Journal of Cleaner Production xxx (2013) 1e8

Contents lists available at SciVerse ScienceDirect

Journal of Cleaner Production

journal homepage: www.elsevier.com/locate/jclepro

Life Cycle Assessment for environmentally sustainable aquaculture management: a case study of combined aquaculture systems for carp and tilapia
Rattanawan Mungkung a, b, c, *, Jol Aubin a, b, Tri Heru Prihadi d, Jacques Slembrouck e, Hayo M.G. van der Werf a, b, Marc Legendre e
a Institut National de la Recherche Agronomique (INRA), UMR 1069, Sol Agro et hydrosystme Spatialisation, 65 rue de Saint Brieuc, CS 84215, 35042 Rennes Cedex, France b Agrocampus Rennes, UMR 1069, Sol Agro et hydrosystme Spatialisation, 35000 Rennes, France c Centre of Excellence on Environmental Strategy for GREEN Business (VGREEN), Department of Environmental Technology and Management, Faculty of Environment, Kasetsart University, 50 Ngamwongwan Road, Chatuchak, Bangkok 10903, Thailand d Research Centre for Aquaculture, Department of Marine Affairs and Fisheries, Bogor, Indonesia e Institut de Recherche pour le Dveloppement (IRD), UMR 226 e ISEM, BP 5095, 34196 Montpellier cedex 05, France

a r t i c l e i n f o
Article history: Received 5 November 2012 Received in revised form 13 May 2013 Accepted 17 May 2013 Available online xxx Keywords: Cage aquaculture Cyprinus carpio Environmental sustainability Indonesia Life Cycle Assessment Oreochromis niloticus

a b s t r a c t
Life Cycle Assessment (LCA) was applied to evaluate the potential environmental impacts associated with two-net cage aquaculture systems of common carp (Cyprinus carpio carpio) and tilapia (Oreochromis niloticus) in the Cirata reservoir, Indonesia. The studied system included ngerling production in hatcheries, sh rearing in cages, and transport of fry and feed as well as that of harvested sh to markets. The environmental impact indicators were calculated based on the annual production in 2006e2007 using the CML2 Baseline 2000 method, and expressed per tonne of fresh sh delivered to the market. The rearing performances and the environmental efciency of the system were highly dependent on the lake water quality. Therefore the location of the cages and associated practices inuenced the environmental impacts. Feed was identied as the major contributor to land occupation, primary production use, acidication, climate change, energy use and water dependence. Those impacts were mainly linked to the production of shmeal followed by the production of crop-based feed materials and the production of electricity for feed processing. Eutrophication was mainly the consequence of the sh growing stage and linked closely to nutrient loading from cages. Better feeding practices to reduce feed conversion ratio (FCR), as well as improvement of feed composition by using less shmeal and more local plant-based materials along with improving energy efciency of feed production processing should be implemented to improve the environmental proles of carp and tilapia production. The reduction of FCR from 2.1 to 1.7 could decrease eutrophication by about 22%. However, it is of rst priority to reduce the number of cages in order to improve the water quality of the reservoir. The comparison of Cirata reservoir sh culture to other sources of animal protein revealed that it generated average energy use but high eutrophication level. LCA was demonstrated to be a useful tool for decision-making when targeting improved environmental sustainability of cage aquaculture. 2013 Elsevier Ltd. All rights reserved.

1. Introduction Indonesia is the 4th aquaculture producer in the world, with 1.7 M tonnes in 2008, and the production increase reaches 10% per year (FAO, 2010). Fish is a major source of animal protein for the
* Corresponding author. Centre of Excellence on Environmental Strategy for GREEN Business (VGREEN), Department of Environmental Technology and Management, Faculty of Environment, Kasetsart University, 50 Ngamwongwan Road, Chatuchak, Bangkok 10903, Thailand. Tel.: 66 2 562 4555x1508, fax: 66 2 942 8715. E-mail address: fscirwm@ku.ac.th (R. Mungkung). 0959-6526/$ e see front matter 2013 Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.jclepro.2013.05.029

population and cultivated sh (mostly carps, tilapias, pangas and gouramis) are highly appreciated because of their freshness, as they are generally sold alive on local markets. Floating net cage aquaculture of common carp (Cyprinus carpio carpio) was implemented on the 6200 ha Cirata reservoir (Citarum River, West Java, Indonesia) in 1988, following the construction of a dam for hydroelectricity production. The cultured species were selected due to a high market demand for carp and its considerable economic value, using technology from previous experiences in other areas in Indonesia. Over the years, a two-net cage aquaculture

Please cite this article in press as: Mungkung, R., et al., Life Cycle Assessment for environmentally sustainable aquaculture management: a case study of combined aquaculture systems for carp and tilapia, Journal of Cleaner Production (2013), http://dx.doi.org/10.1016/ j.jclepro.2013.05.029

