Understanding biorefining efficiency – The case of agrifood waste

Bioresource Technology 135 (2013) 588–597

Contents lists available at SciVerse ScienceDirect

Bioresource Technology journal homepage: www.elsevier.com/locate/biortech

Understanding biorefining efficiency – The case of agrifood waste Miia Kuisma a,⇑, Helena Kahiluoto a, Jouni Havukainen b, Eeva Lehtonen c, Mika Luoranen b, Tuuli Myllymaa d, Juha Grönroos d, Mika Horttanainen b a

MTT Agrifood Research Finland, Lönnrotinkatu 5, FI-50100 Mikkeli, Finland Lappeenranta University of Technology, P.O. Box 20, FI-53851 Lappeenranta, Finland c MTT Agrifood Research Finland, University of Eastern Finland, P.O. Box 1627, FI-70211 Kuopio, Finland d Finnish Environment Institute, P.O. Box 140, FI-00251 Helsinki, Finland b

h i g h l i g h t s " System borders are decisive for biorefining efficiency. " High exploitation of feedstock potential and substitution efficiency are the keys. " Nutrient efficiency depends on efficiency of substitution for mineral fertilisers. " Share of combustion and heat utilisation are decisive for energy efficiency. " Biorefining increases efficiency in comparison to the current situation.

a r t i c l e

i n f o

Article history: Available online 16 November 2012 Keywords: Nitrogen Phosphorus Carbon Energy Systems approach

a b s t r a c t The aim of this study was to determine biorefining efficiency according to the choices made in the entire value chain. The importance of the share of biomass volume biorefined or products substituted was investigated. Agrifood-waste-based biorefining represented the case. Anticipatory scenarios were designed for contrasting targets and compared with the current situation in two Finnish regions. Biorefining increases nutrient and energy efficiency in comparison with current use of waste. System boundaries decisively influence the relative efficiency of biorefining designs. For nutrient efficiency, full exploitation of biomass potential and anaerobic digestion increase nutrient efficiency, but the main determinant is efficient substitution for mineral fertilisers. For energy efficiency, combustion and location of biorefining close to heat demand are crucial. Regional differences in agricultural structure, the extent of the food industry and population density have a major impact on biorefining. High degrees of exploitation of feedstock potential and substitution efficiency are the keys. Ó 2012 Elsevier Ltd. All rights reserved.

1. Introduction Biorefining that produces renewable energy (RE) and organic fertilisers from untapped agrifood wastes has the potential to mitigate climate change and eutrophication (Kahiluoto et al., 2011), and to create new business opportunities in rural regions. While rapidly becoming mainstream, little is known about how the value chain should be designed for the potential of biorefining to be best realised. Efficiency is an important feature of sustainable biorefining (Kokossis and Yang, 2010), nutrient, carbon (C) and energy efficiency being key to both environmental and economic performance. Efficiency is, however, multifaceted. It can crucially mislead decision-making if its dependence on the choices along the entire biorefining supply and demand chain is not revealed. ⇑ Corresponding author. Tel.: +358 29 531 7395; fax: +358 20 772 040. E-mail address: miia.kuisma@mtt.fi (M. Kuisma). 0960-8524/$ - see front matter Ó 2012 Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.biortech.2012.11.038

Currently inefficiently utilised agrifood waste and by-products of primary production, food industry, retail and consumption have great potential as nutrient, C and energy sources (Kahiluoto et al., 2011). They can help mitigate climate change and eutrophication through substituting for non-renewable energy (non-RE) and fertilisers or sequestering C in soil. The products either substitute for imported fuels and fertilisers or represent new exports and provide jobs and finance for public services (Bailey et al., 2011), thus fostering regional economies. In modern agrifood-waste-based biorefining it is possible to integrate all these benefits, if in addition to energy or C, nutrients in biomass are also utilised. The best design of the supply and demand chain depends on benefit prioritisation. Choices for every step from biomass supply (types and quantity covered), collection and conversion (processes, location), to markets (distribution, energy consumption, fields) and demand substituted for (energy, fertilisers) may affect the overall efficiency of

M. Kuisma et al. / Bioresource Technology 135 (2013) 588–597

biorefining. Due to high costs of biomass transportation, biorefining is typically organised on a local or regional scale. Therefore, regional conditions may affect the appropriate design and efficiency of biorefinery systems (Kahiluoto et al., 2011; Kokossis and Yang, 2010). The relevant part of the value chain varies depending on the question and the primary beneficiary or decision-maker posing it. Biorefining is increasingly promoted and shaped by public incentives (e.g. Law of the operating aid for renewable electricity in Finland 1396/2010, Supplementary Information (SI)) and developed by pioneering entrepreneurs. These activities are based on intuitive assumptions rather than on systematic comparisons of biorefinery systems. Despite the policy priority on material use prior to energy use (Directive 2008/98/EC), the studies of material efficiency of agrifood-waste-based biorefining (e.g. Cui et al., 2012; Karp et al., 2011; Leung et al., 2012; Martínez Sabajanes et al., 2012) are in the minority and only few studies of nutrient and C exploitation (Anex et al., 2007; Cayuela et al., 2010) have been conducted. Energy efficiency of feedstock transportation and conversion, and utilisation of products, has been addressed (e.g. Chen et al., 2012; Laser et al., 2009; Pöschl et al., 2010) and a systems approach has been suggested to be applied in development of biorefining (Kokossis and Yang, 2010). However, no comparison of contrasting biorefinery systems covering the entire value chain, i.e. all the system levels (Bunge, 1985) of biorefining, has been previously performed. This study appears to be the first to reveal the dependence of efficiency estimates on the system boundaries set in the supply and demand chain of biorefining. The aim of this study was to assess nutrient, C and energy efficiencies of contrasting agrifood-waste-based biorefinery systems, focusing on several system levels. The biorefinery systems considered form coherent scenarios, each based on a distinct design target. The targets were set to (1) mitigate climate change; (2) mitigate eutrophication or (3) enhance the regional economy. The biorefinery scenarios were developed for two contrasting case regions in Finland. The following research questions were asked: 1. What are nutrient (N, P), carbon (C) and energy efficiencies of agrifood-waste-based biorefining? 2. What are the main determinants of efficiencies that depend on regional conditions and the design of biorefinery systems? 3. How do the efficiencies depend on whether only biorefinery plants or entire supply and demand chains of the biorefinery products are taken into account? 2. Methods 2.1. Contrasting biorefinery scenarios Agrifood-waste-based biorefinery scenarios were designed for contrasting targets. The targets and system boundaries were identified in workshops in discussion with actors representing agrifood and biorefinery systems. The first scenario was designed to mitigate climate change by substituting for non-RE (ENERGY). The second scenario was designed to mitigate eutrophication (WATER) and the third one to enhance the regional economy (ECONOMY). The current situation was described by a baseline scenario (PRESENT). The scenarios comprised several system levels (Fig. SI1): Biorefinery plants level, including conversion within the biorefinery plants; biorefinery chains level, including also transportation, and biorefinery region level, including also the supply of the biomass potential and the demand for the fertilisers and energy, including locations, and the estimated substitution of

