Biennial Variability in an Atmospheric General Circulation Model

Michael A. Alexander, CIRES, University of Colorado, Boulder, CO,

Klaus M. Weickman, CDC, NOAA, Boulder, CO


Recent observational analyses have indicated that tropospheric quasi-biennial oscillations (QBs) may play a fundamental role in regulating the timing and strength of El Nino and the Southern Oscillation (ENSO). We have examined the biennial variability in the tropical troposphere of a 35 year general circulation model (GCM) simulation forced by observed sea surface temperatures (SSTs). The results of spectral analyses and temporal filtering applied to the SST boundary conditions, and the simulated lower and upper tropospheric zonal winds, precipitation, and sea level pressure anomalies, are compared with observations and used to investigate the relationship between variables.

The GCM obtains regions of coherent biennial variability over the tropical Indian and Pacific Oceans in close correspondence with observations. In addition, the evolution of the stronger QBs and the physical relationship between variables are fairly well simulated. Zonal wind anomalies, with a simple baroclinic structure, tend to propagate eastward from the Indonesian region to the central Pacific where they increase in strength. The amplitude of the zonal wind and SST anomalies in the central Pacific vary together, with the largest anomalies occurring during the mid 1960's, mid 1970's, and in the early 1980's. During the time of the warmest SSTs low pressure is found in the east Pacific with high pressure over Indonesia and precipitation is enhanced between the dateline and 120oW. However, the model underestimates the low frequency variability in general and has approximately 1/2 to 2/3 of the observed variability in the biennial range. In addition, the observed phasing of the biennial and annual cycles in the zonal winds over the eastern Indian Ocean is not reproduced by the model.

We have also compared the amount of biennial variability of the near surface zonal winds in the 35 year run with observed SSTs to two 35 year periods in a 100 year control run with climatological SSTs that repeat the seasonal cycle. Only the simulation with observed SSTs has an organized region of enhanced biennial variability near the equator, suggesting a strong oceanic component to the forcing of the QB.