LAND USE CHANGE and AGRICULTURE Program

Assessment of biomass potentials for bio-fuel feedstock production in Europe:

Methodology and results

Günther Fischer, Eva Hizsnyik, Sylvia Prieler, Harrij van Velthuizen

(fisher@iiasa.ac.at, hizsnyik@iiasa.ac.at, prieler@iiasa.ac.at, velt@iiasa.ac.at)

July 2007 Work Package 2 - Biomass potentials for bio-fuels: sources, magnitudes, land use impacts Deliverable D6: Methodology and assessment of biomass potentials in EU27+ under alternative future scenarios

ii

Table of Contents: List of tables, figures and appendixes iii Preface iv Abstract v Acronyms vi

1. Introduction 1

2. Pan-European land resources database 32.1 Geographic details 32.2 Agro-climatic inventory 42.3 Topography (elevation, slope gradient and aspect) 62.4 Land cover / land use 72.5 Soil database 92.6 Protected and designated areas 9

3. Statistical data 102.2 Administrative units 102.3 Agricultural databases 112.4 EUROSTAT land use statistics 12

4. Land for bio-fuel crops - Methodology 134.1 Model overview 134.2 Food and feed area requirements 154.3 Trends of key variables determining food and feed area requirements 224.3.1 Per capita food demand 224.3.2 Yield increases 224.3.3 Intensity of livestock production 234.3.3 Self-sufficiency in agricultural products 244.4 Increases in built-up land 25

5. Land for bio-fuel feedstock production – Scenarios and results 275.1 Scenario storylines and assumptions 275.2 Bio-fuel feedstocks on cultivated land 325.3 Bio-fuel feedstocks on pasture land 34

6. Assessment of land potentials for bio-fuel feedstock production 366.1 Bio-fuel feedstocks 366.2 Agro-Ecological Zones methodology 386.3 Description of data base 42

7. Bio-fuel feedstock potentials on agricultural land 457.1 Agricultural land for bio-fuel feedstocks in 2030 457.2 Production potentials of bio-fuel feedstocks 477.3 Potentials of 1st and 2nd generation bio-fuel production chains 49

8. Crop residues for bio-fuel production 538.1 Methodology 538.2 Results 54 References 57 Appendixes 59

iii

List of tables, figures and appendixes

page Table 1. Variables included in the Pan-European climatology 4 Table 2. Input variables derived from the MOSUS database used in the model ABioE 15 Table 3. Pasture area in 2000, share required for ruminant feed, average dry matter and

feed energy yields by country

21 Table 4. Per capita food consumption (GK$/cap) in Europe for different dietary sub-

categories in 1985-87 and 2000-2002

22 Table 5. Technical progress in yields in Europe between 1985 and 2002 23 Table 6. Trends in livestock production intensity 24 Table 7. Production volume and trade flows of the EU25 agricultural sector in 2000-02 25 Table 8. Self-sufficiency (= ratio production over consumption) of European countries for

the agricultural sector (average 2000-02)

25 Table 9. Built-up and associated land area (BUILT+) increases between 2000 and 2020 26 Table 10. Assumptions driving scenario calculations 29 Table 11. Organic farming area in 2005 and assumptions for 2030 for the ‘low’ scenario 31 Table 12. Europe’s use of cultivated land in 2030 for ‘baseline’ scenario and available area

for bioenergy crops for a ‘low’ and ‘high’ scenario

32 Table 13. Cultivated land use in 2030 (Baseline scenario) 33 Table 14. Pasture land use in 2000-02 and scenario results for 2030 (‘baseline’ scenario) 35 Table 15. Conversion factors from biomass to bio-fuel energy equivalent 42 Table 16. Potential oil crop production in Austria (Example of database) 43 Table 17. Database on Europe’s bio-fuel production potential - Description of tabulated

results

44 Table 18. Land availability in Europe for bio-fuel feedstocks in 2030 46 Table 19. Biomass potentials of bio-fuel feedstocks for selected European countries 48 Table 20. Potential bio-fuel energy production in Europe by 2030 for different scenarios 52 Table 21. Conversion factors used to estimate crop residues for selected major crops 54 Table 22. Agricultural residues estimates potentially available for bio-fuel production 55 Table 23. Agricultural residues of food and feed crops potentially available for bio-fuel use

and associated heating values

56 Figure 1. Data sources of harmonized land use map 7 Figure 2. Major land use categories of the 1x1 km Pan-European land use database 8 Figure 3. Administrative units division for the Pan-European territory (see also Appendix I) 10 Figure 4. Flow chart for food and feed area requirements calculation procedures 16 Figure 5. Flow chart of livestock energy balances for feed requirements calculations 18 Figure 6. Aggregated crop yields (expressed in GK$ per hectare of cultivated land) in

European countries for the base period 2000-02 and assumptions for three scenarios for 2030

31

Figure 7. Regional distribution of cultivated land potentially available for bio-fuel feedstock production in Europe by 2030 (‘baseline’ scenario)

32

Figure 8. Pasture land use in 2030 (‘baseline’ scenario) 35 Figure 9. Agro-Ecological Zones (AEZ) methodology for assessment of bio-fuel feedstock

potentials 39

Figure 10. Potential land use in 2030 for European countries (baseline scenario) 46 Figure 11. Suitability of agricultural land for bio-fuel feedstocks 49 Figure 12. Average potential bio-fuel energy yields of 1st and 2nd generation bio-fuel

feedstocks 51

Appendix 1 Harmonization of a Pan-European land use database – Reclassification of

original land use maps to 12 aggregate land categories LUCL12 59

Appendix 2 Administrative units used in the REFUEL project 61 Appendix 3 Compilation of a new terrain slope database based on SRTM data 64 Appendix 4 Biomass and energy potentials for bio-fuel feedstocks in Europe 66 Appendix 5 Conversion factors for calculation of crop residues 75

iv

Preface This study is part of the REFUEL (www.refuel.eu) project sponsored by the European Commission under the Intelligent Energy – Europe programme. It is designed to encourage a greater market penetration of bio-fuels. To help achieve this goal, a bio-fuels road map is developed that is consistent with EU bio-fuel policies and supported by stakeholders in the bio-fuels field. The project involves seven European partner institutions and runs for two years until begin of 2008.

The land use scenarios and the bio-fuel feedstock potentials presented in this study are closely interlinked with research of other partners in the REFUEL consortium.

v

Abstract The role of bio-fuels has been strongly enhanced by their consideration in the climate change debate, as well as opportunities for rural development and improved energy security. Europe’s bio-fuel production potential has been determined for various land use scenarios up to 2030 using a detailed Pan-European resource database and a spatially explicit feedstock suitability and productivity modelling framework. Both arable land and grassland have been considered for a broad range of bio-fuel feedstock production options.

Europe’s agricultural land today comprises 133 million hectares of cultivated land and 74 million hectares of permanent grassland (EU27 plus Norway, Switzerland, and Ukraine). Future available land for bio-fuel production was estimated while satisfying projected food and feed demand at current aggregate European self-reliance levels for agricultural products. In this land resource based approach, the future land area requirements are estimated for Europe’s agricultural sector. Current use levels of agricultural products and associated land area requirements including trade were determined by country.

Projections of yields and livestock production intensity improvements were used to quantify land that would be freed up for other uses. By 2030, a maximum of 65 million hectares of cultivated land could be used for bio-fuel crop production (‘baseline’ scenario). This land is roughly distributed for one third each in EU15, EU12 and Ukraine. About 24 million hectares of pasture land could become available for principally herbaceous lignocellulosic biomass production. This would set aside some 13 million hectares of grassland for nature conservation and reasons of steep sloping conditions.

IIASA’s agro-ecological zones modelling framework has been updated and expanded for bio-fuel productivity assessments distinguishing five main groups of bio-fuel feedstocks with specific energy production pathways, namely: woody plants, herbaceous plants, oil crops, sugar crops and cereals. A Pan-European land resources database has been compiled at the spatial resolution of 1 km2 enabling results to be aggregated by NUTS2 administrative units. Potential biomass productivity of individual bio-fuel feedstocks and associated energy yields have been calculated for each grid cell and results were tabulated by aggregate land cover classes compiled from CORINE.

For the available land extents, bio-fuel feedstock production potentials have been estimated using two distinct energy conversion scenarios. For the EU27 the analysis shows that after meeting food and feed requirements, if all the remaining agricultural land would be used for bio-fuel feedstocks, by 2030 between 20% to 50% of the projected fuel requirements of the transport sector could be covered by respectively 1st generation conversion (about 4 EJ in the case of ‘baseline’) and 2nd generation conversion (about 8 EJ in the case of the ‘high’ scenario).

Agricultural residues of food and feed crops may provide an additional source of bio-fuel feedstocks. Assuming that up to 50% of crop residues could be removed from the field without significant impacts on soil fertility and erosion, a total of 246 million tons dry matter agricultural residues were generated in Europe in 2000-02, equivalent to about 15% of current transport fuel consumption and about 10% of transport fuels consumed in 2030.

Keywords: bio-fuel feedstock potential, land use scenarios, land resources, agriculture, crop residues

vi

Acronyms ABioE ‘Area available for growing Biomass feedstock for Energy production’ – A

methodology developed to estimate future land areas potentially available for biofuel feedstock production.

AEZ Agro-Ecological Zones BUILT+ Areas occupied for residential, commercial, industrial, and infrastructure

purposes including both built-up land and associated vegetated areas CAP Common Agricultural Policy DEM Digital Elevation Model GIS Geographic Information System GK$ Geary-Khamis Dollars - An international price weight for agricultural

products compiled by the Food and Agriculture Organization IIASA-LUC International Institute for Applied Systems Analysis – Land Use Change

and Agriculture Program LHV Lower Heating Value LUT Land Utilization Type MOSUS “Modeling Opportunities and limits for restructuring Europe towards

Sustainability” (MOSUS) (see www.mosus.net). MOSUS is an R&D project funded under the EU 5th framework programme.

REFUEL A research project sponsored by the European Commission under the Intelligent Energy - Europe program (see www.refuel.eu)

RPR Crop residue to crop main produce ratio SRTM Shuttle Radar Topography Mission SSR Self-sufficiency ratio Energy units: GJ Gigajoules (109 joule) PJ Petajoules (1015 joule) EJ Exajoules (1018 joule) Regional aggregates: EU European Union EU15 EU15 countries of the EU EU15+ EU15 countries plus Norway and Switzerland EU10 Aggregate of ten New Member States joining the EU in 2004 EU12 Aggregate of EU New Member States as of 2007; comprises EU10 and

Bulgaria, Romania EU25 EU15 + EU10 EU27 EU15 + EU12 EU27+ EU27 plus Norway and Switzerland

1

Chapter 1. Introduction

Energy demand for transport, which is currently almost entirely fuelled by petroleum, will continue to experience high growth in the decades ahead. In the EU the transport sector is responsible for around 21% of anthropogenic greenhouse gas (GHG) emissions. In order to curb a fast growing GHG emission profile bio-fuels are considered as a key solution together with fuel saving vehicle technologies. Besides GHG emissions reduction, bio-fuels offer improved energy security through diversification of the fuel mix and economic development prospects in rural regions including job creation.

A recent IEA study on “Energy Technology Perspectives” concludes that bio-fuels may become a key transport fuel option. Through fuel blending and increased use of flexible fuel vehicles, bio-fuels can use the existing fuel distribution infrastructure with only minor adjustments required (IEA, 2006). In the European Union a Bio-fuels directive (CEC ,2003) has set an indicative target of 5.75% bio-fuels market share in 2010. Moreover a biomass action plan (CEC, 2005) and the “EU strategy for Bio-fuels” (CEC, 2006) calls for further promotion of and preparation for large-scale use of bio-fuels. To this end EU member states have began to implement various measures to promote the use of bio-fuels1.

The Common Agricultural Policy (CAP) includes certain schemes to foster energy crop production. In September 2006 the EC proposed to extend the energy crop premium introduced by the 2003 Common Agricultural Policy reform to the eight Member States which currently do not benefit from it. In a further push to encourage the production of feedstocks for renewable energy production, the Commission also proposed allowing the Member States to grant national aid of up to 50 percent of the costs of establishing multiannual crops on areas on which an application for the energy crop aid has been made (CEC, 2006b).

Liquid bio-fuels production costs currently are still high – up to three times the cost of petroleum fuels2 (IEA, 2006b). In most countries embarking on bio-fuels initiatives, the recognition of non-market benefits is often the driving force behind efforts to increase their use. By far the largest production and use is of ethanol in the United States and Brazil. In Europe the focus so far has been on biodiesel production from rape and sunflower but total production is fairly small compared to ethanol production in the US and Brazil (IEA, 2006b). Despite of significant growth rates of biodiesel production3 in the past few years the 2% market share target by 2005 as foreseen in the bio-fuels directive was not met.

Widespread use of bio-fuels will require expanding the range of feedstocks and the introduction of advanced conversion technologies such as Fischer-Tropsch synthesis and ethanol production from lingocellulosic feedstocks. Full development of the bio-fuel option requires also a thorough analysis of possible unintended consequences of a major shift in land use (IEA, 2006).

This study takes a European perspective in assessing Europe’s bio-fuel energy production potential. A certain share of bio-fuels will likely be imported in the future due to lower production costs in developing countries with adequate climates and suitable land resources. While costs and GHG emissions reduction are usually favourable in developing countries other benefits such as rural development and energy security can not be realized by importing bio-fuels.

1 For more information see: http://ec.europa.eu/energy/res/legislation/bio-fuels_members_states_en.htm 2 The exception is Brazil where in recent years the retail price (excluding taxes) of hydrous ethanol (used in dedicated ethanol vehicles) has dropped below the price of gasoline on a volumetric basis (IEA, 2006b). 3 Biodiesel production in Europe reached some 6 million tonnes in 2006 compared to only 2 million tonnes in 2004 (European Biodiesel Board, 2007).

2

Moreover potential detrimental environmental concerns are often raised when bio-fuels are produced in developing countries with often lower environmental standards compared to Europe and risks of triggering deforestation and biodiversity loss.

The main determinants of domestic bio-fuel energy potential relate to land availability, yields of bio-fuel feedstocks, and conversion technologies applied. The focus of this study is on the former two while for conversions technologies results rely on published sources of energy output per unit of feedstock quantity. New technologies in particular the use of lingocellulosic biomass feedstocks is currently one of the cutting edge areas of energy technology research.

Land use competition with food and feed production is considered a potential key barrier to exploiting Europe’s bioenergy production potential. This study aims to assess available land for bio-fuel production for different scenarios in view of satisfying projected food and feed demand at current aggregate European self-reliance levels for agricultural products. A stable pattern of self-reliance implies no changes in trade patterns over time. In this way it is possible to assess Europe’s domestic potential for cultivating energy crops while at the same time sufficient land resources remain for the food and feed sector. As in the past, it is likely that Europe’s agricultural production intensity including yields will further increase. As a result some land will be freed up for cultivating alternative agricultural products. In addition some agricultural land will be lost for urbanization and infrastructure purposes. Scenarios focus on the period 2000 to 2030 and include the territory of the EU27, Norway, Switzerland and Ukraine.

Potential productivity of bio-fuel feedstocks varies with biophysical conditions and management regimes. A spatially explicit feedstock suitability and productivity assessment for a wide range of land utilization types, including feedstocks for first and second generation bio-fuels, provides a regional differentiation of Europe’s bio-fuel production potential. For this purpose the existing agro-ecological zones (AEZ) modelling framework has been updated and expanded for feedstock productivity assessments and a Pan-European natural resources data base was created for a grid cell size of 1 by 1 km.

Chapter 2 describes the Pan-European natural resources database, which was implemented in a Geographic Information Systems (GIS) environment. Chapter 3 introduces the administrative units applied in the study and related databases required for the bio-fuel land availability scenarios. This includes EUROSTAT statistical data and a database created in the MOSUS4 project. The latter includes a detailed account of a country’s agricultural land area requirements separately for food and feed purposes. Chapter 4 describes the methodology applied to estimate future land availability for bioenergy crop production including an analysis of the historic development of key variables relevant for the scenarios. Storylines and assumptions for a ‘baseline’, ‘low’ and ‘high’ bio-fuel energy scenario are outlined in Chapter 5 together with result for cultivated and pasture land. The focus of Chapter 6 is on bio-fuel feedstock types included and the AEZ modelling framework. In Chapter 7 we present spatially explicit results for potential energy from cultivating various bio-fuel feedstocks. Moreover results for combined land use and feedstock type scenarios of Europe’s potential for bio-fuel energy are presented. Chapter 8 describes methodology and results of crop residue potential from food and feed crops for bio-fuel production.

4 “Modeling Opportunities and limits for restructuring Europe towards Sustainability” (MOSUS) – For more information see www.mosus.net. MOSUS is an EU projected funded in the 5th framework program.

3

Chapter 2. Pan-European land resources database

2.1 Geographical details

A Pan-European natural resources database has been compiled for the assessment of land capabilities, productivity and constraints using most recent available geographically explicit data. The different thematic layers in the Geographic Information System (GIS) are now available in an equal area projection (Lambert Azimuth). The projection defined in the CORINE database has been chosen as the common projection for all layers in the REFUEL GIS database. It is an ETRS Lambert Azimutal equal area projection with the following specifications:

Datum D_ETRS_1989 Ellipsoid GRS_1980 Semi-major axis 6378137 Axis units Degrees Flattening ratio 0.00335281068118232 Projection Lambert_Azimutal_Equal_Area False easting 4321000 meters False northing 3210000 meters Central median 10 degrees Latitude or origin 52 degrees

All original maps were projected and converted to a grid cell size of 1x1km for a Pan-European territory extending from longitude -15 to 45 and latitude 35 to 71. Grid maps are generally stored in an ArcGIS environment with the following geographic specifications of the window: Cell Size: 1000 m Number of Rows: 4090 Number of Columns: 3910 Xmin = 2610000 Xmax = 6520000 Ymin = 1350000 Ymax = 5440000

Elevation5 and land cover data sets are also available at a resolution of 100 x 100 m grid cell size using the above geographic window and thus including 40900 rows and 39100 columns.

5 Such high-resolution elevation data is available for territories below 60 degrees north latitude. Above 60 degrees a 1x1 km grid cell raster exists.

4

2.2 Agro-climatic inventory

An agro-climatic inventory has been compiled using the gridded climate parameters available from East Anglia University (CRU Global climatologies) and the VASCLimO global precipitation data from the Global Precipitation Climatology Centre (GPCC). The following two data sets from the CRU global database were selected:

a) Average 1961-90 monthly variables for a 10 x 10 minutes latitude/longitude grid (CRU CL2.0, New et. al 2002)

b) Annual time series for a 0.5° by 0.5° latitude/longitude for monthly climatic variables (CRU TS 2.1, Mitchell and Jones 2005)

For the annual time series precipitation data the VASCLimO data (Beck et. al, 2004) were used.

Original 10 by 10 minutes average climate and 0.5° by 0.5° latitude/longitude time series climatic surfaces were projected and interpolated to a 1 by 1 km grid for the average climatic data and all years between 1961 and 2002. The procedure comprises of two steps. First using a bilinear6 interpolation method within the ArcGIS environment original cell sizes were interpolated to a higher resolution 30 arcsecond grid. Second the resulting grid was projected to the above specified Lambert projection applying a grid cell size of 1 by 1 km. In this way potential inconsistencies along coastlines were kept small.

In the case of temperature a lapse rate of 0.55°C per 100 meter elevation was applied using the respective digital elevation data (DEM). First, a grid of altitude provided by CRU was used to calculate temperature values adjusted to sea level. Bilinear interpolation was performed for temperatures at sea level. Second, a 1 by 1 km DEM, derived from Shuttle Radar Topography Mission (SRTM) data, was used to calculate temperatures for elevations. The 1 km DEM was compiled from SRTM original 3 arc-sec elevations using the median of all 3 arc-second elevation data within each 1 km grid cell.

Table 1 lists the climatic variables included and their respective GIS grid names applied in the 1 by 1 km grids. Table 1. Variables included in the Pan-European climatology Variable Name of GRID units Precipitation PRE mm / month * 10 Vapor pressure1 VAP hecto-Pascals * 10 Relative humidity2 REH percent * 10 Number of wet days WET (no days per month with >0.1mm rain) * 100 Sunshine fraction2 SFR percent of maximum possible * 10 Cloudiness1 CLD percent * 10 10m Wind speed WND m/s * 10 Mean monthly temperature TMP deg C * 100 Diurnal temperature range DTR deg C * 100

1 Variable used in CRU TS 2.1 (time series data); 2 used in CRU CL2.0 (1961-90 average data)

6 A bilinear interpolation uses the value of the four nearest input cell centers to determine the value of the output raster. The new value for the output cell is a weighted average of these four values, adjusted to account for their distance from the center of the output cell.

5

The climatic variables are key inputs to the agro-climatic inventory for the European Agro-Ecological Zones (AEZ) model implementation. The AEZ approach considers detailed thermal and moisture regimes which form the basis of the inventory.

Thermal regimes include a classification of thermal climates, annual temperature profiles, and temperature growing periods and temperature sums (the number of days when mean daily temperature exceeds a given threshold Deg.C).