R. Mungkung et al. / Journal of Cleaner Production xxx (2013) 1e8

system in which carps are stocked in the upper net and tilapias (Oreochromis niloticus) in the lower net, was adopted in an attempt to maximize feed utilization and reduce nutrient loading from cages into the reservoir. The rationale of this carp-tilapia polyculture system is to use the waste output (uneaten feed and sh faeces) from carp and fouling induced by nutrient emissions as feed for tilapia, which also provides supplementary income. Tilapias also help to facilitate water circulation in the cages by grazing on biofouling developing on the nets. Nevertheless, due to their lower market price tilapia are considered as a co-product in this system. The two-net cage aquaculture system has become a common aquaculture practice since 1995, supplying a local livelihood and good investment for income. However, the rapid adoption of the Cirata reservoir as a new aquatic resource for sh production has resulted in the establishment of a massive number of cages. Due to the higher prot gained from cage aquaculture as well as a lower investment requirement compared to pond production, the number of cages has strongly increased from 25,558 cages in 1996 (Costa-Pierce, 1998) to 39,300 cages in 2005 (Abery et al., 2005). Previous studies reported the carrying capacity of the lake based on its capacity for self-purication to be around 10,600 cages (CostaPierce, 1998) and 18,500 cages (Murniyati et al., 2006), whereas it was estimated around 20,300 cages based on empirical results of maximum productivity per cage (Abery et al., 2005). However, after 2003, the number of cages in activity (>38,000) was about twice the estimated carrying capacity of the reservoir. This high cage density has led to a declining productivity per cage, from 3.5 to 7.0 to less than 1.2e1.5 tons per cage per year (Prihadi, 2003). The average weight of harvested carps also decreased from 1000 to 350e500 g, as reported by cage operators surveyed. Water quality analyses from twelve sampling stations across the reservoir found the concentrations of sulphate, hydrogen sulphide, ammonia, nitrite, phosphate and organic matter exceeding water quality standards for aquaculture (Murniyati et al., 2006). In addition, an increased susceptibility to disease and rate of mortality was observed in the reservoir (Bondad-Reantaso, 2004). In this context, this paper explores the environmental implications of cage aquaculture in the Cirata reservoir, through a system analysis of input production (supporting production systems), sh production (main production system), and transport activities (transport of inputs and intermediate products in all stages) associated with the whole supply chain in order to identify areas for environmental sustainability improvement. 2. Methodology Life Cycle Assessment (LCA) was applied to evaluate the environmental consequences of the two-net cage aquaculture production system developed in the reservoirs of West Java. Previous studies have demonstrated the potential uses of LCA in aquaculture management (Papatryphon et al., 2004; Mungkung et al., 2005; Mungkung, 2005; Aubin et al., 2006; Ellingsen and Aanondsen, 2006; Mungkung and Gheewala, 2007; Aubin et al., 2009; Ayer et al., 2007; Pelletier and Tyedmers, 2007; Pelletier et al., 2009; Pelletier and Tyedmers, 2010). LCA standardized methodology is described in the International Organization for Standardization (ISO) standards (ISO, 2006a; ISO, 2006b) and each phase of methodology adopted in this LCA study is explained hereafter. 2.1. System boundary The system evaluated in this study covers the supply chain from hatchery (ngerling production) to farm (sh rearing in cages) and transport of fry and feed as well as that of harvested sh to local markets. The functional unit was one tonne of fresh sh delivered

to the market. Downstream life cycle stages (i.e. consumption and post-consumption, waste management) were not considered, as they are not relevant for aquaculture management decisions. Based on the dened system boundary, the required inventory data included the main inputs (energy and resources) and outputs (wastes and pollutants) in hatchery, cage and transport of live sh to local market by using boats and small trucks. The inventory data requirement also extended to the supporting production systems of the following inputs and outputs: energy resource extraction, energy production, feed ingredients production, chemicals production and infrastructure materials production (Fig. 1). Transport in all stages was also covered. 2.2. Inventory data collection The inventory data for this study were conducted on an annual production basis from 2005 to 2007 by interviewing cage shfarmers representative of typical production practices of different farm sizes. The data collection has been conducted in two stages. The rst stage was a general survey on thirty farms, based on a random sampling within farms of varying technical, social and economic performances. The second stage consisted in a detailed survey on ve farms, to validate the previous data and complete inventories especially on facilities and feed origins. All the collected data were validated by local experts from the Indonesian Department of Marine Affairs and Fisheries. Inventory data for carp ngerling production were gathered by collecting the data from three hatcheries located between 150 and 200 km from the reservoir. All hatcheries used broodstock from the same district; the associated inventory data of broodstock production were obtained through an interview with a representative sh farmer specialised in carp broodstock production. For tilapia ngerlings production, the inventory data were collected by interviewing a tilapia hatchery owner at 80 km from the reservoir. The production of tilapia fry and juveniles was conducted in ponds using natural productivity enhanced by chicken manure fertilisation, and the addition of rice bran or rice meal (homemade enrichment, Table 1). The inventory data for chicken manure production was provided by a study on chicken production in France and Brazil (Da Silva et al. 2010), and adapted to local context by specic data (density, growth, feed conversion) from a chicken grower who sells chicken manure. Environmental impacts associated with products from chicken production (chicken meat and chicken manure) were allocated according to their economic values, estimated by their annual average prices. Commercial pelleted feeds are used for rearing carp broodstock (Feed 1), and for grow-out of both carp ngerlings in the hatcheries and sh in the cages (Feeds 2e4). Their chemical and ingredients compositions (Table 1) were determined on the commercial labels completed by interviews of a feed company nutritionist and the manager of a feed mill. The feed formulations were then validated by nutritional value according to Guillaume et al. (1999). This data was complemented by detailed information from a local feed producer, including the energy use for feed production. The inventory data for local shmeal production was collected from a local shery processing plant dedicated to frozen sh (for export), surimi and shmeal (Palembang, Sumatra Island). Fishmeal was produced from discarded sh from the freezing and surimi activities. Heat energy from wood wastes was used for shmeal processing. Environmental impacts associated with products from trawling sheries (exported sh, surimi raw material sh, and trash sh) were allocated according to their economic values. Data for world market crop-based ingredients were considered according to Boissy et al. (2011). The inventory data for Indonesian electricity production were obtained from Widiyanto et al. (2003).

Please cite this article in press as: Mungkung, R., et al., Life Cycle Assessment for environmentally sustainable aquaculture management: a case study of combined aquaculture systems for carp and tilapia, Journal of Cleaner Production (2013), http://dx.doi.org/10.1016/ j.jclepro.2013.05.029

R. Mungkung et al. / Journal of Cleaner Production xxx (2013) 1e8

Fig. 1. The scope of the double-cage aquaculture production system analysed in this study.

