Resilient systems have been characterised as those that have a higher capacity to absorb shocks and stresses; possess the ability to self-organize into flexible and responsive networks for learning, distribution and change, and; have greater capacity for adaptability. While these concepts have been well developed in the literature as theoretical and conceptual frameworks, there are few examples of operationalizing and empirically applying these concepts, particularly for agroecosystems which are among the most complex of social-ecological systems.

Using a ‘bottom-up' and participatory-based approach, we reviewed and then empirically applied a set of behavioural indicators across three different types of dairy farm systems in eastern New Zealand: organic, low-input or grass-based, and high-input, intensive systems in which supplemental feed is the main input. Results show differences in the resilience and capacity for adaptation for different farm types. The ‘lock in trap' of highly intensive systems, while profitable in the near term, may be less resilient to climate shocks as these are likely to occur in conjunction with changing market and financial risks. Low-input systems are less dependent, in particular, on fossil fuels and were associated with higher levels of farmer satisfaction and well-being. The framework provides a useful template for cross-sector comparison, and demonstrates that in-depth, robust qualitative assessments of resilience can provide a complement to quantitative metrics. The characterisation of resilient farm-systems also has the potential to contribute to broader sustainability frameworks for agriculture. The findings also have implications for the future of the world's largest exporter of dairy products.