The AEZ methodology includes daily soil water balances and estimates actual evapotranspiration (ETa) of specific crops. Calculation procedures follow in principal those described in “CROPWAT” (FAO 1986, 1992) and “Crop Evapotranspiration” (FAO, 1998). ETa is calculated from reference evapotranspiration (ET0), a crop coefficient kC and a water stress coefficient kS. ET0 is calculated according to Penman-Monteith (Monteith, 1965, 1981; FAO, 1992). The moisture regime is calculated from monthly climate parameters, converted to daily data by means of spline interpolations, ensuring consistency of daily levels with monthly means or totals.

The moisture regime determines length of growing period (LGP), i.e. the period during the year when temperature and moisture is conducive to crop growth. For Europe this has been defined as the number of days when temperature is above 5 degree Celsius and ETa is at least 0.4 times ET0.

As a result various derived agro-climatic parameters are available for each 1 by 1 km grid-cell and maps of each item can be produced for spatial verification. They include temperature profile7; begins and ends of thermal growing periods; accumulated temperature during thermal growing periods; mean temperature during thermal growing periods; aridity index (precipitation over reference evapotranspiration); P/ET0 ratio during growing period; total number of growing period days; number of growing period days (days when soil moisture availability can fully meet reference crop water requirements); total number of wet days, i.e., growing period days with excess moisture; begin and end of dormancy period days; begin and end dates of LGP; temperature profile of growing period; accumulated temperatures (above 0◦C, 5◦C, 10◦C) during growing period; average temperature of growing period; multiple-cropping zones classification for rain-fed and irrigated conditions.

7 number of days in intervals of 5◦C-steps from < –5◦C to > 30◦C separately for periods of increasing and decreasing temperatures; thermal growing periods: number of days > 0◦C, > 5◦C, > 10◦C

6

2.3 Topography (elevation, slope gradient and aspect)

The NASA Shuttle Radar Topographic Mission (SRTM) has provided digital elevation data (DEMs) for over 80% of the globe. The SRTM data is publicly available as 3 arc second (approximately 90 meters resolution at the equator) DEMs (CGIAR-CSI, 2006). Original SRTM tiles covering the European continent were imported into ArcGIS, merged8 and projected to a 100 by 100 meters grid cell size.

From this high resolution grid a 1 by 1 km digital elevation model (DEM) was constructed by calculating the median of all 100 grid cells within each 1 km grid cell. SRTM data only extend to 60 degrees north latitude. Therefore for most of Scandinavia, which reaches until over 70 degrees north, elevation data from GTOPO309 (USGS, 2002) was merged to create a European wide 1 by 1 km DEM.

The high resolution SRTM data of 3 arc second have been used for calculating: (i) terrain slope gradients for each 3 arc-sec grid cell; (ii) aspect of terrain slopes for each 3 arc-sec grid cell; (iii) terrain slope class by 3 arc-sec grid cell; and (iv) aspect class of terrain slope by 3 arc-sec grid cell. Products (iii) and (iv) were then aggregated to provide distributions of slope gradient and slope aspect classes by 1 by 1 km grid cell.

Slope gradient and slope aspect were calculated for each 3 arc second grid-cell of the SRTM DEM following procedures used in the Agro-Ecological Zones (AEZ) methodology and described in Appendix 3. As a result distributions of nine slope gradient classes are available for each grid-cell: 0-0.5%, 0.5-2%, 2-5%, 5-8%, 8-16%, 16-30%, 30-45%, and > 45%.

Slope aspect describes the slope direction10. For each 3 arc second grid-cell the respective slope aspect is stored in a five class representation including: One class for slopes below 2%, i.e. indicating no aspect and four classes for slopes facing east (45 to 135 degrees), north, west and south direction.

Original SRTM derived tiles with the above classes showing slopes and aspects were merged and projected to the European Lambert projection with a resulting grid-cell size of 100 by 100 m.

8 For computer space reasons the original 13 tiles covering the European continent were merged into an eastern and western half separately and each projected to the Lambert projection. Finally the projected grids were merged resulting in a large 100 by 100 meters DEM with 40900 rows and 39100 columns. 9 The resolution of GTOPO30 is 30 arc-seconds, i.e. depending on latitude this is approximately a 1 by 1 km cell size. 10 Slope aspect identifies the down slope direction of the maximum rate of change in value from each cell to its neighbors.

7

2.4 Land cover / land use

Three available land cover databases have been reclassified to twelve major land use classes listed in Box 1 for the purpose of determining spatial locations of arable land, grassland, forest and other areas. In this way a harmonized land use map was constructed for the Pan-European territory based on the following sources (Figure 1):

a) the latest CORINE version (CLC 2000) for EU 25 (grid cell size 100 by 100 meters)

(EEA, 2006) b) CORINE Version 1990 (CLC1990) for Switzerland c) JRC’s Global Land Cover for Europe (GLC2000) for the remaining countries

(JRC, 2006)

Details of the reclassification are listed in Appendix 1. Figure 2 shows the resulting Pan-European land use map featuring the twelve major land use classes.

Figure 1. Data sources of harmonized land use map Adjusting spatial databases with statistical information In addition land cover data have been scaled with available sub-national statistics from EUROSTAT. As a result a land share database is available describing for each 1 by 1 km grid cell the percentage distribution of four different land use classes; a) arable, b) grassland, c) forest, and d) other areas. The area sum of each of the first three land use classes over an administrative unit corresponds with the statistical information.

Box 1. Twelve land use categories included in the harmonized Pan-European land use map 1 Forests 2 Natural grassland 3 Wetlands 4 Other natural 5* Arable land 6* Permanent crops 7* Heterogeneous Agriculture 8* Pastures 9 Water, Coastal 10 Bare, sparse vegetation 11 Glaciers, snow 12 Urban & industry * considered for potentially growing bio-fuel crops

8

Figure 2. Major land use categories of the 1x1 km Pan-European land use database

9

2.5 Soil database The soil data are based on the European Soil Database (ESDB) (ESB, 2004). ESDB comprises a polygon vector map at a scale of 1:1 million and the ESDB soil attribute database. The original polygon map was projected and converted to a 1 by 1 km grid-cell size. The soil attribute database has been expanded with additional soil parameters from the ISRIC/FAO/IIASA WISE database (Batjes et.al. 1997).

2.6 Protected and designated areas Protected areas reflect an interpretation of the IUCN-WCMC protected areas inventory11 (at 30-arc seconds separates protected land where cultivation is permitted from areas where cultivation is strictly prohibited. The categories are derived from international and national conventions. These include legally protected areas from World Heritage Convention, Ramsar Wetland Convention, Biogenetic Reserves, European Diploma Type A, Bird Directive and the IUCN Classes I-VI.

11 World Database on Protected Areas (WDPA) – for more information see: http://www.unep-wcmc.org/wdpa/

10

Chapter 3. Statistical data

3.1 Administrative units

In addition to the natural resources data base an administrative layer map has been compiled for the Pan-European territory. Administrative levels are defined as a hierarchic structure down to the NUTS 2 level in the case of EU27+ (EU27 and Switzerland, Norway). For the Ukraine the administrative map includes the oblast level, while for the remaining European countries only national boundaries are available.

Three original administrative maps 12 were projected and geographically merged in a GIS environment. For all resulting 293 administrative units (Figure 3) in addition to the code of the original data source a unique number was assigned to each unit. It provides the link to the GIS map and was defined n such a way that a division of each number by 100 results in the country code used in the FAOSTAT database. Appendix 2 lists all administrative units including the unique number, the area and the original code (mostly one used in the EUROSTAT databases). Through conversion of the original polygon map to the 1x1 km grid cell size map it is possible to aggregate specific information on natural resources for each administrative unit.

Figure 3. Administrative units division for the Pan-European territory (see also Appendix I)

12 1.) EU’s NUTS 3 level map for the EU27+ territory (Source: EUROSTAT); 2.) Ukraine oblast map (Source: ESRI); 3.) National boundaries for remaining countries (Source: FAO global country map)

11

3.2 Agricultural databases In the context of the EU sponsored R&D project MOSUS 13, IIASA-LUC has created a detailed agricultural data set with global coverage:

“Agricultural and forestry products trade balance database including production volumes and land use – a country-specific database from 1980 to 2002”.

The database provides a detailed account of produced and traded agricultural and forestry products by individual countries. Agricultural products include crops, livestock and fisheries, primary as well as processed and derived products. The commodity lists follow FAO’s supply utilization accounts (SUA) and include more than 200 different items. Coefficients and land areas are estimated at detailed commodity level, which in turn are aggregated to selected sub-sectors. For agricultural the following sub-sectors were used:

1) CROPS and crops products 1a) Cereals 1b) Other crops 1c) Fodder crops 2) LIVESTOCK products14 2a) from Ruminants 2b) from Other livestock (mainly pigs and poultry) 3) FISHERIES

Land area requirements were estimated by applying country specific yields to domestic production, imports and exports of individual commodities. For the crops and livestock sector, a detailed and comprehensive data base including multi-cropping index, yields, various use categories, and feed sources for livestock was compiled allowing consistently to deal with production and trade of jointly produced (e.g., flour and bran of wheat) as well as various processed agricultural commodities.

For aggregation the land area associated with crops and livestock products (in ha) can easily be added. However in the case of aggregate physical quantities (tons) of such different commodities like for example sunflower seed, apples, tobacco or cotton in the case of the ‘Other crops’ sub-category an weighted aggregation is required. International price weights of the year 2000 were applied, the so-called Geary-Khamis prices compiled by FAO, to aggregate physical volumes of production, trade and various use categories (e.g. food, feed, waste). Original units of physical production volumes (tons) were multiplied with Geary-Khamis prices and therefore converted into a new unit, the physical production volume in Geary-Khamis dollars equivalent (henceforth referred to as ‘Geary-Khamis production volumes’, short GK$).

In the livestock sector, ruminants (cattle, sheep, goats and horses) have been treated separately from other animals (mainly pigs and poultry). Ruminants rely on pastures, green fodder as well as feed produced on cultivated land while the sub-group other animals rely on the latter only. Attributing land associated with the production of feed crops and by-products from primary crops used in feeding (e.g. brans or soybean cake) was done according to usability of feed sources for different animal types and estimated in proportion to livestock energy requirements. By comparing energy supply from reported feed use and pastures with livestock energy requirements it was possible to allocate feed use separately to ruminants and other animals. 13 “Modeling Opportunities And Limits For Restructuring Europe Towards Sustainability” (MOSUS) – www.mosus.net 14 Trade of livestock animals is included.

12

Pasture land area requirements for ruminants are estimated by converting areas of permanent pasture into a globally comparable unit, the ‘reference pasture unit’. It is calculated by comparing potential grassland productivity data, derived from AEZ methodology (Fischer et.al, 2002), to a reference yield of five tons dry matter per hectare per year.

Annual results are available for the period 1980 to 2002 for the geographic units used in MOSUS, i.e., 51 countries and two country aggregates, namely OPEC (excluding Indonesia, which is included as a separate country) and ‘Rest of the world’. For each country production, trade, and domestic use are recorded together with estimates in Geary-Khamis dollars equivalent (in GK$) land area (in ha) in each component.

The methodology for land appropriation is described in detail in the following publications: “Land associated with production and trade of crops and crop products” (Fischer, et. al. 2006); Land associated with production and trade of livestock products” (Fischer, et. al. 2006); Land associated with production and trade of forestry products” (Prieler, et. al. 2006).

A full accounting of production and trade of agricultural products entails domestic consumption calculated as production plus imports minus exports. Thus each country’s area requirements for its food and livestock products consumption are recorded in the database. These data provide the base information for estimating future food and feed area requirements (see section 4.2).

3.3 EUROSTAT land use statistics

This study relies on EUROSTAT data (www.eurostat.com) for national and sub-national land use statistics, which includes extents of arable land, land under permanent crops and permanent grassland as well as fallow and set-aside land. Agricultural area statistics were assembled for national and NUTS2 level administrative units and merged with the REFUEL administrative codes stored in the GIS database.

Baseline population data and population projections until 2030 were obtained from EUROSTAT as well.

13

Chapter 4. Land for bio-fuel crops - Methodology

4.1 Model overview

IIASA-LUC has developed a methodology, henceforth termed ‘ABioE’ (Area available for growing Biomass feedstock for Energy production), to estimate future land areas potentially available for bio-fuel feedstock production. The methodology builds to a large extent on data created in the context of the MOSUS project. Input data and calculation steps are organized in an MS Excel environment and allow accounting for a scenario approach. Data and calculations are generally performed on a national level for EU27 plus Switzerland, Norway and the Ukraine. In this way most of the Pan-European territory is covered including the land-rich Ukraine.

ABioE first estimates future land area requirements for Europe’s food and livestock sector with the remainder becoming available for bio-fuel crops, grasses and fast-growing trees. Some agricultural land will be lost due to land use change by conversion to built-up and associated land areas. Therefore current (defined as the average of period 2000 to 2002) agricultural area minus the area lost for conversion to built-up and minus future land area requirements for food and feed production define agricultural land area potentially available to grow bio-fuel energy feedstocks. In a general form equation (1) applies.

(1) ctctcct AFoodFeedABuiltUpAAgricABioFuel ,,,022000, −−= − [ha]

ctABioFuel , Available land for bio-fuel feedstock production in future year t of country c [ha]

cAAgric ,022000− Agricultural land in country c in base period 2000-02 [ha]

ctABuiltUp , Increases in built-up and associated land areas between base period 2000-02 and future year t in country c [ha]

ctAFoodFeed , Agricultural land area requirements in country c for domestically produced food and feed (the SSR15-fraction of domestic consumption) [ha]

t Future year t c Country c

Both arable land and grassland are considered as potential areas for bio-fuel feedstock production. Whereas arable land can be used for all types of bio-fuel feedstock, grassland should only become available for producing herbaceous lignocellulosic feedstocks under zero-tillage systems and thus respecting environmental and greenhouse gas concerns. A distinct treatment of cultivated land and permanent grassland implies for feed area calculations differentiating among livestock types: Firstly, ruminants feeding on both grassland and arable feed crops, and secondly other livestock (mainly pigs and poultry), which relies on feed crops only.

Thus requirements of future food and feed areas are calculated separately for cultivated land (defined here as the sum of arable and permanent crops) and permanent grassland. Built-up conversion is assumed to occur at similar rates on cultivated and grassland as outlined in equations 2 and 3.

In the case of pasture land, besides changes in feed area requirements, land use and slope information from the Geographic Information Systems database is integrated to identify potential areas for harvesting bio-fuel crops.

15 SSR refers to self-sufficiency ratio; i.e. domestic production over domestic consumption

14

Population projections and projections for increases in built-up and associated land areas are exogenous to the model and may therefore be changed according to scenario assumption.

Several technical coefficients can be changed to vary assumptions across different scenarios. They include food use per capita, crop yields increases, and livestock production intensity16. In addition the model allows evaluating the impact of varying self-reliance ratios. Thus we may for example assess how much additional land could become available presuming higher imports of feed crops from outside Europe.

Data compilation and calculations of land appropriation for food and feed for the historic period 1985 to 2002 were done in the MOSUS project. It included a comprehensive description of major agricultural sub-sectors’ production and trade for individual countries. As described above (section 1.3), a key element is the aggregation of physical production volumes to a comparable unit (of Geary-Khamis dollars, GK$). Hence all physical units are recorded in GK$ for different sub-categories. Three crop categories are distinguished: 1a) Cereals; 1b) ‘Other crops’; and 1c) Fodder crops. They comprise food and feed items for both primary and processed commodities. Second there are two categories of livestock products: 2a) from ruminant animals and 2b) from ‘Other livestock’. Commodities include all meat, dairy products and eggs. Only ruminant animals rely on two distinctly different land use types. On the one hand cultivated land for growing feed or fodder crops, on the other hand they graze on permanent grassland.

(2) CULTIVct

CULTIVc

BUILTct

CULTIVc

CULTIVct AFoodFeedAAABioF ,,,022000, * −−= +

− α [ha]

(3) PASTUREct

PASTUREc

BUILTct

PASTUREc

PASTUREct AFeedAAABioF ,,,022000, * −−= +

− α [ha]

whereas: 1=+ PASTUREc

CULTIVc αα

CULTIV

ctABioF ,1+ Available cultivated land to grow bio-fuel crops [ha] CULTIV

cA ,022000− Cultivated area of country c in base period 2000-02 [ha] PASTURE

cA ,022000− Permanent grassland area in baseyear period 2000-02 of country c [ha] +BUILT

ctA , Increases in built-up and associated land areas between baseyear and future year t [ha] CULTIVcα Share of cultivated land in total agricultural area in country c [%] PASTUREcα Share of permanent grassland in total agricultural area [%]

CULTIVctAFoodFeed , Cultivated land area requirements in country c for domestically produced food

and feed (the SSR-fraction of domestic consumption) [ha]

PASTUREctAFeed , Permanent grassland area requirements in country n for domestic ruminant

livestock production [ha]

j Future year t c Country c

16 Livestock production intensity is expressed as required energy input per unit of livestock output.

15

4.2 Food & feed area requirements

Key factors for changes in food and livestock sector land area requirements include developments in:

a) Demand, more specifically population and dietary changes; b) Production intensity (crop yields and intensity in livestock production); c) Agricultural trade balance between Europe and the rest of the world.

A key module of ABioE is the estimation of land area requirements for each country’s future food consumption. Current levels of agricultural production and consumption and associated land area requirements, including for trade, are available for each country for the period 1985 to 2002 from the MOSUS land area requirements database (see above section 3.2). The database includes a detailed accounting of area requirements for food and feed production, both for domestic production and consumption17. Thus for a base period (here defined as the average period 2000-2002) food and feed area requirements are known. Moreover the database provides several historic parameters such as per capita demand, yield developments and self-sufficiency ratios, which form the basis for estimating coefficients for future developments. Table 2 summarizes input variables from the MOSUS database used in the calculations for future food and feed area requirements. Table 2. Input variables derived from the MOSUS database used in the model ABioE Variable Variable subcategory* units A. Physical quantities Production volume Cer, Oth.Cr; - RUM, OL mio.GK$ Food use per capita Cer, Oth.Cr; - RUM, OL GK$/p. Domestic USE volumes 1. for domestic food consumption

1a. Crops (Cer, Oth.Cr. FoCr) 1b. Livestock products (from RUM, from OL) 2. for feed to produce domestic livestock

mio.GK$

B. Land area Land in production 1. Cultivated land (Cer, Oth.Cr. FoCr)

2. Permanent pasture hectares

C. Other Population Population numbers Self-reliance Ratio 1. Crops

2. Livestock products from 2a. RUM, 2b. OL % of Use

Crop yield Production volume / Land ratio for CROPS GK$/ha D. Livestock feed energy balances

Energy supply for livestock from FODDER crops separately for RUM, OL from SUA feed items separately for RUM, OL from PASTURES for Ruminants

Gcal

Energy Grassland share for ruminants (GSHR)

Share of feed energy for ruminants supplied from grazing on permanent grassland.

%

* Cer = Cereals; Oth.Cr = Other Crops; FoCr = Fodder crops; RUM = ruminants; OL = other livestock

17 A country’s consumption is calculated as domestic production plus imported commodities (primary and processed) minus exported commodities. Consumption can be described in terms of agricultural products volumes or associated land areas.

16

Figure 4 provides an overview of the food and feed area calculations. Food demand (or domestic use) described as a function of population number and per capita food consumption levels is first converted to domestic production levels using self-sufficiency ratios. Vegetarian food can directly be related to cultivated land area requirements via crop yields. For domestic livestock production feed requirements are calculated with the help of livestock energy balances. Because of different types of land area requirements a distinction of two livestock animal groups is essential. Ruminants relying on both feed crops from cultivated land and feed from grazing on pasture land. In contrast other livestock, primarily pigs and poultry, are only raised with feed crops.

Figure 4. Flow chart for food and feed area requirements calculation procedures

Populationnumber

Food consumptionfood use / capita

fromCereals

from Other crops

Vegetarian food

fromOther LVST

from Ruminants

Livestock products

SSR SSR

Dom.PRODOther LVST

Dom.PRODRuminants

FEED cropsrequirements

Feed fromPASTURE

Livestock ENERGY balances

CROPS Qty.Dom.USE

SSR

CROPS Qty.Dom.PROD

Crop YIELD

CULTIVATEDarea requirements

Pstr YIELD

PASTUREarea requirements

17

Food demand

Calculation procedures start from estimating future food demand based on population projections and anticipated changes in per caput food consumptions separately for a) vegetarian food from cereals, b) vegetarian food from other crops, c) livestock products from ruminants and d) livestock products from other livestock.

(4) xtctc

xtc fuPopFQDomUseVeg ,,, *= [GK$]

(5) ytctc

ytc fuPoptQDomUseLvs ,,, *= [GK$]

tcPop , Population in country c and year t persons xtcFQDomUseVeg , Domestic use of vegetarian products for commodity group x GK$

ytctQDomUseLvs , Domestic use of livestock products country c and year t for commodity

group y GK$

xtcfu , ; y

tcfu , Food use per capita for respective commodity group (CER, OthCr, OL, and RUM) [GK$/person]

x Vegetarian food commodity group – cereals (CER), other crops (OthCr)

y Livestock products commodity group – from other livestock (OL); from ruminants (RUM)

c Country t Year

Self-sufficiency

Domestic consumption levels depend on trade. Commodities may stem from domestic production and from imports. Some commodities are exported. The self-sufficiency ratio (SSR) describes the relationship between domestic production and domestic use and is defined as the ratio of domestic production over domestic use. Historic SSRs change little over time indicating a relatively stable pattern of aggregate trade flows. For the bio-fuel scenarios SSRs are generally kept at the level of the base period 2000-02.

