Making climate model forecasts more useful

Abstract

There is considerable potential for seasonal to inter-annual climate forecasts derived from dynamic models of the earth’s climate to be used widely to help improve management of important real-world issues in a variety of different areas (e.g. disaster management, agriculture, water management, health, natural resource management, food security, and insurance). Unfortunately, several factors currently inhibit this potential, e.g. low skill, low awareness, mismatches in what model forecasts can provide and what users need, and the complexity and probabilistic nature of the information provided. Substantial effort around the world is currently directed towards reducing these impediments. For example, climate model development continues behind the scenes, and techniques such as multi-model ensemble forecasting are progressing rapidly. Communication strategies that enable probabilistic information to be communicated more effectively have been developed and exciting developments such as the emergence of the Argo float program have dramatically improved our ability to initialise forecast systems. We can also look forward to greater computing power in the future, which will allow us to increase the resolution of the models used to perform forecasts. Research on the integration of climate forecasts with risk-management tools more useful to managers is also occurring. The great potential for much wider use of climate model forecasting cannot be denied. However, it will only be realised if models continue to be developed further, if climatic variability continues to be closely monitored from the surface, the atmosphere, the ocean, and from space, and if these data are made readily available to the research community.