Is NHC Intensity Forecast Skill Increasing?

During the past few decades, track forecast errors for tropical cyclones (TCs) have shown a remarkable downward trend at all time periods. It is generally believed, however, that this trend does not apply for intensity forecasts, with a flat trend noted through about 2005. However, intensity errors have shown a notable decline since that time, with a large decrease in errors noted at long range. While some of this decrease is likely due to fewer recent strong or rapidly intensifying hurricanes (which generally have higher errors), a significant portion is probably due to higher intensity forecast skill. This trend is seen in both the Atlantic and eastern Pacific, and can also be seen in comparing intensity errors of 1990-94 versus 2011-15. Probable reasons for the increase in skill will be discussed, which include the introduction of additional skillful intensity models and an intensity consensus, plus better tools to diagnose rapid intensification. Some preliminary results suggest that a large part of the skill increase comes from TCs that change 20 kt or greater within 24h. Other composites of intensity errors will be shown, including how the errors change with latitude, initial intensity and MJO state.