INNOVIM employee Thomas W. Collow summarizes research he and colleagues are conducting at the NOAA Climate Prediction Center.
As we move deeper into the 21st century, the impacts of climate change have become more pronounced across the globe. One region of increased interest is in the Arctic as numerous studies have pointed out that Arctic temperatures are warming at a faster rate than those further south. A rapidly warming Arctic has led to a sharp decline in sea ice coverage over the last decade. It has been suggested that the shrinking sea ice and warming Arctic may induce changes in the jet stream, which is essentially an atmospheric highway that guides storms. This could result in more unusual storm tracks in the future; with Superstorm Sandy’s westward turn into New Jersey in 2012 a recent example.
Loss of sea ice is highly impactful on societal and economic interests across the Arctic. For example, there has been a notable decrease in the polar bear population with dire projections as we move forward, and the once sought after “northwest passage” may become a viable commercial shipping route by mid-century. However, ships with ice breaking capabilities have already been able to successfully navigate the route as the thickness of the ice (vertical depth) has become low enough for such voyages. It is the goal of meteorologists and climatologists to update forecast systems to perform better given the changed environment to allow stakeholders in the Arctic to make more informed decisions. Our work, recently accepted for publication in Monthly Weather Review, demonstrates how we are working to improve Arctic sea ice prediction.
The National Oceanic and Atmospheric Administration (NOAA) Climate Prediction Center is responsible for the prediction of climate parameters such as sea ice on a seasonal timescale. Meteorologists use a variety of numerical computer models to aid in their forecasts, one example being the Climate Forecast System (CFSv2). However, in recent years the sea ice projections from CFSv2 do not match current trends, and continue to predict too high sea ice amounts. Therefore, we set out to address some issues in the model and create experiments to see if we can improve the prediction of sea ice. First, we used an improved sea ice thickness dataset. Specifically, the Pan-Arctic Ice Ocean Modeling and Assimilation System (PIOMAS) data from the University of Washington’s Polar Science Center was substituted for the CFSv2 model’s original thickness. The PIOMAS thickness proved to be more in line with observations and therefore better suited for modeling than the current data being used. We also made some changes to how clouds are generated in the model and how heat is exchanged between the oceans and ice. Ultimately the changes we made greatly improved Arctic sea ice prediction based on simulations of the 2005-2014 sea ice melt seasons and comparisons to observations. Starting in March 2015 we have been issuing experimental sea ice outlooks using this new model configuration to the National Weather Service – Alaska Region. The products have been received favorably and even shared with President Obama on his recent visit to Alaska. Our goal is to continue to refine the system and transition it into an operational product over the next few years.
Citation: Collow, T., W. Wang, A. Kumar, and J. Zhang, 2015: Improving Arctic sea ice prediction using PIOMAS initial sea ice thickness in a coupled ocean-atmosphere model. Mon. Wea. Rev. doi:10.1175/MWR-D-15-0097.1, in press.Share