Model coupling, (the communication and interchange of information between models), is a core component in meeting the challenges of many of today’s complex real-world environmental problems, such as climate change, biodiversity loss and air pollution. Model coupling can facilitate analysis of the interdependencies within and between such complex systems, unlocking new realities about the world and casting new light on areas of science that we are currently blind to.
In many ways, we have the technology for model coupling but still need to overcome several barriers if we are to unlock its full potential. CEEDS recently ran a workshop on this topic, which explored ideas for how some of these barriers could be overcome.
One thing that is clear is that to enable a coupling of models, we must first achieve the coupling of disciplines, including environmental science of different flavours, data science and computer science. In bringing together such disciplines, we often face barriers of different perspectives, disciplinary interests and goals. It’s only natural to focus on the development and goals of your own discipline; it takes time and mindful effort to really recognise others’ disciplinary perspectives as well as our own disciplinary capabilities, biases and limitations. We must also recognise that the coupling of models can be as complex as the models themselves, and devote much more attention to this crucial aspect of integrated modelling frameworks.
These foundational barriers require openness, reflexivity, transparency, time and effort to overcome. Building trust around shared goals and fostering teamwork through collaborative environments, such as virtual labs, are two practical ways in which we can start to overcome the common barriers to model coupling.
Virtual labs can also offer a range of building blocks to support the development of model coupling, making it easier to achieve integrated modelling even for novice programmers: to support shared vocabularies and ontologies, to achieve interoperability between different programming languages, to implement downscaling algorithms, and to reason about end-to-end uncertainty, amongst others.
Collaborative environments that can facilitate long-term interdisciplinary training and learning is one of the keys to overcoming these foundational and language barriers. Developing these environments is not easy; ultimately it will require long-term funding that explicitly recognises the benefits of innovation across different fields.
Although overcoming the above barriers is challenging, we are optimistic in CEEDS that we can learn, grow and enjoy this journey. Collaborating with others should be stimulating and ultimately better approaches to model coupling will lead to better science. We also have the skillsets within CEEDS to do something transformative in this space and call on you, if this excites you, to join us in this venture.
Going forward, we seek further conversation and organisation around this crucial topic and ultimately, we plan for this to lead to funding bids that can address this interdisciplinary challenge. If you are interested, contact the CEEDS leadership on firstname.lastname@example.org. We look forward to hearing from you and planning this exciting journey together.