Federal Department of Economic Affairs FDEA Agroscope Reckenholz-Tänikon Research Station ART How to establish life cycle inventories of agricultural products? Thomas Nemecek Agroscope Reckenholz-Tänikon Research Station ART CH-8046 Zurich, Switzerland http://www.agroscope.ch [email protected] 2 March 2010 Overview Defining system boundaries: temporal and process related How to get the LCI data: data survey vs. modelling ecoinvent database: Version 2.1 Future development to version 3.0 Direct field and farm emissions: how to estimate? Variability and uncertainty: Sources of variability Examples and implications Analysis of variability Assessment of uncertainty How to deal with missing data: generalisation and extrapolation Towards an integrated framework: SALCA Specific aspects of tropical crops Some recommendations How to establish life cycle inventories of agricultural products? T. Nemecek | © Agroscope Reckenholz-Tänikon Research Station ART 2 Defining system boundaries: Temporal system boundaries Annual crops: Starting after harvest of previous crop (including fallow period or catch crop, if no product) Ending with harvest of the considered crop Permanent crops: Annual basis (1st January to 31st December) or Multiannual cropping cycle (distinguishing different phases: planting, young plantation, main yielding phase, eradication) How to establish life cycle inventories of agricultural products? T. Nemecek | © Agroscope Reckenholz-Tänikon Research Station ART 3 Defining system boundaries: Example of crop production Animal production system Animal excrements Products: System boundary Resources Infrastructure: •Buildings •Machinery Manure storage Inputs: •Seed •Fertilisers (min. & org.) •Pesticides •Energy carriers •Irrigation water Field production Catch crops Field work processes: •Soil cultivation •Fertilisation •Sowing •Chemical plant protection •Mechanical treatment •Harvest •Transport Silage maize Sugar beets Fodder beets Beetroot Carrots Cabbage Grain drying Wheat Barley Rye Oats Grain maize CCM Faba beans Soya beans Protein peas Sunflowers Rape seed Potato grading Potatoes Product treatment: Direct and indirect emissions How to establish life cycle inventories of agricultural products? T. Nemecek | © Agroscope Reckenholz-Tänikon Research Station ART Co-Product: Straw © T. Nemecek, ART 2010 4 Defining system boundaries: Where to draw the line between animal and plant production? Animal production (incl. feedstuffs, buildings, emissions, etc.) Manure storage and treatment ? Manure application (incl. machinery use and emissions) Nutrient use in plant production How to establish life cycle inventories of agricultural products? T. Nemecek | © Agroscope Reckenholz-Tänikon Research Station ART © Gaillard & Nemecek, 2006 5 Single crop or cropping system? 2 3 4 5 6 2 7 8 Potatoes Year 4 1 2 3 4 5 6 7 8 Winter wheat 2 3 4 5 6 Spring barley 7 8 9 10 11 12 1 9 10 11 12 1 2 3 4 5 Forage catch crop 5 ... Fallow Month 9 10 11 12 1 3 6 7 8 9 10 11 12 Grain maize Fallow ... 1 ... Grassclover mixture Month 1 Green manure Year 6 2 3 4 5 6 7 8 9 10 11 12 1 2 3 4 5 6 7 8 9 10 11 12 Grass-clover mixture ... How to establish life cycle inventories of agricultural products? T. Nemecek | © Agroscope Reckenholz-Tänikon Research Station ART © T. Nemecek, ART 2001 6 How to get representative LCI data? Two approaches Structural, general production and economic data are regularly recorded in most countries (statistics, FADN, FAO, EUROSTAT) Data on agricultural management are largely missing (fertiliser use, pesticides, use of machinery, timing of interventions, etc.) Two possible solutions: 1. Make a large survey: pilot farm networks one single data source enables to assess the variability preferable, but very expensive! 2. Modelling LCI: based on statistics, FADN, recommendations, expert knowledge, etc. combination of several different data sources difficult to assess the variability most frequently used alternative, much cheaper How to establish life cycle inventories of agricultural products? T. Nemecek | © Agroscope Reckenholz-Tänikon Research Station ART 7 How to get representative LCI data? 1. Example of Swiss farm LCA network Project Life Cycle Assessment – Farm Accountancy Data Network (LCA-FADN) Integrate environmental LCA into FADN Project supported by the Swiss Federal Office for Agriculture Time-frame: 2004 - 2010 with data acquisition from 2006 - 2008 Establish an operating system with 110 farms (during 3 years with 60 in the first year) Establish an information technology infrastructure Training life cycle management principles in practice Develop concepts for evaluation and communication and practice them with farmers and extension services Sectoral monitoring and environmental management of farms How to establish life cycle inventories of agricultural products? T. Nemecek | © Agroscope Reckenholz-Tänikon Research Station ART 8 Farm How to get representative LCI data? 1. Example of farm network / Project LCA-FADN: workflow Farm management software (AGRO-TECH) AccountancySoftware Technical data AccountancySoftware (AGRO-TWIN) LCA centre Trust and accounting office Accountancy data Accountancy Data LCA data FADN evaluation centre Plausibility tests SALCAprep data extraction SALCAcalc LCA calculation SALCAcheck LCA validation and benchmarking Export ÖB-Stelle FADN database Feedback to farmers Existing FADN How to establish life cycle inventories of agricultural products? T. Nemecek | © Agroscope Reckenholz-Tänikon Research Station data ART accountancy Synergies FADN LCA-FADN New FADN 9 Life Cycle Assessment © Agroscope ART 2010 How to get representative