Multi-year Predictability of Temperature and Precipitation over Land


Liwei Jia

12:00:00 - 12:25:00

101 , Mathematics Research Center Building (ori. New Math. Bldg.)

This study provides a basis for multi-year predictions over land by explicitly identifying patterns in multiple climate models that are predictable on multi-year time scales. The patterns are identified by maximizing the Average Predictability Time (APT) of surface air temperature and precipitation in pre-industrial control simulations. Because the patterns are identified from control runs, the predictability arises from internal dynamics that occur in the absence of interannual variations of anthropogenic and natural forcing. The leading two most predictable components of surface air temperature are verified to be significantly predictable for 2-20 years, with one component deriving predictability from the persistence of temperature over the oceans, and the other deriving predictability from evolving ENSO patterns. Global annual mean land precipitation is shown to be significantly predictable for 2-4 years in multiple models. These results contradict the widely held belief that temperature and precipitation is unpredictable beyond seasonal time scales.

For material related to this talk, click here.