Climate Forecasting: Oceans, Droughts, Climate Change and Other Tools of the Trade
At IRI’s monthly climate briefing, talk often focuses on the role that El Niño or La Niña play in driving global climate. With the collapse of La Niñalast month, though, IRI’s forecasters now have to rely on different tools to offer forecasts for the coming year. That’s both good and bad news for forecasting skill.
Climate forecasts are very dependent on tropical ocean temperatures. That’s because changes of a degree or two in tropical ocean temperatures can create large changes in evaporation off of the sea surface. Altered evaporation then starts a process that affects rainfall patterns in the region. These changes in tropical atmospheric conditions ultimately propagate beyond the region, affecting other sensitive areas across the globe.
El Niño and La Niña (together called ENSO) represent one of the most important sources of predictability for climate forecasters. What makes them so important is that their effects are, well, so predictable.
On average, ENSO events happen every three to seven years. The corresponding changes in ocean temperature spread in a similar pattern during each event. They cover a large area, reaching across the Pacific like a tongue for thousands of miles.
ENSO is also well documented. Peruvian fishermen have known about it for centuries. Meteorologists have documented ENSO’s effects beyond South America since the 1890s.
In contrast, other climate phenomena are nowhere near as regular, strong or large. “Other non-ENSO related sea-surface temperature anomalies in the tropics will have some impacts, but their patterns and impacts just aren’t as well-defined as those of ENSO,” said Tony Barnston, IRI’s chief forecaster. “They can help IRI’s forecast skill in specific regions, but they’re not nearly as beneficial to global average skill as ENSO events are.”
While globally averaged precipitation forecast skill is highest during ENSO episodes, this rule does not hold for temperature forecasts. For temperature, skill tends to be highest during El Niño and lowest during La Niña.
This is related to the fact that most of the climate models IRI uses for its forecast have atmospheric greenhouse-gas concentrations set to sometime in the mid-1990s. However, greenhouse gases have risen substantially since then and the Earth’s average temperature has risen with them. Because of this, coupled with La Niña’s tendency to cool global (and especially tropical) temperatures, IRI has tended to overforecast cooler-than-normal temperatures during La Niñas, says Barnston. Thus, the subsidence of La Niña means temperature forecast skill is likely to increase in the coming months.
New climate models with flexible greenhouse gas concentrations have recently become available, and IRI is working to incorporate them into its ensemble of models, which should further improve temperature forecasts.
Brad Lyon highlighted another area of predictability that might be of particular interest to people in the southern United States. As mentioned, climate models generally rely on ocean temperatures as the main driver of climate. However, Lyon talked about the potential of dry land to predict temperatures, particularly heat waves.
In past research, Lyon has shown that droughts in the southern U.S. Plains in 1980 and in the West and Midwest in 1998 in part drove local heat waves. The cause of the droughts was tied in part to tropical climate patterns. However, as the dry weather stretched on, the local ground conditions became a bigger determining factor in temperature.
This kind of analysis could be of use to forecasters in Louisiana, Oklahoma and Texas, where a record drought is still gripping the region. More accurate temperature forecasting could have applications frompublic health to agriculture.