LCI data? 2. Example of modelling LCI Data category Data source(s) Yields for main products FADN ART (weighted means for 1996-2003) Straw yields and crop residues Fertilising recommendations (Walther et al. 2001) Moisture content Quantity of seed Use of machinery (number of passes) Gross-margin catalogue from the extension service (LBL et al. 2000) Sowing and harvest dates Work budget (planning tool, Näf 1996) Quantity of fertilisers Fertilising recommendations (Walther et al. 2001) Types of fertilisers in integrated systems Import statistics (years 1996-98 from Rossier 2000) for mineral fertilisers Pilot farm network (years 1994-96 from BLW et al. 1998) for farmyard manure Types of fertilisers in organic systems Pilot farm network (years 1994-96 from BLW et al. 1998) for farmyard manure Pesticide applications Pilot farm network (years 1994-96 from BLW et al. 1998) Chemical seed dressing Information provided by seed suppliers and experts (survey) How to establish life cycle inventories of agricultural products? T. Nemecek | © Agroscope Reckenholz-Tänikon Research Station ART Source: Nemecek, Erzinger (2004). Modelling representative 10 life cycle inventories for Swiss arable crops. Int J LCA. Sources of LCI data: ecoinvent database v.2.1 More than 4000 generic LCI process datasets on energy supply, resource extraction, material supply, chemicals, metals, agriculture, waste management services, and transport services A joint initiative of the Used by over 1200 members in more than 40 countries ETH domain and Swiss Federal Offices Included in the leading LCA software and eco-design tools Online access to LCI and LCIA results for all datasets Based on industry data, compiled by independent experts Consistent, validated and transparent Continuously maintained International in scope, including e.g. data on US agriculture, worldwide sourcing of raw materials and production of electronics in Asia How to establish life cycle inventories of agricultural products? T. Nemecek | © Agroscope Reckenholz-Tänikon Research Station ART 11 Datasets for the biomass production in ecoinvent: Overview 1. Datasets on agricultural means of production: infrastructure (buildings and machinery) and its usage, fertilisers, pesticides, seed and animal feed 2. Datasets on agricultural and biomass products: • • • • Swiss Centre For Life Cycle Inventories A joint initiative of the ETH domain and Swiss Federal Offices Arable crop products Grass Wood Fibres How to establish life cycle inventories of agricultural products? T. Nemecek | © Agroscope Reckenholz-Tänikon Research Station ART 12 relevant datasests available How to establish life cycle inventories partly of agricultural products? available T. Nemecek | © Agroscope Reckenholz-Tänikon Research Station ART not available Products Asia Products America Products Europe Products CH Inputs Work processes Machinery Production branches Arable crops Fodder crops Horticulture (Field) Horticulture (Greenhouse) Fruit growing Vineyards Cattle production Pig production Poultry production Sheep production Buildings Contents of ecoinvent Version 2.1 What is covered in agriculture? Swiss Centre For Life Cycle Inventories A joint initiative of the ETH domain and Swiss Federal Offices © ecoinvent centre, 2007 13 Contents of ecoinvent version 2.1 Datasets for biomass production Category agricultural means of production agricultural means of production agricultural means of production agricultural means of production agricultural means of production agricultural means of production agricultural means of production agricultural means of production agricultural production agricultural production biomass wooden materials wood energy Total Subcategory Number of datasets buildings 23 machinery 6 work processes 39 mineral fertiliser 24 organic fertiliser 5 pesticides 68 seed 26 feed 10 plant production 120 animal production 4 production 4 extraction 123 fuels 13 465 How to establish life cycle inventories of agricultural products? T. Nemecek | © Agroscope Reckenholz-Tänikon Research Station ART Swiss Centre For Life Cycle Inventories A joint initiative of the ETH domain and Swiss Federal Offices © ecoinvent centre, 2007 14 Contents of ecoinvent version 2.1 Crops and countries Crops barley cotton faba beans fodder beets grain maize grass grass silage green manure hay hemp jute kenaf oil palm potato protein peas ramie rape seed rice rye silage maize soy beans sugar beets sugar cane sunflower sweet sorghum wheat Cereals Oil crops Protein crops Fibre crops Grass How to establish life cycle inventories of agricultural products? T. Nemecek | © Agroscope Reckenholz-Tänikon Research Station ART Countries Brazil Cameroon China Europe France Germany Global India Malaysia Philippines Scandinavia Spain Switzerland Thailand USA Swiss Centre For Life Cycle Inventories A joint initiative of the ETH domain and Swiss Federal Offices © ecoinvent centre, 2007 15 ecoinvent database: online access How to establish life cycle inventories of agricultural products? T. Nemecek | © Agroscope Reckenholz-Tänikon Research Station ART 16 Example: Unit Process Inventory (extract from V1.0) wheat straw IP, at farm CH (kg) SD wheat grains IP, at farm CH (kg) Location/ Uncert Type Unit process inventory for: wheat IP, CH Exchanges Category Unit Value ammonium nitrate, as N, at regional storehouse RER kg 6.71E+01 1 1.07 (2,1,1,1,1,na) 92% 8% pesticide unspecified, at regional storehouse CH kg 2.60E-01 1 1.13 (2,2,3,1,1,na) 92% 8% wheat seed IP, at regional storehouse CH kg 1.80E+02 1 1.07 (2,1,1,1,1,na) 92% 8% tillage, ploughing CH ha 1.00E+00 1 1.07 (2,1,1,1,1,na) 92% 