*2.2. Methods Used in Biofuel Downstream Supply Chain Planning*

Bioenergy plays an essential role in promoting clean energy and securing the future energy supply. This rationale leads to increase global trade of biomass or energy carriers from biomass which has been reported by Schlamadinger et al. [48]. Welfle [49] highlighted that the bioenergy strategies of many countries highly depends on future imported resources to balance the demands. Biomass is unevenly distributed in the view of demand and resources availability. Developed countries and the energy policies drive the increasing reliance on bioenergy pathways to meet the energy demand. Some of the countries are facing insufficient biomass stock to meet the demand, but there are also countries with the potential supply that significantly exceeds the demand. An international biomass supply chain is vital to promote natural symbiosis. Junginger et al. [50] provide a comprehensive discussion on international bioenergy trade, including the drivers and barriers as well as developments in liquid biofuel trade. Europe is one of the prime markets for the trade of biomass for energy generation, with wood pellet as the main substrate (see Figure 3). Lamers et al. [51] show that Brazil is the leading exporter of bioethanol while the United States, Argentina, Indonesia and Malaysia are the major exporters of biodiesel. Based on the forecast by IEA [1], biofuel output is anticipated to reach 1.9 <sup>×</sup> 1011 <sup>L</sup> (+24%) by 2024, owing to better market prospects in Brazil, the United States and China. Asia is expected to lead to biofuel production growth.

**Figure 3.** Wood pellet trade flows, adapted from World Bioenergy Association [52].

Downstream biofuel supply chain planning and management are subjected to a higher level of uncertainty, especially when it involves international trading. The objective function is generally the same as the upstream assessment, where economic and/or environmental aspect is optimised. Physical trade of biomass is not always the optimal solution due to international logistics, which increase the cost and environmental footprints. Laurijssena and Faaij [53], however, suggest that trading biomass is preferential than trading GHG emission credits. Other than the transporting distance, which is the main supply chain issue, the international trade (macro perspective) is affected by incentive-policy (e.g., the EU's Renewable Energy Directive) context and trade tariffs (both import and export tariffs) [54]. The reliability of the assessment results is relying on the accurateness of projection/ prediction as well as the assumptions. The assessment model is usually supported by a range of scenarios representing the optimistic and pessimistic situation. Welfle [49] applied the biomass resource model to evaluate the biomass potential in Brazil. The trading possibility is determined by the availability of the resource, considering the remaining land area, the potential of resource collection, competing for a market of biomass utilisation as well as the conversion pathway. Deng et al. [55] conducted similar research to

identify the potential of trading (import and export) based on resource availability but covering a more extensive range of feedstock and countries. The yield gradient, land-use change and technology development are varying to identify the biofuel potential. The estimation potential for the global scale ranges from 40 to 190 EJ final energy in 2070 where Brazil and Russia are recognised as the prominent exporters, while India and Nigeria are substantial importers. The forecasts mainly identifying the biomass potential rather than the biomass allocation, which considering the detailed costing, travel distance and available market. Lamers et al. [56] assess the potential import streams and supply costs under different sustainability constraints based on a bottom-up global trade model. Figure 4 shows the modelling framework which combining the biomass transport model and biomass allocation model. This modelling considered temporal and logistical determinants without neglecting the market development and time aspects (e.g., delay), where the biomass allocation is suggested.

**Figure 4.** The integrated modelling framework for international biomass (to energy) trade adapted from by Lamers et al. [56].

