Envisioning surprises: How social sciences could help models represent ‘deep uncertainty’ in future energy and water demand

Research output: Contribution to journalArticlepeer-review


Medium- and long-term planning, defined here as 10 years or longer, in the energy and water sectors is fraught with uncertainty, exacerbated by an accelerating ‘paradigm shift’. The new paradigm is characterised by a changing climate and rapid adoption of new technologies, accompanied by changes in end-use practices. Traditional methods (such as econometrics) do not incorporate these diverse and dynamic aspects and perform poorly when exploring long-term futures. This paper critiques existing methods and explores how interdisciplinary insights could provide methodological innovation for exploring future energy and water demand. The paper identifies four attributes that methods need to capture to reflect at least some of the uncertainty associated with the paradigm shift: stochastic events, the diversity of behaviour, policy interventions and the ‘co-evolution’ of the variables affecting demand. Machine-learning methods can account for some of the four identified attributes and can be further enhanced by insights from across the psychological and social sciences (human geography and sociology), incorporating rebound effect and the unevenness of demand, and acknowledging the emergent nature of demand. The findings have implications for urban and regional planning of infrastructure and contribute to current debates on nexus thinking for energy and water resource management.
Original languageEnglish
Pages (from-to)18-28
Number of pages10
JournalEnergy Research & Social Science
Early online date30 Nov 2018
Publication statusPublished - Apr 2019


  • demand forecasting
  • decision-making
  • uncertainty
  • paradigm change
  • energy sector
  • water sector
  • nexus

Research Beacons, Institutes and Platforms

  • Energy
  • Sustainable Consumption Institute


Dive into the research topics of 'Envisioning surprises: How social sciences could help models represent ‘deep uncertainty’ in future energy and water demand'. Together they form a unique fingerprint.

Cite this