8% grain drying, low temperature CH kg 7.64E+01 1 1.07 (2,1,1,1,1,na) 100% resource/land m2a 7.94E+03 1 1.77 (2,1,1,1,1,na) 92% 8% Transformation, from pasture and meadow, intensive resource/land m2 2.90E+03 1 2.67 (2,1,1,1,1,na) 92% 8% Carbon dioxide, in air resource/in air kg 1.39E+04 1 1.07 (2,2,1,1,1,na) 61% 39% Energy, gross calorific value, in biomass resource/biotic MJ 1.67E+05 1 1.07 (2,2,1,1,1,na) 59% 41% 95% Uncert Scores ............................... ............................... Occupation, arable, non-irrigated ............................... air/low population Ammonia density kg 9.06E+00 1 1.30 (2,2,1,1,1,na) 92% 8% Phosphorus water/river kg 2.58E-01 1 1.77 (2,2,1,1,1,na) 92% 8% Nitrate water/ground- kg 1.25E+02 1 1.77 (2,2,1,1,1,na) 92% 8% Isoproturon soil/agricultural kg 1.27E+00 1 1.32 (2,2,3,1,1,na) 92% 8% Cadmium soil/agricultural kg 3.91E-03 1 1.77 (2,2,1,1,1,na) 42% 58% wheat grains IP, at farm CH kg 6.42E+03 wheat straw IP, at farm CH kg 3.91E+03 How to establish life cycle inventories of agricultural products? T. Nemecek | © Agroscope Reckenholz-Tänikon Research Station ART 100% 100% © ecoinvent centre, 2003 17 Plans for the ecoinvent database v.3.0 – release 2011 Co-operation with national database initiatives More detail, more technologies, more completeness: International editorial board and broader supplier base Parameterisation (geography, time, technologies, markets) New data structure based on supply-use framework, allowing easier production of national versions New indicators Sponsor-funded Open Access to individual datasets More frequent updating Improved uncertainty estimation and calculation facilities How to establish life cycle inventories of agricultural products? T. Nemecek | © Agroscope Reckenholz-Tänikon Research Station ART 18 New developments for ecoinvent V3.0: International editorial board and broader supplier base International editorial board Activity editors, for each industry activity and for household activities Cross-cutting editors, to ensure consistency and monitor developments across the entire database, both for specific (groups of) emissions, for geographical areas, scenarios, etc., and for the meta-data fields, e.g. uncertainty Broader supplier base Making it easier for experts and lay users to contribute with new data or corrections to existing data All such contributions will still be subject to our strict quality control, review, and validation procedures before entering into the database How to establish life cycle inventories of agricultural products? T. Nemecek | © Agroscope Reckenholz-Tänikon Research Station ART 19 New developments for ecoinvent V3.0: Parameterisation Geographical parameters: Core international datasets + national differences Using GIS coordinates, all other area parameters can be expressed: Country codes, areas with different population densities, habitat areas, watershed areas, etc. for site-dependent impact assessment Temporal parameters (years) Scenario parameters (e.g. BaU, optimistic, pessimistic) Dataset-internal parameters Inheritance using parent child-relationships How to establish life cycle inventories of agricultural products? T. Nemecek | © Agroscope Reckenholz-Tänikon Research Station ART 20 New developments for ecoinvent V3.0: Better support for alternative modelling options Attributional and consequential modelling: Average versus marginal market modelling Allocation versus substitution (system expansion) Several versions of attributional allocation The unallocated (multi-functional) unit processes are the same for both models How to establish life cycle inventories of agricultural products? T. Nemecek | © Agroscope Reckenholz-Tänikon Research Station ART 21 Estimating direct field and farm emissions Usually no measurement on site possible Two options: 1. Literature values, experiments: take a value for a given situation Specific for the situation Difficult to find Not flexible Mitigation options usually cannot be considered 2. Modelling More flexible Mitigation options can be considered, depending on the model Level of detail should be consistent across the models No globally usable emission models available How to establish life cycle inventories of agricultural products? T. Nemecek | © Agroscope Reckenholz-Tänikon Research Station ART 22 Estimating direct field and farm emissions Ideal emission models should Reflect the underlying environmental mechanisms Be site and time dependent Consider the effect of soil and climate Consider the effect of management Be applicable under a wide range of different situations The different models should have a similar level of detail But also be usable: Parameters are measurable Data can be collected in a reasonable time Calculation is feasible A compromise is needed! How to establish life cycle inventories of agricultural products? T. Nemecek | © Agroscope Reckenholz-Tänikon Research Station ART 23 SALCA emission models Ammonia (NH3) 4 Emissions paths are modelled: 1. Application of farm manure = f(fertiliser amount, NH3 and NH4-concentration, covered area, saturation deficit in the air in function of average monthly temperature) 2. Application of mineral fertiliser = emission factors according to fertiliser type (2-15%, Asman 1992) 3. Emission from pasture = 5% of total N in excrements emitted as NH3 4. Emission from stable = emission factors dependent on animal category, housing system, farm manure type (liquid or solid) Source: Menzi et al. (1997) How to establish life cycle inventories of agricultural products? T. Nemecek | © Agroscope Reckenholz-Tänikon Research Station ART © Agroscope ART, 