You may have noticed that weather forecasts on your cellphone are reasonably accurate for the next three to five days, but very iffy eight to 10 days out. This may not matter much to you, except when you are planning a camping or sailing trip, or an outdoor event, like a wedding. But it matters a lot to a farmer who has to decide on optimal timing for planting, harvesting, and irrigation, or an electric utility manager who needs to plan for the expected supply of solar or wind energy to the grid. It matters to airport, train, and truck operators who worry about major weather events impeding their functions; to those engaged in preparation for and prevention of weather-related disasters; and to insurers who need accurate historical and projected weather and climate data to assess and price risks. Good climate measurements and projections are key to understanding the expected impacts of climate change, such as record-breaking storms, forest fires, droughts, heat waves, and floods and what steps are needed to adapt appropriately.

So, how do we improve weather forecasts—e.g., to get future 10-day forecasts to be as accurate as today’s five-day forecasts—and how do we reach the highest possible prediction accuracy for future trends in global, regional, and local climate change?

The global community has been working to improve the different pieces of this value chain with considerable focus on improved data gathering, especially with the expansion of satellite-based and ocean-based observation, better sharing of information, improvements in weather modeling, better dissemination of forecasts, including via strengthened early warning systems, and better disaster preparedness.

The benefits from improved weather data collection and sharing are significant. The World Bank estimates that the expected improvements in weather and climate prediction models could bring about $5 billion annual benefits in addition to the current $160 billion annual benefits from weather forecasts and climate prediction with a cost-benefit ratio of 1:26.

Read the full article about improving forecasting by Johannes F. Linn, Anthony Rea, Markus Repnik, and Laura Tuck at Brookings.