Environmental Performance of Chocolate Produced in Ghana Using Life Cycle Assessment

: Ghana is an important cocoa producer and exporter and this production is of high economic importance. Increasing interest in the sustainable productions of cocoa/chocolate necessitated the need to assess the environmental impacts associated with the production of different chocolate variants (extra dark (EDC), dark (DC), milk (MC) and ﬂavoured milk (FMC) in Ghana, including the identiﬁcation of environmental hotspots for improvement. The life cycle assessment tool was used following the CML_IA and CED impact assessment methods. EDC had the lowest scores for most of the impact categories while FMC was most impactful. For Global Warming Potential (GWP), EDC and FMC were estimated to be 1.61 kg CO 2 eq. and 4.21 kg CO 2 eq., respectively. CED ranged from 1.44 × 10 2 to 1.50 × 10 2 MJ-eq. Chocolate manufacturing phase was generally more impactful than cocoa cultivation due to high emissions from milk and sugar production. The impact scores for 100 g packaged chocolate bar were the lowest in comparison to 300 g chocolate pouches and 12.5 g packaged chocolate strips. GWP for 100 g and 12.5 g were 0.20 kg CO 2 eq. and 0.39 kg CO 2 eq., respectively. Comparing different destination points for the manufactured chocolate, impact scores for the international destination were similar to those recorded for local destinations. Improvement options are suggested for all phases to ensure more sustainable chocolate production and distribution.


Introduction
Environmental sustainability has been emerging as a pivotal issue in the agri-food sector as it directly impacts food and agriculture. Agriculture is highly vulnerable to climate change, as farming activities directly depend on climatic conditions, especially in developing countries. Climate change is a major contributing factor to the food price crisis, and its negative impacts on agriculture and food security in developing countries are expected to increase. The food sector is estimated to contribute about 25-30% of the total greenhouse gas (GHG) emissions due to anthropogenic emissions from agriculture and land use, storage, transport, processing, packaging, retail, and consumption [1]. Thus, food systems are heavily prioritised on the 2030 Agenda for Sustainable Development [2], a global commitment to eradicate poverty and hunger while ensuring reduction of environmental and socioeconomic concerns. Therefore, in order to advocate for sustainable measures, it is important to accurately assess the impacts of various activities and processes on the environment. Food and energy supply chains are associated with complex and intertwined environmental and socioeconomic impacts [3]. This has led to the use of tools and methodologies for assessing these impacts along various supply chains, including the Life Cycle Assessment (LCA).
Cocoa (Theobroma cacao) is an international cash crop that is mainly cultivated by smallholder farmers in lowland tropics, including parts of West Africa, Latin America, and Asia [4]. Over the past 50 years, world supply and demand for cocoa has been increasing at an annual growth rate of 2.5% [5]. Africa remains by far the most dominant cocoa to conduct further studies to update existing literature as Ghana is fast becoming a major chocolate producing country.
Presently, consumers are more interested in knowing product impacts before making purchasing choices. Thus, key stakeholders, especially manufacturers, within the food sector, need to develop innovative strategies for the improvement of working conditions and efficient use of resources to derive maximum economic benefit with minimum environmental impacts [25]. Manufacturers generally seek to either reduce the pollution caused by their products or highlight their environmental advantages. Governments also need reliable information to refine environmental policies or to devise incentives to promote environmental behaviours [26]. Therefore, the tentative goal of this study was to assess the environmental impacts of the production and distribution of chocolate produced in Ghana. Furthermore, this study sought to provide information geared towards sustaining the environment, particularly on the environmental impacts associated with the local cultivation and processing of cocoa into chocolate in Ghana, and subsequently help in the identification of environmental hotspots along the Ghanaian cocoa value chain.

Materials and Methods
Life cycle assessment (LCA) was used to estimate the impacts of 1 kg of packaged chocolate bar, following the ISO 14040 and 14044 standards. The standard LCA has four distinct methodological phases and are completed in the following order: goal and scope definition, life cycle inventory analysis (LCI), life cycle impact assessment (LCIA), and interpretation of results. LCA is an iterative process, where the different phases can be repeated until the final objective is met [27,28]. LCA is a method that can relate multiple environmental impacts to the function of a product or service. LCA is a decision-making tool which gives a comprehensive approach for evaluating the environmental impacts of a product during the entire production system [27,29]. It is often used to identify the hotspots and the mitigation options of environmental loads associated with a production. LCA is widely accepted and used in the evaluation of activities and processes of the agri-food sector, including agricultural activities, transportation, processing, packaging, storage, and distribution. LCA has numerous applications-it can be used industrially for a variety of purposes, including support of a corporate strategy, research and development, and the design of products or processes. LCA is also used in education as well as for labels and product descriptions [30]. The importance of LCA studies is increasing as companies increasingly apply them to their own products and require LCA data from their suppliers.
The methodology, data, and the assumptions considered in this study are further detailed in the following sections.