In terms of outputs, the amounts of organic and nutrient loading from hatchery and cage production systems were calculated based on a nutrient mass balance method (Cho and Kaushik, 1990). The emission factors associated with engine boats used for transporting
Table 1 Ingredients and chemical composition of the feeds used in the two-net cage aquaculture system in Cirata reservoir. Feed ingredients (origin) Feed content (g/kg) Homemade Commercial feedsb Feed 2 0 300 120 80 20 220 110 100 50 Feed 3 0 220 170 0 30 230 200 100 50 Feed 4 0 230 150 0 50 220 300 0 50 24e26% 6e8% 4e6% 5e8% 1.20% 11e13%

feeds ingredients, commercial feeds, ngerlings and harvested sh were taken from the European Environmental Agency Emission Inventory Guidebook (European Environmental Agency, 2006). For trucks used for the transport of feeds, juveniles and harvested sh from shore to local market, the associated inventory data were taken from the Ecoinvent 2.2 database (Swiss Centre for Life Cycle Inventories, 2010). 2.3. Farms characteristics The production methods and the double cage system have been supported by a Governmental institution (Research Centre for Aquaculture) and standardized for economic reasons and easy dissemination. A typical production unit is composed of a bamboo raft supporting four upper net cages (7 7 3 m) for common carps and one lower net cage (15 15 5 m) surrounding the four upper cages, for tilapias. The units are combined with other facilities (i.e. buoyancy, feeding station, cage pathway, and living house). The farms were classied considering three size classes based on the number of production units per owner: small (<5 units), medium (6e20 units) and large (>20 units). In this study we focussed on small and large farms, in order to compare contrasting practices. Cages are usually operated all year round, with three and two production cycles for carp and tilapia respectively. All farms used commercial pelleted feeds to support carp growth. The feed was distributed manually by cage caretakers on the upper cages, approximately ve times per day. Feed 2 was the most used and corresponded to more than 40% of the total feed used by cage farms on the reservoir (Pk Koko, feed marketing manager, pers. comm.).

Enrichmenta Feed 1 Chicken manure (Indonesia) Rice meal (Indonesia) Fishmeal (Indonesia) Fishmeal (Peru) Fish oil (Peru) Soymeal (Brasil) Wheat bran (Argentina) Maize (USA) Additive Chemical composition Protein Lipid Fibre Ash Phosphorus Humidity 500 500 0 0 0 0 0 0 0 N/A N/A N/A N/A N/A N/A 0 300 0 150 0 300 200 0 50

28e30% 29% 25% 6e8% 5% 6% 4e6% 7.50% 7% 5e8% 12% 12% 1.20% 1.50% 1.30% 11e13% 12% 12%

a Homemade enrichment: ingredients distributed in ponds for tilapia ngerlings production. b Feed 1: used for maintenance of carp broodshes. Feeds 2, 3 and 4: used for grow-out of carps in hatcheries or cage culture.

Please cite this article in press as: Mungkung, R., et al., Life Cycle Assessment for environmentally sustainable aquaculture management: a case study of combined aquaculture systems for carp and tilapia, Journal of Cleaner Production (2013), http://dx.doi.org/10.1016/ j.jclepro.2013.05.029

R. Mungkung et al. / Journal of Cleaner Production xxx (2013) 1e8

Small farms are usually located near to the shore for easy access, whereas large farms tended to position far from the shore for better water quality. Moreover, large farms were likely to give more feed than small farms due to higher stocking density as well as better nancial capacity for buying commercial pelleted feeds. However, feeding management was highly dependent on the farm owners experience and in many cases depended on caretakers skills. These factors have resulted in varied Feed Conversion Ratio (FCR). FCR is dened here as the ratio of the total feed input into cage system over the net gain of cumulated carp and tilapia wet biomass (i.e. the weight of harvested carp and tilapia minus the initial weight of carp and tilapia ngerlings). The FCR from our eld observations ranged between 1.7 and 2.1 and it was noticed that large farms tended to have lower FCRs than small farms. These FCRs were in accordance to those (1.7) reported by Pelletier and Tyedmers (2010) for tilapia lake culture in Indonesia. Consequently, two contrasted practices were dened (Table 2): Practice 1 (large farms) was associated to a high stocking level of carp ngerlings and a moderate FCR level due to good feeding practice or relatively good water quality; Practice 2 (small farms) was associated to lower carp stocking density and higher FCR due to poor feeding management or relatively worst water quality. Three main different types of feed (Feeds 2e4) were used at the cage growing stage. Their compositions are given in Table 1. The shmeal (both local and non-local) contents were 15e20% and the sh oil contents were 2e5%. Fishmeal was produced locally and sourced partially from overseas, whereas sh oil was exclusively imported. Rice meal (22e30%), soy meal (22e23%), and wheat bran (11e30%) were the crop-derived feed materials mainly used in all feeds. Maize was only applied in Feeds 2 and 3 (10%) but not in Feed 4. Rice meal was produced locally whilst the others ingredients (maize, wheat bran) were sourced from abroad and transported to Indonesia by truck and boat. In terms of nutritional composition, Feed 2 had slightly higher protein content compared to Feeds 3 and 4 due to a higher proportion of shmeal which was the main protein source. Associated with a high content of rice meal, Feed 2 had slightly higher bre and ash contents than Feeds 3 and 4. The lipid level in all feeds was in a similar range. The nutritional levels of the three feeds fell into the acceptable range recommended for carp and tilapia (Lovell, 1991). No difference on sh growth was observed by using these different feeds, thus it could be assumed that the quality of these three feeds was relatively equivalent and did not lead to signicant differences of FCR; above all, it is the deteriorating water quality in the reservoir which dominantly inuenced the cage productivity. In a rst step, the two farm practices using the same Feed (n 2) were assessed. In a second step the inuence of feed compositions was analysed.