2010 24 SALCA emission models Nitrous oxide (N2O) N2O in air: adapted method according to IPCC 2006, under consideration of induced N2O-Emissions: Fertilisers: Direct emissions: 1% of available N Symbiotic N-fixation in legumes: no emissions Crop residues: emission factor 1% Storage of farmyard manure: emission factors 0.1% for liquid manure and 2% for dung Pasture: emission factor 2% Induced Emissions: 1% of NH3-N and 0.75% of NO3-N How to establish life cycle inventories of agricultural products? T. Nemecek | © Agroscope Reckenholz-Tänikon Research Station ART © Agroscope ART, 2010 25 SALCA emission models SALCA-nitrate N mineralisation of soil organic matter N uptake plants Non leached N Leaching Leaching + Input of mineral N through fertilisers (NH4, NO3, Amid-N) GRUDAF: Temperature dependent Example: 60 dt yield 80 dt yield 158 kg N uptake N-Uptake functions 211 kg N uptake (STICS) Monthly N-uptake Source: Richner et al. (2006) How to establish life cycle inventories of agricultural products? T. Nemecek | © Agroscope Reckenholz-Tänikon Research Station ART © Agroscope ART, 2010 26 SALCA emission models Methane (CH4) IPCC method 2 (Houghton et al. 1995) currently under revision Animal breading: Emissions from digestion = f(animal category, feeding) Emissions from storage of farm manure = f(animal category, housing system) Emission factors: Liquid manure: 10% Dung and pasture: 1% How to establish life cycle inventories of agricultural products? T. Nemecek | © Agroscope Reckenholz-Tänikon Research Station ART © Agroscope ART, 2010 27 SALCA emission models Phosphorus (P) 4 kinds of P-emissions in water: • • • • Surface run-off in rivers (solved PO43-) Drainage losses in rivers (solved PO43-) Erosion in rivers (P bound to soil particles) Leaching in ground water (solved PO43-) Emissions are dependent of: • • • • • • Soil characteristics (granulation, bulk density, soil water balance) and drainage Quantity of P-fertiliser Type of P-fertiliser (manure, compost, mineral) Field slope and distance to rivers Quantity of eroded soil Plant available P in upper soil How to establish life cycle inventories of agricultural products? T. Nemecek | © Agroscope Reckenholz-Tänikon Research Station ART Source: Prasuhn (2006) © Agroscope ART, 2010 28 SALCA emission models Heavy metals Input-Output-Balance (caused by farmer) per field for: Cd, Cu, Zn, Pb, Ni, Cr, Hg Inputs: - Fertilisers (mineral and organic) Seed Pesticides Feedstuff and auxiliary materials for animal breeding Outputs: - Exported primary products (e.g. grains, meat) Exported co-products (e.g. straw, animal manure) Leaching to groundwater and drainage to surface water Erosion to surface water Allocation for inputs caused by the farmer Source: Freiermuth (2006) The final balance can be negative! How to establish life cycle inventories of agricultural products? T. Nemecek | © Agroscope Reckenholz-Tänikon Research Station ART © Agroscope ART, 2010 29 Variability and uncertainty: Factors influencing environmental impacts Socio-economic conditions Crop management Pedo-climatic conditions Crop yield Life cycle inventory Environmental impacts How to establish life cycle inventories of agricultural products? T. Nemecek | © Agroscope Reckenholz-Tänikon Research Station ART To understand the variability of environmental impacts, we need to look on the variability of the influencing factors © T. Nemecek ART, 2010 30 median q25% q2.5% Yield [t/ha] Global variability of yields Example: potato 50 45 40 35 30 25 20 15 10 5 0 Cumulated potato world production as a function of the yield 0.0 20.0 40.0 60.0 80.0 100.0 Cumulated world production [%] Source: FAOSTAT How to establish life cycle inventories of agricultural products? T. Nemecek | © Agroscope Reckenholz-Tänikon Research Station ART 31 Variability of environmental impacts: Wheat datasets in ecoinvent V2.01 (2007) US w heat grains, at farm Saxony, DE 0.76 Barrois, FR 0.63 CH, org 0.59 0.20 0.40 0.60 6.42 Barrois, FR 0.67 CH, IP 3.49 Castilla, ES 0.59 CH, ext 4.63 Saxony, DE 0.55 Castilla, ES 0.00 US 0.60 0.80 How to establish life cycle inventories of agricultural products? factors| of crop LCI/LCA variability: exampleResearch of wheat Station ART T. Key Nemecek © Agroscope Reckenholz-Tänikon 100a, kg CO2-eq./kg T. Nemecek | © GWP Agroscope Reckenholz-Tänikon Research Station ART 3.58 CH, org 2.31 CH, ext 3.45 CH, IP 3.30 0.00 2.00 4.00 6.00 © ecoinvent centre 2007 w heat grains, at farm 8.00 energy demand, MJ-eq./kg 32 Variability of environmental impacts: Energy demand per ha UAA (62 Swiss farms) M J -E q . Energy demand per ha UAA 300000 280000 260000 240000 220000 200000 180000 160000 140000 120000 100000 80000 60000 40000 20000 0 31 22 21 22 23 11 21 21 15 11 11 14 11 22 21 11 11 21 21 51 13 21 53 51 51 21 21 51 55 51 51 11 11 21 14 21 22 16 53 52 53 21 53 21 55 23 21 21 11 51 53 53 56 51 11 53 53 53 Fa r m ty pe 11 13 14 15 16 21 22 D es c ri pti on a rab le fa rm ing v e ge ta b le c u ltiv at io n f ruit c ultiv at io n v itic ultu re o th er c u ltu re s d airy f arm s u c k ler c o ws Fa rm ty pe How to establish life cycle inventories of agricultural products? T. Nemecek | © Agroscope Reckenholz-Tänikon Research Station ART 23 31 51 52 53 55 56 D es c rip tio n o th er c at tle h ors e s /g o at s /s h e ep d airy fa rm / a ra ble fa rm in g c om bin ed s u c k ler c ow s / a rab le fa rm ing c o m b ine d p ig s a n d po u ltry / a ra ble fa rm in g c om bine d d airy fa rm s / o th e r c o m b in e d c a ttle / ot he r c o m b ine d © Agroscope ART, 2010 33 Variability of environmental impacts: Example: Energy demand per ha UAA (dairy