AI-Driven Oceanography Research

AI finds a missing equation for simulating atmospheric and oceanic turbulence Climate models and weather forecasts simulate turbulent flows spanning scales from thousands of kilometers down to meters. No computer can resolve all of them, so modelers approximate the effect of unresolved small scales on the large-scale dynamics. This approximation is known as a subgrid-scale closure—a model that "closes" the governing equations by filling in what the coarse grid cannot see. Getting closures right matters enormously: their shortcomings are a leading source of uncertainty in climate projections and extreme weather forecasts. For decades, the field has faced a trade-off. One family of closures faithfully reconstructs the small-scale stress patterns but makes simulations blow up. The other keeps simulations stable but oversimplifies the physics—removing too much energy, ignoring backscatter from small to large scales, and underestimating extreme events. Karan Jakhar, Yifei Guan, and Pedram Hassanzadeh break this impasse by changing what equation discovery optimizes for. Previous sparse regression searches consistently landed on the same second-order approximation known since the 1970s, which is accurate but unstable. The key insight: if you also require the discovered equation to reproduce how energy flows between scales—not just match local stress patterns—the algorithm finds something different. Searching 930 candidate terms with this physics-informed dual criterion, Bayesian sparse regression robustly identifies an additional fourth-order term in the Taylor expansion of the subgrid stress (NGM4). NGM4 achieves ~0.99 pattern correlation with reference data, produces stable simulations across four diverse 2D turbulence setups mimicking atmospheric and oceanic dynamics, and accurately captures both bulk statistics and rare extremes. Its coefficients depend only on grid resolution—no tuning for flow regime or Reynolds number—and it needs just 100 training snapshots. The most striking aspect: NGM4 could have been derived analytically decades ago, but because the source of the second-order instability was unclear, higher-order terms were never explored. It took sparse regression guided by the right physics to reveal that the missing piece had been hiding in plain sight. One takeaway that extends well beyond turbulence: the criterion you optimize for determines what you discover. Embedding the right physics into equation discovery can uncover interpretable, generalizable equations that purely data-driven approaches systematically miss. Paper: https://lnkd.in/exGQGaGc #MachineLearning #Turbulence #ClimateModeling #EquationDiscovery #AIforScience #LargeEddySimulation #GeophysicalFluidDynamics #SparseRegression #PhysicsInformedAI #SubgridModeling #DeepLearning #ComputationalPhysics #ExtremeEvents #WeatherPrediction #AIforClimate

AI Unlocks the Ocean: Satellite Data Transformed into Real-Time Current Intelligence A new breakthrough in ocean science is redefining how researchers observe and understand global ocean currents. By applying deep learning to thermal imagery from existing weather satellites, scientists have developed a system that generates detailed, hourly maps of ocean surface movement without requiring new hardware. The method, known as GOFLOW, leverages geostationary satellite data to extract dynamic current patterns at a scale and frequency previously unattainable. Traditional approaches have relied on sparse measurements from drifting buoys or intermittent satellite passes, limiting both coverage and temporal resolution. GOFLOW overcomes these constraints by continuously analyzing thermal signals to infer water motion across vast regions in near real time. This advancement significantly enhances visibility into one of the most critical drivers of Earth’s climate system. Ocean currents regulate global heat distribution, influence weather patterns, and play a central role in carbon cycling by transporting carbon between the atmosphere and the deep ocean. Improved measurement capabilities allow for more accurate climate modeling and forecasting, strengthening the scientific foundation for environmental policy and risk management. Beyond climate science, the operational applications are substantial. High-resolution current data can improve search and rescue missions, optimize maritime navigation, and enhance the tracking of pollutants and marine debris. The ability to monitor ocean dynamics continuously introduces a new level of precision for both civilian and defense-related maritime operations. The broader implication is the emergence of AI as a force multiplier for existing infrastructure. By extracting new intelligence from satellites already in orbit, GOFLOW demonstrates how software innovation can unlock latent value at global scale. This approach reduces the need for costly hardware expansion while accelerating the pace of scientific discovery and operational capability in ocean monitoring. I share daily insights with tens of thousands followers across defense, tech, and policy. If this topic resonates, I invite you to connect and continue the conversation. Keith King https://lnkd.in/gHPvUttw

