An environmental systems analysis of greenhouse horticulture in the Netherlands : the tomato case

<strong>Objective of the thesis</strong>The greenhouse horticulture sector in the Netherlands covers about 10,000 hectares and produces vegetables, cut flowers and pot plants. This agricultural sector is of social and economic importance because of its annual production value, export earnings and the employment it provides. Cultivation in greenhouses, however, is characterised by high inputs of energy, fertilisers and chemical biocides, which contribute to several environmental problems. Through technical options, these environmental problems can be reduced.The general objective of this thesis is to identify technical options to reduce the environmental impact of greenhouse horticulture in the Netherlands and to evaluate their cost-effectiveness. The study focuses on tomato cultivation and on the environmental problems of global warming, acidification, eutrophication, dispersion of toxic biocides and the production of waste. The method of environmental systems analysis is used as a tool for the assessment of technical options for reducing the impact of the sector on multiple environmental problems. A side-objective is to discuss the usefulness of environmental systems analysis in such analyses. <strong>System boundaries and model components</strong>The first step of the analysis is the definition of the system boundaries and the determination of the system components. To this end we carried out a limited environmental life cycle analysis (LCA) for tomato cultivation under glass and its contribution to global warming, acidification and eutrophication. We focused on these three environmental problems because of the interrelations between the underlying processes and the emissions. We quantified the emissions of first-order processes (these are activities such as the use of natural gas and fertilisers) and second-order processes (these are industrial activities such as the production of electricity and fertilisers). Results indicated that, in general, the emissions of first-order processes exceed the emissions of second-order processes. However, in some cases the off-farm emissions were relatively high. For example, the production of electricity and rock wool contribute almost 25% to total acidifying emissions. We concluded that a study of the environmental impact of tomato cultivation in the Netherlands needs to consider CO ₂ emissions from the use of natural gas and the production of electricity, NO _x emissions from the use of natural gas and fertilisers, and from the production of electricity and rock wool, and losses of nitrogen and phosphorus from the use of fertilisers. In addition, we considered biocide use and biocide emissions to the environment and the production of waste. We argue that a profound study of the definition of system boundaries is worthwhile and provides a better understanding of the system.<strong> Model building</strong>A model was developed that can be used to quantify the environmental impact of tomato cultivation in the Netherlands and that can be used to evaluate the cost-effectiveness of technical options for reducing the environmental impact. The model calculates the environmental impact as a function of the activities, emission factors and reduction options applied. The impact is quantified by using environmental pressure indicators mainly. The activities include the use of natural gas, fertilisers, biocides and rock wool, as well as the production of electricity and rock wool. The model calculates the environmental impacts through global warming (emission of carbon dioxide in kg CO ₂ ), acidification (emission of nitrogen oxides in kg NO _x ), eutrophication (emission of eutrophying compound in kg phosphate (PO ₄ )-equivalents), dispersion of toxic compounds (indicated by the use of biocide in kg active ingredients and the emission of biocides quantified by the biocide-air-emission score) and the production of waste (kg waste). The model includes 22 groups of technical options that reduce the activity levels (e.g. the use of gas) and/or the emission factors (e.g. NO _x emissions from gas use). The model accounts for important side effects of the reduction options on the tomato production and on the levels of the activities and other emissions. The reduction costs are calculated as annual costs per hectare and include the annualised investment costs, operational costs and variable costs of the technical options applied (e.g. saving in gas use and effects on the production level). The model selects the most cost-effective combinations from all possible combinations of options. <strong>Exploration of the model</strong>We explored the model for a hypothetical reference situation in which none of the technical reduction options were applied. The model calculations, for instance, showed that most of the profitable options are related to a reduction in the use of gas. The combi-condenser and heat buffer were selected by the model in all cost-effective combinations of options. The cost curves for the hypothetical reference situation illustrate the costs of combinations of options for different reductions in emissions. The cost curves indicated that considerable reduction of the environmental impact could be achieved at net zero costs, but that increasing reduction of the environmental impact resulted in rapidly increasing costs.The model calculations show that reducing the emissions for one compound may also affect the emissions of other pollutants (the above-mentioned side effects). This is especially the case for the reduction of emissions of CO ₂ and NO _x . Furthermore, the options for reducing emissions of eutrophying compounds may affect the emissions of CO ₂ and biocides into the atmosphere and vice versa.To further explore the model we carried out a sensitivity analysis. The model results were found to be sensitive to changes in the values of the emissions factors for biocides into the atmosphere. These emission factors are relatively uncertain. We point out that more research is needed on the quantification of the emission of biocides from greenhouse horticulture. The model results are also sensitive to the inclusion or exclusion of NO _x as an eutrophying compound, and to changes in prices, in particular to changes in the price of natural gas. The higher the gas price the higher the savings in costs for each m <sup>3</SUP>reduction in gas use; consequently, the more expensive reduction options are selected by the model in cost-effective combinations of options.We applied five different methods for multi-criteria analysis (MCA) in order to evaluate the cost-effectiveness of the reduction options in reducing several environmental problems simultaneously. The results of these MCA methods differed considerably. They resulted in different numbers of cost-effective