589

mineral fertilisers and non-RE products, based on the data for the case region. The main features of the scenarios and differences based on the design targets were defined at various system levels, for the aspects (Table 1) having a hypothetical impact on the studied efficiencies. The anticipatory scenarios (Verburg et al., 2006) generated were coherent story lines; all choices (Table 1) were made based on the design target of the scenario. However, a small loss in the target was accepted if substantial benefit for another target was achieved. Conversion processes were chosen so that both nutrients and energy were recoverable in all the three scenarios. Therefore, anaerobic digestion (AD) was a key technology. The sensitivity of the design target of climate change mitigation was analysed through examining another climate change mitigation scenario also designed for recycling and sequestering C (CARBON) (Table SI1) rather than replacing non-RE (ENERGY). The sensitivity of the design target for the choice of the conversion process was analysed in ENERGY. The biorefinery scenarios were formed based on the biomass potential (Table 2) of agrifood wastes and by-products currently available, and within the next 5 years through implementation of the policy targets and regulations already set (see the more detailed biomass assessment in Kahiluoto et al. (2011)). The heating values (LHVd) of the biomass types were based on figures reported in the literature (Mattsson Petersen et al., 2005; Phyllis, 2012). The scenarios were developed for two case regions, for rural South Savo in eastern Finland and for more densely populated Satakunta in south-western Finland (Table SI2). The fresh-weight based density of concentrated biomass was 2.4 t km2 (tonnes per square kilometre) in South Savo and 35 t km2 in Satakunta and scattered biomass 37 t km2 and 115 t km2, respectively. In Satakunta the share of feedstock in ENERGY and ECONOMY, and the energy demand of transportation activities were extrapolated based on the data for South Savo (see detailed below). Feedstock types and quantities for biorefinery plants (Table 3; see CARBON Table SI3) was defined taking into account the design targets (Table 1) and the location of the biomass. Additionally, unprocessed biomass (Table 1, manure) is a part of utilised biomass in PRESENT because it is currently applied in agriculture. In South Savo, the content of N in feedstock is then 3.4 kt a1 and in material products 3.2 kt a1, P 0.9 kt a1 and 0.9 kt a1, and C 56 kt a1 and 54 kt a1, respectively (cf. Table 3). In Satakunta, the content of N in feedstock is 8.1 kt a1 and in material products 5.4 kt a1, P 2.1 kt a1 and 1.7 kt a1, and C 130 kt a1 and 120 kt a1, respectively. Feedstock is agrifood waste and by-product biomass, and therefore production of the main products of the agrifood system, i.e. food, feed, fuel and fibre, is excluded from the scenarios. All harvesting, collecting and transporting of feedstock, and transporting, spreading and application of material products (organic fertilisers), are included. The location of the biorefinery plants and the collection areas of the feedstock (Table 1, Fig. SI2; see CARBON Table SI1) were defined in South Savo based on the location of biomass, animal farms (SI Evira, 2007), district heating networks (SI Energiateollisuus, 2008) and the street and road network (Digiroad, 2009). ArcGIS software and Network Analyst were used to define the collection areas (ESRI, 2010). If the whole biomass potential was not biorefined, as a consequence of the design target, the remaining biomass was included in the scenario and treated as in PRESENT (feedstock for biorefining or utilised as unprocessed biomass); the share of such biomass ranged from 1.2% to 32% (Table 1). In Satakunta, in ENERGY and ECONOMY, the share of feedstock of the biomass potential was estimated based on the district heat demand. First, (Eq. (1)), the heat demand in ENERGY in Satakunta was estimated based on the heat production (supply) in ENERGY in South Savo and the densities and the shares of the biomass:

590

M. Kuisma et al. / Bioresource Technology 135 (2013) 588–597

Table 1 The main differences (supply volumes, types, locations and collection areas of feedstock, location and capacities of biorefinery plants, conversion processes and products, markets (demand), substituted products) at the various system levels of the contrasting agrifood-waste based-biorefinery scenarios depending on the design targets. The shares for the two case regions are given as: South Savo%/Satakunta%. PRESENT Biorefinery plants Biomass potential (% of Feedstock based on design targetsa Biomass treated as in PRESENTa Where feedstock Total feedstock Utilised as unprocessed biomass Unutilised biomass

c d e f g h

ECONOMY

–

Available for centralised biorefinery plants: 71%/99%b 29%/1.2%b

Agricultural biomassc: 93%/ 79% e , 6.9%/21%

Available for centralised biorefinery plants, high quality5: 68%/94%e f , 32%/6.0%b

– 4.9%/21% 85%/64%

6.3%/60% 72%/100% –

64%/89% 98%/97% –

11%/85% 72%/99% –

d

Seea: <1%

Seea: <1%

Seea: <1%

fw) –

, 10%/15%

96%/90% Yes No No 3.7%/10% No No No No

55%/69% Yes No Yes No No No No 45%/31%

95%/90% Yes Yes Yes By-products of biorefining <1%/<1% 3.5%/9.6% <1%/<1% No

No Electricity, heat

No Separated nutrient products

Other products

Precious little energy

Digestate, ash

Electricity, heat

No Traffic fuels, processed organic fertilisers –

Centralised in towns; distributed in villages Limited extent of case region Feedstock, material products

The biggest district heating networks, capacity 100–170 kt a1 fw Collection areas of towns

The biggest animal farms, capacity 5,5 kt a1 fw Complete case region

The biggest district heating networks, capacity 90–160 kt a1 fw Collection areas of towns