Goal and Scope Definition
The goal of the study was to assess the environmental impacts associated with the production of a packaged chocolate produced in Ghana. The study sought to identify environmentally weak points along the cocoa value chain where improvement can be made by farmers, chocolate manufacturers, transporters of cocoa beans and chocolate products to help improve the environmental aspects of the product. The target group for this study includes all stakeholders within the Ghanaian cocoa/chocolate industry, namely: COCOBOD, cocoa farmers, cocoa processors, chocolate manufacturers, environmental authorities and policy makers, the companies involved in storage, packaging, transport, retail, and recycling facilities, researchers, and NGOs.
The system boundary encompasses the essential energy and material inputs/outputs that are related to the processes of producing chocolate. The defined functional unit, based on which the inventory data was normalized for assessing the impacts in this study, was 1 kg chocolate made from cocoa cultivated in Ghana and other ingredients. As shown in Figure 1, the system boundary covers both upstream and downstream processes including the following phases: Cocoa cultivation: materials included were pesticide, insecticide, fertilizer, diesel. Cocoa processing and chocolate manufacturing: raw materials included were cocoa beans, sugar, milk, flavour, and alkaline (Dutching). Electricity, steam, and water consumption in the manufacturing processes, including cleaning activities.
Packaging: Aluminium foil (primary), paper wrapper, and carton boxes. Transport and Distribution: Transport of raw materials as well as packaging materials to the production facility. Distribution of the packaged product from the production facility to major retails centres was also considered in this study.
Other phases including consumption and waste management were excluded from the system boundary as done in other similar studies due to the many different scenarios that needed to be considered, making it difficult for standardization and comparison. Figure 1, the system boundary covers both upstream and downstream processes including the following phases: Cocoa cultivation: materials included were pesticide, insecticide, fertilizer, diesel. Cocoa processing and chocolate manufacturing: raw materials included were cocoa beans, sugar, milk, flavour, and alkaline (Dutching). Electricity, steam, and water consumption in the manufacturing processes, including cleaning activities.
Packaging: Aluminium foil (primary), paper wrapper, and carton boxes. Transport and Distribution: Transport of raw materials as well as packaging materials to the production facility. Distribution of the packaged product from the production facility to major retails centres was also considered in this study.
Other phases including consumption and waste management were excluded from the system boundary as done in other similar studies due to the many different scenarios that needed to be considered, making it difficult for standardization and comparison.

System Description and Data Quality
Data was obtained from different sources to complete the inventory table as shown in Table 1. The foreground data were obtained from a local Ghanaian chocolate manufacturer using a specific questionnaire for the cocoa bean processing, chocolate manufacturing, and packaging phases. The company was established in 2011 on a 158-ha plot of land at the Free zones enclave Tema, and has the capacity to process 60,000 metric tons of cocoa beans per year, making it the largest fully integrated cocoa processor in Ghana. It is involved in the processing of Ghanaian cocoa beans into semi-finished products and confectionery for supply to the worldwide chocolate, ice cream, and bakery industries. Semifinished products include natural and deodorised cocoa butter, specialised cocoa liquor, and natural and alkalised cocoa powder. In confectioneries, the manufacturer is a bean-

System Description and Data Quality
Data was obtained from different sources to complete the inventory table as shown in Table 1. The foreground data were obtained from a local Ghanaian chocolate manufacturer using a specific questionnaire for the cocoa bean processing, chocolate manufacturing, and packaging phases. The company was established in 2011 on a 158-ha plot of land at the Free zones enclave Tema, and has the capacity to process 60,000 metric tons of cocoa beans per year, making it the largest fully integrated cocoa processor in Ghana. It is involved in the processing of Ghanaian cocoa beans into semi-finished products and confectionery for supply to the worldwide chocolate, ice cream, and bakery industries. Semi-finished products include natural and deodorised cocoa butter, specialised cocoa liquor, and natural and alkalised cocoa powder. In confectioneries, the manufacturer is a bean-to-bar producer of refined chocolate, chocolate drinks, and spreads. The background life cycle inventory (LCI) data and data to fill in gaps were obtained from academic peer reviewed publications and Ecoinvent v.3.01 [31], as detailed further below. Product distribution Based on real market scenario Secondary Transport of workers Chocolate producer Primary * The standard agricultural guidelines and activities for cocoa cultivation did not change significantly over the past decade, as the considerations made in the study especially with regards to fertilizers and chemicals are still approved by the COCOBOD. Attempts were however made to obtain primary data from the farmers and other stakeholders, but due to the COVID-19 situation no positive response was received.