2.4. Impact indicators and assessment method The associated inventory data were transformed into impact indicator values and expressed per tonne of harvested sh (carp and tilapia) delivered to the market. The impact indicators considered in this study were: water dependence (m3), energy use (MJ), land occupation (m2 year), climate change potential (CCP, kg CO2-equivalent), acidication potential (AP, kg SO2-equivalent), eutrophication potential (EP, kg PO4-equivalent) and net primary production use (NPP, kg of carbon content). Water dependence included the volume of water owing into cages as proposed for aquaculture production systems by Aubin et al. (2009), as well as the amount of water used in ponds at the hatchery production system. It also includes the water use for irrigation of crops, especially in rice elds. The land areas occupied for crop-based feed material production, ponds, hatchery (but not the oating farm facilities as the lake surface is not is not in competition with other terrestrial activities) were taken into account for the land occupation. NPP refers to the biotic resource use both from plant or animal materials in terms of carbon contents as a result of net carbon ux in trophic levels (i.e. the position it occupies in a food chain to indicate the biomass transfer efciency and net primary production), which was calculated based on the method described in Papatryphon et al. (2004). The characterisation factors for CCP, AP and EP were based on the CML2 Baseline 2000 version 2.03 (Guine et al., 2002). Energy use was calculated according to the Cumulative Energy Demand method, version 1.03 (Frischknecht et al., 2004). 3. Results and discussion The impact analysis of the two rearing practices using Feed 2 is presented in Table 3. The difference between the impacts of carps and tilapias, relatively to the whole sh production, was 30% when considering the economic allocation which associates the production level and the market value of species. This difference illustrates the role of the two species: carp being the target species with the higher market price (10,000 Rp1/kg) and the higher production level (around 94% of total production), tilapia playing the role of secondary species devoted to cleaning, with a lower value (7000 Rp/kg). The lower level for all impact categories for one tonne of sh, was obtained for Practice 1, characterised by higher rearing densities in large farms, better sh feeding practices (indicated by lower FCR) and better water quality environment. The difference in impact values between the two rearing practices was variable. It ranged from 12% for energy use to 40% for eutrophication, with a main group of impact variation around 14% for land occupation, climate change, NPP use and acidication. This result suggests that the difference between impacts of the two practices is less related to variations in input consumption than to the efciency of input use, especially of feed which leads to a high level of eutrophication. The higher FCR observed in practice 2 farms was associated with an increased quantity of feed required for the production of sh and consequently increased quantities of inputs (energy, shery and crop-based sources). On average, one tonne of total sh production (carp and tilapia) required 1000 m3 of water, 71% due to irrigation of crop production (especially for rice), 19% due to water inow into cages during the rearing period, and 10% for ngerlings production. A land area of 1700 m2 year1 was needed, mostly for crop-based feed materials production. As a result of fuel and electricity use, 31,000 MJ of energy (Cumulative Energy Demand) was used. Emissions to air contributed to climate change of 1900 kg CO2 eq. and acidication of 15 kg SO2 eq. per tonne of sh. Eutrophication reached 122 kg PO4 eq. on average, mainly due to the nutrient discharge

Table 2 Main characteristics of the two-net cage aquaculture system according to contrasted practices. Practice 1: large farms, high sh stocking density, generally appropriate feeding practice or water quality; Practice 2: small farms, relatively lower sh stocking density, feeding practice efciency or water quality than for practice 1 (see text for details). Level of inputs and outputs Average values Practice 1 Carp ngerlings (no./unit/year) Tilapia ngerlings (no./unit/year) Pellet feed (kg/unit/year) Harvested carp (kg/unit/year) Harvested tilapia (kg/unit/year) Feed Conversion Ratio 90,000 10,000 33,170 18,450 1640 1.7 Practice 2 72,000 6500 21,226 10,368 728 2.1

Please cite this article in press as: Mungkung, R., et al., Life Cycle Assessment for environmentally sustainable aquaculture management: a case study of combined aquaculture systems for carp and tilapia, Journal of Cleaner Production (2013), http://dx.doi.org/10.1016/ j.jclepro.2013.05.029

R. Mungkung et al. / Journal of Cleaner Production xxx (2013) 1e8

Table 3 Potential environmental impacts associated with the production of one tonne of sh (carp and tilapia), according to practices. Practice 1: large farms, high sh stocking density, generally appropriate feeding practice or water quality; Practice 2: small farms and relatively lower sh stocking density, feeding efciency or water quality than for practice 1 (see text for details). Variation is calculated by dividing the difference of values by the values average of practices 1 and 2. Impact category Unit Per ton of fresh sh delivered to the market Practice 1 Total sh Water dependence Land occupation Energy use Climate Change Acidication Eutrophication NPPU m3 m2 year MJ kg CO2 eq. kg SO2 eq. kg PO4 eq. kg C 877 1585 28,645 1727 13.6 98 14,205 Carp 899 1624 29,680 1747 14.1 100 14,562 Tilapia 629 1138 20,785 1253 9.9 70 10,197 Practice 2 Total sh 1121 1839 32,945 2025 15.9 146 16,462 Carp 1144 1876 33,610 2065 16.1 150 16,794 Tilapia 800 1312 23,501 1444 11.3 105 11,743 Variation

DPractice/Practice
Total sh 24.4% 14.8% 12.9% 14.8% 14.9% 39.3% 14.7% Carp 24.0% 14.4% 12.4% 16.7% 13.2% 40.0% 14.2% Tilapia 23.9% 14.2% 12.3% 14.2% 13.2% 40.0% 14.1%

