Authors' reply to "Comments on model reduction by minimizing the equation error"