistical data due to differences in survey methods.

Environmental impacts and nutrition recommendations Our food choices have an effect on the environment. For example what we eat influence to the energy consumption of food chain. In food chain about 15-20 % of energy consumption is needed for transportation of food. (SwEPA, 1997)

Generally one can say that meat is the most energy demanding food to produce and increased meat consumption is problematic. It is also shown in earlier chapters that local food production and consumption may have environmental gains due to less transportation. Growing vegetables in the open air demand less energy than growing in greenhouses. Thus, local and seasonal vegetables, root crops, fruits and berries decrease the energy consumption needed for food transportation.

96 New Nordic Nutrition Recommendations (NNR) was approved in August 2004. NNR are guidelines for the nutritional composition of a diet which provides a basis for good health (NNR, 2004). In the recommendations there are no instructions for sustainable food choices. Recommendations and instructions for sustainable food choices have been given at least in Sweden and in Germany. (S.M.A.R.T (CTN, 2001), SwEPA 1997, A Sustainable Food Supply Chain. 1998, Green Purchasing of Foodstuffs, 2000)

In Table 7-4 both nutrition and sustainable food choice recommendations are presented and both of them are used in our comparison to evaluate households’ food choices.

Table 7-4. Examples of recommendations Healthy nutrition Environmental perspective Sustainable food choices Fruit, berries and - A high and varied consumption of - A high and varied consumption of vegetables fruit and vegetables is desirable domestic vegetables, fruits and berries on high season and foodstuffs grown in the open air. - If needed off-season, imported fruits or vegetables grown in the open air, give preference to products grown in country near to consumption place Legumes - More leguminous plants instead of meat Potatoes - Traditional use, several nutrients, potatoes have a place in a diet Cereals - An increased consumption of wholegrain cereals is desirable Fish - Regular consumption of fish Milk and milk - Regular consumption of milk and products milk products, mainly lean varieties, is recommended as a part of balanced diet Meat - Consumption of moderate amounts of - Less meat meat, preferably lean varieties, is - Choose meat from animals that have recommended as part of a balanced grazed on natural meadows, e.g. and varied diet cattle and lamb. - Eat less chicken and pork. Edible fats - Soft or fluid vegetable fats, low in - Butter instead of margarine saturated and trans fatty acids, should primarily be chosen Energy-dense and - Food rich in fat and/or refined - Eat less sugar-rich foods sugars, such as soft drinks, sweets, snacks and sweet bakery products should be decreased General - More locally produced food when this is more eco-efficient. - Ecological food - Eat less foodstuffs which include little nutrients, for example: eat fruits instead of sweets - Easy transported foods, eg. juices as concentrate instead of drinking- ready. - Choose the product produced most nearby when there are equal products.

97 The food consumption profile of the Järna households seems to follow the diets suggested in the Nordic Nutrition Recommendations (NNR, 2004) and in the S.M.A.R.T. project (CTN, 2004). These households buy a larger share of vegetables (less meat), less “empty” calories, more ecological food, “right” vegetables (e.g. more legumes and root crops, and less lettuce and cucumbers) and less transported food, compared to the national average food. The only large difference between the results of the Järna survey and the S.M.A.R.T. recommendations is the share of potatoes. The Järna consumers eat substantially less potatoes than the average Swede, while the S.M.A.R.T. project recommends more potatoes. One reason might be recommendations in the antroposophic nutrient concept to minimise intake of solanin producing products like potatoes and tomatoes.

Conclusions We conclude that the consumption profile of the participating households in Sweden differed more from the average Swedish consumers than was the case in the Finish study. However, also the Finnish households participating in this study bought more organic products than ordinary Finnish households. Substantial parts of the food consumed were locally produced but we have not been able to compare that to national averages due to lack of data.

The calculated expenditures on food in the studied group was almost the same as average in Finland, but the Järna consumers lay around 20 % more money on food compared to average Swedes.

The Swedish consumption profile obtained in the study is well suited for the use as a good example in scenario studies of the whole food system reported in the following chapter.