Feedstock, material products

Feedstock, material products

Feedstock

In case region –

In case region The highest primary non-RE consumption Mineral fertilisers

In case region Average non-RE in case region Mineral fertilisers

Export The most expensive non-RE for regional economy Mineral fertilisers

Biorefinery region Markets Substituted energy products Substituted material products a

WATER

Conversion processes of feedstock (% of fw) Anaerobic digestion 9.5%/10% CHP plant Yes Upgrading No Mechanical drying Yes Combustion <1%/<1% Composting 84%/31% Bioethanol No Biodiesel <1%/<1% Mechanical No separation Other techniques 4.5%g/58%h Main products Compost

Biorefinery chains Locations of biorefinery plants Collection areas of feedstock Transportation

b

ENERGY

–

Currently uncollected household biowaste was excluded from feedstock and included in the scenarios treated as in PRESENT. In Satakunta the share of feedstock was estimated based on the district heat demand (see Eq. (1)). Agricultural biomass: green fallow biomass, harvest residues, manure, buffer zone biomass and aquatic biomass. Landfilled, banked, non-harvested. Non-agricultural biomass: on farm animal carcass, food industry wastes and by-products, municipal biowaste and sewage sludge. Low quality: limited utilisation in agriculture (sewage sludge) or impure (inadequate source separated biowaste). Rendering 4.5%. Rendering 16%, feed manufacturing 13%, organic fertiliser manufacturing 28%.

Table 2 The biomass potential of agrifood waste and by-product in the case regions.

hSatakunta ¼ hSouthSav o     cSatakunta sSatakunta   cESatakunta þ  sESatakunta cSouthSav o sSouthSav o

South Savo 1

ð1Þ Here, hSouthSavo is the heat produced in ENERGY in South Savo (TW h a1), cSatakunta is the dry-weight-based density of concentrated biomass in Satakunta (t km2) and, cSouthSavo in South Savo (t km2), cESatakunta is the dry-weight-based ratio of concentrated biomass to biomass potential in ENERGY in Satakunta, sSatakunta is the dry-weight-based density of scattered biomass in Satakunta (t km2) and, sSouthSavo in South Savo (t km2) and sESatakunta is the dry-weight-based ratio of scattered biomass to biomass potential in ENERGY in Satakunta. Secondly, the demand for the feedstock in Satakunta was assessed. In ECONOMY the same feedstock was included; with the exception of biomass types excluded based on the design target (Table 1).

Scattered biomass Green fallow biomass Harvest waste and by-product Manure On-farm animal carcass Buffer zone biomass Aquatic biomass Concentrated biomass Food industry waste and byproduct Municipal biowaste Sewage sludge Total

Satakunta 1

kt a

kt a DM

kt a1

kt a1 DM

120 57 550 0.19 14 7.8

25 28 69 0.055 2.9 3.1

210 260 590 0.16 75 22

42 130 85 0.047 15 9.3

4.8

1.4

120

37

20 26

7.2 3.6

17 71

5.9 11

800

140

1400

340

591

M. Kuisma et al. / Bioresource Technology 135 (2013) 588–597

Table 3 The inputs (feedstock, auxiliary energy) and outputs (energy and material products) of biorefinery plants and chains. Nitrogen (N), phosphorus (P) and carbon (C) are given as kt a1 and lower heating value of dry material (LHVd) and energy (e) as GW h a1. PRESENT

South Savo Biorefinery plants Feedstock: scattereda Feedstock: concentratedb Auxiliary energy: conversion Products: gross Products: netc

ENERGY C

LHVd/e

N

P

C

LHVd/e

N

P

C

LHVd/e

N

P

C

LHVd/e

0.047 0.28

0.008 0.11

1.1 4.2

2.5 0.30

0.58 0.11

40 4.8

0.83 0.11

58 4.2

0.53 0.11

37 4.8

3.4 3.4

2.6 2.6

0.67 0.67

21 21

3.8 3.8

0.93 0.93

41 41

570 41 52 200 67

2.3 0.30

0.10 0.10

400 47 33 220 190

3.6 0.28

0.14 0.14

12 41 0.83 3.7 2.0

2.5 2.5

0.60 0.60

17 17

400 47 150 200 140

1.2 2.6

Satakuntaf Biorefinery plants Feedstock: scattereda,f Feedstock: concentratedb, f Auxiliary energy: conversion Products: gross Products: netc f

ECONOMY

P

Biorefinery chains Auxiliary energy: agricultural workd Auxiliary energy: transportatione

Biorefinery chains Auxiliary energy: agricultural workd, Auxiliary energy: transportatione, f

WATER

N

0.23 2.6

0.085 0.65

5.8 23

0.9 0.9

0.67 0.67

22 22

62 190 54 83 62

12 10

6.2 3.3

1.6 0.75

130 29

8.8 8.8

2.2 2.2

45 45

4.2 26

1300 240 100 1000 920

17 5.5

6.1 3.3

1.6 0.74

120 28

6.9 6.9

2.2 2.2

90 90

43 27

1300 230 180 600 360

8.3 11

6.2 3.3

1.6 0.74

130 29

8.5 8.5

2.0 2.0

39 39

47 46

1300 240 550 950 660 33 51

a

Agricultural and aquatic waste and by-product biomass. Waste and by-product biomass of food manufacturing, grocery stores, food services, households and municipal sewage sludge. c For markets. d Harvesting and collecting of waste and by-product biomass and spreading and application of material products (organic fertilisers). e Collection and road transportation of waste and by-product biomass and material products (organic fertilisers). f In Satakunta the share of feedstock included in ENERGY and ECONOMY and energy demand in transportation activities were estimated based on South Savo (see Section 2.1). b

