On the importance of fine-scale spatial complexity in vegetation restoration studies