Transdisciplinary thinking essential to enable sustainable data-intensive agroecosystems
Transdisciplinary thinking is a fundamental requirement if we are to take meaningful steps towards using big data to solve complex socio-environmental problems.
Photo by Neil Palmer
Data is necessary, but not sufficient, to empower a generation of problem solvers poised to confront socio-environmental challenges related to agroecosystem sustainability.
What is needed, but has been largely absent, is a cadre of problem solvers who are transdisciplinary thinkers proficient in wrangling the multitude of resources available in the open digital commons.
These individuals include, to varying degrees, farmers in the field, government agricultural planners, FAO policy leaders, and just about anybody in the ecosystem of stakeholders who plays a part in the solution space of sustainable agriculture.
The challenge
The informatics community has made substantial progress in technical (and sometimes legal) interoperability and community norms through bodies like the W3C, OGC, GODAN, RDA, the US-based ESIP, and others.
These advances power informatics solutions that better fulfill FAIR (findable, accessible, interoperable, reusable) principles.
Are these advances sufficient to enable science-informed decisions that address pressing issues of agroecosystem sustainability? The answer is: no.
There is little doubt that technical solutions are critical for certain big data applications, such as those routinely used by the high-energy particle physics (e.g., Large Hadron Collider) and astronomy (e.g., Square Kilometer Array) communities.
The data landscape for agroecosystem sustainability challenges is very different in nature.
Such challenges belong to the class of ‘wicked problems’, where socio-environmental challenges with ill-defined scopes evolve over time, where coalescence around a consensus is hard, and where metrics for success are ill-defined.
Sustainability and climate resilience challenges are often classified as wicked problems. These challenges often involve trade-offs between needs that span environmental, economic, and societal dimensions.
For example, rice production and fisheries in the Mekong River Delta face multiple threats including human population change, biodiversity loss, reduction in water flow from upriver damming, saltwater intrusion from sea-level rise, changing patterns of alluvial deposition, shifting profitabilities of agriculture versus aquaculture, an imbalance of political representation on business councils, and many other stresses that operate at various spatial, temporal, and governance scales.
Furthermore, these coupled human-natural stresses and impacts evolve in response to each other.
Data for estimating correlations, much less hinting at causations, is heterogeneous and scant, characterized as ‘miles wide, inches deep’.
This is in stark contrast with data routinely utilized by the physics and astronomy communities mentioned above, which, in relative terms, are inches wide, miles deep. For wicked problems, ‘Variety’ dominates ‘Volume’ and ‘Velocity’ in the 3V formulation of the big data definition.
Before ‘miles wide, inches deep’ data can be applied to problems, a team that finds itself embedded within the wicked-problem landscape needs to formulate a shared understanding of the problem-space.
Only then can the requisite data be identified and informatics tools deployed to find, access, and use relevant data.
Very little of this work happens in a linear fashion: the limited scope and budget of projects typically constrain de-novo data collection, compelling the team to repurpose existing data that vary in fitness-for-use considerations.
The initial problem-space, therefore, needs to be revisited occasionally and team members recalibrated on their shared understanding of the problem.
Transdisciplinary capacity building
The biblical Tower of Babel is used to explain how the discontinuity posed by different languages prevented humanity from building a tower to heaven. This is echoed in the different paradigms and vocabularies that highly trained individuals bring to multidisciplinary teams.
For instance, the term climate change elicits different connotations from a cultural anthropologist, an agronomist, an environmental economist, and an atmospheric scientist.
Each scientist brings a slightly different paradigm – a different conceptual framework – to bear on the problem. These differences ultimately impede the realization of the FAIR data principles because of the different ways we search for and use data, code, and information.
We posit that effective data-intensive approaches for socio-environmental challenges require transdisciplinary – not multidisciplinary – teams.
The distinction between multidisciplinary and transdisciplinary teams is not academic: the authors bring anecdotal evidence of multidisciplinary teams that have fallen short of expectations.
Multidisciplinary teams typically come together with ideas and common methods within their disciplines to solve problems, then work in their disciplinary silos to solve their piece of work (i.e., no knowledge integration across disciplines).
Transdisciplinary teams, on the other hand, normally include topic experts, stakeholders, and decision-makers from the social and natural sciences.
Members of the team undergo training that includes human behavioral aspects, such as common goal-setting exercises and social contracts for resolving differences. The training may also include conceptual tools used to broker different paradigms and vocabularies.
One such tool is the “boundary negotiating object” (BNO). BNOs represent a negotiation of concepts across disciplinary boundaries achieved through iterations of shared representations of the problem (e.g., concept maps, flow charts, etc.) and dialogue.
Transdisciplinary teams are also proficient at identifying socio-environmental frameworks (such as the OECD-derived Drivers, Pressures, State, Impact, Response framework, the UN Millennium Ecosystem Assessment framework, and the recent IPBES Nature’s Contributions to People framework) that can be used to facilitate interoperability between disciplines. BNOs and frameworks are useful, for example, in wicked problems like the Mekong River Basin example mentioned above.
Back to data
The underlying complexity of agroecosystem sustainability wicked problems is a significant hurdle to the effective use of big data for such applications.
From the informatics perspective, that complexity ultimately reinforces the roles of semantic technologies and workflow provenance to enable technically defensible, fully traceable decisions for wicked problems.
With such technologies, decisions, with carefully managed relationships to code, data, peer-reviewed articles, and other information can be shared, repurposed, and contributed to the open digital commons.
GitHub and Jupyter Notebooks are two examples of technologies that facilitate traceable open science and enhance reproducibility.
To fully exploit these technologies, transdisciplinary thinking is a fundamental requirement if we are to take meaningful steps towards using big data for socio-environmental wicked problems.
Aaron J. Piña & Brian Wee
Colorado & Washington, DC, USA
This article was co-authored with contributions from Brian Wee, founder and Managing Director of Massive Connections, a consultancy that connects data, science, technology, and policy to societal needs; and Aaron j. Piña, staff scientist at Aeris LLC, a small business that provides atmospheric science and engineering solutions for aerospace, defense, and intelligence applications.
Blog Competition Entry
This article is published as a part of our publicly open big data blog competition. If you have enjoyed this reading this entry, you can vote by liking, commenting or sharing.
Becky, thank you for your questions. These are topics Brian and I have each extensively considered. Your questions are relevant, in a general sense, to the use of big data in agriculture (i.e. not just this article). We are not ignoring the issues of basic accessibility, which Brian and I are both aware from interactions with the American Indian communities–even here in the US, basic internet community to certain native colleges is far from certain. We are suggesting the need for scientists to get transdisciplinary training to better work with the community stakeholders (of all backgrounds)– transdisciplinary training is an essential component to consider the questions you ask. Accessibility issues of course remain a concern, just as it is a concern here in the US, but that wasn’t the scope of this blog. This 1000-word blog piece was specifically targeted at scientists and professionals who have been carried away with the hype of big data, and we’re saying, without saying that this is hype, that data (big or small) is really not the issue.
How do FAIR principles apply to international settings where trans-disciplinary teams include local government partners with low capacity and engage poor rural farmers? What mechanisms can be used to help local individuals and communities access, utilize, and participate in the accuracy of big data if they don’t have access to smartphones and computers? GitHub and Jupyter Notebooks may be good for inter-disciplinary academic work that involves researchers with coding skills, but translating it into trans-disciplinary work, in many settings, is a completely different story, no? How can we rectify some of these challenges/complexities that wicked problems in agroecology most certainly involve, moving forward?