N, P, C and energy outputs and the energy demand of conversion processes were estimated (Table 3; see CARBON Table SI3). Commercially available technologies, described in Kahiluoto et al. (2011), were adopted. Additionally, the following assumptions were made: for composting of the solid fraction of digestate and biowaste, the larger scale composting reactor was assumed to consume electricity at 37 kW h t1 (SI Lohiniva et al., 2001) and the smaller scale windrows fuel at 18 kW h t1 (SI Myllymaa et al., 2008). The granulation of compost was assumed to consume heat at 950 kW h t1 H2O (SI Pöyry Environment, 2007) and ammonium stripping electricity at 125 kW h t1 (Organics, 2012). The upgrading of biogas by water scrubbing was assumed to consume electricity at 0.36 kW h m3 of biogas and to cause methane loss of 2% (Biogasmax, 2006). The stillage of bioethanol was assumed to incinerate, with the same LHVd as for straw, 0.048 kW h t1 (Kahiluoto et al., 2011). Manure slurry was assumed to separate by decanter centrifuge, with electricity consumption of 6 kW h t1 (Møller et al., 2002). In PRESENT rendering of animal waste and the manufacturing of the feed of plant wastes also occurred. Bioenergy produced in each scenario was the prime energy source for conversion and the substituted energy source was used for complementation, if needed (Table 1). The transportation distances of the feedstock and material products in the scenarios in South Savo were analysed according to road distances (Digiroad, 2009) using ArcGIS software (ESRI, 2010), and to the typical truck capacities. An origin–destination matrix of the shortest path costs using Dijkstra’s (1959) algorithm was formed between the location of feedstock and the biorefinery plant. The transportation distances of PRESENT were defined based on the analysed median distances of farms in South Savo and the transportation distances outside the region, and for biowaste based on the analysed transportation distances in ECONOMY, where the transportation corresponded to the current situation. The energy demand (Table 3; see CARBON Table SI3) for agricultural transportation in harvesting, collecting, and spreading and

application, was based on measurements of performance (SI Palva et al., 2004; SI Palva, 2003; SI Peltonen et al., 2003; SI Peltonen, 1993) and the energy demands for collection and road transportation were analysed using the traffic database (LIPASTO, 2011). Material products (organic fertilisers) were assumed to be transported on the back hauls of feedstock. In Satakunta the energy demand for agricultural transportation, and the transportation of concentrated feedstock and material products was estimated to be equal per mass unit compared to South Savo. The energy demand of transporting scattered feedstock was estimated to be divided by 3.1 per mass unit in Satakunta compared to South Savo, based on the ratios of the biomass densities of biomass types in the case regions. The demand for energy and fertilisers was assessed based on the regional energy (SI Holttinen, 2007; SI Karhunen, 2008) and nutrient balances (Salo et al., 2007; Uusitalo et al., 2007). The primary energy use, i.e. total energy use as energy sources, for transportation and conversion, regional energy demand (Edwards et al., 2006; SI Bröckl et al., 2010; SI Myllymaa et al., 2008) and mineral fertiliser manufacturing (Davis and Haglund, 1999; SI Grönroos and Voutilainen, 2001) was also assessed, in addition to direct energy use. The elemental N, P and C contents of auxiliary non-RE sources were included in the efficiency assessments. 2.2. Efficiency assessment Efficiencies of the contrasting biorefinery scenarios, and the baseline scenario, for the case regions were assessed as output/input ratios, for N, P, C and energy, separately. 2.2.1. Efficiency of biorefinery plants Efficiencies for conversion of feedstock in the biorefinery plants (Fig. SI1) were calculated as:

gBiorefinery plantsN;P;C;e ¼

N; P; C; eP ðN; P; C; eF þ N; P; C; eBPAE Þ

ð2Þ

592

M. Kuisma et al. / Bioresource Technology 135 (2013) 588–597

Here, N, P, C, eP are the net products (the material products as N, P and C (kt a1) and energy (e) as the RE products, as the sum of heat, electricity and traffic fuels (TW h a1)), N, P, C, eF is the feedstock (N, P and C as kt a1 and energy (e) as LHVd (TW h a1)), and eBPAE is the auxiliary energy need in conversion in the biorefinery plants (TW h a1) and N, P, CBPAE is the elemental N, P and C losses of that auxiliary energy need (kt a1). 2.2.2. Efficiency of biorefinery chains Efficiencies of biorefinery chains (Fig. SI1) cover, in addition to biorefinery plants (Section 2.2.1), also auxiliary energy for harvesting, collecting, transporting, spreading and application of the feedstock and material products:

gBiorefinery chainsN;P;C;e ¼

N; P; C; eP ðN; P; C; eF þ N; P; C; eBCAE Þ

ð3Þ

Here, for N, P, C, eP and N, P, C, eF see Eq. (2); eBCAE is the auxiliary energy need in conversion in the biorefinery plants and in agricultural work and transportation in the biorefinery chains (TW h a1) and N, P, CBCAE is the elemental N, P and C losses of that auxiliary energy need (kt a1). 2.2.3. Efficiency of biorefinery regions Efficiency of biorefinery region (Fig. SI1) covers, in addition to biorefinery plants (Section 2.2.1) and biorefinery chains (Section 2.2.2), also substitution, supply and demand in the case regions. Substitution for mineral fertiliser use was based on two complementary fertilisation strategies. The first strategy was based on current mineral fertiliser use in the regions, for the designed and baseline scenarios. The current mineral fertiliser use in South Savo was for N 70 kg ha1 and for P 8 kg ha1 and in Satakunta 73 kg ha1 and P 10 kg ha1, respectively (Salo et al., 2007; Uusitalo et al., 2007). The substitution for mineral fertiliser use by organic fertilisers was assessed as:

gSubstitution of useN;P ¼

N; PP N; PMFU

gSubstitution of needN;P ¼

ð5Þ

where N, PY is the harvested yield (kt a1), and N, PMFN is the sum of the current mineral fertiliser use (in PRESENT, and in the field area fertilised as in PRESENT in the designed scenarios) and the complementary mineral fertiliser need substituting the harvested nutrients (zero field balance, the field area included in the designed scenarios) (kt a1). The N and P in green fallow biomass, and harvest waste and byproducts were considered to substitute for the nutrients taken up by the crop and thus included in the biomass, as part of the internal nutrient cycles of the system (see Kahiluoto et al., 2011). A high nutrient content of field soil, for example P, can also be intentionally reduced through harvesting the biomass for organic fertiliser production. The impact of the latter assumption on nutrient substitution was also studied. Substitution for energy was assessed as the potential to substitute the primary non-RE use, i.e. total energy use as energy sources in the regions by RE products of the scenarios:

gSubstitutione ¼

ðePU  eBCPANONRE Þ  ðePNONRE =eNONRE Þ ðePNONRE þ ePNONREMF Þ

ð6Þ

where ePU are the net RE products, as the sum of heat (according to utilisation degree), electricity and traffic fuels (TW h a1)), eBCPANONRE is the auxiliary primary non-RE need in conversion in the biorefinery plants and in agricultural work and transportation in the biorefinery chains (TW h a1), ePNONRE is the primary non-RE use, as the sum of heat, electricity and traffic fuels (TW h a1), eNONRE is the direct non-RE use, as the sum of heat, electricity and traffic fuels (TW h a1) and ePNONREMF is the auxiliary primary non-RE use in manufacturing the mineral fertilisers based on the current mineral fertiliser use or the zero field balance (TW h a1). Supply efficiency was assessed as a ratio of the biorefinery products to the biomass potentials of the regions:

