A new paper in Remote Sensing is an advancement on a previous study that demonstrated a model that uses generative AI to combine satellite data and physics-based simulations to forecast a wildfire’s path, intensity and growth rate.The original system relied on data from VIIRS, a polar-orbiting satellite that detects heat signatures with relatively high spatial resolution (within a few hundred meters) but revisits the same location only twice a day. The result is precise but intermittent information; a sequence of snapshots separated by long intervals in which the fire continues to evolve. Now, the new paper published in the journal Remote Sensing presents an advancement of the model. The goal is to reduce uncertainty about where and when a fire started, crucial information for predicting its spread. 

Spatially distributed prediction of streamflow and nitrogen (N) export dynamics is essential for precision management of agricultural watersheds. To address this need, a team of researchers led by the Center for Advanced Bioenergy and Bioproducts Innovation (CABBI) propose HydroGraphNet, a knowledge-guided graph machine learning framework integrating process-based knowledge and explicit spatial learning into temporal modeling. HydroGraphNet offers a generalizable framework for distributed modeling to support spatially targeted water quality management in data-scarce watersheds.

Scientists find an exception in the Montreal Protocol for the use of ozone-depleting feedstocks could set the recovery of the ozone layer back seven years.

Restoring both walking and sensation to patients with paraplegia is an ambitious goal—but a team of researchers from the Keck School of Medicine of USC, the University of California, Irvine (UCI) and the California Institute of Technology (Caltech) is now one step closer. With $8 million in funding from the highly competitive National Science Foundation CyberPhysical Systems program, the team is building a fully implantable brain-computer interface (BCI) that allows patients to use their thoughts to control wearable robotic legs, known as a robotic exoskeleton. The system is designed to help patients walk while also restoring the sensation of walking. In the first full test, the BCI was about 92% accurate at both reading step signals from the brain and delivering artificial walking sensation. Existing brain-computer interfaces that restore walking send signals in just one direction, from brain to device. The team’s early proof-of-concept study, done in a patient with epilepsy who had electrodes implanted as part of her medical care, shows it is possible to build a bidirectional, or two-way, system. During the demonstration, the patient sat on her hospital bed with the device by her side (future versions will be small enough to implant inside the body), while one of the researchers wore the robot exoskeleton. When the patient mimed taking a step, the device signaled the exoskeleton, sending the researcher on a walk around the intensive care unit. The system correctly detected brain signals indicating the intent to walk about 92% of the time. The demonstration helped the researchers earn an Investigational Device Exemption from the U.S. Food and Drug Administration, which allows them to test the device in a clinical trial for patients with paraplegia. They aim to implant electrodes for 30 days as a time, using that window to test and refine the system’s capabilities.