Cocoa Cultivation
The production of dried cocoa beans encompasses the production of farm inputs and farm activities carried out during cocoa cultivation such as fertilizer application, pest and disease management, harvesting and breaking of pods, fermentation, drying, and temporary storage of dried beans. The average economic lifespan of cocoa trees was also estimated to be thirty years. Secondary data on the cultivation phase was obtained from [9]. Additionally, background data on production of inputs such as fertilizer, insecticide, fuel were also obtained from Ecoinvent v.3.01 [31].

Transportation of Cocoa Beans to Processing Factories
Dried and bagged cocoa beans are transported by trucks from the farming communities to the warehouses of Cocoa Marketing Company located in Tema, Takoradi, and Kumasi. Afterwards, they are further transported to the processing factories. Inventory data for transportation was calculated based on the average distance of 250 km travelled by engine trucks in Ghana from Kumasi to Accra. The truck chosen was a >32 metric ton based on the average 38-tonne total capacity (and 26 tonne payload) long distance truck-trailer of most cocoa haulage trucks in Ghana as was indicated by the local manufacturer. Data on fuel consumption and emissions for the transportation were obtained from the database based on the selected Euro-3 truck typology as shown in Table 2.

Processing of Cocoa Beans into Chocolate
Cocoa beans are first processed into semi-finished products: cocoa liquor, butter, and powder. These serve as ingredients in the chocolate manufacturing process. Primary data on the sources of the cocoa beans (round trip distance from the cocoa bean warehouse to the manufacturing plant), the type, source, and the transportation links associated with obtaining the other ingredients such as sugar, flavour, and milk, use of electricity to run machines, water use, packaging materials, and the total output of chocolate produced in 2019 were obtained from the manufacturer. An average technology was considered for the processing and manufacturing of chocolate based on information provided by the manufacturer. Machines and equipment were purchased between 2011 and 2017 (70% purchased in 2011, 25% purchased in 2012, and 5% purchased in 2017). Additional background data included production of energy (electricity from hydropower and Liquified Petroleum Gas (LPG)) consumed by the plant's manufacture of key ingredients such as sugar, milk, and flavour.

Packaging
Packaging is key in the food supply chain as it performs many functions including the protection of food products, containment, easy handling, safe transport, extension of shelf life, and marketing [32]. The chocolate products are packaged after production to ensure maintenance of quality, ease of handling and labelling, and protection of the product. Three levels of packaging were considered in this study: aluminium foil (primary packaging) for protection against air and light, printed paper wrapper (secondary packaging) for labelling, and cardboard (tertiary package) for protection and ease of handling during transport. The manufacturer produces three different chocolate product typologies for  Table 4. Primary data on the amount of packaging used was obtained from the manufacturer. In addition, the transportation of the packaging materials from China, where they are manufactured, was also taken in consideration.

Distribution Assumption
Due to the impracticalities related to determining a consumer's intent to specifically leave their house to only purchase a bar of chocolate, assumptions had to be made for the distance consumers would travel to purchase the product. The average transportation distance was calculated from the travel distance to a store within heavily populated areas where the chocolate is mainly sold. The major towns selected were Kumasi, Takoradi, and Koforidua.

Allocation
Allocation procedures are applied in two phases. The first one regards cocoa shells, which constitutes one of the outflows of cocoa bean processing. Shell mass accounts for about 10% of cocoa beans and are used to fuel boilers in the processing plant as it is more economical-thus, no environmental impacts are associated with it. In this study, 100% of the impacts from cocoa bean processing were allocated to the co-products from cocoa processing (liquor, butter, cake, and powder) adopting an economic value allocation criterion as shown in Table 5.

Life Cycle Inventory Table and Scenarios
Based on the assumptions reported in the previous section, the inventory table has been reported separating cocoa bean processing and chocolate manufacturing of four different products to improve transparency. Table 6 reports values referred to the sum of all chocolate products (1 kg of each chocolate product-four kg in total).