Crop quantities required for vegetarian food are readily described by equation (4). In the case of livestock products however consumption levels need to be translated into feed requirements for producing livestock animals. This is achieved via energy requirements of the livestock population. For each country the domestic livestock herd and production determines feed requirements. For both livestock groups domestic production is calculated by applying the respective SSR to projected total use (equation 6).

(6) LVSyc

LVStc

LVStc SSRQDomUseodQ 022000,,, *Pr −=

LVStcodQ ,Pr Domestic production quantity from animal group LVS [GK$]

LVStcQDomUse , Domestic use of livestock products from animals group LVS [GK$]

LVSycSSR 022000, − Self-sufficiency ratio of livestock products from animal group LVS in

country c in period 2000-02 [%]

LVS Livestock animal group: a) RUM = ruminants, b) OL = other livestock (primarily pigs and poultry)

18

Feed crops for domestic livestock production

Domestic livestock production needs to be translated into feed requirements of the domestic livestock herd. This is achieved via calculating energy requirements of various livestock types. In the case of ruminants this is separated into feed from grazing on pasture and, feed from crops grown on cultivated land (Figure 5).

Figure 5. Flow chart of livestock energy balances for feed requirements calculations

Livestock production intensity changes over time due to technical progress, for example in breeding or higher energy supply per feed volume. We describe a technical coefficient LI for livestock intensity as the ratio of feed energy intake over livestock production (i.e. feed energy required per unit of livestock production). The lower the LI the less feed energy is required to produce one volume of livestock quantity, in this definition a decreasing LI represents technical progress (equation 7).

The MOSUS database includes historic records of a country’s livestock production intensity. Its future development is subject to assumptions of different scenarios.

Dom. PROD QuantityOther LVST

Dom. PROD QuantityRuminants

LI OL LI RUM

ENERGY requir.for dom. PRODof Other LVST

ENERGY requir.for dom. PRODof Ruminants

Kcal/GK$

GK$

Mcal

eYldCr

GSHR

Feed fromCROPS

Feed fromPASTURE

Mcal

FEED CROPSrequirements for

domestic LVST prod. GK$

Kcal/GK$

livestock productionIntensity factor

Kcal/HAPstr YIELD

PASTUREarea requirements

HA

Ruminant feed partitioning factor

Energy provisionfrom feed crops

(to Figure 4)Cultivated land

area requirementsHA

19

Feeds of other livestock are solely obtained from crops grown on arable land and the livestock intensity factor can directly be applied to domestic production volumes. For ruminants total energy requirements need to be partitioned into a share derived from feed crops and a share from grazing on pasture land.

(7) LVStc

LVStc

LVStc odQeLI ,,, Pr/= [kcal/GK$]

(8) OLtc

OLtc

OLtc LIodQeCr ,,, *Pr= [Mcal]

(9) RUMtc

RUMtc

RUMtc LIodQe ,,, *Pr= [Mcal]

(10) 022000,, * −= GSHReeCr RUMtc

RUMtc [GJ]

(11) )1(* 022000,, −−= GSHReePstr RUMtc

RUMtc [GJ]

LVS

tcLI , Livestock production intensity for animal group LVS kcal/GK$ LVS

tce , Energy requirements for livestock group LVS Mcal LVStceCr , Energy obtained from feed crops for animal group LVS Mcal

RUMtcePstr , Energy obtained from pasture (for ruminants) Mcal

022000−GSHR Grassland feed share in total feed (only for ruminants), i.e. partitioning factor for feed from crops and pasture for the year 2000-02 %

Total feed crop energy requirements are converted into crop production volumes by applying feed energy yield ( EYldCr ). It describes how much quantity of feed is required to produce one energy unit and reflects a country’s composition of feed crops in energy terms. A higher eYldCr indicates a crop composition favouring crops with high energy contents. Historic data indicate relatively stable feed energy yields over time but varying coefficients across countries.

Total quantities of feed crops required for raising livestock are calculated as the sum of feed crops for both livestock groups divided by ‘feed energy yields’. For the latter the country specific value for the period 2000-02 is applied into the future, thus assuming a rather stable feed composition over time.

(12) tctctc QFeedCrECrEYldCr ,,, /= [GJ/GK$]

(13) 022000,,,, /)( −+= cRUMtc

OLtctc EYldCreCreCrQFeedCr [GK$]

tcEYldCr , Feed crops energy yield GJ/GK$

tcECr , Energy supplied from feed crops GJ

tcQFeedCr , Quantity of feed crops GK$

20

Crop land required for food and feed

A country’s total crop requirements for domestic use include the sum of food and feed crops. Both are traded extensively. The self-reliance ratio calculates domestic production of food and feed crops. Finally, applying specific crop yields to estimated domestic crop production results in cultivated land area requirements for domestic production of food and feed crops.

(14) tcOtherCropstc

Cerealstctc QFeedCrFQDomUseVegFQDomUseVegQCrDomUse ,,,, ++=

(15) Cropsctctc SSRQCrDomUseodQCr 022000,,, *Pr −= [GK$]

(16) tctcCULTIV

tc YCrodQCrAFoodFeed ,,, /Pr= [ha]

tcQCrDomUse , Quantity of crops required for domestic use of food and feed GK$ CerealstcFQDomUseVeg , Quantity of domestic use of vegetarian products from cereals GK$

CerealstcFQDomUseVeg , Quantity of domestic use of vegetarian products from other crops GK$

tcQFeedCr , Quantity of feed crops GK$

tcodQCr ,Pr Quantity of crops in domestic production GK$ CropscSSR 022000, − Self-sufficiency ratio for crops in country c in the period 2000-02 %

tcYCr , Crop yields in country c in year t GK$/ha CULTIV

tcAFoodFeed , Cultivated land area required for food and feed production ha

21

Pasture land

Feed energy requirements for ruminant animals have been calculated separately for feed from crops (equation 10) and feed from grazing (equation 11). Pasture area requirements are calculated by applying an estimated energy yield per hectare of grassland (equation 17). It varies with country and was estimated using potential grassland productivity data calculated as one output from the Agro-Ecological Zones (AEZ) methodology (Fischer et.al., 2002) assuming a digestible energy content of 1.94 Mcal per one ton dry matter (Table 3).

(17) 022000,,, / −= cRUMtc

PASTUREtc eYPstrePstrAFeed [ha]

022000, −ceYPstr Energy yield from pasture for country c for the period 2000-02 kcal/ha PASTURE

tcAFeed , Pasture land area required for ruminant feed ha

Livestock energy balances reveal for some countries the available reported pasture area being considerably larger than the estimated grassland area required for ruminant feed (Table 3). Such “surplus” pasture land may potentially become available for harvesting bio-fuel plants.

Table 3. Pasture area in 2000, share required for ruminant feed, average dry matter and feed energy yields by country

Pasture Area

Feed Use*

Yield DM

eYPstr 2000-02

Pasture Area

Feed Use

Yield DM

eYPstr

2000-02 1000 ha % tons/ha Mcal/ha 1000 ha % tons/ha Mcal/ha

at 1917 68% 5.0 9.7 ch 1039 95% 5.0 9.7 bel 574 100% 8.5 16.5 no 158 81% 4.0 7.8 de 5048 61% 6.5 12.6 cy 4 100% 2.5 4.9 dk 358 70% 7.5 14.6 cz 961 22% 4.5 8.7 es 9396 89% 2.5 4.9 ee 131 42% 4.0 7.8 fi 114 14% 3.5 6.8 hu 1051 20% 3.0 5.8 fr 10087 57% 6.0 11.7 lt 492 32% 4.0 7.8 gr 4675 45% 2.5 4.9 lv 611 16% 4.0 7.8 ie 3333 99% 7.5 14.6 pl 4076 24% 4.5 8.7 it 4353 49% 4.0 7.8 si 308 49% 5.5 10.7 nl 902 100% 8.5 16.5 sk 865 26% 5.0 9.7 pt 1284 82% 3.5 6.8 bg 1616 16% 5.0 9.7 se 447 45% 4.0 7.8 ro 4949 23% 5.0 9.7 uk 10017 82% 5.5 10.7 ukr 7924 38% 3.5 6.8

* Percentage of grassland area estimated providing required ruminant livestock feed energy.

Source: Pasture area (EUROSTAT); in FEED USE (MOSUS database); Yields (AEZ Methodology)

22

4.3 Trends of key variables determining food and feed area requirement

A region’s food and feed area requirements result from the interaction of food demand, food production and trade. Population numbers and per capita food consumption determine demand. On the production side land area requirements depend on production intensity. The MOSUS database allows incorporating two important factors of technological progress. First crop yield developments expressed in GK$ physical output per hectare. Second a coefficient for the intensity of livestock production describing the amount of energy supplied to produce one unit of livestock output.

4.3.1 Per capita food demand

Historic per capita food consumption reveals a fairly stable trend for the consumption of vegetarian products in many countries of Western Europe. Consumption of livestock products apparently is undergoing a shift from ruminant livestock products towards pigs and poultry products. In Eastern Europe the economic changes in the 1990s resulted in decreasing per capita food use (Table 4).

Table 4. Per capita food consumption (in $GK/cap) in Europe for different dietary sub-categories in 1985-87 and 2000-2002

GK$/cap EU15 EU10 Bulgaria Romania Ukraine* 85-87 00-02 85-87 00-02 85-87 00-02 85-87 00-02 90-92 00-02 VEGETARIAN food use CEREALS 26 28 37 35 54 40 40 39 41 34 Other CROPS 163 169 140 121 145 111 138 120 102 89 LIVESTOCK products use from RUMINANTS 165 151 130 78 145 90 83 70 133 81 from Other Livestock 106 119 116 117 98 94 98 80 68 44

* For the Ukraine the first period refers to 1990 to 1992

4.3.2 Yield increases

Yields growth and livestock production intensity increases are likely to free up some land for other uses while satisfying projected consumption levels. The MOSUS database (see section 1.3) features the crop sector of a country in a comprehensive form including all crops as well as multi-cropping. The aggregation of crops is achieved by its conversion to a new unit, the Geary-Khamis dollars (GK$). Yields here represent a countries’ total production volume (in GK$) divided by its cultivated land area. The latter is taken as the sum of arable land and permanent crops18 area. Aggregate yield expressed in GK$/ha captures changes of individual crop yields as well as changes in composition of crops and intensity of arable land use. It provides a comprehensive 18 Permanent crops include olive, vineyards and orchards.

23

measure of a country’s change in agricultural land use efficiency, which is most relevant for the estimation of residual land potentially available for cultivation of bio-fuel crops.

In the 1990s the EU’s Common Agricultural Policy (CAP) introduced a set-aside scheme19. The cultivated land area reported in FAOSTAT, the prime input for the MOSUS database, includes set-aside land. Therefore the MOSUS derived yields for the later years underestimate actual yields. EUROSTAT reports areas for ‘fallow land with no economic use’. The year 2000 numbers were used for yield adjustment for the base period 2000-02. Some 5 percent of the EU15’s 85 million hectares cultivated land is reported to be fallow land with no economic use.

Currently yields in Western Europe are nearly twice as high as in Eastern Europe (Table 5). However in Western Europe there is a wide range of yields from over 3000 GK$/ha in the Netherlands to less than 800 GK$/ha in the Nordic countries (for 2000-02). Eastern Europe’s agricultural decline in output per hectare due to economic restructuring is evident. Table 5. Technical progress in yields in Europe between 1985 and 2002. 1985-87 2000-02 Increase cultivated land GK$/ha GK$/ha percentage/year million hectares EU15 1043 1219 0.93 85 Germany 1178 1347 0.80 12.0 France 1159 1261 0.49 19.5 Spain 652 891 2.04 18.4 …Italy 1565 1617 0.19 11.1 EU10 860 623 -2.12 30 Bulgaria 684 491 -2.64 4.6 Romania 663 512 -1.31 9.9 Ukraine 485 (1990-92) 388 -2.19 33

4.3.3 Intensity of livestock production

Technical progress in the livestock sector can generally be regarded as decreasing input required to produce one unit of livestock product. The MOSUS database records time series of energy supplied for the livestock sectors separated in two sub-groups:

a) ruminants (mainly cattle, sheep and goats) b) other livestock (mainly pigs, poultry)

While feed for the former stems from both grassland and feed crops, the latter is solely fed on feed crops20. Energy intensity in livestock production here is defined as feed energy use per one unit of livestock output (kcal/GK$). Energy efficiency in other livestock production is generally higher than for ruminant production (Table 6).

For ruminants livestock energy intensity in Europe has been fairly stable since 1985 with even a small decreasing trend in Western Europe. This reflects well the extensification of agriculture fostered by the EU’s Common Agricultural Policy. For the EU15 ruminant livestock energy intensity increased from 14.0 kcal energy supplied for one GK$ ruminant output in 1985-87 to

19 The set-aside scheme foresees subsidies to take a certain amount of cultivated land out of production. However, on set-aside land farmers were allowed to grow non-food crops. 20 In Europe important feed crops include maize, wheat and soybeans.

24

14.7 in 2000-02. In Eastern Europe livestock energy intensity is lower amounting to 8.9 kcal/GK$ in 2000-02. The reason is the relatively lower share of energy supply from grassland compared to Western Europe. In the EU10 an estimated 22% of total energy supply stems from grassland compared to 42% for the EU15.

Higher grassland contribution in total feed tends to increase kcal/GK$. However, the treatment of permanent grassland as energy source is somewhat uncertain. In the MOSUS approach energy requirements for reported livestock numbers are compared with energy provided by feed crops grown on cultivated land. Any missing feed requirements are assumed to have been met from permanent grassland as feed source that is not recorded in the same detailed way as crop feeds. Depending on estimated country specific grassland yields, a certain fraction of reported permanent grassland is therefore determined to be required for grazing animals.

In contrast to ruminants, energy intensity in livestock production of pigs and poultry (subsumed in the other livestock sub-group) increased in all countries. As expected, feed energy use per unit of output is lower in Western than in Eastern European countries.

Table 6. Trends in livestock production intensity

Ruminants Other Livestock “Energy intensity in

livestock production*” Production Quantity

“Energy intensity in livestock production*”

Production Quantity

1985-87 2000-02 2000-02 1985-87 2000-02 2000-02 kcal/GK$ kcal/GK$ bio.GK$ kcal/GK$ kcal/GK$ bio.GK$

EU15 14.0 14.7 61 9.1 7.8 48 Germany 11.9 10.9 12.2 9.7 7.8 8.7 France 15.0 15.5 12.4 8.7 8.3 8.3 Spain 19.5 20.8 4.4 9.2 8.6 7.0 Italy 24.5 21.0 7.0 8.3 6.9 5.1 Great Britain 16.9 11.1 7.1 8.2 7.3 4.2 EU10 10.8 8.9 8.5 12.6 11.0 9.0 Bulgaria 14.4 12.1 0.8 11.8 6.6 0.7 Romania 28.0 15.6 2.0 18.1 13.1 1.6 Ukraine 12.7 5.9 21.0 2.1

* feed energy use per one unit of livestock output

4.3.4 Self-sufficiency in agricultural products

The growing importance of international trade is manifested by increasing volumes of agricultural commodity flows. In this way Europe’s land use is influenced indirectly by potential external land resource uses to satisfy EU’s domestic consumption. Trade volumes are significant amounting to more than 40% of production and have increased over the past two decades with annual increase rates of around three percent depending on sub-sector (Table 7).

25

Table 7. Production volume and trade flows of the EU25 agricultural sector in 2000-02

Produc-tion

Growth 85-02 Import Growth

85-02 Export Growth 85-02

Domestic Use SSR*

bio.GK$ % p.a. bio.GK$ % p.a. bio.GK$ % p.a. bio.GK$ % CROPS – primary & processed 117.0 0.08 61.9 2.77 45.4 3.06 133.0 88

Cereals 34.2 0.98 8.6 2.45 10.5 1.42 31.7 108 Other crops 65.9 -0.09 53.3 2.83 34.8 3.65 84.5 78 Fodder crops 16.9 -0.89 n.a. n.a. n.a. n.a. 16.9 100 LIVESTOCK products 126.4 -0.11 37.8 1.77 41.0 1.52 123.0 103

from Ruminants 69.5 -1.07 24.6 0.62 25.0 0.03 68.9 101 from Other livestock 56.8 1.30 13.3 4.68 15.9 4.92 54.1 105

* Self-Sufficiency ratio (Ratio production over consumption)

For the EU25 crops including primary and manufactured products are produced equivalent to 88% of domestic consumption, i.e. 12% of consumed commodities stem from net imports. The feed sector is a major contributor to Europe’s net imports. The EU25 is self-sufficient in ruminant livestock products and a slight net exporter for other livestock products. There are considerable trade flows of agricultural products between European countries (Table 8).

Self-sufficiency rates in Europe have generally remained fairly stable over the past two decades. Any increased reliance on for example feed crops from countries outside Europe would certainly exert influence on land availability for bioenergy crops. Currently of the EU25’s 133 billion GK$ used crops domestically some 47 millions GK$ are used to feed animals. However any major changes in future self-sufficiency ratios are unsupported by historical evidence and thus speculative and should be regarded as more radical scenario assumptions.

Table 8. Self-sufficiency (= ratio production over consumption) of European countries for the agricultural sector* (average 2000-02)

percentage EU25 EU15 de dk es fr ie it nl uk EU10 cz hu pl

CROPS & prod. 88 87 75 74 108 109 56 95 56 58 92 84 118 92

LIVESTOCK & livestock products

from Ruminants 101 99 115 172 89 112 347 58 129 82 118 110 127 122

from Other Lvst. 105 105 86 429 107 113 139 79 236 75 103 98 127 103

* Crops and livestock include the total of primary and manufactured products

4.4 Increases in built-up land In Europe the general trend in land use changes during recent decades has been a decrease in agricultural land and an increase in forest and built-up areas. These trends are expected to remain the main direction of land use change in the future and agricultural land will therefore be lost further and converted to forest and built-up areas. The latter includes all areas occupied for residential, commercial, industrial, and infrastructure purposes including both built-up land and

26

associated vegetated areas (henceforth termed BUILT+). BUILT+ areas are considered not to be any more available for economic forestry or agricultural activities.

Since most new built-up and associated areas are developed at the expense of agricultural land we assume the amount of agricultural areas lost due to land conversion to correspond with estimates for BUILT+ increases.

Future BUILT+ area estimates depend on economic development and land use policies. This rationale was reflected in a study in the context of the MOSUS project (Prieler, 2006). ‘Gross Fixed Capital Formation’ (GFCF) at constant prices, was chosen as driver for economic development. Scenarios of future GFCF expenditure for individual countries were taken from the economic model GINFORS21 (Lutz et al., 2005; Meyer at al., 2004). Land use policies are subsumed in the development of a land resource coefficient LRCBUILT+ defined as the ratio of annual increases in BUILT+ and annual GFCF expenditure. By decreasing LRCBUILT+ developments in land use policies were reflected aiming at lowering annual BUILT+ area consumption over time.

This study estimates for the EU15 by 2020 increases in BUILT+ of 5.3 million hectares, i.e. a 23% increase compared to 22.7 million hectares BUILT+ in the year 2000. This translates into annual per capita increases between 5 and 10 m2 and for the Nordic countries up to 18 m2/cap/year. For the land-scarce Netherlands a much lower increase rate is assumed of only 1.3 m2/cap/year (Table 9).

For the scenarios described in chapter 5 below we have adapted these numbers as the additional BUILT+ area by 2030. This can be interpreted as a conservative approach reflecting minimum land conversion. For Eastern European countries generally an annual increase of 7 m2/cap/year over a 20 year period was anticipated. Again this is perceived as the lower end of potential land conversions from agricultural land to BUILT+.

Table 9. Built-up and associated land area (BUILT+) increases between 2000 and 2020

Country BUILT+ in 2000

Additional BUILT+ until

2020

Per capita increase

2000-2020 Country BUILT+

in 2000 Additional

BUILT+ until 2020

Per capita increase

2000-2020 1000 ha 1000 ha m2/cap/a 1000 ha 1000 ha m2/cap/a at 382 148 9.3 ie 285 77 10.2 bel 592 117 5.5 it 2846 690 6.1 de 4573 1173 7.1 nl 575 41 1.3 dk 366 146 13.7 pt 713 103 5.0 es 2919 644 8.0 se 1352 230 13.0 fi 791 122 11.8 uk 2292 442 3.8 fr 4210 1225 10.4 ch 287 31 2.2 gr 793 144 6.6 no 685 165 18.4

21 Global INterindustry FORecasting System

27

Chapter 5. Land for bio-fuel feedstock production – Scenarios and results

5.1 Scenario storylines and assumptions

Competing land use requirements for Europe’s food and livestock sector as well as land use conversion from agriculture to other uses, in particular built-up and associated land areas, will determine future availability of land for energy crop production. Future food and feed area requirements are the result of developments in food demand combined with changes in production intensity and trade of agricultural products.