ð4Þ

where N, PP are the net material products (organic fertilisers; in PRESENT also unprocessed manure applied to fields) (kt a1), and N, PMFU is the current mineral fertiliser use (kt a1). In PRESENT unprocessed manure is included in the substitution assessment because currently the nutrient content of the manure has to be included in the fertilisation plan when it is applied to fields (SI Agency for Rural Affairs, 2012). The second fertilisation strategy was based on zero field balance (see Kahiluoto et al., 2011), i.e. only nutrients harvested in the region were assumed to be substituted by organic fertilisers in the design scenarios, in the fields included in the scenarios as feedstock source and for organic fertiliser application. Nutrients were harvested in South Savo as N 84 kg ha1 and as P 10 kg ha1 and in Satakunta 70 kg ha1 and 10 kg ha1, respectively (Salo et al., 2007; Uusitalo et al., 2007). If the nutrients recovered in organic fertilisers were not sufficient to substitute for the harvested N and P in the scenarios, complementary mineral fertilisers were used. On the other hand, if there were excess organic fertilisers in the case region, the nutrients were assumed to be utilised beyond the region borders, according to the fertilisation strategy of zero field balance. According to the design targets (Table 1), the scenarios comprised also a varying field area, which was assumed to be fertilised similarly to PRESENT. The avoided mineral fertiliser manufacture in the fertilisation strategy of the zero field balance was assessed as a difference between the current mineral fertiliser use and the complementary mineral fertiliser need. The efficiency of complementary mineral fertiliser need for yield production was assessed as:

N; PY N; PMFN

gSupplyN;P;C;e ¼

N; P; C P ; ePU N; P; C; eBP

ð7Þ

where for N, PP see Eq. (4), for CP see Eq. (2) and for ePU see Eq. (6); N, P, CBP is the waste and by-product biomass supply (biomass potential) as kt a1, and eBP as LHVd (Twha1). Demand efficiency for N and P was assessed as a ratio of the demand to the total inputs available and used in the regions:

gDemandN;P ¼

N; PY ðN; P BP þ N; PMFN Þ

ð8Þ

Here, for N, PY see Eq. (5); for N, PBP see Eq. (7); N, PMFN is the sum of the current mineral fertiliser use (in PRESENT, and in the field area fertilised as in PRESENT in the designed scenarios) and the complementary mineral fertiliser need substituting for the harvested nutrients (zero field balance, the field area included in the designed scenarios) (kt a1). Demand efficiency for energy was assessed as the potential to reduce the use of non-RE resources in the regions:

gDemande ¼

ðeNONRE þ eANONRE Þ ðeRNONRE þ eRANONRE þ eBCANONRE Þ

ð9Þ

where, for eNONRE see Eq. (6); eANONRE is the auxiliary non-RE use in manufacturing the current mineral fertiliser use (TW h a1), eRNONRE is the remaining non-RE need after substitution by the net RE products, as the sum of heat (according to utilisation degree), electricity and traffic fuels (TW h a1), eRANONRE is the remaining auxiliary non-RE need in manufacturing the mineral fertilisers based on the current mineral fertiliser use or the zero field balance (TW h a1) and eBCANONRE is the auxiliary non-RE need in conversion

M. Kuisma et al. / Bioresource Technology 135 (2013) 588–597

in the biorefinery plants and in agricultural work and transportation in the biorefinery chains (TW h a1). 3. Results and discussion The results show that biorefinery scenarios designed for contrasting targets are substantially different in terms of nutrient, C and energy efficiency. The highest system levels, where regional supply, demand and substitution effects are included, reveal more differences among scenarios than conversion alone. 3.1. N, P, C and energy efficiencies Recovery at all the system levels is high for N and P, lower for C and lowest for energy (Figs. 1–4; for CARBON see Figs. SI3–SI6). N losses occurred in combustion and composting (Fig. 1), but could be reduced through the capture of N in flue gases (Anex et al., 2007). Material products substitute for a higher share of non-renewables in the regions than energy products. Biogas output of AD could be 10% higher if feedstock were co-digested (Pöschl et al., 2010), which would increase energy efficiencies in scenarios where AD was the main technology. When the energy content of biomass was exploited efficiently, C recovery in material products, available for sequestration, remained low (Figs. 3 and 4). However, high substitution of non-RE is another way to enhance C efficiency of biorefining. 3.2. Differences among scenarios All the designed scenarios were more efficient than the current situation across the system levels, apart from C (Figs. 1–4; for CARBON see Figs. SI3–SI6). For nutrients, the greatest difference was due to the more efficient substitution for mineral fertilisers, while for energy, the difference was mainly due to a higher share of biorefined biomass relative to the current situation. On the other hand, the current low energy utilisation leads to low C loss (Fig. 3). Nutrient efficiencies were highest in CARBON (Figs. SI3 and SI4), and in South Savo also in WATER and in Satakunta in ENERGY and ECONOMY (Figs. 1 and 2), due to the high coverage of regional biomass potential and use of AD with low nutrient losses. In regions with a high share of agricultural biomass and high quantity of that biomass in relation to field area, such as in South Savo, decentralised biorefining is favoured; centralised ENERGY and ECONOMY were the weakest in terms of nutrient efficiency in South Savo because of the lowest biomass coverage. In Satakunta with a lower share of agricultural biomass, the scenario excluding other biomass types (WATER) had the lowest nutrient efficiencies. The highest energy efficiency was achieved in the ENERGY scenario (Fig. 4) because biorefinery plants were located so as to enable a high degree of heat utilisation and combustion included as a conversion technology. The maximisation of the energy output through combustion of all the feedstock would multiply generation of electricity by 1.3–1.6, heat by 1.3–1.5 and ash by 4.9–11 in comparison with AD included as a partial conversion technology, and all N was lost. For the scenario variant with the design target of climate change mitigation through recycling and sequestering C (SI: CARBON), the benefits, i.e. soil humus formation and C sequestration, were beyond the system boundaries of this study. However, the cost of the design, C release in AD and composting, is included (Fig. SI5). Inclusion of the sequestration efficiency would divide the C efficiency of the scenarios by five to six, depending on the technology (Uhlen, 1976). Only little agrifood waste is currently (PRESENT) biorefined and the conversion processes are diversified. Therefore, the uncertainty in the efficiencies is higher than in the designed scenarios.

