California has lost 7% of its forest cover to climate change over the past 25 years.
A model of the Atlantic Meridional Overturning Circulation (AMOC), pioneered by Henry Stommel over 60 years ago, can exhibit realistic cyclic behavior if the role of Arctic salinity is included.
Using machine learning to represent sub-grid processes in weather and climate models holds promise, but also faces challenges. Incorporating physical knowledge can help.
The implications of nature not conforming to statistical assumptions can be devastating; researchers describe why extreme floods may be bigger than we assume.
The new graph convolutional recurrent neural network (GCRNN) will enable water utilities to forecast water use, even if some sensors fail.
More accurate aftershock zones reveal that the rupture areas of megathrust Aleutian–Alaska earthquakes are larger than we thought and partly overlap, in contradiction with the seismic gap hypothesis.
Researchers present a new approach to modeling the stochastic mixing process of convection using a machine learning technique.
Hurricane winds can lead to coast downwelling, which brings warmer surface water near the coast and can contribute to the intensification of the landfalling hurricane.
As cities face health threats from heat and air pollution—both expected to worsen from climate change—researchers pilot a community scientist effort to map air quality and improve urban health.
Adjoint tomography employing 3D wavefield simulations for 72 well recorded regional earthquakes in the western U.S. yields spectacular improvements to waveform fits.