Calibrating pollen data in climatic terms: Improving the methods