Table 1 Qualitative evaluation of ‘bottom-up’ and ‘top-down’ modelling approaches referred in this paper
From: Approaches for modelling the energy flow in food chains
Models | Brief description | Common benefits | Common limitations | Common assumptions |
|---|---|---|---|---|
Bottom-up approaches | ||||
LCA models | A process or product based evaluation method of the energy use and GHG emissions, typically using a ‘cradle to grave’ approach | - Measure high quality of energy and GHG emissions data | Level of detail required increases the complexity of data collection | - Secondary to primary energy conversion factors |
| Â | Â | - Capture the intricacies and complex nature of the food-energy chain | - Analysis is usually very specific to country and product/process | - Transportation distances and fuels |
| Â | Â | - Allow the identification of energy and GHG emission hot-spots | - Information from different sources cannot be combined, unless uncertainties and assumptions are clearly specified | - Imported energies and GHG emissions |
| Â | Â | Â | - Choice of functional units and demarcations between system add complexity to the method | - Agricultural sector energies and GHG emissions |
| Â | Â | Â | - Ignores the holistic industry impact on the product/process, i.e. static model | - Waste disposal and storage emissions |
| Â | Â | Â | - LCA needs constant update | Â |
MARKAL and sub-models | - Demand-driven multi-period linear programming and cost optimisation tools | - Technologies and processes can be explicitly modelled in detail | - Results depend on the accuracy of demand-inputs and description of technological processes | - Description of technological processes |
| Â | - Simultaneously assess the impact of several technologies through the partial equilibrium of demand and supply of energy | - Currently being employed in various energy research studies, and various extensions/innovations of the model are being developed | - Discount rates of technologies impact the partial equilibrium when forecasting energy demands | - Discount rates and future energy demands |
| Â | Â | - Allows the explicit modelling of the evolution of demand and energy prices | - Does not have information feedback to the wider economy | Â |
| Â | Â | Â | - MARKAL databases have to be regularly updated | Â |
Regression models | Find the causal relationship between the dependent and independent variables in the food-energy chain | - Relatively easy to use and construct | - Require large amount of data to ensure proper correlations | - Determination of errors, and validity of models are subjective to modeller |
| Â | Â | - Provide a simplistic description of problem, and allows quick approximations of different policies | - Provide only a quantitative evaluations, and does not show the factors causing the inefficiencies (depends on the level of disaggregation of the model) | - General assumptions relate to the energy consumption and GHG emissions in the data collection phase of the study |
| Â | Â | Â | - Assumptions are implicit in the model, hence constant updates required | - Data points outside three standard deviations require further investigation |
Top-down approaches | ||||
Economic Input-output (IO) models | Provide the aggregated monetary/energy flow through an economy | - Analyse the impact of the entire economy on each industry, and inter-industry relationships | - Stop at the point of purchase, and ignore waste and imports | - Employ the principle of embodied energy to convert monetary to energetic values |
| Â | Â | - Data are usually obtained from same sources, which provides consistency in analyses | - Aggregated data analysis prevents detection of specific energy/environmental hot-spots | - Economic IO analysis requires energy/GHG assumptions for wastes and imports |
| Â | Â | - Quantify the impact/weight of each sector, and therefore allows identification of low-performing sector | Frequency of national IO tables is low | - Linear production technologies [2] |
| Â | Â | Â | - No capacity constraints [111] | |
| Â | Â | Â | Â | - Sector homogeneity [111] |
| Â | Â | Â | Â | Usually assume a constant level of technologies for future analyses [2,111] |
| Â | Â | Â | Â | - Import emissions usually based on domestic production technologies [2] |
Index decomposition analysis models (IDA) | Decompose aggregate energy and GHG emissions data into pre-defined factors to measure the relative impacts over specific time periods has been used for: (i) energy demand and supply, (ii) energy-related gas emissions, (iii) material flows and dematerialisation, (iv) national energy efficiency trend monitoring and (v) cross-country comparisons [70] | - Method is relatively quick and simple to implement | - Require an adequate level of disaggregation, else actual effects are not clearly identified | Energy and GHG intensities are usually based on monetary outputs, as opposed to physical outputs |
| Â | - LMDI approach has no residuals in decomposition process | - Laspeyres index is simple to implement, but calculates with residuals | Â | |
| Â | Provide quick access to assessing the overall impact of policy measures on the economy | Â | Â | |
Dynamic models | Aim at predicting future energy and GHG expectations of the food-energy chain | Can provide indication of future energy and policy expectations | - Technological effects are often implicitly accounted in models | - Economic growth rates |
| Â | Â | Â | - Can require significant level of assumptions, which questions the validity of such models | - Time preferences |
| Â | Â | Â | - Studies related to the food chain are scarce | - Population growth rates |
| Â | Â | Â | Inflation and depreciation rates | |