Decadal climate forecasting for the energy-sector

Throughout the world, power systems are undergoing massive change in response to the challenge of climate change. Renewable electricity sources - such as wind and solar PV - are playing an increasing role in power systems, fundamentally altering the way power systems operate. In Great Britain, increasing levels of wind power installation now mean that shifts in wind speed are now as important as shifts in temperature in maintaining a secure (i.e., continuous and near-instantaneous) match between electricity-demand and electricity-supply and, as a result, anticipating low-wind-cold-snaps has become a key issue for supply security (Bloomfield et al 2018; Thornton et al 2017). Similar challenges around the integration of renewables into existing power systems are being faced by many countries across the world, raising an important question: how will climate variability and climate change affect the behaviour of power systems in the coming years and decades?

Recent years have seen increased interest from both academia and industry into the risks posed by climate variability and change for power system operations and planning. Much of this research, however, has been based either on an assumption of an unchanging climate (i.e., using historical observations) or else very-long-term climate model projections (relating to 2050 or beyond). There is therefore a significant "decadal-scale information gap" (1-10 years) between these extremes which is significant for power system planning. High quality decadal-scale climate information could, for example, support the provision of early warning systems for stress periods such as low-wind years, or support the identification of optimal "pathways" for integrating new infrastructure over the course of several years.

Decadal-scale climate forecasting is, however, scientifically challenging in that it seeks to make a concrete prediction of near-term future climate (i.e., statements about what will happen based on some initial starting state) as opposed to the long term projections more commonly with climate models (i.e., possible outcomes contingent on the occurrence of a particular greenhouse gas scenario). Until recently, it has been widely believed that the skill of decadal forecasting was rather limited, but recent developments have indicated surprising skill several years ahead (Smith et al, 2020). This PhD will build on these recent developments to explore the extent to which skilful decadal climate predictions can be used to inform the operation and management of the energy system from a season to several years ahead.

This project will therefore seek to pioneer the use of decadal forecasting for energy system applications. A central focus of the work will be to characterize and understand decadal forecast skill and to identify and develop applications through which it can be utilized.