Food demand in Europe is expected to remain relatively stable in the coming decades. As shown in section 4.3.3, principal trade patterns among European countries and between Europe and the rest of the world did not change much in the past decade. Production intensity in general increased in Western Europe with annual approximate growth rates of about one percent. However in certain agricultural sectors there was some reduction in production intensity due to the general socio-economic situation or induced by certain environmental policies. Examples here include Eastern Europe’s declining agricultural production intensity after the economic changes towards a market economy. In Western Europe the decreasing intensity in ruminant livestock production reflects agri-environmental concerns and not declining efficiencies.

The aim of this study is to identify Europe’s potential land area for future bio-fuel feedstock production. Both cultivated land and pasture are considered as areas for growing bio-fuel feedstocks. The time frame is until 2030 and estimates are for individual countries22.

First a reference scenario (‘baseline’) describes ‘likely’ developments under current policy settings. Baseline essentially reflects effects of ongoing trends in food consumption patterns on the one hand and technological progress in food production on the other hand, and it assumes a continuation of current self-reliance levels in Europe’s aggregate food and feed commodities. Continued increases in crops yields and livestock production intensity will free land for bio-fuel production. For Western Europe a continuation of historic trends from 1985 to 2002 is considered to represent the basis for the reference scenario. In 2000 some 4.3 million hectares of cultivated land are reported to be ‘fallow with no economic use’ in the EU15 countries (Eurostat, 2005). For the scenarios it is assumed that these five percent of total cultivated land in EU15 will become available for bioenergy crops production.

The economic changes in Eastern Europe towards a market economy in the early 1990s resulted initially in substantial declines in agricultural output for many sectors. After accession to the European Union with support for restructuring crop yields are increasing again. However they are still at levels well below Western European countries. The ‘baseline’ storyline for Eastern European countries assumes technological progress to continue faster than in Western European countries and to reach Western Europe’s technological level by 2050.

Economic, agricultural and environmental policy may cause an inflection in the developments outlined in ‘baseline’. The relative importance of various policy frameworks can be understood by comparing two contrasting scenarios relative to the reference scenario. The focus is on more or less area becoming available in the future for bio-fuel feedstock production (scenario ‘high’ and

22 The following countries are included: EU25 and Switzerland, Norway, Bulgaria, Romania, Ukraine

28

scenario ‘low’). An assumed precondition for these scenarios is to maintain Europe’s food and feed production to satisfy future consumption at similar self-sufficiency levels as today. There is little reason for food demand to deviate substantially from ‘baseline’. Agricultural production intensity in contrast depends on agricultural and environmental policies as well as technological progress and may vary significantly in different scenarios. Deviations from ‘baseline’ here are realized by different assumptions for crop yield and feeding efficiency developments.

The ‘low’ scenario anticipates a higher share of areas with organic agricultural production. As a consequence yields in Western Europe are lower compared to ‘baseline’. For Eastern Europe we assume ‘baseline’ yield increases can also be realized under conditions of organic agricultural production. The ‘high’ scenario reflects an intensified agricultural production system compared to ‘baseline’. Reasons may be manifold including new varieties, intensified rotation systems or farm restructuring towards larger entities. ‘High’ foresees yields in Eastern Europe to converge faster with EU15 compared to ‘baseline’. More precisely EU12 will reach by 2030 the average yield level of EU15 in the ‘baseline’ variant. For EU15 higher increase rates in yields are assumed. For example for the EU15 aggregate this means by 2030 a seven percent higher yield compared to ‘baseline’.

Table 10 summarizes assumptions and respective technical coefficients underlying the scenario calculations. Linear trends of historic coefficients were calculated for the period 1985 and 2002. Regression slopes of EU27 or EU15 aggregates define future increments starting from base year coefficients of individual countries. All modelling steps are calculated for individual countries in five-year time steps between 2000 and 2030.

Food demand

Future population numbers and per capita consumption of food and livestock products will determine food and feed area requirements. No dietary shifts are foreseen23. Future population numbers are taken from the EUROSTAT projections ‘baseline variant’ except for Norway, Switzerland and Ukraine, which are not included in EUROSTAT. For these countries population projections follow the UN World Population Prospects ‘medium variant’. Projections show an increasing population for most western European countries with some 15 million added by 2030 to the current 384 million of the EU15. In Eastern Europe population is expected to decline from 103 million in 2005 to 96 million by 2030 for the EU12. Significant decreases are projected for the Ukraine (49 million in 2000 and 35 million by 2030),

The historic trend of the EU25’s food consumption per capita is projected into the future for vegetarian products and livestock products from pigs and poultry. In the case of ruminant livestock products consumption (mainly beef and dairy products) historic data reveal for most countries a decreasing trend in consumption. This is in particular true for Eastern Europe where consumption levels today are on average only half the level of Western Europe. In the scenarios some further decrease of per capita consumption is assumed for Western Europe while no further declines are foreseen for Eastern Europe.

23 Shifts towards a more vegetarian diet could release substantial amounts of agricultural land for growing non-food crops. For the EU25 currently (2000-02) about one third of land area requirements are associated with domestic food consumption of crop products compared to two thirds associated with livestock products consumption. From the latter about half of the required area is arable land for the production of feed crops and the other half is pasture.

29

Table 10. Assumptions driving scenario calculations

‘Baseline’ scenario assumptions (if not specified otherwise applied to all countries)

1) Land loss due to increasing built-up and associated land

BUILT+ area increases taken from Prieler (2006)1 – all additional BUILT+ area is converted from agricultural area at the same ratio for cultivated land and permanent pastures.

2) Food DEMAND

2a) POPULATION projections Eurostat scenario ‘baseline variant’ UN projections ‘medium variant’ for no, ch, ukr

2b) Food consumption per capita Cereals Other crops Livestock products Ruminants Other LVST (pigs& poultry)

historic trend (1985-2002) of EU25 historic trend of EU25 Western Europe2 : historic trend of EU15 Eastern Europe3 : no change historic trend of EU25

3) Agricultural PRODUCTION

3a) CROP Yields

Western Europe (EU15, ch, no, cy, mt): historic trend of EU15 + use of set-aside land Eastern Europe (remaining EU12, ukr) EU10 reaches EU15 yields by 2050

3b) Intensity of livestock production4 Ruminants Other livestock

some intensification in feed energy use (own estimate) historic trend of EU15

4) Trade & self-sufficiency constant at the level of the base period 2000-02 ‘Low’ scenario

3aL) CROP Yields Western Europe: Higher share of organic cultivation Eastern Europe: as ‘baseline’

‘High’ scenario

3aH) CROP Yields Western Europe: Higher yields (own estimate) Eastern Europe: EU12 reaches EU15 yields of ‘baseline’ by 2030

1 as described in section 4.4; 2 Western Europe includes EU15, Switzerland, Norway, Cyprus and Malta; 3 Eastern Europe includes EU12 (except Cyprus and Malta) and Ukraine; 4 Intensity is defined as feed energy use per one unit of livestock output.

Self-sufficiency

Self-sufficiency rates are kept constant at the level of the base period 2000-02. Current aggregate trade shares are therefore assumed to remain unchanged for agricultural commodities. In other words if a country is a major exporter of for example livestock products it will also be one in the future scenarios. The same applies to all feed commodities.

Intensity in livestock production

As in the past energy intensity for the production of ‘other livestock’ (mainly pigs and poultry) is expected to further increase with declining feed energy requirements per unit of livestock output. For the future projections the historic level of the EU15 feed energy coefficient (in kcal per volume output) is applied to all countries.

Historic data for the ‘ruminant livestock’ animal sector reveal a decreasing intensity in production for the EU15 aggregate. Over time slightly more feed energy was used for producing one unit of output. However results differ for individual countries. Countries where feed supply per unit of

30

animal output increased (i.e. a decreasing intensity in ruminant livestock production) include Belgium, Spain, France, Ireland, Sweden and the United Kingdom. The other countries show an increasing intensity in production. For the scenarios we assume again a small increase in livestock production intensity.

Crop yields for different scenarios

Increases in future crop yields are essential for freeing up areas for growing bio-fuel feedstock crops. For Western Europe increases are considered to be at the historic level of the EU15 aggregate. The slope (11.5 GK$/ha) of the historic trend is applied for the future. Full use of set-aside land with no economic use is quantified by adjusting statistically reported yields of the base year with set-aside areas. More precisely new (higher) yields are calculated by dividing physical volumes of agricultural production and cultivated land area reduced by the amount of set-aside land. In this way every 10 years yields increase by 115 GK$/ha starting from a base yield in 2000-02 of 1219 GK$/ha, i.e. nearly 1% per annum until 2030.

For Eastern European countries yields in 2000-02 were lower compared to most Western European countries with the EU10 crop yields being only half the level of the EU15. For the ‘baseline’ scenario higher increase rates are assumed compared to Western Europe. The level of increase reflects a closure of the yield gap between EU15 and EU10 by 2050. While in 2000-02 yields in Eastern Europe were only 51% of Western Europe’s yields they reach 78% of Western Europe by 2030.

The ‘low’ scenario anticipates an increasing area of cultivated land with organic farming production schemes. As organic farming does not use chemical fertilizer, herbicides and pesticides the yield level of organic farming is usually less than for conventional farming practices. According to FAO (Bruinsma, 2003) organic farming yields are between 10% and 30% lower than for conventional farming. For the ‘low’ scenario we assume that areas with organic farming will have on average a 20% lower yield.

The extent of organic cultivated land was estimated from EUROSTAT statistical data24 for the base year 2005. Historic data shows that countries reaching a 20% share of organic area in total agricultural land seem to not extend beyond that level. For the ‘low’ scenario it is assumed that organic area doubles in most countries until 2030. Only in Germany and France an even higher increase is assumed due to high historic growth rates. Yield increases of conventional farming are due to higher levels of input or new crop varieties. Organic farming objectives focus on quality rather than increasing physical volumes per hectare. In the ‘low’ scenario yields from organically farmed cultivated land are kept constant between 2005 and 2030. In summary for the organic areas shown in Table 11 a 20% lower yield than for conventional farming in the base year and this yield is kept constant throughout 2030.

Europe’s agricultural sector in the past decades faced overproduction with associated high support costs to maintain farm land. However in view of competing land uses for food, feed, bioenergy crops, nature conservation purposes and human demand for built-up land, cultivated land may become a scarce resource. As a consequence farmers would strive for higher outputs per unit of land, which could be achieved via more use of higher yielding varieties or increased levels of inputs. Yields in Eastern Europe are on average only half of those in Western Europe. Thus there is high potential for improvements in many countries.

24 Statistical data give share of organic area in total agricultural land which includes arable land, permanent crops and permanent grassland. Thus we had to estimate a share for the cultivated land.

31

The ‘high’ scenario foresees higher yields compared to ‘baseline’. For EU15 countries we simply assume a higher coefficient compared to ‘baseline’. As a result average annual yield growth rates are 1.23% in ‘high’ compared to 0.94% in ‘baseline’ (for the EU15 aggregate). In the case of Eastern European countries the assumption is that the EU12 aggregate converges with EU15 ‘baseline’ yields by 2030 (rather than 2050 as anticipated in ‘baseline’).

Table 11. Organic farming area in 2005 and assumptions for 2030 in the ‘low’ scenario

Organic farming area

[1000 ha] Percentage of organic farming areas

in total cultivated land 2005 2030 2005 2030 Austria 148 286 10% 20% Belgium/Lux 16 31 2% 3% Germany 570 2060 5% 18% Denmark 134 255 6% 12% Spain 451 886 2% 5% Finland 153 292 7% 14% France 446 1575 2% 9% Greece 120 237 3% 6% Ireland 8 17 1% 2% Italy 683 1316 6% 12% Netherlands 26 50 2% 5% Portugal 154 300 6% 11% Sweden 189 355 7% 14% United Kingdom 276 536 5% 10% Switzerland 33 65 7% 15% Norway 34 59 4% 8% TOTAL 5446 10349 6% 12%

According to the above assumptions yields were calculated for individual countries throughout the period 2000-02 and 2030 for all three scenarios. Figure 6 highlights regional differences and assumed increases for the different scenarios.

0

500

1000

1500

2000

2500

3000

3500

at bel de dk es fi fr gr ie it nl pt se uk ch no cz ee hu pl si sk bg ro ukr

GK$

/HA

Base period 2000-02 Scenario LOW 2030 Scenario BASELINE 2030 Scenario HIGH 2030

Figure 6. Aggregated crop yields (expressed in GK$ per hectare of cultivated land) in European countries for the base period 2000-02 and assumptions for three scenarios for 2030

32

5.2 Bio-fuel feedstocks on cultivated land

Using the above assumptions the model calculates future cultivated land area requirements for food and feed production. At the same time a certain amount of cultivated land loss occurs due to built-up and associated land area conversion. As a remainder the model calculates surplus agricultural land that can potentially be used for growing bioenergy feedstock crops. Table 12, Table 13 and Figures 7 summarize results for the ‘baseline’ scenario.

According to ‘baseline’ by 2030 cultivated land for growing bioenergy crops amounts to 42 million hectares in the EU27. The crucial role of Eastern Europe is apparent contributing slightly more than half of the EU’s potential. In the EU12 by 2030 a similar amount of cultivated land could be used for growing bioenergy feedstocks as for food and feed crop production. The Ukraine could add an additional 22 million hectares of cultivated land potentially available for bioenergy crops (Figure 7).

Table 12. Europe’s use of cultivated land* in 2030 for ‘baseline’ scenario and available area for bioenergy crops for a ‘low’ and ‘high’ for bio-fuel scenario million hectares EU15+* EU12 Ukraine Total Cultivated land** in 2000-02 85.4 44.6 33.5 162.5 ‘Baseline’ scenario 2030 Food & Feed use 62.5 20.1 10.4 93.0 Conversion to BUILT+ 3.6 1.1 0.6 5.3 Bio-fuel feedstock production 19.3 23.4 22.4 65.1 Bio-fuel feedstock production 2030: ‘low’ scenario 16.9 23.4 22.4 62.7 ‘high’ scenario 23.4 28.3 25.4 77.1

**Cultivated land comprises arable land and permanent crops; * includes Norway and Switzerland Figure 7. Regional distribution of cultivated land potentially available for bio-fuel feedstock production in Europe by 2030 (‘baseline’ scenario) The effect of approximately doubling Western Europe’s area of organically grown food and feed crops to some 10 million hectares in 2030 is relatively small with regard to overall area availability for bio-fuel crops. Estimated areas for bio-fuel feedstock production in the ‘low’ scenario are only 2.4 million hectares less than in ‘baseline’ (in the EU15). This is less than the 3.6 million hectares cultivated area considered to be lost due to built-up land conversion.

Cultivated land losses to built-up and associated areas are significant, in particular in Western Europe. Deviations from ‘baseline’ in potential bioenergy crop area in ‘low’ and ‘high’ scenarios are in the same order or less than cultivated land losses. As outlined above exogenously assumed

EU15+no,ch (19 mio.ha)EU10 (15 mio.ha)Romania+Bulgaria (8 mio.ha)Ukraine (22 mio.ha)

33

built-up area increases may be considered to represent the lower end of potential increases, especially in Eastern Europe.

Table 13. Cultivated* land use in 2030 (‘Baseline’ scenario) 1000 hectares

Food & Feed area requirements

Potential area for bio-fuel feedstock production

Conversion to BUILT+*

Austria 1045 356 64 Belgium-Lux. 795 83 73 Germany 8174 3020 826 Denmark 1498 664 126 Spain 13420 4694 398 Finland 1111 965 116 France 14210 3693 806 Greece 3152 637 65 Ireland 839 219 21 Italy 8784 2002 498 Netherlands 944 76 22 Portugal 1893 735 77 Sweden 1722 786 197 United Kingdom 4023 1100 178 EU15 61610 19030 3467 Switzerland 365 73 9 Norway 582 163 140 Czech Republic 1673 1534 112 Slovakia 760 767 49 Hungary 2343 2242 117 Baltic States 1722 786 197 Poland 7614 6842 422 Slovenia 157 33 11 Cyprus 137 0 3 Malta 8 0 1 Bulgaria 1582 2969 85 Romania 4213 5490 205 EU12 20209 20663 1202 Ukraine 10432 22483 556

* Cultivated land includes arable land and permanent crops; ** Cultivated land lost due to conversion to built-up and associated land between 2000 and 2030

34

5.3 Bio-fuel feedstocks on pasture land Pasture land is considered only for growing herbaceous bio-fuel crops (miscanthus, switchgrass, reed canary grass). In this way no ploughing is required and there is no additional greenhouse gas release.

Livestock energy balances reveal for most European countries the available pasture area being larger than the area required for ruminant feed. In the EU27 total area of permanent pasture amounts to 67 million hectares. In 2000-2002 feed crops provided 60% of energy required for feeding all ruminant animals. Some 41 million hectares pasture area are required to supply the remaining 40% feed. Thus an area of 26 million hectares could be considered for alternative use25 (Table 14). The share of required pasture feed area in total pasture land varies widely across European countries depending on animal density, provision of feed crops and pasture yields. It is highest in the Netherlands, Ireland and Belgium where all pasture area is required for feed while in the remaining Western European countries it is mostly between 50 and 70%. For Eastern European countries generally livestock energy balances calculate only some 30% of pastures in feed use.

Depending on considerations regarding nature conservation and economic accessibility a fraction of this “surplus” pasture land may be used for growing herbaceous bio-fuel crops. To identify non-restricted “surplus” pasture areas additional land use and slope information from the Europe GIS databases was included as follows.

First the extend of area classified as “natural grassland” in the land use data base was considered to represent areas of high nature conservation value and thus excluded from the potential bio-fuel crop area. Second from the remaining ‘surplus’ pasture only the fraction with slopes below 16% was considered for growing bio-fuel crops. This fraction was estimated by overlaying all land use categories containing grassland 26 and the slope distribution map calculated from the high resolution digital elevation SRTM data.

In the scenarios we can therefore distinguish two types of area gained for growing bio-fuel crops. First the feed area which has been used in the base year but has become available due to scenario assumptions on intensified livestock animal production and reduced ruminant consumption. Second potential areas which are not required for animal feed and at the same time no nature conservation or slope limitations apply. In this way future use of permanent pasture can be subdivided into the following four categories.

1. Pasture area required for feeding ruminant animals (FEED) 2. Pasture area becoming available due to technological progress in agricultural production

(i.e. the change in feed area required for ruminant livestock production between the base period and the future) (BioCrops-I)

3. Pasture area not required for livestock feed and not restricted by slope and nature conservation concerns (BioCrops-II)

4. Pasture area not required for livestock feed and reserved for reasons of nature conservation (Natural grassland)

25 However these numbers should be treated with caution since the underlying methodological approach is subject to data limitations and livestock management assumptions. More precisely we do not know how intensely a cow grazes, i.e. whether the animal really consumes all provided energy by each square meter. 26 Four classes from the reclassified 12 classes were included, namely: 2=natural grassland; 4=other natural; 7=heterogeneous agriculture; 8=pastures.

35

In addition like with cultivated land some grassland will be lost due to conversion to built-up land.

Table 14 and Figure 8 summarize baseline scenario results for EU and selected European countries. By 2030 in the EU27 some 33 million hectares will be required for ruminant animal feed. Potential grassland area for growing herbaceous bio-fuel crops amounts to some 18 million hectares. This includes 5 million hectares gained by decreasing consumption levels and increased livestock intensity and 13 million hectares of grassland that has not been used for feed in the base period 2000-02. Another 13 million hectares of grassland remains untouched for reasons of nature conservation and steep slopes. This area may further increase when improved data bases about protected areas could be included in the analysis. This is especially true for Eastern Europe where with the current methodology only 3 million hectares remain as ‘natural grassland’.

Table 14. Pasture land use in 2000-02 and scenario results for 2030 (‘baseline’ scenario)

million hectares EU27 of which EU15 Ukraine

Pasture land in year 2000-02 of which required for ruminant animal feed ‘not in feed use’

67.6

40.9 26.7

52.5

37.4 15.1

7.9

3.0 4.9

Pasture use in 2030 (baseline scenario) Feed requirements for ruminants 33.5 30.7 2.2 Herbaceous bio-fuel crops Type I (Area gained due to technological progress1) 5.2 4.9 0.7

Herbaceous bio-fuel crops Type II (Area not in feed use and no restrictions2) 13.3 5.2 4.8

Natural grassland3 (Reserved for nature conservation) 13.5 9.9 0.1 Conversion to built-up land between 2000 and 2030 2.2 1.8 0.1

1 Area gained because of technological improvements in livestock intensity and lower consumption levels; 2 This includes ‘non-feed’ areas which are not classified as natural grassland in the GIS database and are below a slope of 15%; 3 Natural grassland is considered to be unavailable for bio-fuel feedstock production primarily for reasons of nature conservation

0

2000

4000

6000

8000

10000

12000

at bel de es fr gr ie it nl pt uk ch cz hu bal pl sk bg ro ukr

1000

hec

tare

s

FEED BioCrop-I BioCrop-II Natural Grassland BUILT+

Figure 8. Pasture land use in 2030 (‘baseline’ scenario)

36

Chapter 6. Assessment of land potentials for bio-fuel feedstock production

Potential production for different bio-fuel feedstocks were assessed for previously described scenarios of projected future land availability. For this purpose the existing agro-ecological zones (AEZ) modelling framework has been updated and expanded for feedstock productivity assessments. The AEZ procedures calculate suitability and potential productivity for bio-fuel crops, perennial grasses and trees species on a grid cell basis (Fischer et. al 2006) and are implemented to operate on the European 1 by 1 km GIS land resource data base.