from cages in the reservoir (95%). Linked to the carbon contents of all feed materials, the net primary production inputs were estimated around 15,000 kg of carbon. A contribution analysis was conducted to identify key areas generating signicant impacts. Fig. 2 shows the average of the relative contribution of system components to total impacts associated with one tonne of carp and tilapia production, using Feed 2. Feed production was pinpointed as the major contributor to land occupation (99%), primary production use (100%), acidication (88%), climate change (80%), energy use (75%) and water dependence (71%). The relatively low level of water dependence associated with the cage culture is due to the very low water renewing rate of the reservoir, and high sh density in the cages.1 Feed production and feed use at cages were the primary drivers of sh production impacts. Thus, an analysis of potential impacts associated with different pelleted feeds used at Cirata Lake was conducted. As the impacts associated with feeds used at hatcheries (both for broodstock rearing and ngerling production) were slight, the analysis was based on feeds used during the cage growing stage with Practice 1. From the survey we do not have any information that could suggest a difference between the feed performances, which are more driven by producer practices and water quality. Moreover the difference in chemical composition cannot be the cause of a difference in performances due to the environmental context. Consequently we assumed that the same amount of the different feeds is distributed in the simulation. The results, presented in Table 4, indicate that the use of Feed 2 resulted in the lowest impacts on energy use and climate change. Nevertheless, a major improvement could be obtained for eutrophication (21%) and NPP use (52%), using respectively Feed 4 and Feed 3. Nutrients (especially nitrogen compounds) released from cages are the main cause leading to the high eutrophication level, which is directly related to the protein content of feeds (higher in Feed 2). Moreover, the use, in Feed 2, of high trophic-level trash sh species as raw materials in both local and non-local shmeal processing resulted in a high contribution to net primary production use. Leiognathus sp., Mulloidichthys sp., Caesio sp., Trichiurus haumela, Lethrinus sp., and Escualosa thoracata which are used in local shmeal production have a trophic level of 3.3e3.5 (Pauly and Christensen, 1997) whilst Engrauis encrasicholus, Sardina pilchardus, and Trachurus symmetricus (Papatryphon et al., 2004) which are used for Peruvian shmeal production have a trophic level of 3.1 (Pauly and Christensen, 1997). None of the three feeds resulted in the best or worst environmental performance for all impacts. To choose which feed type to use, a level of importance should be assigned to each impact

category in order to provide a basis of comparison. As eutrophication is the major concern issue among farmers and local people due to the deteriorating water quality in the reservoir, it was considered the most important impact. As a result, Feed 2 performed worst because of its highest impact on eutrophication and net primary production use. The comparison of Feeds 3 and 4 indicated that Feed 3 resulted in a higher requirement of energy, mainly due to the various plant materials required. Therefore, it could be assumed that Feed 4 was the best option to reduce eutrophication. Nevertheless, the relative importance of impacts strongly depends on value choices of decision makers and the result will be inuenced consequently. The results also indicated that the higher protein content of feed was likely to result in a higher eutrophication potential as a result of nutrient released from cages. However, the potential impact on eutrophication resulted not only from animal-derived feed ingredients but also from crop-based feed materials to some extent. The use of plant materials was associated with a high impact on land occupation and on eutrophication to some degree. Transport of non-local plant feed ingredients was associated with impacts on climate change and acidication. This suggests that a substitution of plant materials to reduce the use of shmeal might reduce the impact on eutrophication but possibly shift the environmental burden to climate change and energy use as a trade-off. The ndings on crop-based feed compositions in this study are in close

1 US$ 9600 Indonesian Rp.

Fig. 2. Relative contributions of system components to total impacts associated with one ton of carp and tilapia production using Feed 2 (see Table 1 for composition), average of practice 1 (large farms, high sh stocking density, generally appropriate feeding practice or water quality) and practice 2 (small farms, relatively low sh stocking density, poor feeding practice or water quality; see text for details).

Please cite this article in press as: Mungkung, R., et al., Life Cycle Assessment for environmentally sustainable aquaculture management: a case study of combined aquaculture systems for carp and tilapia, Journal of Cleaner Production (2013), http://dx.doi.org/10.1016/ j.jclepro.2013.05.029

R. Mungkung et al. / Journal of Cleaner Production xxx (2013) 1e8

Table 4 Potential environmental impacts associated with the production of one tonne of sh (carp and tilapia) delivered to market, according to the different types of feed used on sh farms of the Cirata Reservoir (see Table 1 for details on feed composition), maximum variation between feed values and least impacting feed. The maximum variation refers to the difference between extreme values relative to their average: [max(feed) min(feed)]/[(max(feed) min(feed)/2]. Impact category Unit Per ton of fresh sh delivered to the market Practice 1 Feed 2 Water dependence Land occupation Energy use Climate Change Acidication Eutrophication NPPU m3 m2 year MJ kg CO2 eq. kg SO2 eq. kg PO4 eq. kg C 877 1585 28,645 1727 13.6 98 14,205 Feed 3 711 1648 36,662 1762 12.9 80 8319 Feed 4 729 1603 32,188 1925 12.3 79 9292 Maximum variation 21% 4% 25% 11% 10% 21% 52% Least impacting Feed Feed Feed Feed Feed Feed Feed Feed 3 2 2 2 4 4 3

agreement with previous studies on trout and salmon feeds (Papatryphon et al., 2004; Pelletier and Tyedmers, 2007; Pelletier et al., 2009; Boissy et al., 2011). 4. Application of results

these recommendations are in line with Tacon et al. (2006) and Tacon and Metian (2008). The improvement of energy efciency for feed production processing should be done at the same time to improve the environmental proles of feed. 4.2. Environmental performance benchmarking

