Tracking integration of climate change adaptation using machine learning

09:00 Thursday 30 May

OC255

Room S16

 

Robbert Biesbroek (Netherlands) 1; Ioannis Athanasiadis (Netherlands) 1; Shashi Badloe (Netherlands) 1

1 - Wageningen University & Research

Global environmental governance frameworks increasingly call for ‘global stocktake’ to track how governments, business and civil society organizations are taking action and transforming their practices to address the critical challenges the Anthropocene poses. How to track progress under the 2030 Agenda for Sustainable Development, the UNFCCC Paris Agreement and the Sendai framework for Disaster Risk Reduction, for example has become an emerging field of research. One critical issue often reported is the challenge of tracking the process of integrating, synchronization and harmonizing global environmental objectives into specific country and sector specific contexts.

Although much literature has described integration as a ‘holy grail’ for policy makers, we know very little about integration beyond single or small-n cases. This is partly because: a) vagueness of the concept of integration in scholarship; b) lack of coherent, consistent and comprehensive frameworks to assess integration; c) intensive data collection necessary for in-depth investigation of integration over time and across context; d) challenge of dealing with strongly varying contextual settings which hamper meaningful understanding of integration.

In this presentation we propose a conceptual model and test its value to systematically track policy integration. The conceptual model makes use of four dimensions of policy integration: frames, subsystems, goals, instruments. We apply the conceptual model to the question of how countries across the globe are integrating climate change adaptation policies, strategies and actions in their existing sectoral policy subsystems. To test the framework we explore the value of supervised machine learning tools as new ways for tracking progress on integrating. Using data science methods we can parse through many text sources and identify patterns of how different policy sectors in a country frame adaptation in similar or different ways; to understand how the concept of adaptation has been diffused and integrated in existing policy sectors; and which policy instruments are proposed to integrate adaptation in each sector. For the pilot, policy and legislative data from the United Kingdom will be used to see how the UK (globally seen as a forerunner country on adaptation) has integrated climate change adaptation in their sectoral policy areas. Using the experiences of the pilot, we will reflect on the conceptual framework and the value of using machine learning tools to track adaptation policy integration globally.