6.1 Bio-fuel feedstocks

We consider two main types of bio-fuels, bioethanol and biodiesel, and differentiate between so-called 1st generation conversion, based respectively on biochemical conversion of biomass to ethanol via intermediates such as sugar or based on vegetable oil for biodiesel, and 2nd generation bio-fuels based on biochemical processes or thermochemical conversion using combustion, gasification and conversion of syngas, or pyrolysis.

While first generation bio-fuel technologies have reached an advanced stage and are widely used in many countries, second generation technologies are still mainly applied in experimentation and demonstration projects. To prepare for the large-scale use of cost-competitive second generation bio-fuels, continued research and development is needed to make the new technologies successful. Moreover significant changes in management and infrastructure requirements are necessary to introduce bio-fuel economy based on second generation conversion technologies. Forecasts for the latter are therefore more speculative.

At the same time second generation bio-fuel technologies based on lignocellulosic processing are widely regarded as the most promising route to large scale bio-fuel production. Energy scenarios and policy alternatives that favour second generation bio-fuels are clearly superior in terms of land use implications. First, energy yields (and greenhouse gas balances) of feedstocks for lignocellulosic processing are expected to be much higher than for first generation technologies. Second, and of great importance in the land use discussion, feedstocks for second generation bio-fuels would allow a wider spectrum of land to be considered for production. Notably, grassland that is not viable for first generation bio-fuels due to environmental and greenhouse gas implications, would also become a valuable resource for bio-fuel production, e.g., producing high-yielding lignocellulosic feedstocks without having to plough the land.

For the suitability and productivity assessments with the AEZ modelling framework, five main groups of land utilization types with specific bio-fuel production pathways are distinguished, namely: Woody plants, herbaceous plants, oil crops, sugar crops and cereals.

37

(1) Woody plants – (2nd generation bio-fuels)

These LUTs include short rotation forestry management systems. Tree species considered include poplars, willows and eucalypts. The selected tree species cover a wide range of ecological regions of Europe.

• Poplar (Populus nigra, Populus euramericana cv rob, Populus alba, Populus tremula, Populus balsamiferas Populus maximowiczii, Populus tomentosa, Populus euphraetica)

• Willow (Salix alba, Salix viminalis) • Eucalypt (E. globulus, E. camaldulensis, E. viminalis)

(2) Herbaceous lingocellulosic plants - (2nd generation bio-fuels)

The herbaceous plants selected are productive in terms of lignocellulose and cover a wide range of ecologies. Included are:

• Miscanthus (Miscanthus sinensis) • Switchgrass (Panicum virgatum) • Reed canary grass (Phalaris arundinaceae)

(3) Oil crops – (1st generation bio-fuel for production of biodiesel)

The two selected oil crops are widely grown in respectively southern and central, and northern and central Europe.

• Sunflower (Helianthus annuus) • Rapeseed (Brassica napus oleifera)

(4) Cereals - (1st generation bio-fuel for production of bioethanol)

Selected cereals are wheat, maize, rye and triticale. Wheat and maize are widely grown, rye and triticale are (currently) much less grown but have similar potential for starch to energy conversion as wheat.

• Wheat (Triticum aestivum) • Rye (Secale cereale) • Triticale (Tritico secale) • Maize (Zea mays)

(5) Sugar crops - (1st generation bio-fuel for production of bioethanol)

Sugar beet is a widely grown crop in Europe, while sweet sorghum is regarded as a potential energy crop for the sugar to energy production pathway.

• Sugar beet (Beta vulgaris) • Sweet sorghum (Sorghum bicolor)

38

6.2 Agro-Ecological Zones methodology

The Food and Agriculture Organization of the United Nations (FAO), in collaboration with the International Institute for Applied Systems Analysis (IIASA), has developed the agro-ecological zones (AEZ) methodology (Fischer et al, 2002) for the assessment of agro-ecological potentials of agricultural crops as well as for specific bio-fuel crops and perennial grasses. For bioenergy assessments a companion model of AEZ has been developed that enables assessments of potential productivity of tree species as well (Fisher et al. 2005).

AEZ follows an environmental approach; provides a standardized framework for the characterization of climate, soil and terrain conditions relevant to crops, perennial grasses and forest species production, and uses environmental matching procedures to identify limitations of prevailing climate, soil and terrain for assumed management objectives.

For the purpose of the assessment of bio-fuel potentials in Europe, the AEZ model is including inventories of ecological adaptability characteristics as well as inventories of specific ecological and environmental requirements for crop, perennial grasses and tree species. The natural resources inventory is based on an up-to-date pan-european GIS database of climate, soil, terrain and land cover. Figure 9 and accompanying text present a schematic overview of the flow and integration as implemented.

Feedstock Land Utilization Types (LUT) The land utilization types include definitions and descriptions of feedstocks bio-fuel crops, perennial grasses and tree species. The LUT attributes include characteristics of the feedstock species and information on management practices, inputs and utilization of produce.

Feedstock/LUT catalogue The catalogue database provides a quantified description of LUTs including adaptability characteristics such as: rotation length, vegetation period, photosynthetic pathway, photosynthesis temperature relationships, maximum leaf area index, partitioning coefficients, and parameters describing ecological requirements of the selected bio-fuel crops, perennial grasses and tree species. Climate database

Gridded climate parameters from East Anglia University (CRU global climatologies) and VASCLimO global precipitation data from the Global Precipitation Climatology Centre (GPCC) are used (see section 2.2).

Soil association composition database

The soil association composition database contains ESDB soil attributes database expanded with additional soil parameters from the ISRIC/FAO/IIASA WISE database (Batjes et.al. 1997).

Land resources database

The land resources database includes layers of the Soil Map of Europe, a slope distribution database, land cover layers, a protected areas layer and administrative areas, all with associated attribute databases.

39

FeedstockLand Utilization

Types (LUT)

FeedstockLUT Catalog

Feedstock species characteristicsBiomass and yield parameters

Rotation characteristics,Partitioning coefficients, Ecological requirements

Climate Analysis

Biomass and Yield Calculation

Climate Database

Land Resources Database

Soil AssociationComposition

Database

Biofuel Feedstock Species Potentials

by Grid-cell

Administrative Areas

Protected areas

Land Use/Land Cover

Terrain slopes

Soils

GIS

Feedstock Selection Feedstock SpecificEnergy Content

Land use Scenarios

Ecologically and Agro-economically viable

Bio-fuel Potentials

Climatic and EdaphicMatching

Procedures

Assessment of Feedstock Potentials

Figure 9. Agro-Ecological Zones (AEZ) methodology for assessment of bio-fuel feedstock potentials

Land characteristics (GIS)

Soils, elevation, physiography, present land use/cover and administrative divisions are kept as individual layers in the geographical information system:

• The soil data are based on the European Soil Database (ESDB) (ESB, 2004).

• The NASA Shuttle Radar Topographic Mission (SRTM) has provided digital elevation data. The SRTM data is available as 3 arc second DEMs (CGIAR-CSI, 2006). Original SRTM tiles covering the European continent were used. For areas north of 70 degrees north, elevation data from GTOPO30 (USGS, 2002) was used. Slope gradients were calculated; resulting in distributions of eight slope classes for each grid-cell: 0-0.5%, 0.5-2%, 2-5%, 5-8%, 8-16%, 16-30%, 30-45%, and > 45%.

40

• Three available land cover databases have been reclassified to twelve major land use classes for the purpose of determining spatial locations of arable land, grassland, forest and other areas. A harmonized land use map was constructed for the Pan-European territory based on the following sources:

a) CORINE version (CLCC 2000) for EU 25 b) CORINE Version 1990 (CLC1990) for Switzerland c) JRC’s Global Land Cover for Europe (GLC2000)

• Protected areas reflect an interpretation of the IUCN-WCMC protected areas inventory at 30-arc seconds separates protected land where cultivation is permitted from areas where cultivation is strictly prohibited.

• Administrative areas, an administrative layer map has been compiled for the Pan-European territory. Administrative levels are defined as a hierarchic structure down to the NUTS 2 level in the case of EU27+ (EU27 and Switzerland, Norway). For the Ukraine the administrative map includes the oblast level, while for other remaining European countries only national boundaries have been used

Climate data analysis

From basic climatic data, monthly reference evapotranspiration (ETo) has been calculated according to Penman-Monteith. A water-balance model provides estimations of actual evapotranspiration (ETa) and length of growing period (LGP). Temperature and elevation are used for the characterization of thermal regimes (TR) as follows: thermal climates, representing major latitudinal climatic zones; winter and summer temperatures and extreme temperatures; temperature growing periods (LGPt), and accumulated temperatures. Temperature requirements of individual LUTs are matched with temperature regimes prevailing in individual grid-cells. For grid-cells with an optimum or sub-optimum match, biomass increment calculations are performed.

Biomass increment and yield calculations

The methodology for the calculation of potential biomass for crops is based on the eco-physiological biomass model of Kassam. For tree species it is based on the combined Chapman-Richard biomass increment model, and the biomass model of Kassam. It provides temperature and radiation based biomass production of individual crops and tree species.

Climatic suitability

Climatic constraints cause direct or indirect losses in the biomass increment. Climatic constraints are influenced by the following conditions:

Crops:

• The variability and degree of water-stress during the growing period; • the yield-quality reducing factors of pests, diseases and weeds; • the climatic factors, operating directly or indirectly, that reduce yield and quality of

produce mainly through their effects on yield components and yield formation • the climatic factors which effect the efficiency of farming operations and costs of

production, and • the risk of occurrence of late and early frost. Tree species:

• The variability and degree of water-stress during the growing period;

41

• constraints indirectly related to climatic conditions (e.g., pests, diseases and invasion of unwanted species or weeds);

• the climatic factors which effect the efficiency of forestry operations and costs of production, and

• the risk of occurrence of late and early frost, and disturbance by fire Edaphic suitability

The edaphic suitability assessment is based on matching of soil and terrain requirements of biomass plant species with prevailing soil and terrain conditions. These are management and input specific.

Bio-fuel feedstock potentials

The suitability and productivity assessments were carried out by matching climate characteristics, calculating gross biomass increments or yields, and subsequently, by matching soil and terrain characteristics of each grid-cell with ecological requirements of the crops and tree LUTs considered. Results are stored in a database that contains distribution of land suitability classes and attainable yields or biomass increments for each 1 km grid-cell for each of the feedstock LUTs.

Land use scenarios up to 2030

Three scenarios have been considered for available land for bio-fuel cultivation (see section 5.1). Modeling steps are calculated for individual countries27 in five years steps between 2000 and 2030. Feedstock selector

On the basis of land use scenarios and feedstock specific energy content for each grid-cell a representative feedstock is selected to represent bio-fuel suitability.

Ecologically and agro-economically viable bio-fuel potentials

This database contains bio-fuel potentials for the land-use scenarios by grid cell

The procedures are implemented to operate on a GIS grid-cell database. For Europe the model performs on a 1 by 1 kilometre grid. The grid matrix extends to 4090 rows and 3910 columns, of which more than eight million grid cells cover Europe’s continental area. Diverse agro-climatic indicators as well as potential yields of the above outlined bio-fuel crops were calculated for Europe’s territory on the 1- by 1-km grid.

27 The following countries are included: EU25 and Switzerland, Norway, Bulgaria, Romania, Ukraine

42

6.3 Description of data base

Maps showing potential productivity for different bio-fuel feedstocks

Potential biomass productivity for individual bio-fuel feedstocks has been estimated for 1 by 1 km grid cells and is stored in a GIS data base. Biomass yield calculations assume adequate fertilization and full mechanization. Yields represent current available farm technology and current available feedstock varieties28.

Within each of the above outlined five bio-fuel feedstock groups the most productive feedstock was selected in each grid-cell to represent a particular feedstock group. Selection criteria were tons dry weight for harvested woody and herbaceous celluloses and cereals, or oil and sugar yields of oil- and sugar crops, respectively. For oil crops an extraction rate of 40% of rapeseed and 44% of sunflower were assumed. Sugar beets are assumed to contain 75% water and 15% sugar, while sweet sorghum is assumed to contain 78% water and 10.5% sugar. For each grid cell the best performing crop within a group was determined.

Yields of each feedstock group were converted to bio-fuel energy equivalents (see Table 15 for conversion factors) and the spatial distribution of bio-fuel potential for available land resources was mapped for 1st and 2nd generation conversion technologies.

Table 15. Conversion factors from biomass to bio-fuel energy equivalent

Bio-fuel feedstock group Bio-fuel crop Conversion factor Woody Poplar, Willow, Eucalyptus 9.6 GJ / ton dry weight

Herbaceous lignocellulosic Miscanthus, Switchgrass Canary reed 9.3 GJ / ton dry weight

Oil crops Rapeseed Sunflower

14.4 GJ / ton seed 16.2 GJ / ton seed

Cereals Maize Wheat, Barley, Rye, Sorghum

7.9 GJ / ton grain 7.5 GJ / ton grain

Sugar crops Sugar beet Sweet sorghum

2.1 GJ / ton harvested (fresh weight) 7.2 GJ / ton harvested (dry weight)

In the various maps, physical output (tons biomass/ha, kg oil/ha, kg sugar/ha) and energy yields (GJ/ha) have been classified according to suitability classes as follows:

• VS very suitable (80-100% of maximum yields) • S suitable (60-80%) • MS moderately suitable (40-60%) • mS marginally suitable (20-40%) • vmS very marginally suitable (less than 20% of maximum yields) • NS not suitable

Suitability is described by the land utilization type, i.e. a particular crop variety such as winter wheat (not the aggregate crop group). The maximum suitability is defined as the 99 percentile of yield of a particular land utilization type over Europe. Therefore grid cells in different countries

28 Yields in particular of second generation bio-fuel feedstocks are expected to improve rapidly with the deployment of modern breeding technologies (e.g. Heaton and Long).

43

may have a similar suitability rating but different yields because they refer to different land utilization types. Annex 4 presents potential productivity maps of bio-fuel feedstocks.

Aggregation and tabulation of suitability and productivity results by administrative units

Results were tabulated by aggregate CORINE land cover classes (as described in section 2.4), by country and by NUTS2 administrative level. Lignocellulosic feedstock potentials (woody and herbaceous) were assessed for available arable land and pastures (LC classes 5 - 8). Other feedstocks were assessed for arable land only (LC classes 5, 6, and 7). Tabulated results are presented in terms of five suitability classes fully compatible with mapped outputs.

For each country and land cover class the area coverage of the best performing feedstock type in each bio-fuel feedstock group was selected. Table 16 shows an extract for the potential of oil crop production in Austria. The total arable area as given by the GIS land cover database is 11,000 km2. Of those, some 2100 km2 is not suitable for oil crop production. The remaining 8900 km2 arable land is variously suitable. In 80% of this area rapeseed is outperforming sunflower. On the best land, yields range between 1.0 and 1.3 kg oil per hectare for the suitability classes VS and S.

Table 16. Potential oil crop production in Austria (Example of database) Suitable area (km2) NAME LC TY (defining LUT) TOTAL VS S MS mS vmS NS Austria 5 (Arable) 1 (Rapeseed) 7063 3391 2621 940 90 21 Austria 5 2 (Sunflower) 1772 1104 562 101 5 0 Austria 5 11 (Total) 11004 4495 3183 1041 95 21 2170 Average class yield (kg oil/ha) Austria 5 (Arable) 1 (Rapeseed) 1342 1136 819 471 234 Austria 5 2 (Sunflower) 1610 1211 845 491 271 Austria 5 11 (Total) 1408 1149 822 472 234

Table 17 summarizes coverage and contents of the tabulated results. Results are presented for four different land cover classes (arable, permanent crops, heterogeneous agriculture, and pasture) as well as for total land area. Two files describe the specific land utilization types defining respective results of the highest yielding crop within each bio-fuel group. Land utilization types (LUTs) are described in Box 1.

44

Table 17. Database of Europe’s bio-fuel production potential - Description of tabulated results.

File AGGREGATION LEVEL

Administrative Land Cover class2 – Land Utilization Type (LUT) 3

Bio-fuel_h0.xls National (Country) total land area Bio-fuel_h1.xls National (Country) by LC class Bio-fuel_h2.xls NUTS21 by LC class Bio-fuel_h3.xls National (Country) by LC class & LUT Bio-fuel_h4.xls NUTS21 by LC class & LUT Worksheets in each excel file Description Units

E0_all_* Best performing crop in energy terms of all 5 bio-fuel feedstock groups GJ / ha (bio-fuel equivalent)

E0_crops_* Best performing crop in energy terms of bio-fuel feedstocks for 1st generation prod. GJ / ha (bio-fuel equivalent)

E1_woody_* Woody lignocelluloses kg / ha (dry weight biomass) E2_herbaceous_* Herbaceous lignocelluloses kg / ha 9dry weight biomass) E3_vegoil_* Oil crops kg oil / ha E4_starchy_* Cereals kg / ha E5_sugar_* Sugar crops kg sugar / ha

1 as defined in Appendix II and described in section 3.2; 2 LC classes: 5 = arable, 6 = permanent crops; 7 = heterogeneous agriculture; 8 = pastures; 3 as described in Box 1 Box. 1 Codes for specific land utilization types in the five different bio-fuel land utilization groups

All bio-fuel feedstocks (Worksheet E0) - TY 1 = Cereals, 2 Sugar crops, 3 Oil crops, 4 Herbaceous cellulosic, 5, Woody cellulosic, 11 Totals Woody cellulosic (Worksheet E1) - 1 Poplar, 2 Willow, 3 Eucalyptus, 11 Totals Herbaceous cellulosic (Worksheet E2) - 1 Miscanthus, 2 Switchgrass, 3 Canary reed, 11 Totals Oil crops (Worksheet E3) - 1 Rapeseed, 2 Sunflower, 11 Totals Cereals (Worksheet E4) - 1 Wheat, 2 Barley, 3 Rye, 4 Maize, 5 Sorghum, 11 Totals Sugar crops (Worksheet E5) - 1 Sugar beet, 2 Sweet sorghum, 11 Totals

45

Chapter 7. Bio-fuel feedstock potentials on agricultural land

The main determinants for bio-fuel energy potentials from feedstock grown on Europe’s agricultural areas are: (a) available agricultural land for bio-fuel feedstock production, (b) bio-fuel feedstock species grown and their respective yields, and (c) conversion technologies applied. The main focus is on land availability and physical bio-fuel feedstock production potential. Energy yields of different conversion technologies rely on published sources.

7.1 Agricultural land for bio-fuel feedstocks in 2030

Land availability for growing bio-fuel feedstock has been discussed in detail in the previous section. While bio-fuel feedstocks may create an economically attractive alternative to growing crops for conventional uses, we are primarily interested to assess availability of agricultural land in excess of land required for food and feed uses. Therefore it is assumed that EU27 will maintain its current (period 2000-02) level of self-sufficiency for food and feed crops and for livestock products. As a result of future consumption changes as well as technological progress and associated yield increases, arable land and grassland will gradually be set free and can be used for bio-fuel production without affecting required food and feed production quantities.

Agricultural land freed up by efficiency improvements can be used for growing bioenergy feedstocks or other agricultural products, such as additional feed and fiber crops for increasing self-sufficiency rates, thus reducing imports or increasing commodity export. Alternatively, the freed-up land could be used for growing biomass for the heat and power sector. Though these alternative uses may compete with bio-fuel feedstocks, here we assume that all the land not required for food and livestock production (at base period 2000-02 aggregate self-sufficiency level) will be available for bio-fuel feedstock production.

The ‘baseline’ scenario reflects a continuation of past trends in yields and per capita consumption. It assumes use of the year 2000 set-aside areas for bio-fuel feedstocks. The ‘low’ scenario features assumed increases of organic farming, which approximately doubles and reaches by 2030 about 10.3 million hectares in EU15, i.e. some 16% of cultivated land then used for food and feed production or about 12% of year 2030 projected total arable land in EU15 (including arable land available for bio-fuel feedstocks). Lower yields of organic farming increase the land required for food and feed production. For Eastern Europe the same yields are assumed as in ‘baseline’. By 2030 still a significant share of cultivated land in EU27, almost two-thirds of the total, is required for food and feed production.

Future use of grassland areas distinguishes between pastures and natural grassland. On the basis of the available GIS data, the latter was defined as areas designated as ‘natural grassland’ in CORINE and other grassland areas with terrain slopes exceeding 16%. Future use of pasture areas for grazing and fodder production is determined by changes in the projected size of ruminant herds, which is modelled in accordance with projected changes in livestock product demand and assumed (modest) improvements in ruminant feeding efficiencies. Remaining pasture land is assumed to be available for growing herbaceous lignocellulosic bio-fuel feedstocks. Especially for Eastern Europe, with relatively low current intensities of pasture use, pasture land potentially freed up for producing bio-fuel feedstocks is substantial.

46

Future balances of agricultural land by individual countries are presented in Figure 10 (‘baseline’ scenario in 2030).

The integration of Eastern European countries has significantly increased the EU’s agricultural land resource base. Large areas in Eastern Europe potentially become available for bioenergy feedstocks. These area gains are a consequence of the large potential for yield improvements in the food and feed sector combined with only modest increases in domestic demand.