4.1. Key areas for environmental improvement The number of cages has been exceeding the estimated carrying capacity of the Cirata reservoir since several years. This has resulted in a deteriorating water quality in the reservoir and a declining productivity. This was obvious when comparing the FCR observed from the present study (1.7e2.1) to those, around 1.4, regularly obtained on most farms of the reservoir in the past (Costa-Pierce, 1998; Abery et al., 2005; Murniyati et al., 2006). The cage number is the main driver of the environmental performance of the cage production system in the Cirata reservoir. The increasing occupation and aquaculture production in the reservoir carries a risk of total collapse of the sh production (accumulation of organic material and irregular occurrences of upwelling from bottom anoxic water). Therefore, the rst action that should be undertaken is the application of the existing regulation on cage number limitation as recommended previously (Costa Pierce, 1998), in accordance with the protected and production area dened. In parallel, the water quality monitoring of the reservoir has to be pursued. The second type of environmental improvement to be undertaken is an increased retention of nutrients by the production system. In this regard the addition of tilapias to carp culture is an important innovation. It permits a supplementary income together with an increased efciency of the system through cleaning of the nets and reusing of carp cage losses. Nevertheless, the tilapia yield remains low in comparison to carp, and could be increased by improving the design of the tilapia cages (size, access to water surface.). The combination with other types of production like molluscs and aquatic plants might offer an interesting potential too (Troell et al., 2009). Feed was identied as an important cause of impacts and thus a key area for environmental improvement. The amount and type of feed given are strongly linked to the impacts of carp and tilapia production, a lower FCR can result in considerable improvement of the environmental proles of culture of these sh. Better feeding management should be implemented: the amount of feed given should take into account the water quality surrounding the cages, and the feeding should be adjusted according to sh health and uctuating weather conditions. With respect to the feed choice, based on the current feed types, use of Feed 3 or 4 could reduce eutrophication by about 25%. For further improvement, reduction of shmeal content in feed should be considered and local cropbased feed materials should be used rather than non-local ones; Apart from the identication of relevant areas for environmental improvement, the LCA results can also be useful to compare the environmental performance of carp and tilapia with other aquaculture and agriculture products. Having provided the magnitude of impacts per one ton of sh produced within the same boundary of analysis (i.e. the farm gate or the landed site especially for cod), the products can be compared. Energy use, climate change, and eutrophication were selected for this comparison, because they are the most relevant issues at the global and local scales. The environmental impacts were calculated for the total sh production (carp and tilapia) to avoid allocation issues and using the Practice 1 and Feed 2 (representative of the main aquaculture practice on the Cirata Reservoir). The transport of sh to market was not included. The potential impacts of carp and tilapia were compared to those of other sources of animal protein (calculated for one tonne of raw animal as functional unit) reported in previous studies (Ziegler et al., 2003; Basset-Mens and van der Werf, 2005; Williams et al., 2006; Pelletier and Tyedmers, 2010; Boissy et al., 2011; Ewoukem et al., 2012; Mungkung et al., 2005) and presented in Fig. 3 as percentages. The allocation rules for co-products were different among the studies. Most of studies used economic based allocation, while Mungkung et al. (2005) used mass based allocation and Pelletier and Tyedmers (2010) used gross energy content based allocation. These studies are therefore not fully comparable, nevertheless comparing the orders of magnitude makes sense. The three case studies in Indonesia (tilapia and carp-tilapia cultures; Pelletier and Tyedmers, 2010, and present study) display similar levels of potential impacts. Nevertheless the impacts of Cirata Reservoir cultures are higher in eutrophication. The potential impact on climate change of the different tilapia and carp cases studied have the lowest levels. The lowest impacts on climate change of tilapia resulted from a lower energy use than in cod and salmon farming, especially in the Cameroon case study. A lower greenhouse gas emission level was also found during the tilapia growing stage compared to pork and chicken. However, carp-tilapia culture presents a high potential impact on eutrophication followed by prawn, but doesnt reach the level of the tilapia-clarias case study, which suffers from poor nutrient and water management. The eutrophication performance of aquaculture productions is mostly associated to the nutrient use efciency and to a direct release of nutrients in the aquatic environment, whereas pork and

Please cite this article in press as: Mungkung, R., et al., Life Cycle Assessment for environmentally sustainable aquaculture management: a case study of combined aquaculture systems for carp and tilapia, Journal of Cleaner Production (2013), http://dx.doi.org/10.1016/ j.jclepro.2013.05.029

R. Mungkung et al. / Journal of Cleaner Production xxx (2013) 1e8

Fig. 3. Relative potential impacts of the production of one ton of carp and tilapia in the Cirata cage-aquaculture system compared to one tonne of other animal products. Allocation methods: Economic allocation: Basset-Mens and van der Werf, 2005; Williams et al., 2006; Boissy et al., 2011; Ewoukem et al., 2012; Mass based allocation: Mungkung et al., 2005; Gross energy based allocation: Pelletier and Tyedmers, 2010.

chicken manure are spread on land. Chicken and pork perform better than aquatic products, especially than cod, in terms of energy use (tilapia e Clarias case excepted). The highest energy use in cod was linked to the diesel use for the trawling operation. This comparison between production systems should be interpreted with caution, because of differences in LCA methodologies: allocation rules and characterisation methods. Nevertheless, the orders of magnitude are robust. It should be noted here that a different basis could be used to compare the environmental performance of different sources of animal protein, for instance, the normalisation of environmental indicator values with the essential nutrient levels given by different meats (Mungkung and Gheewala, 2007) or the edible protein energy output per energy input, the so called edible-protein energy return on investment ratio (Tyedmers, 2004; Tyedmers et al. 2005). However, different functional units may lead to different results. 5. Conclusions This study provided a comprehensive evaluation of potential environmental impacts associated with the two-net cage aquaculture systems of carp and tilapia operated in the Cirata reservoir. LCA results have identied the practices linked to water quality (distance to lake shore) and feeding management as the main causes of impact. The ability of this production system to valorise the feed and to retain nutrients should be improved. The carptilapia complementarity should be enhanced and completed by other practices (i.e. mollusc culture). Production of feed ingredients and their origins, energy use in feed manufacture, and nutrients released from feeding sh in the cages, were identied as the main factors contributing to impacts. To improve the environmental proles of carp and tilapia production, these factors should be systematically taken into account for improving the feed quality along with better feeding management in order to lower the FCR. Nevertheless, these improvements will not be sufcient to upgrade the water quality of the reservoir, and to establish a virtuous circle, giving good conditions to improve the sh production efciency and diminish the efuent level. We therefore recommend a reduction of the number of cages in accordance with the lakes carrying capacity and to survey the water quality of the whole lake.