farms) Energy demand of dairy farms 80000 Eutrophication of dairy farms 70000 other inputs 350 emissions of animals purchase of foodstuff 300 purchase of animals 50000 PPP 250 40000 kg N-Eq./ha UAA MJ-Eq./ha UAA 60000 30000 20000 10000 200 150 100 other inputs emissions of animals purchase of foodstuff fertiliser / nutrients purchase of animals seeds PPP energy carriers fertiliser / nutrients machines seeds buildings / equipment energy carriers machines 0 50 1 2 3 4 5 6 buildings / equipment 7 8 9 1 2 3 10 11 12 13 14 15 0 4 5 6 7 How to establish life cycle inventories of agricultural products? T. Nemecek | © Agroscope Reckenholz-Tänikon Research Station ART 8 9 10 11 12 13 14 15 © Agroscope ART, 2010 34 (i) Energy use (MJ eq./$) 6.0 Variance control as a basis for environmental management y = 3.79x - 0.46 r = 0.73, P = 0.007 4.0 M 2.0 An balanced use of energy and fertilisers improves ecoefficiency. The best farms (1, 2) had the lowest pesticide use per area unit. The orchards have high yields (high labour input) and a good physiological and ecological equilibrium. Farm No. 1 Farm No. 2 0.0 0% 20% 40% 60% 80% 100% 120% 140% 160% Aq. (iii) Eutrophication (PO4 eq./$) 0.15 y = 0.09x - 0.01 r = 0.77, P = 0.003 0.10 M 0.05 Farm No. 1 Farm No. 2 0.00 How to establish life cycle inventories of agricultural products? T. Nemecek | © Agroscope Reckenholz-Tänikon Research Station ART 0% 20% 40% 60% 80% 100% 120% Coefficent of Variance 140% 160% 35 Source: Mouron et al. (2006) 16 0.8 14 0.7 12 0.6 10 0.5 8 0.4 6 0.3 NH3 emission (kg NH3/ha) 4 0.2 Relative emission rate (kg NH3-N/kg TAN) 2 0.1 0 kg NH3-N/kg TAN kg NH3/ha Variability and non-linearity Averages may lead to wrong results 1x40 m3 slurry 13.5 kg NH3 2x20 m3 slurry 17.4 kg NH3 0.0 0 10 20 30 m3 of slurry 40 50 Ammonia emission as a function of quantity of slurry applied. TAN = total ammonia N in the slurry (after Menzi et al. 1997) How to establish life cycle inventories of agricultural products? T. Nemecek | © Agroscope Reckenholz-Tänikon Research Station ART © T. Nemecek ART, 2010 36 Uncertainty assessment in ecoinvent V2.1: Pedigree matrix Indicator score Reliability Completeness Temporal correlation Geographical correlation Further technological correlation Sample size 1 2 3 4 Qualified estimate (e.g. Verified data partly by industrial expert); data Non-verified data partly Verified data based on based on assumptions derived from theoretical based on qualified measurements OR non-verified data information estimates based on measurements (stoichiometry, enthalpy, etc.) Representative data Representative data Representative data Representative data from all sites relevant for from >50% of the sites from only some sites from only one site relevant for the market (<<50%) relevant for the relevant for the market the market considered market considered OR considered OR some over an adequate period considered over an adequate period to even >50% of sites but from to even out normal sites but from shorter out normal fluctuations shorter periods fluctuations periods 5 Non-qualified estimate Remarks verified means: published in public environmental reports of companies, official statistics, etc unverified means: personal information by letter, fax or e-mail Representativeness unknown or data from a Length of adequate period depends on small number of sites process/technology AND from shorter periods less than 3 years means: data measured in 1997 or later; Less than 15 years of Less than 10 years of Less than 6 years of Less than 3 years of score for processes with investment cycles difference to our difference to our difference to our difference to our of <10 years; reference year (2000) reference year (2000) reference year (2000) reference year (2000) for other cases, scoring adjustments can be made accordingly Similarity expressed in terms of Data from unknown OR enviornmental legislation. Suggestion for grouping: distinctly different area Average data from larger Data from smaller area (north america instead of North America, Australia; Data from area under area in which the area than area under study, or European Union, Japan, South Africa; middle east, OECDstudy under study is included from similar area South America, North and Central Africa Europe instead of and Middle East; Russia) Russia, China, Far East Asia Examples for different technology: Data on related - steam turbine instead of motor propulsion processes or materials Data on related in ships processes or materials Data on related but same technology, Data from enterprises, - emission factor B(a)P for diesel train but different technology, processes or materials OR processes and materials based on lorry motor data but on laboratory scale of OR data on laboratory Data from processes under study (i.e. identical Examples for related processes or different technology scale processes and and materials under technology) materials: study but from different same technology - data for tyles instead of bricks production technology - data of refinery infrastructure for chemical >100, continous > 10, aggregated figure sample size behind a figure reported in the >=3 unknown measurement, balance >20 in env. report information source of purchased products Age of data unknown or more than 15 years of difference to our reference year (2000) How to establish life cycle inventories of agricultural products? T. Nemecek | © Agroscope Reckenholz-Tänikon Research Station ART © ecoinvent centre, 2007 37 Uncertainty assessment for French wheat 95% confidence interval How to establish life cycle inventories of agricultural products? T. Nemecek | © Agroscope