The design of cost effective power systems with high shares of variable renewable energy technologies (VREs) requires a modelling approach that simultaneously represents the whole energy system combined with the spatiotemporal and inter-annual variability of VREs. Here we soft-link a long term energy system model, that explores new energy systems configurations from years to decades, with a high spatial and temporal resolution power system model that captures VRE variability from hours to years. Applying this methodology to Great Britain for 2050, we find that VRE focused power system design is highly sensitive to the inter-annual variability of weather and that planning based on a single year can lead to operational inadequacy and failure to meet long term decarbonisation objectives. However, some insights do emerge that are relatively stable to weather year. Reinforcement of the transmission system consistently leads to a decrease in system costs while electricity storage and flexible generation, needed to integrate VREs into the system, are generally deployed close to demand centres.

Global ambition to limit anthropogenic warming to 2°C requires a radical transformation of the energy system to one that produces 'net-zero' GHG emissions before 2100 1 . For a 1.5°C limit, action has to be even more rapid, with net-zero emissions achieved much earlier 2 . The goal of net-zero GHG emissions is expressed in the Paris Agreement as a system that achieves 'a balance between anthropogenic emissions by sources and removals by sinks' 3 . In

Energy modelling has a crucial underpinning role for policy making, but the modelling-policy interface currently faces several limitations. Therefore a reinvention of this energy modellingpolicy interface is detailed to better provide timely, targeted, tested, transparent and iterated insights from such complex multidisciplinary tools. Energy models provide the integrating framework that assists energy policy and industrial energy decision makers. By applying data to a coherent theoretical structure and using computer modelling software, they provide essential quantitative insights into alternative energy system design under conditions of pervasive uncertainty.