Benchmarking the environmental performance of dairy farming systems

Summary

Milk production has a major impact on the environment and competes increasingly for scarce resources. As the demand for milk is expected to increase, these issues are likely to worsen. Benchmarking the environmental performance of dairy farming systems offers the opportunity to identify best farm practices and to provide guidance for reducing the environmental impact. Currently, benchmarking is hampered by the lack of an effective method that results in a set of indicators that is easily quantifiable and detects variations in environmental performance between farms. The aim of this thesis, therefore, was to develop a sound method to benchmark the environmental performance of dairy farming systems. This thesis focuses is on specialized dairy farming systems in Europe.

The first challenge in benchmarking the environmental performance of dairy farming systems is to select a set of indicators that are relevant, measurable, valid, timely and understandable. Environmental indicators can be derived from various approaches, including a nutrient balance (NB) approach and a life cycle assessment (LCA). An NB is generally applied at farm level, and yields indicators that are relatively easy to quantify and communicate. We found that an NB at farm level can be used to benchmark dairy farming systems, if differences in on-farm losses are large and off-farm losses are relatively unimportant. Only if farms differ largely in the amount and/or type of purchased inputs, such as feed, the farm-based NB should be extended to a chain based NB or an LCA. An LCA, however, requires extensive data information, which can be difficult to collect. We, therefore, explored correlations between eight commonly used NB and LCA indicators with the system boundary from cradle-to-farm gate. We found that a set indicators, consisting of the nitrogen surplus, the phosphorus surplus, land use and energy use can be used as a proxy to benchmark the environmental performance of dairy farming systems, representing also global warming potential, acidification potential, freshwater eutrophication potential and marine eutrophication potential.

The second challenge in benchmarking the environmental performance of dairy farming systems is to cope with data uncertainties. We therefore first evaluated the effect of epistemic uncertainty on benchmarking the nitrogen use efficiency of dairy systems. We found that ranking of farms based on this single indicator is not possible when the epistemic uncertainty of parameters is large and differences in N use efficiency are small. We, furthermore, identified the most influential parameters (e.g. input of concentrates, mineral fertilizer ) and found that reducing epistemic uncertainty of those parameters improved benchmarking results significantly. Afterwards, we demonstrated how to use fuzzy data envelopment analysis (DEA) to account for uncertainties of multiple indicators in benchmarking the eco-efficiency of dairy farming systems. With fuzzy DEA, the number of farms receiving the highest efficiency score was lower compared to standard DEA. In addition, fuzzy DEA identified a different set of peers than standard DEA. By taking uncertainty into account during the quantification processes, fuzzy DEA can contribute to increasing the reliability of results and prevent biased conclusions.

Exploring correlations between environmental indicators can facilitate decision-makers to derive an effective set of indicators that can be used as proxies for benchmarking. In addition, decision-makers should acknowledge the effect of epistemic uncertainty on benchmarking results. When setting up reference values for penalty, for example, this value should be based on a range rather than a single value in order to account for epistemic uncertainty.