This paper employs a stochastic and dynamic intermediate storage model to estimate the optimal stockpile levels at both a mine and port for a coal supply chain. The optimisation model demonstrates that the principle costs incurred from high inventories of coal include working capital, storage costs and double handling costs, whilst costs incurred from low inventories are dominated by train cancellations, spot price purchases of coal to make up shortfalls and demurrage. The optimisation model allows for the dynamic interaction of cost functions across the supply chain and results in optimal inventories that are typically lower than intuitively assumed by logistics managers.