Scenarios
In LCA studies, a scenario describes a possible future situation relevant for specific LCA applications based on specific assumptions about the future, and may also include the presentation of the development from present to future. Scenarios provides possibilities to prepare for alternative and uncertain future options without knowing anything about the probability of the possible outcomes [33]. In this study, the impacts associated with several scenarios were considered and examined, to help in the suggestion of useful and relevant mitigation strategies.
The first scenario examined was the impacts associated with different destination for 1 kg (10 × 100 g) packaged chocolate bars to different destinations. The major cities in Ghana such as Kumasi, Koforidua, and Takoradi were considered as the baseline destinations while Ancona, Italy was considered as a scenario for the export of the product. In the scenario for local destination, a transport-mean which was a truck with a load capacity >32 metric ton was selected, and the distances were estimated using Google maps. In the scenario for international destination, the distance from the company to Ghana port, and from there to the port to Hamburg and by road to Ancona in Italy is illustrated in Figure 2 below. The sea distances were also obtained from an online world seaports catalogue, marine, and port info website (http://ports.com/ (accessed on 18 February 2021)).
The first scenario examined was the impacts associated with different destination for 1 kg (10 × 100 g) packaged chocolate bars to different destinations. The major cities in Ghana such as Kumasi, Koforidua, and Takoradi were considered as the baseline destinations while Ancona, Italy was considered as a scenario for the export of the product. In the scenario for local destination, a transport-mean which was a truck with a load capacity >32 metric ton was selected, and the distances were estimated using Google maps. In the scenario for international destination, the distance from the company to Ghana port, and from there to the port to Hamburg and by road to Ancona in Italy is illustrated in Figure  2 below. The sea distances were also obtained from an online world seaports catalogue, marine, and port info website (http://ports.com/ (accessed on 18 February 2021)). Another scenario considered was the transportation of workers by company buses to the factory. This scenario is not directly linked to the process flow for chocolate production, however due to the possible importance of transportation of workers in Ghana it was considered. The company has seven regular buses with a capacity of 15 persons per bus for 105 estimated workers and runs a double shift system. Thus, the average distances were calculated, and the impact associated with the transportation of workers were determined as shown in Table 7. This was compared to transportation of workers by private cars in Europe covering the same distance as shown in Table 2.

Life Cycle Impact Assessment
The collected and aggregated data were input in the PRéConsultants SimaPro 8.2.3 software to construct all the significant process flows (inputs and outputs for each life cycle phase) to model the product systems. The CML_IA baseline V3.01 method was applied to estimate the environmental impacts based on the problem-oriented (midpoint) approach for 1 kg chocolate. In this study, the impact categories examined for CML_IA included: Abiotic Depletion Potential (ADP), Abiotic Depletion Potential (fossil fuels) (ADP) Acidification Potential (AP), Eutrophication Potential (EP), Global Warming Potential (GWP 100yr), Ozone layer depletion (ODP), Human Toxicity Potential (HTP), Fresh Water Aquatic Ecotoxicity Potential (FAETP), Marine Aquatic Ecotoxicity Potential (MAETP), Terrestrial Ecotoxicity Potential (TETP), and Photochemical Oxidation Potential (POCP) [34]. Additionally, the Cumulative Energy Demand (CED V1.08) single issue Another scenario considered was the transportation of workers by company buses to the factory. This scenario is not directly linked to the process flow for chocolate production, however due to the possible importance of transportation of workers in Ghana it was considered. The company has seven regular buses with a capacity of 15 persons per bus for 105 estimated workers and runs a double shift system. Thus, the average distances were calculated, and the impact associated with the transportation of workers were determined as shown in Table 7. This was compared to transportation of workers by private cars in Europe covering the same distance as shown in Table 2.

Life Cycle Impact Assessment
The collected and aggregated data were input in the PRéConsultants SimaPro 8.2.3 software to construct all the significant process flows (inputs and outputs for each life cycle phase) to model the product systems. The CML_IA baseline V3.01 method was applied to estimate the environmental impacts based on the problem-oriented (midpoint) approach for 1 kg chocolate. In this study, the impact categories examined for CML_IA included: Abiotic Depletion Potential (ADP), Abiotic Depletion Potential (fossil fuels) (ADP) Acidification Potential (AP), Eutrophication Potential (EP), Global Warming Potential (GWP 100yr), Ozone layer depletion (ODP), Human Toxicity Potential (HTP), Fresh Water Aquatic Ecotoxicity Potential (FAETP), Marine Aquatic Ecotoxicity Potential (MAETP), Terrestrial Ecotoxicity Potential (TETP), and Photochemical Oxidation Potential (POCP) [34]. Additionally, the Cumulative Energy Demand (CED V1.08) single issue was used. The calculation of CED included: non-renewable (from fossil and nuclear) and renewable (wind, solar, geothermal, and water) energy sources [35].

