No small feat is assessing the risk of extreme weather in our communities. Policymakers take on this task by relying heavily on global climate models, a challenging thing to do. These models have their limitations though they are strong. Regarding what Boston might experience, they cannot give precise details while providing a broader overview of the future climate conditions that may prevail in entire regions.
To fill in these gaps, policymakers often combine these broad-scale models with more detailed, finer-resolution models. By the changes in climate, such fine-tuned models help estimate the likelihood of events like flooding hitting a city like Boston. However, there is a catch; these assessments are only as good as the forecasts made from initial global models.
MIT professor Themistoklis Sapsis and his team have been working on this problem. Their approach involves using machine learning and dynamical systems theory to refine predictions produced by these coarser climate models. In other words, researchers have altered simulations so that they become closer approximations for larger scale real world settings. This correction not only improves predictions about specific locations but also enriches our understanding of how extreme weather patterns will change over time during next decades.
The broad impacts of their work were emphasized by Sapsis, underlining its ability to assist in every aspect from biodiversity and food security to economic planning. It is a game changer having abilities to predict with accuracy how specific regions will have their extreme weather change.
Their results are shown to be promising as they publish them in the Journal of Advances in Modeling Earth Systems. This led them to refine simulations from one of the most used climate models making it more compatible with observed climate patterns hence improving future projections.
This should be seen as a novel approach owing to their refinement of the model’s output through data-driven algorithms rather than rewriting its underlying equations that might be complex and unstable. Such an innovative step has the potential for bettering our knowledge on climate dynamics, improve chances for future advancements.
This study has wider implications than just academia. As the world inevitably gets influenced by climatic changes, accurate predictions for extreme weather events are very important. By fine-tuning our models we can better prepare ourselves for what lies ahead.
The research would not have been possible without the support of the US Defense Advanced Research Projects Agency (DARPA). This was a plain manifestation of how crucial it is for research institutions to partner with governments in order to solve complex global issues.