References Cameron, M.E. & Van Staveron, W.A. 1988. Manual on Methodology for Food Consumption Studies.Oxford University Press, Oxford. CTN. 2001. Ät S.M.A.R.T. – ett utbildningsmaterial om maten, hälsan och miljön (in Swedish). Centrum för tillämpad näringslära (Centre for Applied Nutrition), Samhällsmedicin. Stockholms läns landsting. Stockholm. Available on www.sll.se/w_ctn/3938.cs. Green Purchasing of Foodstuffs. 2000. Swedish Environmental Protection Agency. Report 5128. Stockholm. Handla för framtiden! Om mat och miljö I det hållbara samhället. 1997. Naturvårdsverket. Rapport 4900. Stockholm. Irish, M. 1982. On the interpretation of budget surveys: purchases and consumption of fats. Applied Economics. 14, 15-30. Jordbruksverket. 2004. Konsumtionen av livsmedel och dess näringsinnehåll. Uppgifter t.o.m. år 2002. (In Swedish). Rapport 2004:7. Tables in chapter 5. Männistö, S., Ovaskainen, M-L. & Valsta, L. (ed.) 2003. FINRAVINTO 2002-TUTKIMUS. THE NATIONAL FINDIET 2002 STUDY. Publications of the National Public Health Institute B3/2003. Nielsen, A.C. 2004. Kotitalouspaneeli 12 kk/30.6.04 ??? File from webpage NNR. 2004. Nordic Nutrition Recommendations, 4th edition. Nordic Council of Ministers. Copenhagen. Pena, D. & Ruiz-Castillo. 1998. The Estimation of Food Expenditures From Household Budget Data in the Presence of Bulk Purchases. American Statistical Association Journal Business & Economic Statistics. 16, 292-303.

98 Seppänen, L (ed.). 2004. Local and Organic Food and Farming around the Baltic Sea. Ekologiskt lantbruk nr 40. SLU. Statistics Sweden. 2004. Hushållens utgifter (HUT) PR 35 SM 0401 (in Swedish: Household expenditures). File from webpage www.scb.se/statistik/HE/PR0601/2003A01/8%20Hushållsgrupp%20och%20miljömärkta %20och%20ekologiska%20varor%20-andel%20per%20utgiftsgrupp.xls Statistical Yearbook of Finland 2004. 2004. Statistics Finland. SwEPA. 1997. Att äta för en bättre miljö. Slutrapport från systemstudie Livsmedel. (In Swedish: Eating for a better environment. Final report from systems study Food.) Rapport 4830. Naturvårdsverket. Stockholm. SwEPA. 1998. A Sustainable Food Supply Chain. A Swedish Case Study. The Swedish Environmental Protection Agency. Report 4966. Stockholm. Tennilä, L. 2000. Elintarvikkeiden kulutus kotitalouksissa [Households Food consumption]. Tilastokeskus [Statistics Finland]. Tulot ja kulutus [Income and Consumption] 2000:18.

99 8. Food basket scenarios In order to put the environmental studies presented earlier in this report in relation to the consumers consumption of food, two scenario studies were performed; one in Järna and one in Juva. Food basket scenario, Järna Sweden The eco-local food baskets presented in Chapter 7 in this report give examples of food consumption profiles consisting of more ecological and locally produced food than is common in the Nordic countries. Especially the Swedish example shows that it is possible in practise to buy such food when the demand exists. This case also represents a consumption profile that consists of less meat and more vegetables and root crops, which seems to be in accordance with nutritional and environmental recommendations given in for example the S.M.A.R.T. material (CTN, 2001).

The aim with the scenario study presented here was to investigate if, and in that case how much, the environmental impacts are reduced. The objective was to answer three main research questions: 1. Which environmental impact has the choice of food produced on ERA (ecological recycling agriculture) farms? 2. Which environmental impact has the choice of food produced locally? 3. Which environmental impact has the alternative food consumption profile obtained in the consumer survey in Järna?

Methodology The environmental impacts assessed were nitrogen surplus in agriculture, and global warming and consumption of primary energy resources in the agricultural, transporting and processing parts of the food system. The environmental impact data were obtained from results presented in Chapter 2 and 5 in this report. The food consumption profile data were derived from the results presented in Chapter 7.

Four scenarios with different combinations of food consumption profiles, agricultural systems, and processing and transportation systems were put together to answer the research questions. The scenarios are: 1. Average Swedish food consumption, Average Swedish agriculture 2002-2004, and Conventional food processing and transports 2. Average Swedish food consumption, ERA farms, and Conventional food processing and transports 3. Average Swedish food consumption, ERA farms, and Local (small-scale) food processing and transports 4. Eco-local food consumption, ERA farms, and Local (small-scale) food processing and transports

Pooling technique To achieve a good comparison between conventional agriculture and ERA we used a pooling technique, where it was calculated how much agricultural land that was necessary to use from each of four different production-type farm groups respectively to fulfil the food consumption. A detailed description of the stepwise technique based on the average Swedish food consumption per year and the alternative food basket in the Järna study will be presented

100 on the BERAS homepage www.jdb.se/beras. In order to get the consumption and production to match when the ERA farms should produce the food for the average Swedish food consumption profile, some assumptions were necessary. The most important was that the consumption volumes of ruminant meat (beef and lamb) and monogastric meat (pork and poultry) had to be exchanged. In other words, the ruminant meat consumption was increased to the level of today’s monogastric meat consumption but the total meat consumption was kept on the same level. The main reason for this was that ERA farming need more grassland growing (at least 40 % of the acreage under tillage), and it is the ruminants that can utilise that crop. The grain production consequently has to be decreased, which give less fodder for pigs and poultry.