According to the ‘Baseline’ scenario in 2030, arable land use required for agricultural production in the EU27 is estimated to be some 82 million hectares leaving about 42 million hectares arable land for growing bio-fuel feedstocks. Use of grassland in the EU27 in 2030 was estimated as follows: use of pasture land for livestock feeding amounts to some 33 million hectares, leaving around 18 million hectares for herbaceous and woody bio-fuel feedstocks, while 13 million hectares remain ‘untouched’ as natural grasslands and pastures.

Table 18. Land availability in Europe for bio-fuel feedstocks in 2030

EU15+1 mln ha

EU12 mln ha

Ukraine mln ha

ARABLE land

Baseline (trend) 19.3 23.4 22.4 Low (more organic cultivation) 16.9 23.4 22.4 High (higher yields) 23.4 28.3 25.4 PASTURE Baseline 4.8 0.3 0.7

Scen

ario

High (as in baseline + partial use of grassland not required for feed) 10.1 8.4 5.5

1 EU15+ (EU15 plus Norway and Switzerland)

0

5000

10000

15000

20000

25000

30000

at bel de dk es fi fr gr ie it nl pt se uk ch no cz ee hu lt lv pl si sk bg ro

1000

ha

Cultivated for food & feed Cultivated for bioenergyPasture for feed & natural Pasture for bioenergy

Figure 10. Potential land use in 2030 for European countries (‘baseline’ scenario)

Table 18 summarizes the scenario results of land availability for bio-fuel feedstocks for the EU15, EU12 (New Member States, including Bulgaria and Romania) and the Ukraine, separately for arable land and pastures. Results highlight the large contribution of Eastern European

47

countries. By 2030 more than half of EU’s 40 to 50 million hectares cultivated land potentially available for bio-fuel feedstock production will be found in Eastern European countries. The Ukraine could contribute another 20 million hectares.

7.2 Production potentials of bio-fuel feedstocks

Bio-fuel feedstock potentials were calculated for a 1 by 1 km raster throughout Europe. Results include yields and suitable areas for 15 individual bio-fuel feedstock types. The bio-fuel feedstock groups used with 1st generation conversion technologies are: (a) oil crops (rapeseed, sunflower), (b) cereals (maize, wheat, barley, rye, sorghum), and (c) sugar crops (sugar beet, sweet sorghum). The 2nd generation bio-fuel feedstock groups are: (d) woody (poplar, willow, eucalyptus), and (e) herbaceous lignocellulosic (miscanthus, switchgrass, canary reed).

Within each of the five feedstock groups the most productive feedstock was selected in each 1 by 1 km grid-cell to represent a particular feedstock group. Results refer to rainfed29 conditions and assume adequate fertilization and full mechanization. Scenarios do not consider possible increases in bio-fuel feedstock yields over time.

Five aggregate current (agricultural) land use classes, as described in the GIS land resource database, are considered for bio-fuel feedstocks: arable land, land under permanent crops, heterogeneous agriculture, pasture land, and natural grassland. Land use categories taken into account for 2nd generation bio-fuel feedstocks include all five land use categories; for 1st generation bio-fuels only arable land, land in use for permanent crops and the class of heterogeneous agriculture are considered.

Table 19 summarizes for selected European countries suitability distributions and associated yields. For instance, in Germany some 40% of the agricultural land is assessed as very suitable (VS) for herbaceous lignocellulosic feedstocks with an average yield of 16.6 t/ha dry weight. Some 22% are considered suitable (S) with an average yield of about 13 t/ha, etc. The aggregate suitability distributions shown in Table 19 reflect the spatial distribution and specific suitability of the best-performing sub-types in each feedstock group and grid-cell. It should be pointed out that land in the category ‘not suitable’ also includes various non-agricultural areas, e.g., in the mixed land cover class of ‘heterogeneous agriculture’.

Results show the spatial variation of bio-fuel feedstock yields across Europe and provide a basis for assessing comparative advantages of specific bio-fuel feedstocks. Due to possible competition with the food and feed sector, it is sometimes suggested to use only the more marginal agricultural areas for bio-fuel feedstock production. While we have assumed in our estimations that bio-fuel feedstocks, in particular 1st generation types, will be included as part of a multi-year crop rotation (and therefore average yields apply), the availability of biomass yields for different suitability classes allow preliminary conclusions on the effect of targeting specific types of agricultural land for bio-fuel feedstock production.

29 Results are also available for irrigated conditions but are not discussed in this report.

48

Table 19. Biomass potentials of bio-fuel feedstocks for selected European countries

SUITABILITY distribution on agricultural area* (%)

Average YIELD (rainfed) in suitability class

Unit of Yield

Suitability index** VS S MS ms NS VS S MS ms GERMANY Herbaceous 40 22 16 1 22 16.6 12.9 9.0 5.2 ton d.w./ha Woody 19 37 29 10 5 13.4 10.4 7.1 4.0 ton d.w./ha Cereals 45 17 17 4 17 8.5 6.5 4.6 2.8 ton d.w./ha Sugar crops 28 22 13 9 28 8.3 6.5 4.5 2.6 ton sugar/ ha Oil crops 45 19 15 5 17 1.5 1.1 0.8 0.5 ton oil / ha FRANCE Herbaceous 37 23 9 1 30 18.5 14.4 9.9 5.9 ton d.w./ha Woody 53 24 9 3 12 15.4 10.8 7.1 3.5 ton d.w./ha Cereals 39 21 13 6 12 7.2 5.9 4.0 2.7 ton d.w./ha Sugar crops 17 25 15 10 32 8.0 6.2 4.3 2.4 ton sugar/ ha Oil crops 24 28 14 6 28 1.4 1.1 0.8 0.5 ton oil / ha ITALY Herbaceous 14 23 15 6 43 19.5 14.7 10.2 6.3 ton d.w./ha Woody 29 27 11 2 30 15.1 10.8 7.1 3.5 ton d.w./ha Cereals 14 25 17 7 36 7.0 5.4 3.8 2.3 ton d.w./ha Sugar crops 8 18 14 11 49 8.0 6.0 4.1 2.4 ton sugar/ ha Oil crops 11 23 19 8 40 1.3 1.0 0.7 0.5 ton oil / ha GREAT BRITAIN Herbaceous 17 30 18 3 32 14.0 11.6 8.4 4.5 ton d.w./ha Woody 15 23 17 11 34 13.2 10.0 6.7 3.6 ton d.w./ha Cereals 10 34 25 6 26 7.1 6.2 4.4 2.7 ton d.w./ha Sugar crops 13 19 16 10 41 7.8 6.2 4.5 2.8 ton sugar/ ha Oil crops 20 27 19 9 24 1.4 1.1 0.8 0.5 ton oil / ha SWEDEN Herbaceous 22 32 24 6 16 10.9 9.6 6.9 4.1 ton d.w./ha Woody 0 45 22 11 22 9.8 6.7 3.5 ton d.w./ha Cereals 31 25 21 11 12 6.6 5.3 3.2 2.0 ton d.w./ha Sugar crops 1 6 17 24 53 7.5 6.3 4.4 2.7 ton sugar/ ha Oil crops 28 29 10 7 25 1.2 1.0 0.7 0.4 ton oil / ha POLAND Herbaceous 33 10 18 0 39 17.1 13.3 9.4 5.4 ton d.w./ha Woody 14 37 31 10 7 13.3 10.6 7.2 4.1 ton d.w./ha Cereals 34 11 16 4 35 8.6 6.5 4.5 2.9 ton d.w./ha Sugar crops 25 17 14 6 38 8.6 6.7 4.5 2.6 ton sugar/ ha Oil crops 35 11 15 4 34 1.5 1.2 0.8 0.5 ton oil / ha BULGARIA Herbaceous 50 23 4 0 22 19.2 14.6 10.1 5.6 ton d.w./ha Woody 29 47 9 1 14 13.8 9.6 6.8 3.5 ton d.w./ha Cereals 65 11 3 1 20 8.0 6.4 4.5 2.7 ton d.w./ha Sugar crops 26 35 14 3 23 5.9 5.2 4.3 2.3 ton sugar/ ha Oil crops 64 12 3 1 20 1.3 1.0 0.7 0.4 ton oil / ha UKRAINE Herbaceous 55 15 15 4 11 18.0 13.0 8.9 5.2 ton d.w./ha Woody 19 36 23 5 17 13.4 10.0 6.5 4.2 ton d.w./ha Cereals 36 35 17 1 9 8.5 5.7 4.6 2.3 ton d.w./ha Sugar crops 29 45 14 2 11 7.9 5.0 3.5 2.2 ton sugar/ ha Oil crops 29 13 39 7 12 1.6 1.1 0.7 0.4 ton oil / ha

* Agricultural areas for herbaceous and woody lignocellulosic feedstocks include the following land use categories (as defined in the GIS land use database): arable, permanent crops, heterogeneous agriculture, permanent pastures, and natural grassland. For all other feedstocks only arable land, land used for permanent crops and the class of heterogeneous agriculture is included. **Suitability classes: VS=very suitable; S=suitable; MS=moderately suitable; ms=marginally suitable; NS=not suitable

49

7.3 Potentials of 1st and 2nd generation bio-fuel production chains

A comparison of land suitability and productivity across feedstock groups was achieved via conversion into ‘bio-fuel equivalents’ energy units (in GJ) using published conversion factors (see Table 15).

Suitability maps for bio-fuel feedstocks in Europe were generated for all individual feedstock types as well as for the five bio-fuel feedstock groups. Figure 11 presents land suitability calculated across potential bio-fuel feedstocks sources measured in bio-fuel equivalent. Results are shown for grid-cells included in the land mask of current agricultural areas30.

Figure 11. Suitability of agriculture land for bio-fuel feedstock production

30 All 1x1 km grid-cells used in this map include at least one 100x100 m grid cell belonging to one of the aggregate agricultural land cover/use categories (i.e. cultivated land, land under permanent crops, heterogeneous agriculture, pasture land).

50

While for some countries such as the Ukraine, Bulgaria, Romania and France large tracts were assessed as very highly or highly suitable for bio-fuel feedstocks, results for other countries point towards a considerable spatial diversity.

Most current processes for bio-fuel production follow the so-called 1st generation production pathway relying on the sugar, starch, and oil components of crops. As a consequence “well-to-wheels31” greenhouse gas reductions achieved by substituting fossil fuels are in the order of 20% to 70% compared with using only fossil fuels. Some argue that greenhouse gas benefits are even lower depending on production pathways and crop types used. New processes under development, so-called 2nd generation production pathways, are converting as well the cellulosic components of plants to bio-fuels and produce very low net greenhouse gas emissions.

In the climate change debate and the discussion of mitigation options, the possible reduction of GHG emissions are considered as a key driving force for replacing fossil fuels with bio-fuels. Therefore, policy development for Europe’s domestic bio-fuel production potential requires a careful distinction of feedstocks of the 1st and 2nd generation production pathways.

To capture differences of bio-fuel potentials in terms of these broad technology pathways, for each grid-cell the most productive feedstock (in terms of bio-fuel energy equivalent) within the 1st generation bio-fuel feedstocks (oil cops, cereals, sugar crops) was selected as and vice versa for the 2nd generation bio-fuel feedstocks (herbaceous and woody lignocellulosic). Map 1 to 4 in Annex 4 show attainable energy yields of 1st respectively 2nd generation bio-fuel feedstocks. Energy yields reflect agricultural management with adequate fertilization and full mechanization.

Figure 12 compares average potential yields (in bio-fuel energy equivalent) of bio-fuel feedstock production for 1st and 2nd generation production pathways by country. Average yields were calculated based on the following assumptions. First, bio-fuel feedstock production would be integrated into regular multi-year crop rotations and occur proportionally in all arable land classes. Second, average yields were calculated in accordance with extents of arable land given in the statistics, i.e., using the calculated suitability distribution up to the extents indicated by statistics. This approach takes into account that less suitable areas, especially in grid-cells of the class ‘heterogenous agriculture’, are usually not used for arable farming.

In large parts of Europe potential energy yields of rainfed production range between 60 and 120 GJ/ha for 1st generation and between 100 and 180 GJ/ha for 2nd generation bio-fuel feedstocks respectively. Map 5 in Annex 4 shows energy yields of 1st generation bio-fuel feedstocks relative to (as a percentage of) 2nd generation feedstocks. In some marginal rainfed areas of Spain and southern Ukraine, 1st generation energy yields are similar to 2nd generation feedstocks, yet at a relatively low level of around 40 GJ/ha. In most of Europe energy yields of 2nd generation feedstocks are substantially higher than 1st generation feedstocks, by some 40% to 80%.

The analysis of feedstocks in terms of land use efficiency, i.e. yields expressed as bio-fuel equivalent per ha, within the 1st generation feedstock group indicates the relative advantage of sugar crops (sugar beet and sweet sorghum) for bioethanol production over oil crops for biodiesel production (Map 7 in Annex 4). A similar map was produced for 2nd generation feedstocks (Map 8 in Annex 4). While showing a mixed picture due to the heterogeneity of biophysical resources, it should nevertheless be noted that in many areas potential energy yields of best herbaceous and woody lignocellulosic feedstocks do not differ substantially.

31 “Well-to-wheels” refers to the complete chain of fuel production and use, including feedstock production, transport to the refinery, conversion to final fuel, transport to refueling stations, and final vehicle tailpipe emissions.

51

0

20

40

60

80

100

120

140

160

180

at bel

de dk es fi fr gr ie it nl pt se uk ch cy cz ee hu lt lv pl si sk bg ro ukr

GJ

/ hec

tare

Best of 1st generation Best of all (mainly 2nd generation)

Figure 12. Average potential bio-fuel energy yields of 1st and 2nd generation bio-fuel feedstocks Current production of bio-fuels in the EU is clearly dominated by biodiesel, representing 80% of total bio-fuel production. Rapeseed is the main feedstock for biodiesel production, providing 84% of total bio-fuel feedstock, followed by sunflower with 13%. Biodiesel production has grown strongly in the past years in response to the EU’s bio-fuel directive. There are however a number of good reasons why biodiesel does not represent Europe’s most effective and sustainable path toward achieving a higher share of bio-fuels in total transport fuels. As only a limited fraction of the feedstock plant is used for producing vegetable oil, domestic feedstocks for biodiesel are relatively land inefficient in terms of energy output per hectare. Also, due to high fertilizer and pesticide requirements rapeseed performs rather poorly with regard to GHG emission reductions.

As shown in Map 7 (see Annex 4) land productivity (expressed in GJ/ha) for oil crops is generally inferior to cereals and sugar crops. Energy yields of oil crops are typically ranging between 20% and 40% of 2nd generation bio-fuel feedstock yields (Map 6, Annex 4).

Land use and feedstock scenarios – 1st versus 2nd generation bio-fuel production chain

For the available arable and pasture land areas in 2030, bio-fuel feedstock production potentials have been estimated for two scenarios. They consider alternatives for land availability, differentiate by feedstock type, and portray two extremes with respect to choices of 1st and 2nd generation feedstocks. The ‘1st generation only’ scenario considers only conventional (1st generation) bio-fuel feedstocks (cereals, oil crops, sugar crops) and conversion chains, while the ‘2nd generation’ scenario also considers herbaceous and woody lignocellulosic plants and selects the best-yielding feedstock (in energy terms) for each land unit. In this case, significantly higher energy output per hectare implies in most areas cultivation of 2nd generation energy feedstocks.

Table 20 summarizes potential bio-fuel feedstock production (in EJ bio-fuel equivalent). The main assumptions are: First, projected food, feed and livestock product demand is satisfied and the aggregate level of self-sufficiency of the base period 2000-02 is maintained, i.e., no additional net imports of food and feed products are caused by bio-fuel production. Second, all then available remaining agricultural land is used for bio-fuel feedstock production. Third, evaluation of feedstocks is done with current conversion technologies and efficiencies.

Results highlight: (i) the importance of Eastern Europe, (ii) the potential contribution from the Ukraine, and (iii) substantial differences in the potential energy generated through respectively the 1st and 2nd generation bio-fuel feedstock production chains. In the EU more than half of the

52

bio-fuel potential is located in the New Member States, especially when considering 2nd generation feedstock conversion pathways.

Table 20. Potential bio-fuel energy production in Europe1 by 2030 for different scenarios

EJ bio-fuel equivalent Scenario: FEEDSTOCK types

1st generation only (oil crops, sugar, cereals)

2nd generation (mostly herbaceous and woody

lignocelluloic) EU15+ EU12 Ukr Total EU15+ EU12 Ukr Total

ARABLE LAND Baseline 1.5 2.1 2.3 5.9 2.3 3.2 3.4 8.9 Low 1.3 2.1 2.3 5.6 2.0 3.2 3.4 8.6 High 1.8 2.5 2.6 6.9 2.8 3.8 3.8 10.4 PASTURE LAND 1st generation only 2nd generation Baseline and Low Not used Not used

Land

ava

ilabi

lity

Scen

ario

High Not used 1.3 1.0 0.8 3.1 1 EU15+ includes EU15 and Norway, Switzerland. EU12 refers to New Member States. Total is the sum of results for EU15+, EU12 and Ukraine.

On the high end, the scenarios indicate a bio-fuel potential of some 13 EJ of the estimated available cultivated and pasture land for Europe (including Ukraine). According to the PRIMES baseline scenario (DG TREN 2006) primary energy demand for the EU25 by 2030 is projected to be 79 EJ. Of this total, transport fuel use is 17.6 EJ increasing by about 30% from a current 13.4 EJ.

For the EU27+ the analysis shows that if all the agricultural land potentially available for bio-fuels would be used for cultivation of the most energy-efficient bio-fuel feedstocks, then by 2030 up to 50% of projected transport fuel consumption could be produced within the EU. Applying 1st generation technologies on arable land only, around 20% of projected transport fuel consumption could be covered from domestic sources.

The relatively low contributions of pasture areas compared to potentials from arable land are due to the large extents of land that will continue to be required for animal grazing and fodder production32 and due to areas reserved for nature conservation and environmental protection.

Ukraine is contributing about a third of Europe’s overall bio-fuel feedstock potential, which is explained by the large and productive arable land area and relatively high possible yield increases. Anticipated doubling of yields in the baseline scenario combined with comparatively high potential feedstock yields explains the substantial biophysical potential of Ukraine. Projected yield increases may be constrained due to degrading social and demographic conditions in rural areas, lack of capital for investments, and uncertain policy and economic conditions.

32 Scenarios foresee no further replacement of feed from grazing with feed crops.

53

Chapter 8. Crop residues for bio-fuel production

Agricultural residues can provide an additional contribution of biomass to ethanol production. Agricultural residues include straw, stubble, stalk, cob, husks and peelings. Residues from fruit trees and nuts include fibres, husks and shells. The availability of these residues for energy purposes is restricted by technical, environmental and economic factors rendering it difficult to precisely quantify. Crop residues fulfil also important ecosystem services essential to maintenance of soil fertility and erosion protection.

Agricultural residues, processing by-products and wastes are readily available and can contribute significant amounts of feedstocks during the introduction phase for second generation bio-fuel production chains.

8.1 Methodology

This study produced estimates of the potential availability and use of agricultural residues for bio-fuel production. Factors that determine the amount of residues include crop type and yields, the biomass ratio of crop residue to crop main produce (RPR), and percentage of residues removed from the field and potentially used as bio-fuel feedstock.

Cultivars of higher yielding varieties aim at a higher share of the total biomass to be stored in the harvested parts. As a consequence the relative amount of crop residues (the RPR factor) is less than for lower yielding breeds. For major cereals, potatoes and sugar beet a linear relationship is assumed between the upper and lower bounds of RPR values relative to the yield of the main produce. RPR estimates for individual crops are derived from literature (Koopmans and Koppejan, 1998; Ryan and Openshaw, 1991; Jölli and Giljum, 2005).

The maximum amount of crop residues that can be removed from the field without significantly affecting soil fertility is debated. Some consider crop residues as currently unused waste material and make a strong case for its use for bio-fuel production (e.g. Sommerville, 2006). Others perceives crop residues as a valuable resource that provides irreplaceable environmental services (Smil, 1999) and argue removal of crop residues would exacerbate risks of soil erosion by water and wind, deplete soil organic matter, degrade soil quality, increase non-point source pollution, decrease agronomic productivity, and reduce crop yields per unit input of fertilizers and water (Lal, 2007). Moreover the importance of retaining residues on fields depends largely upon specific local conditions.

For the calculations we applied a widely adopted assumption that up to 50% of crop residues could be removed without significant impacts on soil fertility or soil erosion (e.g. Lynd et.al, 2002). For vegetables, roots and tubers a use factor of only 5% is assumed due to bulkiness and high water content of residues.

Estimates of the available amount of crop residues are based on FAOSTAT production data by country and year. FAOSTAT data provide harvested quantities of individual agricultural commodities in fresh weight. A conversion factor of fresh weight to dry weight and the respective RPR values were applied to sum up the amounts of crop residues. Table 21 summarizes conversion factors for commodities where RPR depends on yield. For the remaining commodities a constant RPR is assumed (Appendix 4).

54

For each country and year between 1985 and 2002 agricultural residues potentially available as bio-fuel feedstock was calculated. For the scenarios to 2030 projected agricultural residues are linked to estimated quantities of domestic crop production. The assumed yield increases of agricultural production lead to decreases of the average RPR value between the base period 2000-02 and 2030.