Impacts Associated with 1 kg Chocolate Using CML_IA Baseline Method
The characterisation results (overall impact scores) for different impact categories for 1 kg chocolate produced in Ghana are shown in Figure 3a-k. From the results obtained, extra dark chocolate (EDC) predominantly recorded the least scores for most impact categories while milk chocolate (MC) had the highest scores for most impact categories. Impact scores for EDC and dark chocolate (DC) were not significantly different while MC and flavoured milk chocolate (FMC) also showed little differences. Significant differences were however observed between EDC and DC and FMC and MC for several impact categories including GWP, EP, POCP, and AP. The differences were mainly due to the type and amount of different ingredients used for the different chocolate product types. Results showed different relative contribution for the two phases considered: CCP (cocoa cultivation and processing) and CM (chocolate manufacturing) for different impact categories. ADP, HTP, TETP, and FAETP were largely influenced by the CCP phase while GWP, ODP, EP, and AP were significantly influenced by the CM phase. extra dark chocolate (EDC) predominantly recorded the least scores for most impact categories while milk chocolate (MC) had the highest scores for most impact categories. Impact scores for EDC and dark chocolate (DC) were not significantly different while MC and flavoured milk chocolate (FMC) also showed little differences. Significant differences were however observed between EDC and DC and FMC and MC for several impact categories including GWP, EP, POCP, and AP. The differences were mainly due to the type and amount of different ingredients used for the different chocolate product types. Results showed different relative contribution for the two phases considered: CCP (cocoa cultivation and processing) and CM (chocolate manufacturing) for different impact categories. ADP, HTP, TETP, and FAETP were largely influenced by the CCP phase while GWP, ODP, EP, and AP were significantly influenced by the CM phase.    Abiotic depletion is generally related to the consumption of non-b and its value is a measure of the scarcity of a substance which is affect available in nature and its rate of extraction [36]. From this study, AD × 10 −5 kg Sb eq. (EDC) to 2.43 × 10 −5 kg Sb eq. (MC) per kg chocolate as s ADP was mainly influenced by emissions from the cocoa bean cultiv 95.6%) due to pesticide (67.5-72%), fertilizer (11.4%) and insecticide (1 tion and use for MC and EDC, respectively. The differences were beca of cocoa liquor, butter, and powder used. Recanati et al. [15] reporte value of 1.11 × 10 −5 kg Sb eq. for 1 kg dark chocolate while Ntiamoah Mial et al. [18] obtained comparatively estimated higher values of 1.62 cocoa production and processing and 7.00 × 10 −3 kg Sb eq. per unit for b chocolate confectionary, respectively.

Abiotic Depletion (ADP)
Abiotic depletion is generally related to the consumption of non-biological resources and its value is a measure of the scarcity of a substance which is affected by the quantity available in nature and its rate of extraction [36]. From this study, ADP ranged from 3.48 × 10 −5 kg Sb eq. (EDC) to 2.43 × 10 −5 kg Sb eq. (MC) per kg chocolate as shown in Figure 3a. ADP was mainly influenced by emissions from the cocoa bean cultivation phase (89.7-95.6%) due to pesticide (67.5-72%), fertilizer (11.4%) and insecticide (10.7-12.2%) production and use for MC and EDC, respectively. The differences were because of the amount of cocoa liquor, butter, and powder used. Recanati et al. [15] reported a slightly lower value of 1.11 × 10 −5 kg Sb eq. for 1 kg dark chocolate while Ntiamoah and Afrane [9] and Mial et al. [18] obtained comparatively estimated higher values of 1.62 × 10 −3 kg Sb eq. for cocoa production and processing and 7.00 × 10 −3 kg Sb eq. per unit for both milk and dark chocolate confectionary, respectively.

Abiotic Depletion (ADP) Fossil Fuels
As shown in Figure 3b, EDC recorded the least score of 11.20 MJ while FMC recorded the highest score of 1.75 MJ for ADP fossil fuels. The CM phase was more impacting with emissions resulting from mainly milk powder (50%) and sugar (18%) while the CCP phase was influenced by LPG (12.5%) and pesticide (10%) for FMC and MC. For EDC, impacts were mainly associated with the production and use of LPG (30%), pesticide (24%), and transported sugar (26%).

Human Toxicity
HTP indicates the potential amount of a chemical that can cause harm when released into the environment. It is calculated based on both the inherent toxicity of a compound and its potential dose. These chemicals are usually by-products, and include arsenic, sodium dichromate, hydrogen fluoride, phosphorus, manganese, zinc, and chlorine and are often released during electricity production from fossil sources [17]. From this study, HTP ranged from 0.55 kg 1,4-DB eq. (EDC) to 0.41 kg 1,4-DB eq. (MC) as shown in Figure 3e. HTP was mainly influenced by emissions from the CCP phase (80-91%) due to pesticide (about 42-47%) and LPG (33-38%) production and use for MC and EDC, respectively. Ntiamoah and Afrane [9] reported a similar value of 5.11 kg 1,4-DB while Konstansas et al. [17], reported 1.66 kg 1,4 DB eq. and 2.03 kg 1,4 DB eq. for chocolate countlines and chocolates in bag, respectively.

Photochemical Oxidation
Photochemical oxidation is the type of smog created from the effect of sunlight, heat, and volatile non-methane volatile organic compounds (NMVOC) and NOx. POCP primarily depends on the concentration of CO, SO 2 , NO, NH 4 , and NMVOC emitted into the air [36]. EDC recorded the least while FMC recorded the highest of 3.07 × 10 −4 kg C 2 H 4 eq. and 5.69 × 10 −4 kg C 2 H 4 eq., respectively, as shown in Figure 3i. CM was more impacting than CCP due to emissions from transported milk powder (57%) and transported sugar (14%) for FMC and MC. Emissions from CCP resulted from pesticide production and use. Recanati et al. [15], reported a higher value of 1.07 × 10 −3 kg C 2 H 4 eq. for dark chocolate while Ntiamoah and Afrane [9] discovered a similar value of 8.09 × 10 −4 kg C 2 H 4 eq. for processed cocoa.