593

3.3. Differences between regions Differences among regions in nutrient and energy efficiency of agrifood biorefining were mainly attributable to differences in agricultural production, in importance of the food industry and in population density. Energy content is high in by-products of the food industry and in crop residues from grain production for food and fodder for monogastrics (pork, chicken). A higher population density allowed a higher degree of heat utilisation decisive for energy efficiency. In regions where more untapped biomass is available, the increases in nutrient and energy efficiencies through biorefining are higher. The potential for substituting nutrients is substantially higher in a case region where more non-agricultural waste and by-product biomass is available for biorefining. Differences in efficiency were greater between scenarios and system levels in Satakunta than in South Savo (Figs. 1, 2 and 4; for CARBON see Figs. SI3, SI4 and SI6). This is due to a single scenario with a poor performance in terms of nutrients and energy efficiency and a single scenario superior in terms of energy efficiency in Satakunta. In terms of nutrients there is a crucial difference between the case regions in WATER, which is the weakest scenario in Satakunta (Figs. 1 and 2). Agricultural biomass contains less N per unit of field area than spatially concentrated waste biomass, and the share of agricultural biomass is smaller in Satakunta compared to South Savo. Nutrient substitution did differ between the current situation and the designed scenarios in South Savo only if the nutrients were used as efficient substitutes for mineral fertilisers, unlike in Satakunta, where new non-agricultural biomass was applied (Figs. 1 and 2; for Carbon see Figs. SI3 and SI4). Anex et al. (2007) reported corresponding efficiencies of substitution for current mineral fertiliser N use by bioethanol by-products (straw-based 0.23 and switchgrass-based 0.78). In South Savo, N in organic fertilisers substitutes for roughly one third of N in harvest, while in Satakunta for two thirds, apart from WATER. Substitution is higher for P, allowing regional nutrient sovereignty (self-sufficiency), and in Satakunta even exceeding the demand in the region and allowing export from the region. If also green fallow biomass and harvest waste and by-products are biorefined, the nutrient substitution potential is multiplied by 1.2–1.4 in South Savo and by 1.1–2.3 in Satakunta. 3.4. Differences among system levels How system boundaries are set is decisive for the efficiency estimate of agrifood biorefining. It affects both the absolute and relative differences among scenarios. The key determinants for efficiency can only be detected through investigating the impact of various choices at all the relevant levels of the biorefinery system; taking into account regional characteristics, utilisation degree of biomass potential and efficiency of substitution in addition to conversion and transportation. Nutrient efficiencies depend on substitution for mineral fertilisers by organic fertiliser products in agriculture, and not on biorefinery plants and chains (Figs. 1 and 2; for Carbon see Figs. SI3 and SI4). For example, in Satakunta, the highest N losses in conversion through combustion were 0.6 kt a1, while avoided N use in mineral fertilisers was simultaneously 7.1 kt a1. N efficiencies of biorefinery chains are 0.85– 1.8% smaller compared to those for biorefinery plants, so the effect of transportation is small. The corresponding efficiency difference for C is 2.7–5.4%, with no effect for P. Also for energy, the higher system levels revealed more differences between scenarios than biorefining alone (Fig. 4; for CARBON see Fig. SI6). The efficiency of complementary mineral fertiliser need for yield production was marked in the designed scenarios compared to current use (PRESENT) (Table 4; for CARBON see Table SI4),

594

M. Kuisma et al. / Bioresource Technology 135 (2013) 588–597

SOUTH SAVO 1.40 1.20

BIOREFINERY PLANTS

B

I

O

R

E

F

I

N

E

R

Y

R

E

G

I

O

N

1.00 0.80 0.60 0.40 0.20 0.00

SUBSTITUTION Eq. 2 products/feedstock

SUPPLY

Eq. 4

DEMAND

Eq. 7

Eq. 8

products/mineral fertiliser use products/biomass potential

yield/(biomass potential + mineral fertiliser need)

SATAKUNTA 1.40 1.20

BIOREFINERY PLANTS

B

I

O

R

E

F

I

N

E

R

Y

R

E

G

I

O

N

1.00 0.80 0.60 0.40 0.20 0.00

SUBSTITUTION

SUPPLY

DEMAND

Eq. 2

Eq. 4

Eq. 7

Eq. 8

products/feedstock

products/mineral fertiliser use

products/biomass potential

yield/(biomass potential + mineral fertiliser need)

PRESENT

ENERGY

WATER

ECONOMY

Fig. 1. The nitrogen (N) efficiencies at various system levels of the biorefinery scenarios in the case regions. For PRESENT, the current mineral fertiliser use is assumed (Eq. (8)).

and higher in Satakunta, with a regional surplus for P in the biorefinery products, than in South Savo. Substitution for harvested nutrients rather than for current mineral fertiliser use is justified by the high amount of organically bound nutrients in organic fertilisers. A new fertilisation strategy, which takes into account the full nutrient potential of organic fertilisers, is needed to achieve efficient recycling of renewable nutrients (Drinkwater and Snapp, 2007). Currently nutrients in agrifood waste (manure) are used inefficiently, i.e. in addition to mineral fertilisers, which already substitute for nutrients in harvested yield. Regarding energy (Fig. 4), highly upgraded products (ECONOMY) led to the lowest efficiency in biorefinery plants and chains due to high use of auxiliary energy. Conversion process efficiency (gross products/feedstock, cf. Eq. (2)) of bioethanol and combustion processes together (0.69) was comparable to efficiencies reported by Laser et al. (2009) for more diverse bioethanol and thermochemical processes. At higher system levels of the biorefinery scenarios energy

efficiencies are lowest, when low biomass coverage, energy recovery from only a part of biorefined biomass and low heat demand due to the decentralised location of biorefinery plants were features of the scenario (WATER). The impact of transportation on energy efficiencies was minor when studied through a systems perspective in comparison with higher system levels (Fig. 4; cf. Table SI5). 3.5. Generality The two case regions represent a broad range in structure of agricultural production, the extent and structure of the food industry, and population density and thus in the main regional determinants for biorefining efficiency. Efficiencies would be greater in regions of higher population density and more extensive food industry. The studied biorefinery scenarios broadly cover the variation in biorefinery systems according to the classification of Cherubini et al. (2009). The lessons learned from agrifood-waste-based

