Once every few years, the normally cool waters of the eastern tropical Pacific become unusually warm in a phenomenon known as El Niño. During El Niño years, disruptive weather patterns often occur over wide regions of the globe, including North America. For example, flooding rains can strike California, producing coastal erosion, mudslides, and crop damage, while droughts may occur in Australia and other regions. Along the South American coast, the local fishing industry is severely disrupted by the unusually warm ocean waters.

GFDL and Princeton University maintain active collaborative efforts aimed at understanding and predicting El Niño. George Philander of Princeton University is a pioneer in the field of El Niño simulation and theory. His work with GFDL's Ron Pacanowski established the viability of simulating El Niño using coupled ocean-atmosphere models and set the stage for attempts to use such models for El Niño prediction. At GFDL, Kikuro Miyakoda and Tony Rosati have led efforts to build a system to predict both El Niño and its impact on large-scale weather patterns. Miyakoda has long been a leader in efforts to extend the limits of useful weather forecasts. Many of his research group's earlier modeling innovations had been incorporated into operational weather forecasting models at the National Meteorological Center, leading to significant improvements in the 3-5 day forecasts now available to the public.

Modeled ocean temperature and surface current distribution over the tropical Pacific Ocean region obtained from a GFDL coupled ocean-atmosphere model used to predict El Niño/Southern Oscillation (ENSO). Lower (blue surface): three-dimensional depiction of the surface on which the ocean temperature is 20°C. Undulations of this surface can be used to monitor ocean heat content changes associated with ENSO. During the cold phase of ENSO, this surface is deep in the warmer western tropical Pacific but rises toward the sea surface in the cooler eastern equatorial Pacific, as in the example shown. Upper: corresponding distribution of sea surface temperature and surface ocean currents. Gray regions depict land areas of New Guinea and northern Australia.

Using coupled ocean-atmosphere models developed at GFDL and other research centers, scientists now have begun to make physically based predictions of El Niño conditions, with lead times of a year or more. These models have the potential to predict El Niño's effects on weather patterns over North America and other regions far removed from the tropical Pacific. Farmers, energy planners, and water resource managers are examples of those who would benefit from improved extended-range weather and short-term climate predictions.

Model skill in the prediction of eastern tropical Pacific sea surface temperature anomalies for different forecast lead times. Higher values indicate greater skill. Based on an ensemble of 14 retrospective forecasts for 1982-1988 using a GFDL coupled ocean-atmosphere model. Forecast Method 1 uses both surface and sub-surface ocean data for initialization, whereas Forecast Method 2 uses only surface data. The results show the crucial importance of subsurface data for the success of the forecasts with this prediction system. [Source: Anthony Rosati, et al., Monthly Weather Review, submitted.]

Observed and modeled atmospheric circulation anomalies over the Pacific Ocean and North America during six El Niño winters. The model simulations illustrate how long-range (~1 year) predictions of tropical Pacific sea surface temperature may ultimately lead to improved long-range predictions of seasonal weather anomalies, even in regions remote from the tropical Pacific such as the continental United States. Shown are composite 500mb geopotential height anomalies in meters. The three bottom diagrams are from atmospheric model experiments in which observed sea surface temperatures have been specified for different domains. [Source: Ngar-Cheung Lau and Mary Jo Nath, Journal of Climate, August 1994.]

GFDL'S MODULAR OCEAN MODEL

Detailed predictions of El Niño are made possible by the "Modular Ocean Model," a product of extensive research at GFDL. Through the efforts and foresight of Mike Cox, Keith Dixon, Ron Pacanowski, and Tony Rosati at GFDL, this ocean model is now a shared resource, used by hundreds of researchers worldwide. The model has become the mainstay of the oceanographic modeling community, not just in new operational El Niño predictions at the National Centers for Environmental Prediction, but for global climate change research and other applications as well.