Simulating Historical Precipitation Variability Associated With ENSO

It has been suggested that warmer temperatures associated with increasing greenhouse gas emissions will increase precipitation intensity and result in a more vigorous hydrologic cycle (IPCC, 1996; p. 7 and references therein). Given the potential social, environmental, and economic consequences of such a scenario, developing a better understanding of the mechanisms which force changes in the hydrologic cycle and assessing the current skill of climate models in predicting such changes are of obvious importance. Climate variability associated with the El Nino/Southern Oscillation (ENSO) provides coherent signatures in surface temperature and related hydrologic quantities that, while not necessarily a surrogate for global warming, do serve as a useful testbed for assessing climate model performance. Furthermore, while considerable attention has been devoted to describing the spatial pattern and regional changes of precipitation associated with ENSO, little is known regarding how ENSO modifies the net precipitation over the tropics as a whole, or to understanding how the variations in precipitation relate to changes in other aspects of the hydrologic cycle.

In this study, an atmospheric general circulation model (GCM) is integrated with observed sea surface temperatures for the period 1946-1988. The ability of the model to reproduce observed patterns of precipitation variability associated with the ENSO is then assessed by comparing the model simulations to several decades of rain-gauge measurements and nearly two decades of satellite observations. Finally, the interannual coupling between sea surface temperatures, column integrated water vapor, and precipitation is examined to characterize the role of ENSO in forcing changes in the intensity of the tropical hydrologic cycle.

When forced with the observed SSTs, the GCM successfully reproduces the ENSO-driven patterns of precipitation variability observed from both rain gauge measurements over land and satellite observations over oceans. This close agreement provides confidence in the ability of the GCM to predict variations in precipitation resulting from changing SST patterns, such as those which may occur from increasing greenhouse gases. However, the changes in tropical-mean precipitation simulated by the GCM differ substantially from those observed. In particular, the satellite observations indicate a distinct increase (decrease) in precipitation during the warm (cold) phase of ENSO, whereas the GCM simulates very little interannual variability in precipitation. The observed increase in tropical-mean precipitation is also shown to coincide with a decrease in the mean residence time of water vapor in the tropics, suggesting an enhanced cycling of water vapor and strengthened hydrologic cycle during a warm, El Nino. The GCM simulations, on the other hand, imply a slower cycling of water during El Nino and a weaker hydrologic cycle. The implications of these results for understanding past and future changes in precipitation and the hydrologic cycle will be discussed at the conference.