New Academy Report Reviews the State of Climate Forecasting

Published: September 15th, 2010

Heidi Cullen

It takes just one person to make a prediction, but it takes an entire community to make a forecast. That’s what Lisa Goddard, a scientist at the International Research Institute for Climate & Society (IRI) and a member of the Committee on Assessment of Intraseasonal to Interannual Climate Prediction and Predictability recently explained during a public lecture. Goddard is one of the co-authors of a new report from the National Research Council (NRC) that examines current climate forecasting capabilities such as seasonal hurricane and drought forecasts.

Intraseasonal to interannual (ISI) climate forecasts make up the middle ground between what you might think of as a classic weather forecast — one that typically looks out over the next several days — and global climate change simulations that look at trends in long-term climate out to the end of the 21^st century. More specifically, ISI forecasts focus on weekly, monthly, seasonal, and annual timescales. These timescales are important for planning in sectors like water resource management, agriculture, and energy.

However, making accurate forecasts on these timescales is a challenge in part because they require high-quality global observations, state-of-the-art computer models, and a deep understanding of sources of predictability within the climate system. The NRC report examines current capabilities for making ISI predictions and recommends opportunities for future improvement.

The report does a nice job of explaining the sources of predictability. Such sources include processes at work within the atmosphere, ocean, and land that affect the state of the climate on weekly, monthly, seasonal and annual timescales. According to the report, “the sources of predictability are measured, represented, and simulated by ISI forecast systems through an assemblage of “building blocks,” namely observational systems, statistical and dynamical models, and data assimilation schemes.” There are three interrelated categories of predictability that exist within the climate system. The first is related to particular variables that exhibit inertia or "memory," such as ocean heat content, in which unusual conditions can take relatively long periods of time (days to years) to decay.

The second source of predictability is related to patterns of variability or feedbacks, like the El Niño-Southern Oscillation (ENSO), where anomalous conditions in the tropical Pacific Ocean influence seasonal climate in the mid-latitudes around the globe. The third source of predictability is due to external factors or "forcing." Volcanic eruptions, changes in solar activity, and the accumulation of greenhouse gases in the atmosphere are all examples of external forcing. These events or processes can affect the climate on ISI timescales in predictable ways that can be used to make climate predictions. Ultimately, a good climate forecast comes directly from our understanding of these specific sources of predictability.

Overall, the report concludes that there is no silver bullet that will result in a magic leap forward in the skill of ISI forecasts. Instead, the report makes it clear that operational forecast centers around the world (including the National Centers for Environmental Prediction (NCEP), the European Centre for Medium Range Forecasting (ECMWF), IRI (Goddard's home institute), as well as many universities and research laboratories) will need to take a steady, thoughtful, and long-term approach to improving observational capabilities, computer models, and data assimilation systems.

The report also demonstrates that improved observational systems not only help increase our understanding of the complex workings of the climate system, and thus fine tune the models, but also lead to better initial conditions for the forecasts, and can improve skill. Most recent forecast improvements have come from both improved models and improved ways of feeding observations into the models for their initial conditions — in almost equal parts. In other words, as Goddard says, it will take a community to build the routine, high-quality forecasts society wants and needs.