2018pcpi banner

3: Improve understanding of polar predictability on seasonal to decadal timescales

Co-Leads: Ed Hawkins (University of Reading, UK) and John Fyfe (CCCma, Canada)

The goal of theme 3 is to coordinate research that investigates the physical mechanisms that give rise to polar climate predictability on seasonal to multi-decadal timescales. The intent is to ascertain the potential for improving the actual skill of operational climate predictions for polar regions. The participants of this group will coordinate and collaborate with operational forecasters wherever possible. An important focus would be the impacts of polar variability on lower latitudes, for which there is growing evidence. This theme aims to apply understanding of the system gained through investigating seasonal to decadal prediction to guide observational network design and deepen our understanding of longer timescales and validate and/or improve models for all timescales.

Some key questions for the development of GCM-based operational systems that we seek to investigate:
1)    How important is sea-ice thickness assimilation for skilful predictions? [very?]
2)    What is the seasonal and state dependence of sea-ice predictability? [early summer barrier?]
3)    What do users of the forecasts actually need? Concentration, thickness, storminess, waves?
4)    Will initialisation of sea-ice improve forecasts of trends in extent, or improve skill outside the Arctic?
5)    If low skill is found, is this due to: inadequacies in the model, uncertainty in initial conditions or inherent lack of predictability?
6)    How will predictability change in a warmer climate with less ice?

Possible next steps:
1)    Coordinate with the numerous nascent projects to investigate this topic through a workshop and follow-up with teleconferences and other communication devices.
2)    Encourage forecast exchange for seasonal to inter-annual predictions of pan-Arctic and regional concentration (and thickness, ice age?) using range of approaches from statistical to GCM-based
3)    Include a focus on ‘case-studies’, including 1996, 2007, 2012 & 2013
4) Encourage operational centres to produce parallel “perfect” predictability experiments to try and answer point (5) above
5)    Encourage parallel atmospherically forced ice-ocean hindcasts?
6)    Engage stakeholders and understand their requirements