Application of machine learning can optimize hurricane track forecast

When a hurricane approaches, providing a few extra hours’ notice can be the difference between life and death. Now, Penn State researchers report that applying a machine learning technique to a group of possible storm paths could help meteorologists provide more accurate medium-term forecasts and issue timely warnings to communities in the path of these potentially deadly storms.

Read more

Ocean acidification prediction now possible years in advance

CU Boulder researchers have developed a method that could enable scientists to accurately forecast ocean acidity up to five years in advance. This would enable fisheries and communities that depend on seafood negatively affected by ocean acidification to adapt to changing conditions in real time, improving economic and food security in the next few decades.

Read more

Energy Regulation Rollbacks Threaten Progress Against Harmful Ozone

The fight against harmful ozone is under legal threat. Air quality and carbon emissions regulations are currently in limbo in courts and congress, from core legislation from the 1970s to rules from the last U.S. administration. This study models the future losses in the fight to drive down respiratory-damaging, ground-level ozone if the regulations go away.

Read more