Effects of Extreme Weather Events on Time-Mean Wind and Derivative Wind Fields

Extreme weather events, defined as greater than three standard deviations from the mean, significantly affect time-mean winds. These effects are especially pronounced in time-mean divergence, vorticity, and vertical velocity fields. This talk presents analysis of a variety of wind sources over mid-latitude ocean regions to determine the degree that these wind and derivative wind fields are affected by extreme weather events. We analyze 10 years of QuikSCAT scatterometer surface wind observations, and winds from a long integration of high resolution COAMPS mesoscale model simulations and NCEP reanalysis fields, the latter two of which include three dimensional winds throughout the depth of the atmosphere. We show that the signs of the time-mean divergence, vorticity, and vertical velocity change in many regions depending on whether or not extreme events are included in the time means. Storms are often marked by strong convergence, cyclonic vorticity, and upward motion. We show that storms contribute to generating highly skewed distributions of divergence, vorticity, and vertical velocity. Skewness inherent in these distributions is often sufficient to dramatically alter their time means, so that the time means are not indicative of prevailing conditions. These results have implications for interpreting some hypotheses of ocean-atmosphere interaction based mostly on analysis of time-mean wind and derivative wind fields. Most often, these hypotheses involve assumptions that synoptic weather variability is a source of noise in winds which can be mitigated through time averaging, and that the resultant time means represent typical conditions. In essence, we argue that this is not necessarily the case. An example demonstrating this counterpoint will be shown concerning bands of time-mean convergence and upward vertical velocity coincident with western boundary currents which have been hypothesized to be a steady response to mesoscale SST forcing.