Reckenholz-Tänikon Research Station ART © T. Nemecek ART, 2010 38 Potential use of multivariate statistics in LCA explain variability Multivariate statistics (like principal component analysis, PCA) can be used to show similarities between environmental impacts It can be also used to group environmental profiles, e.g. of crops Analysis based on a set of midpoint LCIA indicators In the study applied to crop inventories from SALCA (Switzerland) and ecoinvent (global) How to establish life cycle inventories of agricultural products? T. Nemecek | © Agroscope Reckenholz-Tänikon Research Station ART 39 Principal component analysis of SALCA inventories Eigenvalues of correlation matrix Active variables only 5.0 4.5 52.73% 4.0 3.5 Eigenvalue 3.0 2.5 27.63% 2.0 1.5 1.0 6.18% 5.07% 4.51% 0.5 2.24% .94% .70% 0.0 -0.5 -1 0 1 2 3 4 5 6 7 8 9 10 Eigenvalue number 80% of variance explained by first two principal components How to establish life cycle inventories of agricultural products? T. Nemecek | © Agroscope Reckenholz-Tänikon Research Station ART © T. Nemecek ART, 2010 40 Principal component analysis of SALCA inventories Projection of the variables on the factor-plane ( 1 x 2) 1.0 Acidi Eutro 0.5 Factor 2 : 27.63% GWP Ozone 0.0 AET_EDIP Energy HTP_CML -0.5 TET_EDIP -1.0 -1.0 -0.5 0.0 0.5 1.0 Factor 1 : 52.73% Relationship between impact indicators and factors 1 and 2 How to establish life cycle inventories of agricultural products? T. Nemecek | © Agroscope Reckenholz-Tänikon Research Station ART © T. Nemecek ART, 2010 41 Factor 1: - can group crops - related to the yield 5 Data for Swiss crops from SALCA database: grouping by crop group (CER = cereals, LEG = legumes, MAI = maize, OIL = oil crops, ROOT = root crops, VEG = vegetables). 4 3 Factor 2 2 1 0 -1 -2 -3 -4 -6 -4 -2 0 2 4 6 CER LEG MAI OIL ROOT VEG Factor 1 How to establish life cycle inventories of agricultural products? T. Nemecek | © Agroscope Reckenholz-Tänikon Research Station ART © T. Nemecek ART, 2010 42 Factor 2: - related to the farming system and the intensity 5 4 Data for Swiss crops from SALCA database: grouping by farming system (Conv=conventional, IPint = integrated intensive, IPext = integrated extensive, Org = organic). 3 2 Factor 2 1 0 -1 -2 -3 -4 -6 -4 -2 0 2 4 6 Conv Ipint Ipext Org Factor 1 How to establish life cycle inventories of agricultural products? T. Nemecek | © Agroscope Reckenholz-Tänikon Research Station ART © T. Nemecek ART, 2010 43 Principal component analysis of SALCA inventories Scatterplot (FALSR58_Res 14v*246c) Factor 1 = -5.9426-2.1271*x 2 1 0 Factor 1 -1 -2 -3 -4 -5 -6 -7 -3.0 -2.8 -2.6 -2.4 -2.2 -2.0 LnInvYield:Factor 1: r2 = 0.4561 -1.8 -1.6 -1.4 -1.2 -1.0 -0.8 -0.6 -0.4 LnInvYield Yield is a key factor How to establish life cycle inventories of agricultural products? T. Nemecek | © Agroscope Reckenholz-Tänikon Research Station ART © T. Nemecek ART, 2010 44 Principal component analysis of ecoinvent inventories 3.0 2.5 2.0 1.5 1.0 0.5 0.0 -0.5 -1.0 -1.5 -2.0 -2.5 -8 -6 -4 -2 0 2 4 6 CER FIB LEG MAI OIL ROOT Factor 1 Effect of the crop group (factor 1) How to establish life cycle inventories of agricultural products? T. Nemecek | © Agroscope Reckenholz-Tänikon Research Station ART © T. Nemecek ART, 2010 45 Principal component analysis of ecoinvent inventories 3.0 2.5 2.0 1.5 1.0 0.5 0.0 -0.5 -1.0 -1.5 -2.0 -2.5 -8 -6 -4 -2 0 2 4 6 Conv IPint IPext Org Factor 1 Effect of the farming system (factor 2) How to establish life cycle inventories of agricultural products? T. Nemecek | © Agroscope Reckenholz-Tänikon Research Station ART © T. Nemecek ART, 2010 46 Principal component analysis of ecoinvent inventories 3 CHCH CH 2 1 CH CH CH CH CH CH 0 ESES -1 FRFR US DE DE RER -2 -8 -6 -4 -2 0 2 4 6 w heat barley rye Factor 1 Cereals in different countries How to establish life cycle inventories of agricultural products? T. Nemecek | © Agroscope Reckenholz-Tänikon Research Station ART © T. Nemecek ART, 2010 47 Potential use of multivariate statistics in LCA to explain variability Between 76 and 80% of the variability could be explained by the first two principal components. Factor 1 crop (group) and yield Factor 2 farming system (conventional, integrated, extensive, organic) More data are needed for more systematic analyses The analysis helps to show similarities and differences between environmental profiles to find suitable proxies to derive simplified methods for extrapolations and approximations How to establish life cycle inventories of agricultural products? T. Nemecek | © Agroscope Reckenholz-Tänikon Research Station ART 48 How to fill data gaps in agricultural LCI? The classical approach: 1. Establish detailed and specific inventories for each situation Currently used alternatives: 2. Use proxies: what you think is the closest LCI (generalisation) 3. Streamlined LCA models New approaches: 4. Extrapolation by yield correction 5. Modular extrapolation method geographical extrapolation product extrapolation How to establish life cycle inventories of agricultural products? T. Nemecek | © Agroscope Reckenholz-Tänikon Research Station ART 49 Extrapolation by yield correction c' c' E E Product extrapolation: E cp e p ac ' (1 e p ) ac Y Y l' l' E E l a a Geographical extrapolation: E p e p l ' (1 e p ) l Y Y Impacts related to the yield (constant per kg) Impacts not related to the yield (constant per ha) ep Fraction of the impacts related to the yield Estimation of this fraction: • 0.7 for cereals from the ecoinvent datasets • 0.5 as default value How to establish life cycle inventories of agricultural products? T. Nemecek | © Agroscope Reckenholz-Tänikon Research Station ART © Roches & Nemecek ART, 2010 50 Modular EXtrapolation for Agricultural LCA (MEXALCA) Basic idea: It is possible to split an inventory into different independent modules. This enables easier adaptation of an existing inventory to a new situation. Working procedure: 1. 