The analysis on feed type inuence was useful to determine a quantitative potential improvement by means of feed composition. Nevertheless, this result is not univocal, and shows a trade-off between energy use and eutrophication. The environmental performance comparison facilitated the benchmarking of carp and tilapia with other cultivated species. For instance, the Cirata reservoir production systems have a better efciency for energy use and a lower emission of greenhouse gas than salmon farming. Nevertheless, their hot spot remains eutrophication. This study thus showed the usefulness of LCA for better understanding the potential environmental impacts associated with aquaculture product/production systems and providing supporting information to make decisions for more sustainable aquaculture management. The data produced by the study should be used in participatory approaches in order to build a common vision between farmers and civil service of the problems of Cirata Reservoir to nd shared solutions, as proposed from the Evad Project (ReyValette et al., 2008; Lazard et al., 2010).

Acknowledgement This paper forms part of the Agriculture and Sustainable Development (Agriculture et Dveloppement Durable) project EVAD nanced by the French National Research Agency (ANR Agence Nationale de la Recherche) e Contract ANR-05-PADD-008.

References
Abery, N.W., Sukadi, F., Kartamihardja, E.S., Koeshendrajana, S., Buddhiman, S., de Silva, S.S., 2005. Fisheries and cage culture of three reservoirs in west Java, Indonesia: a case study of ambitious development and resulting interactions. Fish. Manage. Ecol. 12, 315e330. Aubin, J., Papatryphon, E., Van der Werf, H.M.G., Petit, J., Morvan, Y.M., 2006. Characterisation of the environmental impact of a turbot (Scophthalmus maximus) re-circulating production system using Life Cycle Assessment. Aquaculture 261, 1259e1268. Aubin, J., Papatryphon, E., Van der Werf, H.M.G., Chatzifotis, S., 2009. Assessment of the environmental impact of carnivorous nsh production systems using life cycle assessment. J. Clean Prod. 17 (3), 354e361. Ayer, N.W., Tyedmers, P.H., Pelletier, N.L., Sonesson, U., Scholz, A., 2007. Co-product allocation in life cycle assessments of seafood production systems: review of problems and strategies. Int. J. Life Cycle Assess. 12, 480e487.

Please cite this article in press as: Mungkung, R., et al., Life Cycle Assessment for environmentally sustainable aquaculture management: a case study of combined aquaculture systems for carp and tilapia, Journal of Cleaner Production (2013), http://dx.doi.org/10.1016/ j.jclepro.2013.05.029

R. Mungkung et al. / Journal of Cleaner Production xxx (2013) 1e8 Mungkung, R., Gheewala, S., 2007. Use of life cycle assessment (LCA) to compare the environmental impacts of aquaculture and agri-food products. FAO/WFT Expert Workshop. 24e28 April, 206, Vancouver, Canada. In: Bartley, D.M., Brugre, C., Soto, D., Gerber, P., Harvey, B. (Eds.), Comparative Assessment of the Environmental Costs of Aquaculture and Other Food Production Sectors: Methods for Meaningful Comparisons, FAO Fisheries Proceeding. No. 10. FAO, Rome, pp. 87e96. Murniyati, A.S., Tri Heru, P., Priyono, B., 2006. Management of sustainable oating net cage aquaculture on reservoir. The case study of Cirata reservoir, west Java Province, Indonesia. In: Paper Presentation at the Second International Symposium on Cage Aquaculture in Asia (CAA2), Zhejiang University, Hangzhou, Zhejiang Province, China, 3e8 July. Papatryphon, E., Petit, J., Kaushik, S.J., van der Werf, H.M.G., 2004. Environmental impact assessment of salmonid feeds using Life Cycle Assessment (LCA). Ambio 33 (6), 316e323. Pauly, D., Christensen, V., 1997. Trophic levels of shes. Box 16. In: Froese, R., Pauly, D. (Eds.), FishBase 97: Concepts, Design and Data Sources. ICLARM, Manila, p. 127. Pelletier, N., Tyedmers, P., 2010. Life cycle assessment of frozen tilapia llets from Indonesian lake-based and pond-based intensive aquaculture systems. J. Ind. Ecol. 14, 467e481. Pelletier, N., Tyedmers, P., 2007. Feeding farmed salmon: is organic better? Aquaculture 272, 399e416. Pelletier, N., Tyedmers, P., Sonesson, U., Scholz, A., Ziegler, F., Flysjo, A., Kruse, S., Cancino, B., Silverman, H., 2009. Not all salmon are created equal: life cycle assessment (LCA) of global salmon farming systems. Environ. Sci. Technol. 43 (23), 8730e8736. Prihadi, Tri H., 2003. Upaya Perbaikan Lingkungan Untuk Menunjang Kesinambungan Budidaya Ikan Dalam keramba Jaring Apung (KJA). Pengembangan Budidaya Perikanan di Waduk. In: Suatu Upaya Pemecahan Masalah Budidaya Ikan dalam KJA. Pusat Riset Perikanan Budidaya. DKP, pp. 45e56. Rey-Valette, H., Clment, O., Aubin, J., Math, S., Chia, E., Legendre, M., Caruso, D., Mikolasek, O., Blancheton, J.-P., Slembrouck, J., Baruthio, A., Ren, F., Levang, P., Morissens, P., Lazard, J., 2008. Guide to the Co-construction of Sustainable Development Indicators in Aquaculture e EVAD, p. 144. Swiss Center for Life Cycle Inventories, 2010. Ecoinvent v2 Database. Ecoinvent Center 2010. Tacon, A.G.J.., Hasan, M.R., Subasinghe, R.P., 2006. Use of Fishery Resources as Feed Inputs to Aquaculture Development: Trends and Policy Implications. In: FAO Fisheries Circular No. 1018. Rome. Tacon, A.G.J., Metian, M., 2008. Global overview on the use of sh meal and sh oil in industrially compounded aquafeeds: trends and future prospects. Aquaculture 285, 146e158. Troell, M., Joyce, A., Chopin, T., Neori, A., Buschmann, A.H., Fang, J.G., 2009. Ecological engineering in aquaculture e potential for integrated multi-trophic aquaculture (IMTA) in marine offshore systems. Aquaculture 297, 1e9. Tyedmers, P., 2004. Fisheries and Energy Use. In: Encyclopedia of Energy, vol. 2, pp. 683e693. Tyedmers, P., Watson, R., Pauly, D., 2005. Fueling global shing eets. Ambio 34 (8), 619e622. Widiyanto, A., Kato, S., Maruyama, N., 2003. Environmental impact analysis of Indonesian electric generation systems (Development of a life cycle inventory of Indonesian electricity). JSME. Int. J. Ser. B 46 (4), 650e659. Williams, A.G., Audsley, E., Sandars, D.L., 2006. Determining the Environmental Burdens and Resource Use in the Production of Agricultural and Horticultural Commodities. Main Report, DEFRA Research Project IS0205. Craneld University and DEFRA, Bedford. Ziegler, F., Nilsson, P., Mattsson, B., Walther, Y., 2003. Life cycle assessment of frozen cod llets including shery-specic environmental impacts. Int. J. Life Cycle Assess. 8 (1), 39e47.