595

M. Kuisma et al. / Bioresource Technology 135 (2013) 588–597

SOUTH SAVO 1.40 1.20

BIOREFINERY PLANTS

B

I

O

R

E

F

I

N

E

R

Y

R

E

G

I

O

N

1.00 0.80 0.60 0.40 0.20 0.00

SUBSTITUTION

SUPPLY

DEMAND

Eq. 2

Eq. 4

Eq. 7

Eq. 8

products/feedstock

products/mineral fertiliser use

products/biomass potential

yield/(biomass potential + mineral fertiliser need)

SATAKUNTA 1.40 1.20

BIOREFINERY PLANTS

B

I

O

R

E

F

I

N

E

R

Y

R

E

G

I

O

N

1.00 0.80 0.60 0.40 0.20 0.00

SUPPLY

DEMAND

Eq. 2

SUBSTITUTION Eq. 4

Eq. 7

Eq. 8

products/feedstock

products/mineral fertiliser use

products/biomass potential

yield/(biomass potential + mineral fertiliser need)

PRESENT

ENERGY

WATER

ECONOMY

Fig. 2. The phosphorus (P) efficiencies at various system levels of the biorefinery scenarios in the case regions. For PRESENT, the current mineral fertiliser use is assumed (Eq. (8)).

SOUTH SAVO

SATAKUNTA

1.40

1.40

1.20

BIOREFINERY PLANTS

BIOREFINERY REGION

1.00

1.20

BIOREFINERY PLANTS

1.00

BIOREFINERY REGION

0.80

0.80

ENERGY

0.60

0.60

WATER

0.40

0.40

ECONOMY

0.20

0.20

0.00

0.00 Eq. 2 products/feedstock

SUPPLY

SUPPLY

Eq. 2

Eq. 7

Eq. 7

products/feedstock

products/biomass potential

products/biomass potential

PRESENT

Fig. 3. The carbon (C) efficiencies at various system levels of the biorefinery scenarios in the case regions.

596

M. Kuisma et al. / Bioresource Technology 135 (2013) 588–597

SOUTH SAVO 1.40 1.20

BIOREFINERY PLANTS

BIOREFINERY CHAINS

B I O R E F I N E R Y

R E G I O N

1.00 0.80 0.60 0.40 0.20 0.00

SUBSTITUTION Eq. 2 products/(feedstock + auxil. energy need)

Eq. 3

Eq. 6

products/(feedstock + auxil. energy need)

products/primary non-RE need

SUPPLY

DEMAND

Eq. 7

Eq. 9

products/biomass potential

non-RE use/(biomass potential + non-RE need)

SATAKUNTA 1.40 1.20

BIOREFINERY PLANTS

BIOREFINERY CHAINS

B I O R E F I N E R Y

R E G I O N

1.00 0.80 0.60 0.40 0.20 0.00

SUBSTITUTION Eq. 2

Eq. 3

products/(feedstock + products/(feedstock + auxil. energy need) auxil. energy need)

PRESENT

SUPPLY

Eq. 6 products/primary non-RE need

ENERGY

DEMAND

Eq. 7

Eq. 9

products/biomass potential

WATER

non-RE use/(biomass potential + non-RE need)

ECONOMY

Fig. 4. The energy efficiencies at various system levels of the biorefinery scenarios in the case regions.

biorefining are applicable to other biorefining systems, yet dedicated energy crops are an exception, with their impacts on crop production and land use change. However, the conceptual systems framework developed and applied here would also fully serve better understanding of efficiency and the impacts of dedicatedenergy-crops-based biorefining. Irrespective of the designs and products of biorefinery systems, in addition to high efficiency in conversion processes and transportation, are high exploitation efficiency of the untapped biomass potential and high substitution efficiency the keys to sustainable biorefining. 4. Conclusions Design and system boundaries decisively influence the relative efficiency of biorefinery scenarios. Biorefining of agrifood waste increases nutrient and energy efficiency in comparison with the current situation. Nutrient efficiency is increased through comprehensive biomass coverage, AD and efficient substitution of mineral fertilisers. Proximity of conversion to heat utilisation and a high share of combustion enhance energy efficiency. Regional differences in agrifood-waste-based biorefining efficiency depend on agricultural structure, the extent of the food industry and population density. In conclusion, high degrees of exploitation of feedstock potential and substitution efficiency, rather than efficiency in conversion, are the keys to sustainable biorefining.

Table 4 The efficiency of complementary mineral fertiliser (nitrogen (N) and phosphorus (P)) need for yield production (yield/mineral fertiliser need, Eq. (5)). For PRESENT, the current mineral fertiliser use is assumed. PRESENT South Savo Biorefinery region N 1.20 P 1.31 Satakunta Biorefinery region N 0.96 P 1.04

ENERGY

WATER

ECONOMY

1.57 4.40

1.71 15.7

1.56 3.29

2.87 –a

1.61 –a

3.91 –a

a In Satakunta, there is no P fertiliser need; the organic fertilisers will totally meet the P demand, and even create a surplus.

Acknowledgements The authors wish to acknowledge Professor Markku Virtanen, Small Business Center of Aalto University, School of Business, for his contribution to the scenario design. The authors are grateful for financial support from the Finnish Funding Agency for Technology and Innovation, the European Social Fund of the EC (the ELY Centre of South Savo), the South Savo Regional Fund of the Finnish Cultural Foundation and MTT Agrifood Research Finland.