2. 3. 4. Establish a base inventory for one or several typical situations Split the inventory into independent modules Calculate unit inventories/impacts per module and input unit Determine amount of inputs used in each country (using global estimators derived from FAOSTAT) 5. Extrapolate inventory to any producing country 6. Estimate global/regional impacts (medians, means, distribution) How to establish life cycle inventories of agricultural products? T. Nemecek | © Agroscope Reckenholz-Tänikon Research Station ART 51 Extrapolation using MEXALCA Impacts per input unit Base crop inventory Basic cropping operations Soil tillage Variable machinery operations N fertilisation, including N-emissions P fertilisation, including P-emissions K fertilisation Pesticide application Irrigation Product drying Splitting Calculation of unit impacts Input parameters: •yield per area unit •Mechanisation index •% of no-till area •kg N, P2O5, K2O applied •kg pesticide active ingredient •m3 water used •kg water evaporated Global estimators (based on FAOSTAT) Good quality data available for some or all inputs Total impacts for extrapolated country x GWP 100 a [kg CO2-eq/kg] 0.3 Global distribution of impacts 0.25 Extrapolation 0.2 Total impacts for extrapolated country y 0.15 0.1 Impacts for extrapolated situation Total impacts 0.05 How to establish life cycle inventories of agricultural products? extrapolated T. Nemecek | © Agroscope Reckenholz-Tänikon Research for Station ART 0 0% 20% 40% 60% 80% Percentage of the world potato production 100% 52 country z © T. Nemecek ART, 2010 MEXALCA results: impacts per input unit Modules Potatoes Impacts MachFix MachTill MachVar Nfert 13604.50 1818.25 1074.68 photochemic O3 formation [kg ethylene-eq] non-renewable Energy [MJ-eq] Pfert Kfert Pestic Irrigat Drying 4621.45 70.91 31.26 10.69 341.5 9.988 0 118.49 272.66 0.614 15.127 0.247 0 0.65 0.08 0.23 0.001 6E-04 2E-04 0.0092 2E-04 0 Nutrient enrichment [kg N-eq] 12.65 0.34 0.60 0.917 0.126 7E-04 0.023 2E-04 0 Acidification [kg SO2-eq] 9.38 0.95 1.80 0.282 0.039 0.003 0.099 9E-04 0 Aquatic ecotoxicity 100a [kg 1,4-DCB-eq] 56.92 0.13 0.45 0.015 0.404 0.007 114.99 4E-04 0 Terrestrial ecotoxicity 100a [kg 1,4-DCB-eq] 0.99 0.01 0.05 7E-04 0.009 3E-04 80.696 1E-04 0 460.52 38.32 209.11 1.216 0.97 0.337 337.68 0.181 0 GWP 100a [kg CO2-eq] Human toxicity 100a [kg 1,4-DCB-eq] How to establish life cycle inventories of agricultural products? T. Nemecek | © Agroscope Reckenholz-Tänikon Research Station ART 13.45 2 © Roches & Nemecek ART, 2010 53 MEXALCA results: impacts per kg of potato in the world QUANTILES Energy [MJ-eq] GWP [kg CO2-eq] O3 form. [kg ethylene-eq] IMPACTS Nutr. enrich. [kg N-eq] Acidific. [kg SO2-eq] Aquat. Ecotox.[kg 1,4-DCB-eq] Terr. Ecotox. [kg 1,4-DCB-eq] Human tox.[kg 1,4-DCB-eq] 2.5% 9.11E-01 7.38E-02 2.84E-05 1.85E-03 9.44E-04 10.0% 9.77E-01 8.58E-02 3.13E-05 1.92E-03 1.14E-03 6.91E-02 6.96E-02 25.0% median 75.0% 1.27E+00 1.72E+00 3.00E+00 1.11E-01 1.23E-01 1.91E-01 4.75E-05 6.59E-05 8.50E-05 2.41E-03 3.44E-03 5.54E-03 1.23E-03 1.49E-03 2.27E-03 1.18E-02 1.65E-02 2.30E-02 5.41E-03 9.15E-03 1.26E-02 7.26E-02 8.34E-02 1.01E-01 90.0% 3.05E+00 1.92E-01 8.53E-05 5.61E-03 2.30E-03 3.06E-02 1.89E-02 1.40E-01 97.5% 4.15E+00 2.05E-01 1.07E-04 7.52E-03 2.82E-03 5.24E-02 3.50E-02 2.00E-01 The modular inventory system enables us to calculate the inputs and impacts in any producing country and to calculate median and quantiles for the inputs and for the impacts for the global production (per kg of product or per cultivated ha). How to establish life cycle inventories of agricultural products? T. Nemecek | © Agroscope Reckenholz-Tänikon Research Station ART © Roches & Nemecek ART, 2010 54 Results: estimated distribution of GWP of the potato production GWP 100 a [kg CO2-eq/kg] 0.3 0.25 0.2 0.15 0.1 0.05 0 0% 20% 40% 60% 80% 100% Percentage of the world potato production How to establish life cycle inventories of agricultural products? T. Nemecek | © Agroscope Reckenholz-Tänikon Research Station ART © Roches & Nemecek ART, 2010 55 First validation: impacts per kg Global Warming Potential 100 years [kg CO2-eq] Acidification [kg SO2-eq] Colours barley wheat ry e potato pea 2 4 6 8 10 0.2 0.4 0.6 0.8 1.0 1. 2 ecoinvent ec oinvent Photochemical ozone formation [kg ethylene-eq] Nutrient enrichment [kg N-eq] r2 0.43 5 0.006 0. 6 r2 0.493 y 1.1 03x 0.00 2 Colours b arley whea t ry e p otat o p ea 0.002 r 2 0.493 modular inventory 1. 0 y 0.386x 0.198 0. 2 4 6 r2 0.796 modular inv entory 8 y 1.163 x -0.042 2 modular inventory 10 0.010 Non renewable energy demand [MJ-eq] 0.002 0.004 0.006 0.008 0.010 barley wheat ry e pot ato pea 0.00005 0.00015 0. 00025 0. 02 0. 03 y - 0.158x 0.021 r 2 0.022 Colours barley wheat ry e potato pea 0. 01 Colours modular inventory 0.00020 y 0.973x 0 r2 0.939 0.00005 modular inventory 0. 04 ecoinvent 0.01 0. 