Workshop: Stratospheric predictability and impact on the troposphere

Processes

• Which stratospheric processes are important to represent accurately for improved tropospheric predictability on medium-range to seasonal timescales?

• Which tropospheric processes are important to represent accurately for improved stratospheric predictability?
   
• Can we quantify the strength of stratospheric impact on the troposphere (e.g., following weak and/or strong vortex events) versus the impact of other remote forcers such as MJO and ENSO?

• What are the best tools and metrics to objectively evaluate the strength of the stratosphere-troposphere coupling, and, the skill provided by the stratosphere for tropospheric predictability?

Model design and development

• What can be done to eradicate persistent stratospheric biases shared by many models? Is there an indication that certain biases are more important than others for tropospheric predictability?
   
• What is the most efficient testing protocol for assessing if model changes result in an enhanced tropospheric skill as a result of a better stratospheric representation? For example, should we be focusing on the time periods around “windows of opportunity” (e.g., SSWs, final warmings in the Southern Hemisphere) when stratosphere is known to impact the troposphere?

• How important is adding more complexity to parametrization design is for stratospheric circulation? E.g., should we be building source dependency into nonorographic gravity wave drag parametrizations in high resolution models? Should we be incorporating 3D propagation into gravity wave drag schemes?
   
• Are there certain dynamical core designs that are known to better represent stratospheric processes (e.g., gravity waves, transport, QBO) and/or stratosphere-troposphere coupling?
   
• Does the location of model top matter for stratospheric and/or tropospheric predictability? If so, where should the model top be located?

Role of observations

• Are there any additional observations which are needed to enable model improvements relevant for weather prediction?
   
• Are observational data sufficient to initialize the stratosphere?
   
• How much do (or should) modellers trust stratospheric re-analyses?