Table 21. Conversion factors used to estimate crop residues for selected major crops

%water content

USE factor

lower yield boundary

higher yield

boundary

RPR at lower yields

RPR at higher yields

perc. perc tons/ha tons/ha ratio ratio

Wheat 15% 50% 1.50 9.00 1.75 0.70 Rice, Paddy 15% 50% 2.50 7.00 2.00 1.00 Barley 15% 50% 1.00 7.00 2.50 0.90 Maize 15% 50% 1.50 9.00 2.00 1.00 Rye 15% 50% 1.00 6.00 2.50 1.50 Oats 15% 50% 1.00 6.00 2.50 1.50 Millet 15% 50% 0.40 2.50 4.00 2.00 Sorghum 15% 50% 1.00 6.00 4.00 1.25 Potatoes 65% 5% 7.50 45.00 1.00 0.50 Sugar Beets 75% 50% 10.00 75.00 0.70 0.40 Soyabeans 15% 50% 0.50 3.00 3.50 1.50 Sunflower Seed 40% 50% 0.50 3.00 3.50 1.75 Rapeseed 40% 50% 1.00 3.50 3.50 2.00

8.2 Results The amount of crop residues generated during food and feed harvesting was calculated for individual countries for the period between 1985 and 2002 and for the different scenarios up to 2030 (Table 22 and 23).

According to this estimation, in the base period 2000-02 a total of 457 million tons agricultural residues (in dry matter) were produced in EU27 of which some 215 million tons (applying use factors as shown in Table 21) could potentially be used as feedstock for energy production. This would leave about 242 million tons of residues on the fields to fulfil ecosystem functions such as erosion control and soil fertility enhancement. About three-quarters of available agricultural residues are generated from cereals. Assumed yield increases and consequent lowering of RPR values will gradually decrease the amount of agricultural residues towards 2030 (see Table 22).

Currently agricultural residues are mainly used for heating and biogas production. The lower heating values (LHV or net calorific value) of most agricultural residues vary between 15 and 17 GJ/tonne. The LHV describes the amount of heat released during combustion. Table 23 presents estimated LHV of available agricultural residues by country.

Beyond heating values of feedstocks, the energy content in terms of bio-fuel equivalent depends on technologies and processes used to convert agricultural residues to bio-fuels. It will generally

55

be significantly lower than the LHV. Assuming for agricultural residues a similar bio-fuel equivalent as for herbaceous lignocellulosic plants (9.3 GJ per ton biomass, dry weight), in the base period 2000-02 a total energy of 2.3 EJ (bio-fuel equivalent) could be derived from agricultural residues of food and feed crops in Europe. More than 60% of these residues were produced in EU15. By 2030 this potential will have decreased to about 1.7 EJ (bio-fuel equivalent, baseline scenario).

In the EU15 the analysis shows that agricultural residues could provide a significant additional feedstock for bio-fuel production, particularly in the period before 2020. For 2030, the EU15 baseline scenario results for 1st generation bio-fuel feedstocks amount to a potential of 1.5 EJ compared to a potential of 1.2 EJ from agricultural residues (see Table 20). In the EU12 and Ukraine, with rich and currently less intensively used land resources, the possible contribution of agricultural residues to bio-fuel production is relatively small compared to the potential of agricultural land that could be freed-up for bio-fuel feedstock production.

Table 22. Agricultural residues potentially available for bio-fuel production

Estimates based on FAOSTAT data million tons DM

Baseline scenario million tons DM

1985 1990 1995 2000-02 2010 2020 2030 EU-15 Cereals 114 107 95 106 Other crops 41 48 44 47 CROPS total 155 155 140 153 149 140 130 EU-12 Cereals 53 53 55 51 Other crops 11 10 10 10 CROPS total 64 63 66 61 52 44 36 Ukraine Cereals n.d. n.d. 25 26 Other crops n.d. n.d. 8 6 CROPS total n.d. n.d. 33 32 26 21 16

56

Table 23. Agricultural residues of food and feed crops potentially available for bio-fuel use and associated heating values

Agricultural residues (1000 tons dry matter)

Energy content (in PJ)

2000 2030 2000 2030 Austria 3354 2977 54 48 BELUX 1534 1058 25 17 Germany 28468 22719 455 364 Denmark 5193 4453 83 71 Spain 24562 22045 393 353 Finnland 3300 3004 53 48 France 39657 34836 635 557 Greece 6644 5376 106 86 Ireland 963 1016 15 16 Italy 20520 15667 328 251 Netherlands 1349 687 22 11 Portugal 1916 1355 31 22 Sweden 3858 3598 62 58 United Kingdom 12140 10866 194 174 EU15 153387 129772 2455 2075 Switzerland 870 691 14 11 Norway 937 850 15 14 Cyprus 191 69 3 1 Czech Republic 5489 3342 88 53 Estonia 586 280 9 4 Hungary 9433 5999 151 96 Lithuania 2196 1175 35 19 Latvia 861 185 14 3 Malta 11 1 0 0 Poland 21536 12393 345 198 Slovenia 422 201 7 3 Slovakia 2498 1678 40 27 Bulgaria 5069 2602 81 42 Romania 12894 7713 206 123 EU12 61158 35816 979 570 Ukraine 31446 15665 503 251

57

References

Batjes, N.H., G. Fischer, F.O. Nachtergaele, V.S. Stolbovoi, and H.T. van Velthuizen (1997): Soil Data Derived from WISE for Use in Global and Regional AEZ Studies. FAO/IIASA/ISRIC Interim Report IR-97-025, International Institute for Applied systems Analysis, Laxenburg, Austria.

Beck, C., J. Grieser and B. Rudolf (2004): A New Monthly Precipitation Climatology for the Global Land Areas for the Period 1951 to 2000 (published in Climate Status Report 2004, pp. 181 - 190, German Weather Service, Offenbach, Germany)

Bruinsma J (Editor) (2003): World agriculture towards 2015/2030; an FAO perspective. 2003. chapter 11.3.2 page 311, second paragraph.

Commission of the European Communities (CEC) (2003): Directive 2003/30/EC of the European Parliament and of the Council of 8 May 2003 on the promotion of the use of bio-fuels and other renewable fuels for transport (OJEU L123 of 17 May 2003).

Commission of the European Communities (CEC) (2005): Biomass Action Plan, SEC(2005) 1573, Communication from the Commission, COM(2005)628 final, Brussels, 25 October 2003.

Commission of the European Communities (CEC) (2006): An EU strategy for Bio-fuels, Communication from the Commission, COM(2006)34 final. Brussels.

Commission of the European Communities (CEC) (2006b): Report from the Commission to the Council - on the review of the energy crops scheme. Communication from the Commission, COM (2006) 500 final, Brussels

CGIAR-CSI (2006): NASA Shuttle Radar Topographic Mission (SRTM). The SRTM data is available as 3 arc second (approx. 90m resolution) DEMs. The dataset is available for download at: http://srtm.csi.cgiar.org/

European Biodiesel Board (2007): http://www.ebb-eu.org/stats.php

European Environment Agency (EEA) (2006): Corine Land Cover 2000. Available at: http://terrestrial.eionet.europa.eu/CLC2000.

ESB (2004): "European Soil Database (v 2.0), European Soil Bureau Network and the European Commission, EUR 19945 EN, March 2004"

Fischer G., van Velthuizen H.T., Shah M., Nachtergaele F.O. (2002): Global Agro-Ecological Zones Assessment: Methodology and Results, IIASA Research IR-02-02. International Institute for Applied systems Analysis, Laxenburg, Austria.

Fischer G, Prieler S., van Velthuizen H.T. (2005): Biomass potentials of miscanthus, willow and poplar: results and policy implications for Eastern Europe, Northern and Central Asia, Biomass and Bioenergy 28 page 119-132, Elsevier.

Fischer G, Shah M., van Velthuizen H.T., Nachtergaele F (2006). Agro-Ecological Zones Assessment. IIASA Reprint RP-06-003, from Encyclopedia of Life Support Systems (EOLSS), EOLSS Publishers, Oxford, UK [2005]

IEA (2006): Energy Technology Perspectives 2006. Scenarios & Strategies to 2050. OECD/IEA. Paris

IEA (2006b): Bio-fuels for transport. An international perspective. OECD/IEA Paris.

58

Joint Research Centre (JRC) (2006): Global 2000 Land Cover (GLC2000). Available at: http://www-gvm.jrc.it/glc2000/defaultGLC2000.htm.

Jölli D., Giljum S. (2005): Unused biomass extraction in agriculture, forestry and fishery. Sustainable Europe Research Institute (SERI). SERI studies. www.seri.at

Koopmans A., Koppejan J (1998): Agricultural and forest residues – generation, utilization and availability. Paper presented at the Regional Consultation on Modern Applications of Biomass Energy, 6-10 January 1997, Kuala Lumpur, Malaysia. http://www.fao.org/docrep/006/AD576E/ad576e00.pdf

Lal R., Pimentel D. (2007): Bio-fuels from crop residues. Soil & Tillage Research 93, p.237-238.

Lutz Christian, Bernd Meyer, Marc Ingo Wolter (2005): MOSUS Scenarios and Simulation Results. Gesellschaft für wirtschaftliche Strukturforschung mbH (GWS), Germany Link: www.gws-os.de

Lynd L.R., Haiming J., Joseph G.M., Charles E.W.1, and Bruce D. (2002): “Bioenergy: Background, Potential, and Policy”, Policy briefing prepared for the Center for Strategic and International Studies. http://i-farmtools.org/ref/Lynd_et_al_2002.pdf (pg. 53)

Meyer B., Christian Lutz, Marc Ingo Wolter (2004): Economic growth of the EU and Asia. A first forecast with the global econometric model GINFORS. Paper presented at the 1st KEIO-UNU-JFIR Panel Meeting. Economic Development and Human Security. Tokyo, 13.-14. February, 2004.

Mitchell Timothy D. and Jones Philip D. (2005): An improved method of constructing a database of monthly climate observations and associated high-resolution grids. International Journal of Climatology Vol.25 Issue 6, p.693-712.

New, M., Lister, D., Hulme, M. and Makin, I., 2002: A high-resolution data set of surface climate over global land areas. Climate Research 21.

Ryan Paul, Openshaw Keith, (1991): Assessment of Biomass Energy Resources – A Discussion on its Need and Methodology. The World Bank Industry and Energy Department, Paper No. 48. http://ideas.repec.org/p/fth/wobaie/48.html

Sommerville C. (2006): The billion-ton bio-fuel vision. Science 312, 1277.

USGS (2002): GTOPO30 – Global 30 arc second elevation data. U.S. Geological Survey, National Mapping Division, EROS Data Center; for download available at: http://edcdaac.usgs.gov/gtopo30/gtopo30.html

59

APPENDIX 1. Harmonization of a Pan-European land use database – Reclassification of original land use maps to 12 aggregate land categories LUCL12

Label in original source map LUCL12 Value in orig. map

AREA* [mio.ha]

Original map: CLC2000 (used for EU27) Continuous urban fabric 12 1 0.6 Discontinuous urban fabric 12 2 13.3 Industrial or commercial units 12 3 2.0 Road and rail networks and associated land 12 4 0.2 Port areas 12 5 0.1 Airports 12 6 0.3 Mineral extraction sites 12 7 0.6 Dump sites 12 8 0.1 Construction sites 12 9 0.1 Green urban areas 12 10 0.3 Sport and leisure facilities 12 11 0.8 Non-irrigated arable land 5 12 107.3 Permanently irrigated land 5 13 3.2 Rice fields 5 14 0.6 Vineyards 6 15 3.9 Fruit trees and berry plantations 6 16 2.5 Olive groves 6 17 4.0 Pastures 8 18 37.6 Annual crops associated with permanent crops 7 19 1.0 Complex cultivation patterns 7 20 25.9 Land principally occupied by agriculture, with significant areas of natural vegetation 7 21 20.7

Agro-forestry areas 7 22 3.2 Broad-leaved forest 1 23 45.5 Coniferous forest 1 24 64.9 Mixed forest 1 25 29.2 Natural grasslands 2 26 11.9 Moors and heathland 4 27 9.1 Sclerophyllous vegetation 4 28 9.8 Transitional woodland-shrub 4 29 24.0 Beaches, dunes, sands 10 30 0.4 Bare rocks 10 31 2.1 Sparsely vegetated areas 10 32 3.9 Burnt areas 10 33 0.1 Glaciers and perpetual snow 11 34 0.2 Inland marshes 3 35 1.2 Peat bogs 3 36 7.2 Salt marshes 3 37 0.3 Salines 3 38 0.1 Intertidal flats 3 39 1.0 Water courses 9 40 1.1 Water bodies 9 41 10.0 Coastal lagoons 9 42 0.5 Estuaries 9 43 0.3

* Area refers to the extent of the particular class in the harmonized Pan-European land use map.

60

APPENDIX 1. Harmonization of a Pan-European land use database – Reclassification of original land use maps into 12 aggregate land categories LUCL12

Label in original source map LUCL12 Value in orig. map

AREA [mio.ha]

Original map: CLC1990 (used for Switzerland) Urban fabric 12 11 0.1 Industrial 12 12 0.0 Mine 12 13 0.0 Artificial non-agricultural vegetated areas 12 14 0.0 Arable land 5 21 0.5 Permanent crops 6 22 0.0 Pastures 8 23 0.5 Heterogeneous agricultural areas 7 24 0.0 Forests 1 31 1.1 Shrub and/or herbaceous vegetation associations 4 32 0.8 Open spaces with little or no vegetation 10 33 0.7 Inland wetlands 3 41 0.0 Inland waters 9 51 0.1 Original map: GLC2000 (used for rest) Tree Cover, broadleaved, deciduous, closed 1 2 34.7 Tree Cover, broadleaved, deciduous, open 1 3 2.8 Tree Cover, needle-leaved, evergreen 1 4 55.1 Tree Cover, needle-leaved, deciduous 1 5 0.0 Tree Cover, mixed leaf type 1 6 40.2 Mosaic: Tree Cover / Other natural vegetation 1 9 2.7 Tree Cover, burnt 10 10 0.1 Shrub Cover, closed-open, evergreen 4 11 6.7 Shrub Cover, closed-open, deciduous 4 12 13.2 Herbaceous Cover, closed-open 2 13 15.8 Sparse herbaceous or sparse shrub cover 10 14 13.4 Regularly flooded shrub and/or herbaceous cover 3 15 11.8 Cultivated and managed areas 5 16 94.9 Mosaic: Cropland / Tree Cover / Other natural vegetation 7 17 20.7 Mosaic: Cropland / Shrub and/or grass cover 7 18 21.0 Bare Areas 10 19 5.1 Water Bodies 9 20 13.0 Snow and Ice 11 21 1.6 Artificial surfaces and associated areas 12 22 1.4 Irrigated Agriculture 5 23 0.4

61

APPENDIX 2. Administrative units used in the REFUEL project Code No. Name SQKM UNITED-KINGDOM ukc 22901 North East 8601

ukd 22902 North West (incl. Merseyside) 14164

uke 22903 Yorkshire & The Humber 15438

ukf 22904 East Midlands 15641 ukg 22905 West Midlands 13013 ukh 22906 Eastern 19128 uki 22907 London 1579 ukj 22908 South East 19109 ukk 22909 South West 23883 ukl 22910 Wales 20800 ukm 22911 Scotland 79195 ukn 22912 Northern Ireland 14147 AUSTRIA at11 1111 Burgenland 3966 at12 1112 Niederösterreich 19194 at13 1113 Wien 415 at21 1121 Kärnten 9545 at22 1122 Steiermark 16415 at31 1131 Oberösterreich 11991 at32 1132 Salzburg 7163 at33 1133 Tirol 12640 at34 1134 Vorarlberg 2601 BELGIUM

be10 25510 Région de Bruxelles-Capitale/Brussels Hoofdstedelijk Gewest

163

be21 25521 Prov. Antwerpen 2875 be22 25522 Prov. Limburg (B) 2427 be23 25523 Prov. Oost-Vlaanderen 3008 be24 25524 Prov. Vlaams Brabant 2119 be25 25525 Prov. West-Vlaanderen 3168 be31 25531 Prov. Brabant Wallon 1097 be32 25532 Prov. Hainaut 3814 be33 25533 Prov. Liège 3856 be34 25534 Prov. Luxembourg (B) 4460 be35 25535 Prov. Namur 3676 DENMARK dk00 5400 Denmark 43066 ESTONIA ee00 6300 Estonia 45306 FINLAND fi13 6713 Itä-Suomi 85179 fi18 6718 Etelä-Suomi 45034 fi19 6719 Länsi-Suomi 64586 fi1a 6730 Pohjois-Suomi 141526 fi20 6720 Åland 1466 FRANCE fr10 6810 Île de France 12068 fr21 6821 Champagne-Ardenne 25720 fr22 6822 Picardie 19506 fr23 6823 Haute-Normandie 12355

fr24 6824 Centre 39531 fr25 6825 Basse-Normandie 17758 fr26 6826 Bourgogne 31753 fr30 6830 Nord - Pas-de-Calais 12445 fr41 6841 Lorraine 23669 fr42 6842 Alsace 8327 fr43 6843 Franche-Comté 16303 fr51 6851 Pays de la Loire 32376 fr52 6852 Bretagne 27465 fr53 6853 Poitou-Charentes 25968 fr61 6861 Aquitaine 41803 fr62 6862 Midi-Pyrénées 45601 fr63 6863 Limousin 17055 fr71 6871 Rhône-Alpes 44965 fr72 6872 Auvergne 26171 fr81 6881 Languedoc-Roussillon 27766

fr82 6882 Provence-Alpes-Côte d'Azur 31836

fr83 6883 Corse 8727 GERMANY de11 7911 Stuttgart 10557 de12 7912 Karlsruhe 6918 de13 7913 Freiburg 9403 de14 7914 Tübingen 8917 de21 7921 Oberbayern 17529 de22 7922 Niederbayern 10326 de23 7923 Oberpfalz 9691 de24 7924 Oberfranken 7230 de25 7925 Mittelfranken 7243 de26 7926 Unterfranken 8530 de27 7927 Schwaben 9992 de30 7930 Berlin 890 de41 7941 Brandenburg - Nordost 15592 de42 7942 Brandenburg - Südwest 14062 de50 7950 Bremen 397 de60 7960 Hamburg 748 de71 7971 Darmstadt 7444 de72 7972 Gießen 5382 de73 7973 Kassel 8293

de80 7980 Mecklenburg-Vorpommern 23071

de91 7991 Braunschweig 8119 de92 7992 Hannover 9065 de93 7993 Lüneburg 15594 de94 7994 Weser-Ems 14968 dea1 7951 Düsseldorf 5293 dea2 7952 Köln 7364 dea3 7953 Münster 6919 dea4 7954 Detmold 6525 dea5 7955 Arnsberg 8013 deb1 7956 Koblenz 8097 deb2 7957 Trier 4929 deb3 7958 Rheinhessen-Pfalz 6830 dec0 7959 Saarland 2571 ded1 7961 Chemnitz 6120 ded2 7962 Dresden 7943

62

ded3 7963 Leipzig 4399 dee1 7964 Dessau 4307 dee2 7965 Halle 4445 dee3 7966 Magdeburg 11799 def0 7967 Schleswig-Holstein 15679 deg0 7968 Thüringen 16201 GREECE

gr11 8411 Anatoliki Makedonia, Thraki 14196

gr12 8412 Kentriki Makedonia 19179 gr13 8413 Dytiki Makedonia 9470 gr14 8414 Thessalia 14049 gr21 8421 Ipeiros 9154 gr22 8422 Ionia Nisia 2302 gr23 8423 Dytiki Ellada 11315 gr24 8424 Sterea Ellada 15558 gr25 8425 Peloponnisos 15511 gr30 8430 Attiki 3811 gr41 8441 Voreio Aigaio 3845 gr42 8442 Notio Aigaio 5304 gr43 8443 Kriti 8346 IRELAND

ie01 10401 Border, Midlands and Western 33251

ie02 10402 Southern and Eastern 36999 ITALY itc1 10601 Piemonte 25298 itc2 10602 Valle d'Aosta 3253 itc3 10603 Liguria 5415 itc4 10604 Lombardia 23403

itd1 10605 Prov. Autonoma Bolzano 7400

itd2 10606 Provincia Autonoma Trento 6189

itd3 10607 Veneto 18231 itd4 10608 Friuli-Venezia Giulia 7863 itd5 10609 Emilia-Romagna 22156 ite1 10610 Toscana 22981 ite2 10611 Umbria 8460 ite3 10612 Marche 9714 ite4 10613 Lazio 17215 itf1 10614 Abruzzo 10791 itf2 10615 Molise 4442 itf3 10616 Campania 13605 itf4 10617 Puglia 19365 itf5 10618 Basilicata 9997 itf6 10619 Calabria 15072 itg1 10620 Sicilia 25728 itg2 10621 Sardegna 24095 LUXEMBOURG lu00 25600 Luxembourg 2596 NETHERLANDS nl11 15011 Groningen 2406 nl12 15012 Friesland 3537 nl13 15013 Drenthe 2679 nl21 15021 Overijssel 3421 nl22 15022 Gelderland 5137 nl23 15023 Flevoland 1563

nl31 15031 Utrecht 1449 nl32 15032 Noord-Holland 2877 nl33 15033 Zuid-Holland 3229 nl34 15034 Zeeland 1927 nl41 15041 Noord-Brabant 5083 nl42 15042 Limburg (NL) 2209 PORTUGAL pt11 17411 Norte 21284 pt15 17415 Algarve 4993 pt16 17416 Centro (PT) 28198 pt17 17417 Lisboa 3001 pt18 17418 Alentejo 31593 SPAIN es11 20311 Galicia 29569 es12 20312 Principado de Asturias 10602 es13 20313 Cantabria 5320 es21 20321 Pais Vasco 7234

es22 20322 Comunidad Foral de Navarra 10390

es23 20323 La Rioja 5045 es24 20324 Aragón 47720 es30 20330 Comunidad de Madrid 8028 es41 20341 Castilla y León 94225 es42 20342 Castilla-la Mancha 79462 es43 20343 Extremadura 41634 es51 20351 Cataluña 32114 es52 20352 Comunidad Valenciana 23256 es53 20353 Illes Balears 4991 es61 20361 Andalucia 87602 es62 20362 Región de Murcia 11313 SWEDEN se01 21001 Stockholm 7051 se02 21002 Östra Mellansverige 43235 se04 21004 Sydsverige 14377 se06 21006 Norra Mellansverige 72031 se07 21007 Mellersta Norrland 77173 se08 21008 Övre Norrland 165087 se09 21009 Småland med öarna 35974 se0a 21010 Västsverige 34508 CYPRUS cy00 5000 Cyprus 9248 CZECH-REPUBLIC cz01 16701 Praha 496 cz02 16702 Strední Cechy 11015 cz03 16703 Jihozápad 17613 cz04 16704 Severozápad 8639 cz05 16705 Severovýchod 12453 cz06 16706 Jihovýchod 13982 cz07 16707 Strední Morava 9103 cz08 16708 Moravskoslezko 5567 HUNGARY hu10 9710 Közép-Magyarország 6917 hu21 9721 Közép-Dunántúl 11239 hu22 9722 Nyugat-Dunántúl 11211 hu23 9723 Dél-Dunántúl 14160 hu31 9731 Észak-Magyarország 13450 hu32 9732 Észak-Alföld 17711 hu33 9733 Dél-Alföld 18332