Results Table 8-1 presents results of the pooling technique for nitrogen surplus in agriculture based on four production-type groups of BERAS-farms compared to the average Swedish agriculture. The average Swedish agriculture does not fully reflect the average Swedish food consumption but when using the pooling technique on the conventional farms presented by Myrbeck (1999) we got almost the same values of nitrogen surplus (not shown).

To interpret this result, it is important to take the land used outside Sweden in consideration to produce fodder which mainly is included in the BERAS farms. With this in mind we can see that the surplus of nitrogen total and per capita is 45 % with food from BERAS farms with this restriction that the conventional food basket is based on more ruminant meat than in the conventional food consumption of today. Table 8-1. Agricultural area need and Nitrogen surplus for three scenarios:´Swedish average (mainly conventional) agriculture, BERAS farms producing the average Swedish food-basket, and BERAS farms producing an alternative (mainly ecological) food-basket (presented in Chapter 10 in this report). Average Swe. agric. Swedish consumption & Eco-local consumption & 2002-2004 BERAS farms 2002-2004 BERAS farms 2002-2004 % % % Agr. area, million ha 2.45 100 4.76 194 1.70 69 Agr. area, ha/capita 0.27 100 0.53 194 0.19 69 Capita/ha 3.67 100 1.93 52 5.29 144 N-surplus, kg/capita 22 100 14 63 8 36 N-surplus, kg/ha 80 100 26 32 42 52 N-surplus, million kg 196 100 123 63 71 36 in Sweden

Concerning the agriculture the three scenarios presented in Table 8-1 are the relevant since that compares conventional and ecological (recycling) agriculture, and an alternative more vegetarian consumption. In Figure 8-1, Figure 8-2 and Figure 8-3, presenting results for Nitrogen surplus, Global Warming Potentials and Consumption of primary nutrient resources, four scenarios are relevant since also the dimension locally produced is compared.

A scenario based on BERAS-farms, with the same total meat consumption (but with a higher share of ruminant meat) should have a not realistic size of arable land in Sweden as a consequence of the lower productivity on the studied BERAS-farms. However, the conventional system also use agriculture outside Sweden but that is not included here. In the alternative Järna food consumption scenario based on BERAS-farms (Scenario 4), the need of agricultural arable land would on the other hand decrease with almost 40 % compared to the situation of today.

101

90 6 80 80 N-surplus/capita 5 70 N-surplus/ha

60 4

50 42 3 40 kg N surplus 30 26 26 2 22 Agricultural area (Mha) 20 14 14 1 8 10

0 0 1 2 3 4 Scenario

Figure 8-1. N-surplus in four scenarios, kg N per capita and kg N per ha.

4500 6 GWP/capita 4000 3700 GWP/ha 5 3500 3200

3000 4

2500 3 equivalents 2 2000 1700 1500 1500 2 kg CO

1000 900 Agricultural area (Mha) 1000 800 600 1 500

0 0 1 2 3 4 Scenario

Figure 8-2. Global warming potentials in four scenarios, kg CO2 equivalents per capita and kg CO2 equivalents per ha.

102 35 6 Primary energy/capita

30 Primary energy/ha 28 5

25 4

20 17 3 15

10 10 2

10 8 Agricultural area (Mha) 5 5 GJ primary energy consumption 1 5 3

0 0 1 2 3 4 Scenario

Figure 8-3. Consumption of primary energy resources in four scenarios, GJ primary energy resources per capita and GJ primary energy resources per ha. Figure 8-4 present a summary of the results. It shows the relative difference between the environmental impacts in the four scenarios. Scenario 1 represents the situation of today.

2,0

N-surplus/capita 1,5 GWP/capita

1,0 Primary

relativ value energy/capita 0,5 Agricultural area in Sweden, million ha 0,0 1 2 3 4 Scenario

Figure 8-4. N-surplus, Global warming potentials and Primary energy resources consumption per capita and agricultural area need in four scenarios, relative values.

Conclusions The results clearly show that changes in our food consumption has possibilities for making the environmental impacts studied lower. A large reduction of the environmental impacts would occur if the Swedes change their food consumption preferences in accordance with the eco-local food basket presented in Chapter 7 in this report.

103

A complete change to ERA (ecological recycling agriculture) would also decrease the environmental impacts, even when the food consumption profile remains as Swedish average of today. The agricultural area needed would, however, increase to almost the double why that combination is unrealistic. The conversion to 100 % ERA produced food would demand a change also in people’s food consumption profile.