Basset-Mens, C., van der Werf, H.M.G., 2005. Scenario-based environmental assessment of farming systems: the case of pig production in France. Agr. Ecosyst. Environ. 105, 127e144. Boissy, J., Aubin, J., Drissi, A., van der Werf, H.M.G., Bell, G.J., Kaushik, S.J., 2011. Environmental impacts of plant-based salmonid diets at feed and farm scales. Aquaculture 321, 61e70. Bondad-Reantaso, M.G., 2004. Trans-boundary aquatic animal diseases focus on Koi Herpes Virus (KHV). Aquacult. Asia 9, 24e28. Cho, C.Y., Kaushik, S.J., 1990. Nutritional energetics in sh: protein and energy utilization in rainbow trout. In: Bourne, G.H. (Ed.), Aspects of Food Production, Consumption and Energy Values. World Rev. Nut. Diet 61, 132e172. Costa-Pierce, B.A., 1998. Constraints to the sustainability of cage aquaculture for resettlement from hydropower dams in Asia: an Indonesian case study. J. Environ. Devel 7, 333e336. Da Silva, V.P., Van der Werf, H.M.G., Soares, S.R., 2010. LCA of French and Brazilian poultry broiler production systems. Working towards a more sustainable agrifood industry. In: Proceedings of the 7th International Conference on Life Cycle Assessment in the Agri-food Sector, Bari, Italy, 22e24 September 2010. Ellingsen, S., Aanondsen, S.A., 2006. Environmental impacts of wild caught cod and farmed salmon e a comparison with chicken. Int. J. Life Cycle Assess. 1 (1), 60e65. European Environmental Agency (EEA), December, 2006. EMEP/CORINAIR Emission Inventory Guidebook. Available online at: http://reports.eea.europa.eu/ EMEPCORINAIR3/en/page002.html (last accessed March, 2008). Ewoukem, T.E., Aubin, J., Mikolasek, O., Corson, M.S., Eyango, M.T., Tchoumboue, J., van der Werf, H.M.G., Ombredane, D., 2012. Environmental impacts of farms integrating aquaculture and agriculture in Cameroon. J. Clean. Prod. 28, 208e214. Food and Agriculture Organization of the United Nations (FAO), 2010. State of World Fisheries and Aquaculture 2010. FAO, Rome, p. 224. Frischknecht, R., Jungbluth, N., Althaus, H.J., Doka, G., Dones, R., Hirschier, R., Hellweg, S., Humbert, S., Margni, M., Nemecek, T., Speilmann, M., 2004. Implementation of Life Cycle Impact Assessment Methods (Version 1.1). EcoInvent Report No. 3. Swiss Centre for Life Cycle Inventories, Dbendorf, p. 116. Guillaume, J., Kaushik, S.J., Bergot, P., Mtailler, R. (Eds.), 1999. Nutrition et alimentation des poissons et crustacs. INRA Editions, Paris, p. 489. Guine, J.B., Gorre, M., Heijungs, R., Huppes, G., Kleijn, R., de Koning, A., van Oers, L., Wegener Sleeswijk, A., Suh, S., Udo de Haes, H.A., de Bruijn, H., van Duin, R., Huijbregts, M.A.J., 2002. Handbook on Life Cycle Assessment. An Operational Guide to the ISO Standards. Kluwer Academic Publishers, Dordrecht, The Netherlands, p. 692. ISO (The International Organization for Standardisation), 2006a. Environmental Management e Life Cycle Assessment e Principles and Framework. ISO 14040. ISO, Geneva, p. 32. ISO (The International Organization for Standardisation), 2006b. Environmental Management e Life Cycle Assessment e Requirements and Guidelines. ISO 14044. ISO, Geneva, p. 58. Lazard, J., Baruthio, A., Math, S., Rey-Valette, H., Chia, E., Clment, O., Aubin, J., Morissens, P., Mikolasek, O., Legendre, M., Levang, P., Blancheton, J.P., Ren, F., 2010. Aquaculture system diversity and sustainable development: sh farms and their representation. Aquatic Living Resour. 23 (2), 187e198. Lovell, R.T., 1991. Nutrition of aquaculture species. J. Anim. Sci. 69, 4193e4200. Mungkung, R., 2005. Shrimp Aquaculture in Thailand: Application of Life Cycle Assessment to Support Sustainable Development (PhD Dissertation). Centre for Environmental Strategy (CES), School of Engineering, University of Surrey, UK, p. 360. Mungkung, R., Udo de Haes, H.A., Clift, R., 2005. Potentials and limitations of Life Cycle Assessment in setting ecolabelling criteria: a case study of Thai shrimp aquaculture product. Int. J. Life Cycle Assess. 11 (1), 55e59.

Please cite this article in press as: Mungkung, R., et al., Life Cycle Assessment for environmentally sustainable aquaculture management: a case study of combined aquaculture systems for carp and tilapia, Journal of Cleaner Production (2013), http://dx.doi.org/10.1016/ j.jclepro.2013.05.029