Acidification
Some anthropogenic activities that release emissions have the potential to reduce pH due to the acidifying effects. These emissions which are gases such as ammonia (NH 3 ), nitrogen oxides (NOx) and sulphur oxides (SOx) can cause acid deposition [36]. From this study, AP estimated was least for EDC (5.44 × 10 −3 kg SO 2 eq.) and highest for FMC (2.64 × 10 −2 kg SO 2 eq.) as shown in Figure 3j. CM was the most impacting phase with transported milk powder accounting for 80-83%. Konstansas et al. [17] also reported 6.50 × 10 −2 kg SO 2 eq. for moulded chocolate with milk powder contributing 90%.

Eutrophication
Eutrophication values are often influenced by the estimated amounts of ammonia, nitrates, nitrogen oxides and phosphorous emitted to both air and water [36]. From this study, EP estimated was least for DC (2.13 × 10 −3 PO 4 3− eq.) and highest for FMC (1.41 × 10 −2 PO 4 3− eq.) as shown in Figure 3k. CM was the most impacting phase with transported milk powder accounting for 87% for FMC and MC. Sugar had a net positive credit of −2% to −24% (FMC and DC, respectively), mainly due to the difference between impacts of the co-product (molasses) and the final sugar product. Molasses is considered to be an alternative to spring barley (global product) used as animal feed through system expansion with a comparatively lower impact which accounts for the net positive impact on the environment. CM was mainly influenced by direct and indirect emissions associated with fertiliser production and use. Ntiamoah and Afrane [9] and Recanati et al. [15] reported comparable scores of 1.05 × 10 −3 kg PO 4 3− eq. and 2.54 × 10 −2 kg PO 4 3− eq. for processed cocoa and dark chocolate, respectively.

Impact Assessment for Different Packaged Chocolate Products
From the results obtained, the impact scores for 65 g packaged chocolate strips were highest across all impact categories while 100 g packaged chocolate bar and 300 g packaged chocolate pouch were relatively identical as shown in Table 8. The differences are mainly due to the amount of packaging materials used and the different primary package typologies; 65 g packaging recorded the highest scores due to the comparatively higher amount of paper used for secondary packaging. Chocolate is a typical confectionary product with a longer shelf-life, and thus exposure to air and light can result in loss of taste and flavour and loss of surface gloss [17,22]. Carton boxes are used as tertiary packaging for distant transportation of chocolate. The GWP results obtained in this study which range between 0.20 kgCO 2 eq. and 0.39 kgCO 2 eq. are comparable to values of 0.20 kg CO 2 eq., 0.34 kg CO 2 eq., 0.81 kg CO 2 eq., and 1.91 kgCO 2 eq. per kg chocolate bar (FU) reported by Konstantas et al. [17], Recanati et al. [15], Allione et al. [23], and Perez Neira [16], respectively.

Impacts Associated with 1 kg Chocolate Based on CED Method
The characterisation results (overall impact scores) for different impact categories for the different chocolate types and packages are shown in Table 9. From the results obtained, there was little difference between the chocolate types with the most impacting energy source being renewable (water) with a score of 1.29 × 10 2 MJ eq. (86-90%) for the generation of hydroelectricity. A substantially lower value of 33.75 MJ/FU was reported by Recanati et al. [15] for dark chocolate produced in Italy. This was primarily due to the use of a highly efficient trigeneration system for electricity, heating, and cooling. Over 70 MJ/kg chocolate was also estimated for milk chocolate produced in Ecuador and Indonesia by Bianchi et al. [24]. The impact score trend for the different packaging types based on the CED method was similar to that observed using the CML_IA method. The 65 g packaged chocolate strips recorded the highest score, followed by the 300 g packaging with the 100 g packaged chocolate bar recording the least as shown in Table 9. For all the packaging typologies, non-renewable (fossil) contributed the most (83-88%) mainly due to the energy for the production of paper and cardboard; 65 g packaging recorded the highest scores due to the comparatively higher amount of paper used for secondary packaging, with relative contribution of paper (70%) and cardboard (22%). For 300 g packaging, paper contributed 62.7%, cardboard 25%, and aluminium foil 12.3%; 100 g packaging was least impacting with paper contributing 38%, cardboard 35%, and aluminium foil 27%.