combinations of options and in different options selected in the cost-effective combinations. On the other hand, the use of multi-criteria analysis appeared to be useful for tracing robust reduction options. These robust reduction options are options that were always selected in the cost-effective combinations of options, independent of the multi-criteria method used. Robust reduction options for tomato cultivation include the combi-condenser, heat buffer (both improving energy efficiency), high pressure cleaner (reducing the biocide use) and regional schemes for composting organic waste (reducing the amount of waste disposed). Many different MCA methods are available and the choice of the methodology is subjective. For this reason, and because of the observed differences in the results of the five MCA methods, it may be recommended to use more than one method for MCA in (environmental) research to gain insight into the effect of the choice of MCA method used and to trace robust reduction options. <strong>Model application and optimisation analysis</strong>We analysed cost-optimal strategies to meet national environmental targets for tomato cultivation in the Netherlands. We accounted for some of the heterogeneity of Dutch tomato cultivation (1220 hectare) by defining three farm types: Innovators (245 hectare), In-Betweens (730 hectare) and Low-Costs (245 hectare). The main differences between these farm types are the size of the greenhouse, the production volume, the intensity of the production (and therefore the activity levels) and the application of emission reduction options. Innovators produce tomatoes in relatively large new greenhouses and apply several technical reduction options. Low-Costs farmers produce tomatoes in relatively small old greenhouses and apply few options. The In-Betweens occupy the middle ground between the two other farm types. Data were based on the 1995 situation.Two types of optimisation analysis were performed. The first type aimed at minimising the costs to achieve selected environmental targets for the tomato cultivation sector as a whole. The environmental targets for tomato cultivation are based on current policy for Dutch greenhouse horticulture. In the second type of optimisation analysis we analysed the extent to which the environmental impact could be reduced under different cost constraints. In both types of optimisation analysis we calculated the areas on which cost-effective combinations of options are applied on the different farm types to achieve the optimal situation.The results of the optimisation analysis indicate that current policy targets for the tomato sector for global warming (24% reduction in CO ₂ emission) and eutrophication (45% reduction in the emissions of eutrophying compounds) can be achieved at net negative costs. The targets for biocide emissions to the atmosphere (a reduction of 38% from 1995 levels) and energy efficiency improvement (65% over 1980 levels) can be achieved at relatively high costs. The target for the reduction in biocide use in 2010 cannot be achieved by the technical reduction options analysed. When model results indicate that the costs of achieving a certain target are negative at the national level, this may not be necessarily the case for all individual firms. Model results indicate that the most cost-effective solutions are often achieved at the national levels by applying relatively expensive options to part of the tomato cultivation area and relatively cheap options to other parts of the area. In these solutions, the costs fall mainly on the farm types Low-Costs and In-Betweens.We also calculated cost-optimal ways to achieve the above mentioned policy targets for tomato cultivation sector simultaneously. The model results indicate that the net costs of achieving all targets simultaneously are negative. We observed that the model selects other combinations of reduction options in the multiple target optimisation than for the optimisation of the individual targets. Most noticeable in this respect are the selection of a fixed screen and double glass in the roof in combination with the no-windows options, in addition to the more generally applied options such as the combi-condenser, heat buffer, Econaut, strips around window panes and regional composting.The results of the second type of the optimisation analysis were used to develop cost curves. These curves illustrate the costs of different levels of national emission reduction for tomato cultivation and were developed for all individual environmental problems analysed as well as for the integrated environmental impact using the five multi-criteria methods. The cost curves illustrate that, for the 1995 situation, for each environmental problem considered about 20% of the impact from tomato cultivation in the Netherlands can be reduced at net zero costs. <strong>Extrapolating the results to total greenhouse horticulture</strong>Extrapolating the results of the tomato case to the whole greenhouse sector is not easy because of crop specific characteristics. Most importantly, tomato cultivation uses relatively more natural gas, fertilisers and rock wool than the greenhouse horticulture sector as a whole, but less electricity. Furthermore, tomatoes are relatively sensitive to a reduction in radiation (light) that may result from the application of some reduction options, such as screens. These differences may have an important effect on the cost-effectiveness of the options. Despite these differences we discussed the possibilities of achieving the environmental targets for greenhouse horticulture in 2010, based on the tomato case. We argue that most of the environmental targets probably can be achieved, but that targets for the emission of eutrophying compounds and biocides to the atmosphere will probably be difficult to achieve. A more thorough analysis is needed to draw more definite conclusions on the reduction of the environmental impact of greenhouse horticulture as a whole. <strong>Environmental systems analysis</strong>This thesis shows that environmental systems analysis can be useful for analysing complex problems concerning economic and environmental aspects. The environmental systems analysis procedure involves six steps: 1. definition of the system boundaries and the system components, 2. description of the objectives, 3. model building, 4. systems analysis, 5. selection of the optimal system, and 6. conclusions and documentation. It is our experience that these steps are performed interactively rather than in a strict sequence. Iteration and feedback occur to help refine the objectives, improve the model and adapt the constraints. In the analysis we applied a combination of tools, including environmental life cycle analysis, environmental indicators, cost-effectivety analysis and optimisation analysis, multi-criteria analysis and sensitivity analysis.