The review by Diesenreiter and Kranzl [57] suggested that no customised models are available for incorporating global import/export potentials and international trade. The evaluated approaches in the review study are divided into basic modelling approaches for analysing the macroeconomic effect of international trade, computable general equilibrium models and geographic information system. A similar conclusion has been drawn by Solberg et al. [58] in a study for IEA bioenergy task 40 where none of the existing models is capable of performing good analyses of international trade of biomass and bioenergy products. Figure 5 summarises the assessed models in both studies in identifying the weakness and strengths. There is a common characteristic where all the presented models are customised more to the case of the EU. Green-X model [59], which allows the consideration of different energy policy instruments is also for the application of the European level. The gravity model of trade [60] is among the standard model in predicting bilateral trade flows according to the economic sizes and distance between two units despite the argument on the identified results [61]. Röttgers et al. [62] analyse the effect the EU imposes on the trade of the biofuel commodities and identify the drives (e.g., trade regulation or bioenergy regulation) of biofuel trade. The assessment suggests that EU trade integration has no enabling effect on canola oil trade where the import from outside of the EU is preferable. The result warrants a closer look at the political measures and its effectiveness, especially the green investment subsidy. However, other factors such as economies of scale, resource scarcity and value chain structure have to be taken into account as well for a conclusive picture.

Rentizelas et al. [63] stressed the need for a decision support tool to facilitate the supply chain design rather than assessing the supply chain of specific origin and destination location. A multicriteria tool based on data envelopment analysis which considers the environmental impact and cost is conducted to identify the efficiency of alternative pathways (Up to 56 pathways between Brazil and the UK) of international biomass supply chains are developed. Three models for bioenergy trade analysis, include TIMER (dynamic energy system model), GFPM (spatial partial equilibrium model based on price endogenous linear programming) and POLES (dynamic partial equilibrium model), have been reviewed by Matzenberger et al. [64]. It was concluded that further integration of international bioenergy trade, emerging barriers and drivers into the existing models is essential for a more realistic answer regarding the future role of the bioenergy system.


**Figure 5.** Model for international trade of biomass for energy. Please refer to Diesenreiter and Kranzl [57] and Solberg et al. [58] for the detailed discussion on each approach.

Maximising the economic performance is the common objective function in most of the models. However, the trade-offs between economic and environmental performance as well as the other factors have to be also considered. Total footprints-based multi-criteria optimisation is proposed by Cuˇ ˇ cek et al. [65] to consider the economic performance, environmental footprint as well as the social footprint in determining an optimal regional biomass energy supply chain. Jonkman et al. [66] propose a decision support tools with the advantages that it can take into account the goals of individual actors of the supply chain than only optimising the economic and environmental performance. This is a significant development and beneficial for supply chain with the involvement of different countries. The multicriteria approach is potential to adapted for global supply chain problem, although it is demonstrated through a case study in the Netherlands. Lee et al. [67] proposed a global supply chain optimisation framework supported by two-stage stochastic linear programming model (TRMISP) to identify the supply chain design (Southeast Asia to Europe and North America) under the price and demand uncertainty. This is important as uncertainties are one of the main challenges in modelling and optimisation of the international supply chain. Transfer pricing, currency exchange and taxation rates have to consider in the global supply chain planning. The studies which include one or more of the factors include de Matta and Miller [68] (Transfer price- generalised Benders decomposition approach), and Gonela et al. [69] (Tax-credit- stochastic mixed-integer linear programming model). Razm et al. [70] proposed a multi-objective mathematical model with the aids of GIS to design a global sustainable bioenergy supply network. This is a comparatively comprehensive model which considered all the crucial components at the international level.

The methodological challenges such as uncertainties of international statistics, inconsistent data on trade volumes and final use of traded products [71] persist despite advancing. It is expected to enhance the development of IoT and big data in the near future. Fingerman et al. [72] assessed the opportunities and risks for sustainable biomass export, particularly to Europe from the South-Eastern of United States. The long-term strategies assessment by Pelkmans et al. [73] for European bioenergy markets considered North America, South America, East Europe, Africa and Southeast Asia as the potential sourcing regions. It is concluded that policies should be stable and consistent within a long-term vision. Macro perspective assessments specifically done for Asia countries are generally lesser, especially compared to EU and South America. It deserves more research attention by adapted to the localised condition and forecast towards a close to the optimal global solution. Different approaches can be fitted for solving the problem related to international biofuel trade by integrating to the existing energy models. However, the considered variables are not consistent for a robust solution, and data availability for modelling is still one of the critical issues.