Impact Assessment Based on Characterization for Different Transport Destinations
The relative contribution of impacts associated with the distribution scenarios for 1 kg packaged EDC considered in this study are shown in Figure 4. The total impact scores for the three destination points (Takoradi, Kumasi, and Koforidua) did not show many differences in impact scores across all categories. This could imply that the midpoint impact categories for local distribution in Ghana are more likely to be influenced by the transport mean and its characteristics than the estimated distances. The impact scores for the international destination (Ancona, Italy) were slightly higher than those recorded for the local destinations for most of the impact categories assessed. For GWP, Koforidua had the least impact score of 1.86 kg CO 2 eq. while Ancona recorded 2.13 kg CO 2 eq., while for AP Koforidua and Ancona had scores of 6.93 × 10 −3 kg SO 2 eq. and 9.71 × 10 −3 kg SO 2 eq., respectively. This could be attributed to the general high efficiency of transoceanic ships, though it was estimated to be responsible for about 2.1% of global greenhouse gas (GHG) emissions annually [38]. Therefore, this accounts for the relative low impact scores associated with this transport scenario. In addition, the truck means selected, EURO 4, is more efficient with less emission as compared to EURO 1-3.

Impacts Associated with Transportation of Workers
The total impacts associated with the transportation of workers for this scenario using the CML_IA baseline method is shown in Figure 5.

Impacts Associated with Transportation of Workers
The total impacts associated with the transportation of workers for this scenario using the CML_IA baseline method is shown in Figure 5.
From the results obtained, there were no significant differences for the impact categories, with transportation by regular buses being slightly more impacting. The relative differences ranged between (0-10%) with ADP fossil fuels and TETP recording the most differences. The differences were due to the transport means since the same distance and number of workers were considered. LCA studies often do not take some important but indirect impacts such as transportation of workers into consideration. However, due to the high environmental impacts associated with daily transportation of workers, who are essential for the production of the chocolate, it was prudent to compare the different transport scenarios and suggest appropriate mitigation strategies. According to the emission inventory analysis, road transportation including private cars, trucks, and buses is the main contributor to local air pollution [39]. The activities of the transport industry release several million tons of gases each year into the atmosphere which contribute to global warming and stratospheric ozone layer depletion. These include lead (Pb), carbon monoxide (CO), carbon dioxide (CO 2 ; not a pollutant), methane (CH 4 ), nitrogen oxides (NOx), nitrous oxide (N 2 O), chlorofluorocarbons (CFCs), perfluorocarbons (PFCs), silicon tetraflouride (SF6), benzene and volatile components (BTX), heavy metals (zinc, chrome, copper and cadmium), and particulate matters (ash, dust) [40].

Impacts Associated with Transportation of Workers
The total impacts associated with the transportation of workers for this scenario using the CML_IA baseline method is shown in Figure 5. From the results obtained, there were no significant differences for the impact categories, with transportation by regular buses being slightly more impacting. The relative differences ranged between (0-10%) with ADP fossil fuels and TETP recording the most differences. The differences were due to the transport means since the same distance and number of workers were considered. LCA studies often do not take some important but indirect impacts such as transportation of workers into consideration. However, due to the high environmental impacts associated with daily transportation of workers, who are essential for the production of the chocolate, it was prudent to compare the different transport scenarios and suggest appropriate mitigation strategies. According to the emission inventory analysis, road transportation including private cars, trucks, and buses is

