electrical engineering

Scientists harness AI to reveal forces behind glacier surges

Charles Ferrer — Thu, 05 Mar 2026 22:12:42 +0000

Scientists harness AI to reveal forces behind glacier surges Charles Ferrer Thu, 03/05/2026 - 15:12

Charles Ferrer

Ute Herzfeld (PI), Harald Sandal (pilot), Gustav Svanstroem (helicopter technician) and Matthew Lawson (research assistant) during the Negribreen Glacier System Airborne Geophysical campaign (Photo Credit: Thomas Trantow).

Glaciers are constantly changing and reshaping the Earth’s surface.

��Ҵ�ý�ƽ�� researchers have developed a new machine learning tool to better understand how Arctic glaciers suddenly accelerate or “surge”.

The team, led by Ute Herzfeld, a research professor in the Department of Electrical, Computer and Energy Engineering, created an open-source cyberinfrastructure called GEOCLASS-image, designed to decode the physical processes behind glacier motion using high-resolution satellite imagery and machine learning.

Glacier surges are sudden bursts of movement in otherwise slow-flowing ice.

Normally, glaciers move at a steady pace, but during a rare “surge”, that rate can accelerate up to 200 times faster than usual. The ice fractures into deep crevasses and pushes large volumes of ice toward the ocean. These dramatic events provide scientists with new insight into the unpredictable drivers of sea-level rise.

“Most deep machine learning systems don’t know what to look for in images,” said Herzfeld, who is also the director of the Geomathematics, Remote Sensing and Cryospheric Sciences Laboratory. “We have built a system that understands the physics of ice deformation, so the classifications actually mean something.”

Understanding how a glacier surges

Unlike traditional artificial intelligence systems that often struggle to interpret complex natural phenomena, the team created a new neural network approach—VarioCNN—to better understand glacial acceleration.

“Surging glaciers are one of the deep uncertainties in sea-level rise projections,” Herzfeld said. “They can move much faster than normal and current earth system models do not yet have the ability to account for them.”

To tackle this problem, Herzfeld and her team merged two powerful approaches: a deep convolutional neural network (CNN), common in the field of computer science and remote sensing and a physics-informed neural network model that captures how crevasses in the ice form, widen and intersect during motion.

“Think of neural networks as Lego blocks,” Herzfeld said. “We’ve taken some from physically informed models, some from deep learning and built a new kind of AI that’s meaningful.”

Putting AI to the test

The team tested their approach on a real-world event: the unexpected 2016 surge of Negribreen, a glacier located in the Arctic archipelago of Svalbard a 1,000 km south of the North Pole.

This isn’t just another AI model but one that understands the physics of glacial acceleration.
~Ute Herzfeld

Using Maxar WorldView satellite imagery collected in 2016-2018, the researchers tracked subtle changes across the glacier’s surface with remarkable detail.

They discovered that crevasse patterns, which change dramatically during a surge, hold information about surge dynamics that can be retrieved using their neural network approach.

One-dimensional crevasses appeared at the leading edge of the surge, while deeper within the surge area, complex patterns tell the story of the transformation and deformation of the ice, which can be of use in numerical modeling of the glacial acceleration.

Shear, a type of deformation that plays a key role in glacial acceleration, is easily misclassified in deep learning, but correctly identified using VarioCNN.

With their new VarioCNN model, they classified different types of crevasses from satellite images and used those patterns to interpret how the glacier moved and changed.

Results of the classification were then used to understand how the surge expanded and affected the entire Negribreen glacier system. Ultimately, ice mass equivalent to 1% of global annual sea-level rise transferred to the ocean.

Published in Remote Sensing, their results demonstrated how integrating physical knowledge into a neural network model, carried out at the computational level, can advance machine learning and glaciological understanding of glacier surges. The paper was selected as the cover story of Remote Sensing receiving record downloads during the first two weeks after publication.

Student Connor Meyers setting up a GPS station at the edge of Negribreen (Photo Credit: Ute Herzfeld).

“The problem of task-oriented machine learning is especially intriguing to me,” said Silas Twickler (Phys’25) who was a research assistant on the project. “While simply applying pre-existing neural networks may be sufficient for certain applications, the augmentation of these networks can allow for a drastic improvement in machine learning.”

AI for the geosciences

A major hurdle in applying machine learning to studying glaciers is the limited amount of labeled data. To overcome this, Herzfeld’s team developed a way that allows scientists to gradually refine the model using a relatively small number of hand-labeled satellite images.

VarioCNN was trained on just a few thousand of examples, far fewer than the 100,000 images than typical deep learning models require. Due to its modular design, the GEOCLASS cyberinfrastructure can be adapted to study other geophysical processes and potentially surfaces of other planets.

“Our tool is not just for glaciologists, but for anyone working with remote sensing and physical systems,” Herzfeld said. “Ultimately, we hope to give scientists better tools to understand how the Earth is changing.”

This research was funded by the National Science Foundation Office of Advanced Cyberinfrastructure and NASA Earth Sciences Division.

Off

Traditional

White

Negribreen glacier during an ice surge in 2017 (Credit: Ute Herzfeld).

Researchers build ultra-efficient optical sensors shrinking light to a chip

Charles Ferrer — Mon, 23 Feb 2026 16:37:42 +0000

Researchers build ultra-efficient optical sensors shrinking light to a chip Charles Ferrer Mon, 02/23/2026 - 09:37

Charles Ferrer

Lu at the new electron beam lithography system used to develop microresonators at COSINC.