Locally produced food showed a somewhat reduced global warming impact in the cases studied but nutrient surplus and consumption of primary energy resources did not change. Food basket scenario, Juva Finland Pentti Seuri, Miia Kuisma, Hanna-Riikka Tuhkanen

Introduction The production of the BERAS farms in Juva is concentrated on milk and beef production. Therefore it is not possible to assemble a reliable whole food basket scenario according to the data of these BERAS farms. From Finnish food basket bread cereals, milk and beef were selected.

Methodology The average Finnish food consumption of selected food items was described according to the official Finnish statistics (Ravintotase 2003). Only selected food items, which include 55 % of the total energy input through food, were considered (Figure 8-x). The rest of the food consumption was kept as before.

Figure 8-5. The Finnish food consumption 2002. Conventional agricultural production in Finland was described according to i.a. the official Finnish statistics (Maatilatilastollinen vuosikirja 2003, Lötjönen et al 2004). Average nutrient loadings of conventional agriculture practises in Finland were described according to national nitrogen balance (Lemola and Esala 2004). Nitrogen surplus were estimated both as an average surplus for whole agricultural area and as an average surplus for plant and animal production hectares separately.

Agricultural land need abroad for fodder production was not included in the scenario.

The BERAS farms in Juva presented organic agriculture practises. Farm data were collected from two specialized crop production farms, three milk farms and three beef farms. However, crop production farms produced mainly fodder grains (oat, barley), which were anyway used as a data source.

Results and discussion The agricultural land need on the situation where all bread cereals, milk and beef were produced according to the methods of organic agriculture would be 7 % bigger than on the present stage (Table 8-2). Table 8-2. Agricultural area demand (ha) and nitrogen (N) surplus (kg/product unit (1 000 kg) or kg/ha) on production of the average Finnish food consumption (bread cereals, milk and beef). Two agricultural production methods were compared with: conventional agricultural production in Finland and organic agriculture in BERAS farms in Juva.

104 Conventional Finnish BERAS farms 2002 agriculture 2002 % % Agricultural area (ha) 2 240 000 100 2 394 615 107 Agr. area demand / product unit (ha/1000 kg) bread cereals 100 012 100 152 400 152 milk 629 250 100 833 021 132 beef 330 095 100 340 283 103 N-surplus (kg/ha) a b bread cereals 60 30 100 36 60/120 milk 60 90 100 27 45/30 beef 60 90 100 29 48/32 N-surplus (kg/ Food basket 1) 63 561 420 89 341 410 100 37 846 174 60/42 a) The division between the agricultural products was made on the grounds of mean agricultural hectare and surplus. b) The division between the agricultural products was made on the grounds of animal production hectare, which was estimated to be larger than the surplus of crop production hectare. 1) The food basket includes 55 % of the energy content of the average Finnish food consumption for one year.

Nitrogen surpluses or loading potentials of the selected raw materials of food were smaller in organic production than in average agricultural production except for bread cereals when surplus differences between plant and animal production were included (Figure 8-6). 100000 beef 80000 milk bread cereals 60000

40000

N-surplus (1000 kg) 20000

0 conventional agriculture conventional agriculture BERAS farms (average surplus) (surplus based on production lines )

Figure 8-6. Nitrogen surplus in the agriculture of the average food consumption in Finland.

References Lemola, R. & Esala, M. 2004 Lötjönen, T., Muuttomaa, E., Koikkalainen, K., Seuri, P & Klemola, E. 2004. Laajamittaisen luomutuotannon teknologia – taloudellinen toteutettavuus ja ekologinen kestävyys. (Technology of large-scale organic production.) Maa- ja elintarveketalouden tutkimuskeskus. Maa- ja elintarviketalous 44. 131 pp. http://www.mtt.fi/met/pdf/met44.pdf Maatilatilastollinen vuosikirja 2003. 2003. (Year book of farm statistic 2003.) Maa- ja metsätalousministeriön Tietopalvelukeskus. Helsinki. 266 pp. Ravintotase 2002 ja 2003 (ennakko). 2003. (Food balance 2002 and 2003 (advance).) Maa- ja metsätalousministeriön Tietopalvelukeskus. Helsinki. 27 pp.

105 Appendix 1. Farm gate nutrient balances

Table A1-1. Characteristics of the BERAS-farms

Table A1-2. Farm gate balances with calculated surplus of N, P and K and primary nutrient balances for N on the Swedish BERAS-farms 2002-2004.

Table A1-3. Farm gate balances with calculated surplus of N, P and K and primary nutrient balances for N on the Finnish F-BERAS farms 2002-2004.

Table A1-4. Farm gate balances with calculated surplus of N and primary nutrient balances for N on the Finnish J-BERAS farms 2002-2004. Farm 8b and 9b are simulated with animals and 8a and 9a are excluded from the average values calculated for all farms.

106