Evaluation of the ability of CFSv2 to forecast polar vortex displacements in the stratosphere of the Northern hemisphere

Abstract

In this paper, the ability of the CFSv2 model to forecast the polar vortex displacements in the stratosphere of the Northern hemisphere is evaluated. We use the ERA-I reanalysis daily data of the potential vortex (pv) and the horizontal wind; vertical wind in the 430, 475, 530, and 600 K in the potential temperature field; and the temperature in the geopotential height at 50 hPa. And the 1–850 hPa isobaric surface temperature and u, v wind field daily data from the 9-month business forecast data of the Climate Forecast System Version 2 (CFSv2). The two polar vortex displacement events shift to Eurasia that occurred in Feb 2008 and Jan 2012. The results show that the polar vortex displacement events in February 2008 and January 2012 shifted toward the European region. During the polar vortex displacements, the westerly wind at 60°N latitude decreases, the temperature gradient reverses at the north polar and 60°N latitudes at 50 hPa, and a weaker stratospheric warming occurs in this process. For two of the polar vortex displacement events, the CFSv2 predictions of the start, largest, and end times of the polar vortex displacement are accurate at forecast lead times of 0 and 5 days, and the prediction results for the westerly winds and temperature gradients are also excellent. However, the CFSv2 prediction ability at forecast lead times exceeding 20 days is poor. For the polar vortex displacement event in February 2008, the CFSv2 predictions are good for the start and largest times but are not consistent with the actual situation for the end time at a forecast lead time of 10 days. The CFSv2 predictions are good for the end time but are not accurate for the start and largest times at a forecast lead time of 15 days. For the polar vortex displacement event in January 2012, the CFSv2 predictions are poor for the start, largest and end times and for the zonal wind, but are good for the temperature gradient during the displacement at forecast lead times of 10 days. The change in the polar vortex boundary and zonal wind and temperature gradients cannot be predicted at forecast lead times of 15 days.