��Ҵ�ý�ƽ�� researchers have built high performing optical microresonators opening the door for new sensor technologies.

At its simplest form, a microresonator is a tiny device that can trap light and build up its intensity.

Once the intensity is high enough, researchers can perform unique light operations.

“Our work is about using less optical power with these resonators for future uses,” said Bright Lu, a fourth-year doctoral student in electrical and computer engineering and a lead author on the study. “One day these microresonators can be adapted for a wide range of sensors from navigation to identifying chemicals.”

For this endeavor published in Applied Physics Letters, the team focused on ‘racetrack’ resonators, named for their elongated shape that resembles a running track.

Specifically, researchers used ‘Euler curves’ — a type of smooth curve also found in road and railway design. Just as cars can’t make sharp right-angle turns in motion, light can not be forced into abrupt bends.

“These racetrack curves minimize bending loss,” said Won Park, Sheppard Professor of Electrical Engineering, a co-advisor on the study. “Our design choice was a key innovation of this project.”

By guiding light smoothly through the resonator, they dramatically reduced light loss, allowing photons to circulate longer and interact more strongly inside the device.

If too much light is lost, Lu says, high light intensities can’t be achieved for these microresonators to operate at the needed performance.

Made in Colorado

Incredibly small in size, the microresonators were built using the Colorado Shared Instrumentation in Nanofabrication and Characterization (COSINC) clean room’s new electron beam lithography system.

The facility provides a highly-controlled environment required to work at the microscopic scales that can lead to reliable device performance.

Optical waveguide microresonators on a chip created in this effort, which are ten times thinner than human hair.

Many optical and photonic devices are smaller than the width of a piece of paper, meaning even tiny dust particles or surface imperfections can disrupt how light travels through a material.

“Traditional lithography uses photons and is fundamentally limited by the wavelength of light,” Lu said. “However, electron beam lithography has no such constraint. With electrons, we can realize our structures with sub-nanometer resolution, which is critical for our microresonators.”

For Lu, the hands-on fabrication process was a fulfilling aspect of the project.

“Clean rooms are just cool and you’re working with these massive, precise machines and then you get to see images of structures you made only microns wide. Turning a thin film of glass into a working optical circuit is really satisfying.”

A key success from the work was the ability of the researchers to use chalcogenides, a broad term encompassing a family of specialized semiconductor glasses.

“These chalcogenides are excellent materials for photonics because of their high transparency and nonlinearity,” said Park. “Our work represents one of the best performing devices using chalcogenides, if not the best.”

Chalcogenides were helpful since they have strong transparency for light to pass through the device at high intensities needed for microresonators.

However, the materials are not easy to process for the device, so there’s a balancing act to tread.

“Chalcogenides are difficult, but rewarding materials to operate for photonic nonlinear devices,” said Professor Juilet Gopinath, who has worked on this project with Park for more than ten years. “Our results showed that minimizing the bend loss enables ultra-low loss devices comparable to state-of-the-art in other materials platforms.”

Measuring light at the microscale

Erikson with the optical setup for capturing data measuring absorption and thermal effects.

Once fabricated, the microresonators were handed off for testing, work led by James Erikson, a physics PhD student specializing in laser-based measurements. He carefully aligned lasers with microscopic waveguides, coupling light into and out of the device while monitoring how it behaved inside.

They looked for ‘dips’ within the data in transmitted light that indicate resonance as photons get trapped. By analyzing the shape of those dips, they were able to extract properties like absorption and thermal effects.

“The most obvious indicator of device quality is the shape of the resonances and we want them to be deep and narrow, like a needle piercing through the signal background,” said Erikson. “We’ve been chasing this kind of resonator for a long time, and when we saw the sharp resonances on this new device we knew right away that we’d finally cracked the code.”

Erikson added, to make a good device you need to know how much light will be absorbed versus transmitted. Thermal effects become important when adding laser power as you run the risk of damaging the device.

“The way most materials interact with light also changes depending on the temperature of the material,” said Erikson, “so as a device heats up its properties can change and cause it to work differently.”

In the future, the microresonators could be used for compact microlasers, advanced chemical and biological sensors and even tools for quantum metrology and networking.

“Many photonic components from lasers, modulators and detectors are being developed and microresonators like ours will help tie all of those pieces together,” said Lu. “Eventually, the goal is to build something you could hand to a manufacturer and create hundreds of thousands of them.”

��Ҵ�ý�ƽ�� researchers have built high performing optical microresonators opening the door for new sensor technologies. In the future, the microresonators could be used for compact microlasers, advanced chemical and biological sensors and even tools for quantum metrology and networking.

Off

Traditional

White

The fabrication cleanroom facility provides state-of-the-art instrumentation including lithography, thin-film deposition and among others. (Credit: COSINC)

An earthquake on a chip: New tech generates tiny waves, could make smartphones smaller, faster

Charles Ferrer — Wed, 14 Jan 2026 21:32:04 +0000

An earthquake on a chip: New tech generates tiny waves, could make smartphones smaller, faster Charles Ferrer Wed, 01/14/2026 - 14:32

A team of engineers has developed a new device that works like a laser but, instead of light, generates incredibly small vibrations called surface acoustic waves.

Off

Traditional

White