The effect of variability in renewable input sources on the optimal design and reliability of an integrated energy system designed for off-grid mining operation is investigated via a two-stage approach. Firstly, possible energy system designs are generated by solving a deterministic non-linear programming (NLP) optimization problem to minimize the capital cost for a number of input scenarios. Two measures of reliability, the loss of power supply probability (LPSP) and energy index of reliability (EIR), are then evaluated for each design based on the minimization of the external energy required to satisfy load demands under a variety of input conditions. Two case studies of mining operations located in regions with different degrees of variability are presented. The results show that the degree of variability has an impact on the design configuration, cost and performance, and highlights the limitations associated with deterministic decision making for high variability systems. (C) 2016 Elsevier Ltd. All rights reserved.