M. Kuisma et al. / Bioresource Technology 135 (2013) 588–597

Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j.biortech.2012.11. 038. References Anex, R.P., Lynd, L.R., Laser, M.S., Heggenstaller, A.H., Liebman, M., 2007. Potential for enhanced nutrient cycling through coupling of agricultural and bioenergy systems. Crop Sci. 47, 1327–1335. Bailey, C., Dyer, J.F., Teeter, L., 2011. Assessing the rural development potential of lignocellulosic biofuels in Alabama. Biomass Bioenergy 35, 1408–1417. Biogasmax, 2006. Report on Technological Applicability of Existing Biogas Upgrading Processes. European Commission. Bunge, M., 1985. Treatise on Basic Philosophy, vol. 7. Epistemology & Methodology III: Philosophy of Science and Technology Part II. Reidel, Dordrecht. Cayuela, M.L., Oenema, O., Kuikman, P.J., Bakker, R.R., van Groenigen, J.W., 2010. Bioenergy by-products as soil amendments? Implications for carbon sequestration and greenhouse gas emissions. GCB Bioenergy 2, 201–213. Chen, S., Chen, B., Song, D., 2012. Life-cycle energy production and emissions mitigation by comprehensive biogas–digestate utilization. Bioresour. Technol. 114, 357–364. Cherubini, F., Jungmeier, G., Wellisch, M., Willke, T., Skiadas, I., Van Ree, R., de Jong, E., 2009. Toward a common classification approach for biorefinery systems. Biofuels Bioprod. Bioref. 3, 534–546. Cui, Z., Shi, J., Wan, C., Li, Y., 2012. Comparison of alkaline- and fungi-assisted wetstorage of corn stover. Bioresour. Technol. 109, 98–104. Davis, J., Haglund, C., 1999. Life Cycle Inventory (LCI) of Fertiliser Production, Fertiliser Products Used in Sweden and Western Europe. Master’s Thesis. SIKReport No. 654. Chalmers University of Technology. Digiroad, 2009. A National Road and Street Database. ESRI Finland, the Finnish Transport Agency. Dijkstra, E.W., 1959. A note on two problems in connexion with graphs. Num. Math. 1, 269–271. Directive 2008/98/EC of the European Parliament and of the Council of 19 November 2008 on Waste and Repealing Certain Directives. Drinkwater, L.E., Snapp, S.S., 2007. Nutrients in agroecosystems: rethinking the management paradigm. Adv. Agron. 92, 163–186. Edwards, R., Larivé, J.-F., Mahieu, V., Rouveirolles, P., 2006. Well-to-Wheels Analysis of Future Automotive Fuels and Powertrains in the European Context, WELL-toTANK Report, Version 2b (May 2006). ESRI, 2010. ArcGIS Desktop: Release 9.3. Redlands, CA, Environmental Systems Research Institute.

597

Kahiluoto, H., Kuisma, M., Havukainen, J., Luoranen, M., Karttunen, P., Lehtonen, E., Horttanainen, M., 2011. Potential of agrifood wastes in mitigation of climate change and eutrophication – two case regions. Biomass Bioenergy 35, 1983– 1994. Karp, S.G., Igashiyama, A.H., Siqueira, P.F., Carvalho, J.C., Vanderberghe, L.P.S., Thomaz-Soccol, V., Coral, J., Tholozan, J.-L., Pandey, A., Soccol, C.R., 2011. Application of the biorefinery concept to produce L-lactic acid from the soybean vinasse at laboratory and pilot scale. Bioresour. Technol. 102, 1765–1772. Kokossis, A.C., Yang, A., 2010. On the use of systems technologies and a systematic approach for the synthesis and the design of future biorefineries. Comput. Chem. Eng. 34, 1397–1405. Laser, M., Larson, E., Dale, B., Wang, M., Greene, N., Lynd, L.R., 2009. Comparative analysis of efficiency, environmental impact, and process economics for mature biomass refining scenarios. Biofuels Bioprod. Bioref. 3, 247–270. Leung, C.C.J., Cheung, A.S.Y., Zhang, A.Y.-Z., Lam, K.F., Lin, C.S.K., 2012. Utilisation of waste bread for fermentative succinic acid production. Biochem. Eng. J. 65, 10– 15. LIPASTO, 2011. Traffic Emissions. VTT Technical Research Centre of Finland. (accessed 01.06.12). Martínez Sabajanes, M., Yáñez, R., Alonso, J.L., Parajó, J.C., 2012. Pectic oligosaccharides production from orange peel waste by enzymatic hydrolysis. Int. J. Food Sci. Techol. 47, 747–754. Mattsson Petersen, C., Berg, P.E.O., Rönnegård, L., 2005. Quality control of waste to incineration – waste composition analysis in Lindköping, Sweden. Waste Manage. Res. 23, 527–533. Møller, H.B., Sommer, S.G., Ahring, B.K., 2002. Separation efficiency and particle size distribution in relation to manure type and storage conditions. Bioresour. Technol. 85, 189–196. Organics, 2012. Ammonia Stripping, Systems for Ammonia Removal from Landfill Leachate. Organics Limited. (accessed 04.07.12). Phyllis, 2012. Database for Biomass and Waste. Energy Research Centre of the Netherlands. (accessed 01.06.12). Pöschl, M., Ward, S., Owende, P., 2010. Evaluation of energy efficiency of various biogas production and utilization pathways. Appl. Energy 87, 3305–3321. Salo, T., Lemola, R., Esala, M., 2007. National and regional net nitrogen balances in Finland in 1990–2005. Agric. Food Sci. 16, 366–375. Uhlen, G., 1976. Effects of nitrogen, phosphorus and potassium fertilizers and farm manure in longterm experiments with rotational crops in Norway. Ann. Agron. 27, 547–564. Uusitalo, R., Turtola, E., Grönroos, J., Kivistö, J., Mäntylahti, V., Turtola, A., Lemola, R., Salo, T., 2007. Finnish trends in phosphorus balances and soil test phosphorus. Agric. Food Sci. 16, 301–316. Verburg, P.H., Rounsewell, M.D.A., Veldkamp, A., 2006. Editorial. In: Verburg, P.H., Rounsewell, M.D.A., Veldkamp, A. (Eds.), Scenario-based Studies of Future Land Use in Europe, Special Issue, vol. 114. Agric. Ecosyst. Environ, pp. 1–6.