Dynamical Climate Predictability of the NCEP CFS in the Stratosphere and the Statistical Downscaling for Climate Prediction in the Troposphere

The primary objective of our proposed study is to explore a new source for the prediction skill of climate variability from intra-seasonal to interannual time scales, namely the stratospheretroposphere coupling. We envision that in addition to well researched ocean-atmosphere interaction (ENSO) and land-surface interaction, this may be the new avenue of progress through modes of variability that involve the stratosphere, or more generally the hemispheric mass circulation from the tropical troposphere into the high latitude stratosphere, and back through the troposphere.

The three main proposed research activities are

Task A: To systematically evaluate the prediction skill for stratospheric anomalies in the 25-year (1982-2006) retrospective seasonal climate predictions made by NCEP’s Climate Forecast System (CFS).

Task B: To explore the way of ‘specifying’ surface weather or its statistics (over a month) from (predicted values of) a few predictable modes in the stratosphere. In other words, we propose to identify some statistical diagnostic tools on the relationships between stratospheric circulation anomalies and the statistical distribution (or changes in the statistics) of tropospheric/surface synoptic scale weather events.

Task C: To explore how to utilize the information derived from the model-based stratosphere prediction and statistical-based stratosphere-troposphere coupling to improve the climate prediction skill of surface weather beyond the lead time of 2 weeks. We will develop a new set of products for winter season climate predictions at various lead times from a month to a season or longer.

The overall goal of the proposed research is to utilize the extra NWP skill in the stratosphere beyond inherent predictability time scale to be identified in “Task A” for climate predictions of tropospheric circulation anomalies by “downscaling” the stratospheric climate prediction (Task C) via the statistical diagnostic relations between the stratospheric and tropospheric anomalies to be identified in Task B. In essence, we propose a new hybrid climate prediction strategy: predicting the stratospheric circulation anomalies by a dynamically based general circulation model beyond the inherent predictability time scale and using the simultaneous diagnostic relation between the stratospheric and tropospheric circulation anomalies for climate predictions of the tropospheric climate variability. The proposed hybrid prediction strategy for troposphere/surface climate variability at seasonal and interannual time scales is akin to the strategy for weather forecasts in 1960/70s, which was to predict 500 hPa circulation using dynamical models and to forecasts surface conditions using (simultaneous) statistical downscaling from the 500 hPa to the surface (e.g., MOS).