Improvement Options for Sustainability
Several improvement options along the cocoa/chocolate supply chain can be suggested to improve the environmental performance of chocolate as all the phases were equally impacting for different impact categories. Most cocoa farmers in Ghana use traditional farming methods for the cultivation of cocoa. This method involves relying on the rainfall for irrigation, less use of agricultural machinery and sun drying of harvested cocoa beans. Emissions associated with energy use and water that could potentially impact global warming, acidification, and abiotic depletion are reduced. As shown in this study, most of the environmental impacts associated with cocoa cultivation are related to the amount and type of inorganic fertilisers, pesticides, and insecticides used. Therefore, substituting the use of inorganic fertiliser with organic fertiliser would decrease the emissions and their related impacts. Cocoa pod husk, which is generally considered as waste product after pod splitting, can also be used as organic fertiliser during cultivation [41]. Another option that could be considered is inter-cropping cocoa trees with other agroforestry crops like coconut. According to Utomo et al. [19], cocoa-coconut agroforestry recorded 3.67 × 10 −1 kg CO 2 -eq. and 4.31 × 10 −2 kg SO 2 -eq. for GWP and AP, while cocoa monoculture recorded substantially higher values of 7.06 × 10 1 kg CO 2 -eq and 8.11 × 10 −2 kg SO 2 -eq, respectively. The use of low input systems which rely on integrated pest management, that involve the use of biological agents for pest and diseases control, adequate soil fertility management, including the cultivation of high yielding and more disease and pest resilient cocoa varieties developed by the CRIG and other research bodies could also enhance the environmental performance of cocoa cultivation in Ghana. Also, advanced farming systems could be employed to increase cocoa yield. Production of dried cocoa using a technical management system yielded 1400 kg/year as compared to a traditional farming method of 300 kg/year with comparable GHG emissions of 2.49 and 2.82 kg CO 2 -eq. [16].
The production and transport of milk powder and sugar were the main impacting materials for the chocolate manufacturing phase. Substitution of milk powder and sugar with other less impacting ingredients able to perform similar functions could be useful. Milk powder impacts are mainly associated with cattle rearing, thus possible improvement options could include the modification of the diet of cattle to reduce methane emissions from enteric fermentation [42]. Roibas et al. [43] considered the effect of modified cow diet through the inclusion of high-quality forages (linseeds) on the emission of GHGs and found that the GWP of milk was reduced by 10%. Increasing the efficiency of energy use in ingredients production and chocolate manufacturing through implementation of integrated energy management systems or more advanced and innovative technologies such as trigeneration systems could potentially reduce emissions from LPG and hydroelectricity [17].
Packaging was also identified as a major contributor to several of the impact categories and should also be subject to improvements. Most of the impacts in this phase were associated with the extraction, manufacturing, and use of aluminium foil and paper. Data for alternative packaging that could be substituted with aluminium foil while preserving functionality is currently unavailable. Therefore, the use of recycled aluminium and a general reduction in the weight of the packaging materials used are recommended.
In relation to impacts associated with the transportation of workers, the manufacturer could consider using more efficient transport means to reduce fuel consumption and emissions. Also, the company could consider changing the transport means to larger buses to reduce the number of buses for one destination. A more efficient bus schedule could also be developed to reduce the number of times the buses travel the same distance in relation to the shift system. Furthermore, the company may consider hiring workers who stay close by or perhaps a future consideration to provide accommodation for workers closer the company to reduce the distance. Better work management could also be suggested to improve worker efficiency.
Stakeholders could adopt these feasible strategies to improve efficiency and to mitigate potential negative environmental impacts. Local chocolate manufacturers can improve their carbon footprint while enhancing their competitive urge on both the local and international markets. This would be considered a shared value by consumers as environmental protection awareness increases and could be beneficial for the marketing of chocolate produced in Ghana for higher profit gains. Furthermore, governing bodies could also develop guidelines and policies to ensure a more environmentally sustainable cocoa value chain.

Some Limitations of the LCA
The Life Cycle Assessment methodology, although very useful, is also characterized by some limitations. Although the ISO standard gives a consensus definition for LCA and provides a general framework for conducting an assessment, it leaves much to interpretation by the person conducting the assessment, which may lead to high variation among results for even the same product. Other limitations include time and resource constraints in gathering inventory data, missing impact data and models for LCIA, challenges with data uncertainty, allocation of environmental burden across co-products, and assigning credit for avoided burden. The standard LCA methodology can also be regarded as an incomplete tool for measuring sustainability since it does not take social welfare into consideration and is unable to provide sustainability thresholds or acceptable limitations for the entire society [44]. Today, there is a burgeoning interest in updating and standardizing social and economic life cycle assessment for universal application. Despite these limitations, LCA can be improved through research, inherent in the methodology and alternate modelling choices. LCA offers a strong environmental tool in the movement toward sustainability.

Conclusions
The tentative goal of this LCA study was to assess the environmental impacts of the production and distribution of chocolate produced in Ghana. Generally, the environmental impact scores obtained were consistent with other similar studies conducted, particularly in Europe. Along the cocoa/chocolate supply chain, the manufacturing phase was found to be most impacting with milk and sugar accounting for most of the environmental burden. Consequently, darker chocolates were found to be more environmentally sustainable than milk chocolate derivatives. Although the 300 g packaged chocolate was expected to have the least environmental impact due to its larger size, the 100 g chocolate was rather found to have the least environmental burden, while the smaller 65 g chocolate was most impacting. With regards to transportation, impacts associated with the export of chocolate to Europe was not substantially different from distribution within Ghana. Furthermore, the comparison of impacts for the transportation of workers by buses in Ghana to passenger cars in Europe did not show any significant differences for the same distance. Therefore, the location of the processing and manufacturing plant in either Ghana or Europe does not confer any additional advantage from the environmental point of view. Plant siting in Ghana may be favoured due to availability of cocoa while Europe may be favoured by proximity to milk, sugar, and more efficient energy systems. Mitigation strategies should therefore be more focused on the cultivation phase, especially on the use of fertilizers and pesticides, and the manufacturing phase, particularly milk and sugar production. The quantification of the environmental impacts associated with chocolate could effectively support cocoa farmers, chocolate manufacturers, policy makers, and consumers in their pathway towards sustainable production, distribution, and consumption of chocolate products.