63

LATVIA lv00 11900 Latvia 64603 LITHUANIA lt00 12600 Lithuania 64890 MALTA mt00 13400 Malta 316 POLAND pl11 17311 Lódzkie 18215 pl12 17312 Mazowieckie 35604 pl21 17321 Malopolskie 15177 pl22 17322 Slaskie 12325 pl31 17331 Lubelskie 25119 pl32 17332 Podkarpackie 17848 pl33 17333 Swietokrzyskie 11677 pl34 17334 Podlaskie 20183 pl41 17341 Wielkopolskie 29837 pl42 17342 Zachodniopomorskie 22444 pl43 17343 Lubuskie 13976 pl51 17351 Dolnoslaskie 19936 pl52 17352 Opolskie 9404 pl61 17361 Kujawsko-Pomorskie 17970 pl62 17362 Warminsko-Mazurskie 24025 pl63 17363 Pomorskie 18167 SLOVAKIA sk01 19901 Bratislavský kraj 2052 sk02 19902 Západné Slovensko 14987 sk03 19903 Stredné Slovensko 16239 sk04 19904 Východné Slovensko 15738 SLOVENIA si00 19800 Slovenia 20274 BULGARIA bg01 2711 BG1 10266 bg02 2712 BG2 18221 bg03 2713 BG3 20015 bg04 2721 BG4 20296 bg05 2722 BG5 27512 bg06 2723 BG6 14723 ROMANIA ro01 18301 RO1 37312 ro02 18302 RO2 35751 ro03 18303 RO3 34477 ro04 18304 RO4 28796 ro05 18305 RO5 32025 ro06 18306 RO6 34167 ro07 18307 RO7 34111 ro08 18308 RO8 1805 NORWAY no01 16201 NO1 5371 no02 16202 NO2 52584 no03 16203 NO3 36568 no04 16204 NO4 25589 no05 16205 NO5 49306 no06 16206 NO6 41182 no07 16207 NO7 112907 SWITZERLAND ch01 21101 Region_lemanique 8376 ch02 21102 Espace_Mittelland 10066 ch03 21103 Nordwestschweiz 1960

ch04 21104 Zuerich 239 ch05 21105 Ostschweiz 11352 ch06 21106 Zentralschweiz 5972 ch07 21107 Ticino 2811 UKRAINE CA 23001 Kharkivs'ka 31434 CC 23002 Chernivets'ka 8240 CE 23003 Khersons'ka 25391 CH 23004 Chernihivs'ka 31875 CK 23005 Cerkas'ka 20924 CM 23006 Khmel'nyts'ka 20615 DN 23007 Dnipropetrovs'ka 31880 DO 23008 Donets'ka 26433 IF 23009 Ivano-Frankivs'ka 13846 KI 23010 Kirovohrads'ka 24516 KR 23011 Respublika Krym 25755 KY 23012 Kyyivs'ka 28914 LU 23013 Luhans'ka 26658 LV 23014 L'vivs'ka 21454 MY 23015 Mykolayivs'ka 23983 OD 23016 Odes'ka 33256 PO 23017 Poltavs'ka 28667 RI 23018 Rivnens'ka 20019 SU 23019 Sums'ka 23799 TE 23020 Ternopil's'ka 13846 VI 23021 Vinnyts'ka 26464 VO 23022 Volyns'ka 19945 ZK 23023 Zakarpats'ka 12593 ZP 23024 Zaporiz'ka 27099 ZY 23025 Zhytomyrs'ka 29805 Other Nations AD 600 Andorra 465 CL 90002 Channel Islands 200 FO 6400 Faeroe Islands 1402 GI 90003 Gibraltar 6 IS 9900 Island 102908 IM 26400 Isle of Man 575 LI 12500 Liechtenstein 161 MC 14000 Monacco 2 SM 19200 San Marino 62 SJ 26000 Jan Mayen 377 VC 90004 Vatican City 0 AL 300 Albania 28543 BY 5700 Belarus 206122

BA 8000 Bosnia and Herzegovina 51129

HR 9800 Croatia 56458

MK 15400 Macedonia,The Fmr Yug Rp 25540

MD 14600 Moldova, Republic of 33139 CS 18600 Serbia and Montenegro 103107 RU 18500 Russian Federation 1916346 TR 22300 Turkey 451844

64

APPENDIX 3. Compilation of a new terrain slope database based on SRTM data Under an agreement with the National Aeronautics and Space Administration (NASA) and the National Geospatial Intelligence Agency (NGA) of the Department of Defense, the U.S. Geological Survey (USGS) is now distributing elevation data from the Shuttle Radar Topography Mission (SRTM). The SRTM is a joint project between NASA and NGA to map the Earth’s land surface in three dimensions at a level of detail unprecedented for such a large area. Flown aboard the NASA Space Shuttle Endeavour February 11-22, 2000, the SRTM successfully collected data from over 80 percent of the Earth’s land surface, for most of the area between 60º N and 56º S latitude.

The data currently being distributed by NASA/USGS (finished product) contains “no-data” holes where water or heavy shadow prevented the quantification of elevation. These are generally small holes, which nevertheless render the data less useful, especially in fields of hydrological modelling. Dr. Andrew Jarvis of the CIAT Land Use project, in collaboration with Dr. Robert Hijmans and Dr. Andy Nelson, have further processed the original DEMs to fill in these no-data voids. This involved the production of vector contours, and the re-interpolation of these derived contours back into a raster DEM. These interpolated DEM values were then used to fill in the original no-data holes within the SRTM data.

The DEM files have been mosaiced into a seamless global coverage, and are available for download as 5˚ x 5˚ tiles, in geographic coordinate system - WGS84 datum. The available data cover a raster of 24 rows by 72 columns of 5˚ x 5˚ latitude/longitude tiles, from latitude 60º N to 56º S.

These processed SRTM data, with a resolution of 3 arc second (approximately 90m at the equator), i.e. 6000 rows by 6000 columns for each 5˚ x 5˚ tile, have been used for calculating: (i) terrain slope gradients for each 3 arc-sec grid cell; (ii) aspect of terrain slopes for each 3 arc-sec grid cell; (iii) terrain slope class by 3 arc-sec grid cell; and (iv) aspect class of terrain slope by 3 arc-sec grid cell. Products (iii) and (iv) were then aggregated to provide distributions of slope gradient and slope aspect classes by 30 arc-sec grid cell and for a 5’x5’ latitude/longitude grid as used in global AEZ.

The computer algorithm used to calculate slope gradient and slope aspect operates on sub-grids of 3 by 3 grid cells, say grid cells A to I:

A B C D E F G H I

SRTM data are stored in 5˚x5˚ tiles33. When E falls on a border row or column (i.e., rows or columns 1 or 6000 of a tile) the required values falling outside the current tile are filled in from the neighbouring tiles.

To calculate terrain slope for grid cell E, the algorithm proceeds as follows:

1) If the altitude value at E is ‘no data’ then both slope gradient and slope aspect are set to ‘no data’.

2) Replace any ‘no data’ values in A to D and F to I by the altitude value at E. 3) Let Px, Py and Pz denote respectively coordinates of grid point P in x direction (i.e.

longitude in our case), y direction (i.e. latitude in our application), and z in vertical direction (i.e., altitude), then calculate partial derivatives (dz/dx) and (dz/dy) from:

(dz/dx) = - ((Az-Cz) + 2· (Dz-Fz) + (Gz-Iz)) / (8·size_x)

(dz/dy) = ((Az-Gz) + 2· (Bz-Hz) + (Cz-Iz)) / (8·size_y)

33 For the globe the computer program processes 36 million sub-grids, in total 32.4 billion sub-grids are considered.

65

When working with a grid in latitude and longitude, then size_y is constant for all grid cells. However, size_x depends on latitude and is calculated separately for each row of a tile.

The slope gradient (in degrees) at E is

slgE = arctan( 2 2( dz / dx ) ( dz / dy )

and in percent is given by

slpE = 100 2 2( dz / dx ) ( dz / dy )

The slope aspect, i.e. the orientation of the slope gradient, starting from north (0 degrees) and going clock-wise, is calculated using the variables from above, as follows:

aspE = arctan ( ( dz / dx ) /( dz / dy )

The above expression can be evaluated for (dz/dy) ≠ 0. Otherwise aspE = 45˚ (for (dz/dx) < 0) or aspE = 270˚ (for (dz/dx) > 0)

4) To produce distributions of slope gradients and aspects for grids at 30 arc-sec or 5 min latitude/longitude, slope gradients are groups into 9 classes:

C1: 0 % ≤ slope ≤ 0.5 % C2: 0.5 % ≤ slope ≤ 2 % C3: 2 % ≤ slope ≤ 5 % C4: 5 % ≤ slope ≤ 10 % C5: 10 % ≤ slope ≤ 15 % C6: 15 % ≤ slope ≤ 30 % C7: 30 % ≤ slope ≤ 45 % C8: Slope > 45 % C9: Slope gradient undefined (i.e., outside land mask)

Slope aspects are classified in 5 classes:

N: 0˚ < aspect ≤ 45˚ or 315˚ < aspect ≤ 360˚ E: 45˚ < aspect ≤ 135˚ S: 135˚ < aspect ≤ 225˚ W: 225˚ < aspect ≤ 315˚ U: Slope aspect undefined; this value is used for grids where slope gradient is undefined or

slope gradient is less than 2 %.

66

APPENDIX 4. Biomass and energy potentials for bio-fuel feedstocks in Europe Source: IIASA-LUC GIS database

Maps show selected results of the bio-fuel feedstock assessment for Europe.

Calculations were performed on a resource database with a raster of 1 by 1 km grid cells.

CORINE land use data at a resolution of 100 by 100 m results were used to define two type of masks:

1) agricultural land (including arable land, permanent crops, heterogeneous agriculture, pasture); and

2) cultivated land (including arable land, permanent crops, heterogeneous agriculture)

67

APPENDIX 4 - MAP 1.

Energy yieldsGJ/HA biofuel equiv.

9 - 4041 - 6061 - 8081 - 100101 - 120121 - 140141 - 160161 - 180181 - 200201 - 242

Map 1. Potential energy yields of 2nd generation bio-fuel feedstocks on agricultural land The map shows energy yields (in bio-fuel equivalent) of the best-yielding 2nd generation bio-fuel feedstock. They include a) woody plants (Poplar, Willow, Eucalyptus), and b) herbaceous lignocellulosic plants (miscanthus, switchgrass, reed canary grass).

Results are only shown for pixels with at least one 100x100m pixel defined as one of the following land use classes in the land use map: arable land, permanent crops, heterogeneous agriculture, pasture.

68

APPENDIX 4 - MAP 2.

Energy yieldsGJ/HA biofuel equiv.

9 - 4041 - 6061 - 8081 - 100101 - 120121 - 140141 - 160161 - 180181 - 200201 - 242

Map 2. Potential energy yields of 1st generation bio-fuel feedstocks (cereals, sugar crops, oil crops) on arable land The map shows energy yields (in bio-fuel equivalent) of the best-yielding 1st generation bio-fuel feedstock. They include a) oil crops (sunflower, rapeseed), b) cereals (wheat, maize, rye, triticale), and c) sugar crops (sugar beet, sweet sorghum).

Results are only shown for pixels with at least one 100x100m pixel defined as one of the following land use classes in the land use map: arable land, permanent crops, heterogeneous agriculture.

69

APPENDIX 4 - MAP 3.

Energy yieldsGJ/HA biofuel equiv.

9 - 4041 - 6061 - 8081 - 100101 - 120121 - 140141 - 160161 - 180181 - 200201 - 242

Map 3. Potential energy yields of 2nd generation bio-fuel feedstocks on agricultural land (Zoom Central Europe) The map shows energy yields (in bio-fuel equivalent) of the best-yielding 2nd generation bio-fuel feedstocks. They include a) woody plants (poplar, willow, eucalyptus), and b) herbaceous lignocellulosic plants (miscanthus, switchgrass, reed canary grass).

Results are only shown for pixels with at least one 100x100m pixel defined as one of the following land use classes in the land use map: arable land, permanent crops, heterogeneous agriculture, pasture.

70

APPENDIX 4 - MAP 4.

Energy yieldsGJ/HA biofuel equiv.

9 - 4041 - 6061 - 8081 - 100101 - 120121 - 140141 - 160161 - 180181 - 200201 - 242

Map 4. Potential energy yields of 1st generation bio-fuel feedstocks (cereals, sugar crops, oil crops) on cultivated land (Zoom Central Europe) The map shows energy yields (in bio-fuel equivalent) of the best-yielding 1st generation bio-fuel feedstocks. They include a) oil crops (sunflower, rapeseed), b) cereals (wheat, maize, rye, triticale), and c) sugar crops (sugar beet, sweet sorghum).

Results are only shown for pixels with at least one 100x100m pixel defined as one of the following land use classes in the land use map: arable land, permanent crops, heterogeneous agriculture.

71

APPENDIX 4 - MAP 5.

< 20 %21 - 30 %31 - 40 %41 - 50 %51 - 60 %61 - 70 %71 - 80 %81 - 90 %> 90 %

Map 5. Potential energy yields – Ratio of 1st over 2nd generation bio-fuel feedstocks The map shows a yield comparison of the best-yielding 1st generation bio-fuel feedstocks relative to the best-yielding 2nd generation bio-fuel feedstocks.

A mask for arable land has been applied.

72

APPENDIX 4 - MAP 6.

< 20 %21 - 30 %31 - 40 %41 - 50 %51 - 60 %61 - 70 %71 - 80 %81 - 90 %> 90 %

Map 6. Energy yields - Ratio of 1st generation oil crops over 2nd generation bio-fuel feedstocks

The map shows a yield comparison of the best-yielding oil crop (sunflower or rapeseed) relative to the best-yielding 2nd generation bio-fuel feedstock.

A mask for arable land has been applied.

73

APPENDIX 4 - MAP 7.

Map 7. Best-yielding feedstock (in terms of bio-fuel equivalent) within the 1st generation feedstock groups (cereals, oil crops, sugar crops)

74

APPENDIX 4 - MAP 8.

Map 8. Best-yielding feedstock (in terms of bio-fuel equivalent) within the 2nd generation feedstock groups (woody and herbaceous lignocellulosic feedstocks)

75

APPENDIX 5. Conversion factors for calculation of crop residues (see also Table 15)

% water content USE RPR

% water content USE RPR

Cassava 65 0.05 0.20 Carrots 85 0.05 1.25 Yautia (Cocoyam) 65 0.05 0.20 Okra 85 0.05 1.25 Taro (Coco Yam) 65 0.05 0.20 Green Corn (Maize) 85 0.05 1.25 Yams 65 0.05 0.20 Chicory Roots 85 0.05 1.25 Tallowtree Seeds 40 0.50 0.20 Carobs 85 0.05 1.25 Sugar Cane 75 0.50 0.30 Vegetables Fresh nes 85 0.05 1.25 Roots and Tubers nes 65 0.05 0.50 Pop Corn 15 0.50 1.50 Sweet Potatoes 65 0.05 0.60 Buckwheat 15 0.50 1.50 Coconuts 10 0.50 0.60 Quinoa 15 0.50 1.50 Palm Oil 10 0.50 1.00 Fonio 15 0.50 1.50 Bananas 40 0.50 1.00 Triticale 15 0.50 1.50 Plantains 40 0.50 1.00 Canary Seed 15 0.50 1.50 Oranges 40 0.50 1.00 Mixed Grain 15 0.50 1.50 Tang.Mand.Clement.S. 40 0.50 1.00 Cereals nes 15 0.50 1.50 Lemons and Limes 40 0.50 1.00 Olives 35 0.50 1.50 Grapefruit and Pomelos 40 0.50 1.00 Karite Nuts (Sheanuts) 40 0.50 1.50 Citrus Fruit nes 40 0.50 1.00 Castor Beans 40 0.50 1.50 Apples 40 0.50 1.00 Tung Nuts 40 0.50 1.50 Pears 40 0.50 1.00 Jojoba Seeds 40 0.50 1.50 Quinces 40 0.50 1.00 Safflower Seed 40 0.50 1.50 Apricots 40 0.50 1.00 Poppy Seed 40 0.50 1.50 Sour Cherries 40 0.50 1.00 Melonseed 40 0.50 1.50 Cherries 40 0.50 1.00 Linseed 40 0.50 1.50 Peaches and Nectarines 40 0.50 1.00 Hempseed 40 0.50 1.50 Plums 40 0.50 1.00 Oilseeds nes 40 0.50 1.50 Stone Fruit nes, Fresh 40 0.50 1.00 Brazil Nuts 15 0.50 2.00 Figs 40 0.50 1.00 Cashew Nuts 15 0.50 2.00 Mangoes 40 0.50 1.00 Chestnuts 15 0.50 2.00 Avocados 40 0.50 1.00 Almonds 15 0.50 2.00 Pineapples 40 0.50 1.00 Walnuts 15 0.50 2.00 Dates 40 0.50 1.00 Pistachios 15 0.50 2.00 Persimmons 40 0.50 1.00 Kolanuts 15 0.50 2.00 Cashewapple 40 0.50 1.00 Hazelnuts (Filberts) 15 0.50 2.00 Kiwi Fruit 40 0.50 1.00 Areca Nuts (Betel) 15 0.50 2.00 Papayas 40 0.50 1.00 Nuts nes 15 0.50 2.00 Fruit Tropical Fresh nes 40 0.50 1.00 Sesame Seed 40 0.50 2.00 Fruit Fresh nes 40 0.50 1.00 Mustard Seed 40 0.50 2.00 Cocoa Beans 15 0.50 1.00 Kapokseed in Shell 40 0.50 2.00 Tobacco Leaves 75 0.50 1.00 Jute 15 0.50 2.00 Grapes 40 0.50 1.20 Jute-Like Fibres 15 0.50 2.00 Cabbages 85 0.05 1.25 Ramie 15 0.50 2.00 Artichokes 85 0.05 1.25 Sisal 15 0.50 2.00 Asparagus 85 0.05 1.25 Agave Fibres nes 15 0.50 2.00 Lettuce 85 0.05 1.25 Fibre Crops nes 15 0.50 2.00 Spinach 85 0.05 1.25 Coffee, Green 15 0.50 2.10 Cassave Leaves 85 0.05 1.25 Beans, Dry 10 0.50 2.50 Tomatoes 85 0.05 1.25 Broad Beans, Dry 10 0.50 2.50 Cauliflower 85 0.05 1.25 Peas, Dry 10 0.50 2.50 Pumpkins, Squash, 85 0.05 1.25 Chick-Peas 10 0.50 2.50 Cucumbers and Gherkins 85 0.05 1.25 Cow Peas, Dry 10 0.50 2.50 Eggplants 85 0.05 1.25 Pigeon Peas 10 0.50 2.50 Chillies&Peppers, Green 85 0.05 1.25 Lentils 10 0.50 2.50 Onions+Shallots, Green 85 0.05 1.25 Bambara Beans 10 0.50 2.50 Onions, Dry 85 0.05 1.25 Pulses nes 10 0.50 2.50 Garlic 85 0.05 1.25 Groundnuts in Shell 15 0.50 2.50 Leeks and Oth.Alliac.Veg 85 0.05 1.25 Cottonseed 15 0.50 3.50 Beans, Green 85 0.05 1.25 Peas, Green 85 0.05 1.25 Broad Beans, Green 85 0.05 1.25 String Beans 85 0.05 1.25