Purpose This paper aims to consider both the greenhouse gas (GHG) emissions and behavioural response in the optimal sizing of solar photovoltaic systems (PV modules and batteries) for energy communities. The objective is to achieve a high self-sufficiency rate whilst taking into account the grid carbon intensity and the global warming potential of system components. Design/methodology/approach Operation and sizing of energy communities leads to optimization problems spanning across multiple timescales. To compute the optimisation in a reasonable time, the authors first apply a simulation periods reduction using a clustering approach, before solving a linear programming problem. Findings The results show that the minimum GHG emissions is achieved for self-sufficiency rates of 19% in France and 50% in Germany. Research limitations/implications The analysis is restricted to specific residential profiles: further work will focus on exploring different types of consumption profiles. Practical implications This paper provides relevant self-sufficiency orders of magnitude for energy communities. Originality/value This paper combines various approaches in a single use case: environmental considerations, behavioural response as well as multi-year energy system sizing.