Nonstationarity of turbulent heat fluxes at Summit, Greenland

Turbulence data collected over a total of 25 days during two summers are used to describe processes responsible for the nonstationarity of turbulent sensible heat fluxes at Summit, Greenland. A stationarity test shows that about 40% of the data are classified as nonstationary. Three main factors are explored to account for the large fraction of nonstationary runs: (1) intermittency of turbulence in stable conditions, (2) changes in net all-wave radiation in response to cloud forcing, and (3) diurnal trends in stability. A classification procedure that accounts for the intermittent nature of turbulence shows that during stable, nonstationary conditions 50% of the total sensible heat flux is realized in 22% of the sampling time. Intermittency often occurs at Summit during periods characterized by weak and irregular horizontal winds in combination with strong stability. Rapid changes in net all-wave radiation in response to cloud forcing results in nonstationarity during unstable conditions. Between 0930-1130 and 1900-1930 UTC turbulent heat fluxes are not only small in magnitude but also typically change sign, with nonstationarity during these periods often as high as 65%. These results should help resolve some of the present uncertainties in obtaining reliable fluxes at this site, in particular under stable atmospheric conditions.