Did a skillful prediction of near-surface temperatures help or hinder forecasting of the 2012 US drought?