02 How to establish life cycle inventories of agricultural products? T. Nemecek | © Agroscope Reckenholz-Tänikon Research Station ART 0.03 0.04 56 © Roches & Nemecek ART, 2010 Sensitivity analysis Performed considering the median (=q50%), q10% and q90% of each input (estimated variability of the inputs) POT AT O Q uantiles IMPACTS non-renewable energy [MJ-eq] G WP 100a [k g CO 2-eq] photo. oz one formation [kg ethylene-Eq] nutrient enrichm ent [kg N -eq] Acidification [kg SO2-Eq] Aquatic ecotoxic it y, 100a [kg 1,4-DCB-Eq] Terres trial ec otoxic it y, 100a [kg 1,4-DCB-Eq] Human toxicity, 100a [k g 1, 4-DC B-Eq] Variation: 5 to 10% MachVar Nfer t q10% q90% q10% -1% -1% -1% 0% 0% 0% 0% -1% 7% 5% 11% 0% 3% 0% 0% 7% -11% -28% -5% -64% -47% 0% 0% -4% Variation: 10 to 50% q90% 22% 55% 11% 125% 93% 1% 0% 8% INPUTS Pfert K fert Pestic Irrig at q10% q90% q10% q90% q10% q90% q10% -2% -2% -1% -4% -3% -3% 0% -1% 3% 2% 1% 5% 3% 4% 0% 2% -1% -1% -1% 0% 0% 0% 0% -1% Variation: 50 to 100% How to establish life cycle inventories of agricultural products? T. Nemecek | © Agroscope Reckenholz-Tänikon Research Station ART 4% -2% 7% -27% 3% -1% 4% -9% 3% -2% 6% -15% 0% 0% 0% 0% 1% -1% 2% -3% 0% -76% 288% 0% 0% -99% 377% 0% 3% -43% 165% -11% Drying q90% q10% q90% 62% 21% 34% 1% 6% 0% 0% 25% 0% 0% 0% 0% 0% 0% 0% 0% 0% 0% 0% 0% 0% 0% 0% 0% Variation: > 100% © Roches & Nemecek ART, 2010 57 Potentials of extrapolation Extrapolation cannot replace data collection and the establishment of detailed and specific inventories Very important time saving possible Allows to create generic data sets on global and multinational level Assessment of global variability Fairly good estimates possible for energy demand, global warming and ozone formation, land occupation Difficult for eutrophication and acidification (no site-specific parameters considered) and toxicity (no detailed information on pesticide active ingredients) Can be used as first approximation and where ingredients is not so relevant How to establish life cycle inventories of agricultural products? T. Nemecek | © Agroscope Reckenholz-Tänikon Research Station ART 58 SALCA: An integrated concept for agricultural environmental assessment SALCA = Swiss Agricultural Life Cycle Assessment SALCA consists of the following elements: Database for life cycle inventories for agriculture (in collaboration with ecoinvent) Models for the calculation of direct emissions from field and farm A selection of impact assessment methods (midpoints) Methods for the assessment of impacts on biodiversity and soil quality Calculation tools for agricultural systems (farm, crop) Interpretation scheme for agricultural LCA Communication concept for the environmental management of farms How to establish life cycle inventories of agricultural products? T. Nemecek | © Agroscope Reckenholz-Tänikon Research Station ART 59 SALCA calculation tools Large variability large number of calculations automation required Generic parametrised system modelling for farms and crops: SALCA-farm: generic LCA system for farms SALCA-crop: generic LCA system for arable crops and forage production systems The templates are designed in order to cover all farms/crops All elements, which occur in at least one system must be included Variables are defined, which can describe the different quantities of inputs The variables that are not relevant for a particular system are set to zero Modular structure How to establish life cycle inventories of agricultural products? T. Nemecek | © Agroscope Reckenholz-Tänikon Research Station ART 60 Modular architecture of the tool SALCA-crop V3.1 Produktionsinventar.xls Input data SALCA-field SALCA-Nitrate Calculations SALCA-Field (other direct emissions) Calculations Input data SALCA-nitrate SALCA (TEAM/SimaPro) LCI Calculations SALCA-Erosion Calculations Data transfer by macros Input data SALCA-erosion Input data SALCA-soil quality Input data SALCA (TEAM/SimaPro) SALCA-Heavy metals Calculations SALCA-soil quality Calculations SALCA (TEAM/SimaPro) LCIA Calculation SALCA-biodiversity Data entry Calculations (separate tool) 61 How to establish cycle inventories of agricultural products? Transfer LCIlife data T. Nemecek | © Agroscope Reckenholz-Tänikon Research Station ART 6 separate tools in EXCEL: data entry can be done through the common production inventory or directly in the tool LIFE CYCLE IMPACT ASSESSMENT (LCIA) Input data SALCA heavy metals Production inventory: Common data entry of all parameters for all tools LIFE CYCLE INVENTORY (LCI) Internal Links in EXCEL-sheet Data entry © R. Freiermuth, T. Nemecek, ART 2010 Specific aspects of tropical production systems: relevant LCI aspects Less managed production higher variability more dependent on the environment Labour input instead of machinery how to consider manpower? Use of draught animals how to consider? Reconsider the delimitation between plant and animal production Adaptation of emission models to the conditions of the tropics and subtropics (soil, climate) How to establish life cycle inventories of agricultural products? T. Nemecek | © Agroscope Reckenholz-Tänikon Research Station ART 62 Recommendations for agricultural LCI Large variability many observations needed Collect detailed farm management data Standardised methodology Automated calculation Use of standard LCI formats (EcoSpold, ILCD) Need for a standardised format for agricultural management data Regionalisation, use of GIS Variability and uncertainty should be assessed as standard Infrastructure should be included Development of globally applicable emission models How to establish life cycle inventories of agricultural products? T. Nemecek | © Agroscope Reckenholz-Tänikon Research Station ART 63 Thanks to My colleagues: Gérard Gaillard, Ruth Freiermuth, Martina Alig, Daniel, Baumgartner, Anne Roches, Katharina Plassmann Ecoinvent centre Unilever: Llorenç Milà i Canals, Sarah Sim, Tirma Garcia-Suarez You for your kind attention! How to establish life cycle inventories of agricultural products? T. Nemecek | © Agroscope Reckenholz-Tänikon Research Station ART 64
© Copyright 2024