🌊 Inspired by AI Trends for 2025 and Their Potential for Ocean Data! 🌍🤖 As I read all the exciting predictions about general-purpose AI for 2025, I couldn’t help but wonder: how could these trends be applied to my domain—ocean data—to improve insights and decision-making for sustainable oceans? 🌱🐟 Here’s my take on 8 AI trends and how they could shape the future of working with ocean data: 1️⃣ Agentic AI => AI systems (sets of “agents”) that can make decisions and act independently. Example: Autonomous underwater drones (AUVs) or surface vessels (USVs) that monitor marine ecosystems, detect illegal fishing, or track changes in the ocean. The recent incidents of suspicious cable cuts in the Baltic Sea or the North Sea are also relevant here, as big data analysis of AIS and DAS is being utilized. 2️⃣ Inference Time Compute => Enables LLMs to iterate on their outputs during execution, improving reasoning, accuracy, and context comprehension. Example: Rapid analysis of satellite data for real-time detection of algal blooms, temperature anomalies, CO2 uptake etc 3️⃣ Very Large Models => Extremely powerful AI models with billions (or even trillions) of parameters. Example: Predicting long-term ocean trends, like sea level rise or species migration, by processing massive datasets. 4️⃣ Very Small Models => Compact AI systems that run on small devices like buoys or sensors. Example: Low-power ocean sensors detecting changes in water quality (e.g., pH or pollution levels) and sending alerts in real-time. 5️⃣ More Advanced Use Cases => Expanding AI’s potential to tackle complex real-world problems. Example: Simulating and optimizing marine protected areas (MPAs) to ensure biodiversity and sustainable fisheries. 6️⃣ Near Infinite Memory => AI systems with massive memory to recall and learn from vast datasets. Example: Building a historical database of oceanographic data to identify long-term trends and correlations, like warming oceans and fish stocks. 7️⃣ Human-in-the-Loop Augmentation => AI tools that collaborate with humans to enhance decisions. Example: Assisting marine scientists in interpreting complex ocean data, like predicting coral bleaching events, with actionable insights. 8️⃣ Rise of New Multimodal and Foundation Models trained on Earth Observation Data => AI models that combine multiple data types, such as satellite images and underwater acoustic data. Example: Combining geospatial and hydroacoustic data to monitor illegal fishing, map biodiversity, or assess the impact of ocean noise on marine life. 🌟 What excites me most? These trends could help us better understand and protect our oceans while driving sustainable practices globally. I’d love to hear your thoughts! 💡 How do you think AI could revolutionize ocean data or your own domain? 🌊 Drop your ideas, share related projects, or comment below! Let’s explore the possibilities together! 👇✨

Researchers at Stanford University are revolutionizing how we understand the massive Antarctic ice sheet and its potential impact on global sea levels. Using advanced machine learning techniques, they've revealed hidden physics of ice movements that could reshape climate change predictions. 🌊📉 Key points: 1) Massive impact: Antarctica's ice holds enough water to raise sea levels by a staggering 190 feet! Accurate predictions are crucial for preparing coastal areas worldwide. 🏝️ 2) Machine Learning magic: By analyzing satellite and radar data, researchers are using AI to dive deeper into how ice moves and melts. This approach offers insights that traditional models have missed. 🚀 3) Revealing complexity: Most ice models assume uniform properties, but this study shows that ice behaves differently in various directions. Imagine trying to cut a log along its grain vs. against it. ✂️ 4) Future predictions: With better understanding, these advanced models can help predict how Antarctic ice will evolve as the planet warms, influencing sea-level rise predictions and more! 🔮 #AntarcticIce #AIResearch #ClimateChange #MachineLearning #EarthScience #SeaLevelRise #StanfordUniversity

🐋 Perch 2.0, a bioacoustics model trained primarily on bird calls, is delivering high performance in a surprising place: the ocean. Despite having no underwater audio in its training data,  Perch 2.0, Google Deepmind bioacoustic foundation model, can enable and scale insights for underwater marine ecosystems, particularly for identifying whale vocalizations. This breakthrough enables an "agile modeling" workflow where researchers can build custom classifiers for new marine species using only a few labeled examples. Why this matters: 🐋 Efficiency: Drastically reduces the compute and time needed to turn acoustic data into scientific insights. 🐋 Generalization: Shows that mastering detailed terrestrial sounds (like bird "coos") builds a universal feature set applicable to marine life. 🐋 Open Science: We’re sharing a new end-to-end tutorial to help the cetacean research community leverage these tools via Google Cloud. Scaling our understanding of marine ecosystems is vital for conservation. By building generalist AI, we can accelerate discovery for researchers everywhere. Read the blog by Lauren Harrell and explore the end-to-end demo: goo.gle/4alKOXf Research paper: https://lnkd.in/djFS4VTa

In my recent efforts of expanding research efforts here at Six Senses Kanuhura, I have been working on AI-driven automated detection of fish to measure fish abundance on our surrounding seagrass meadows (a strong indicator of ecosystem health). Early efforts show promising signs of detection, even with fish that are seemingly inconspicuous in the seagrass! This could prove to be a hugely valuable tool that saves hours of time flicking through RUV (Remote Underwater Video) footage, whilst also removing an element of experimenter bias and compiling crucial data to support the protection of these misunderstood ecosystems in the Maldives. For context, I am using a pre-trained model, namely the MegaFishDetector that runs through YOLOv5. I know that for many Marine Biologists working in resorts they can find it hard to dedicate time (or even find the motivation) to analyse data, well these days we have the technology at our fingertips to make this often arduous task much easier, and also kinda fun- think outside the box and be inspired by the possibilities!

🧠 This AI Glider is Mapping the Ocean 100x Faster Than Humans Flying Fish Technologies Pty Ltd are transforming marine monitoring. Their AI-powered underwater gliders are capturing the ocean like never before: 🚤 100x faster than traditional methods 📍 15+ geotagged data points per second 📸 6 million images analysed by machine learning 🐠 Mapping everything from fish to fragile benthic habitats in clear 3D photogrammetry Recent highlights from their mission to the Red Sea: → 350km of continuous reef surveyed → 3.5M images captured in under a month → 200M datapoints generated in just 2 days They provide real-time, high-resolution, high-impact intelligence, powering decisions for ocean conservation and climate resilience, enabling: ⤷ Photorealistic digital twins to track change over time ⤷ AI-driven habitat classification and species detection ⤷ Driverless, boat-based gliders that follow terrain and depth This is the kind of NatureTech that moves us from scattered data to smart, systemic ocean protection. Excited to follow David Kettle and the team at FFT’s progress on this! #OceanTech #MarineScience #NatureTech #AIforNature #BlueCarbon

Modeling Talk: Oct 28, 2025 Climate modeling in the era of AI Laure Zanna, New York University Video Recording: https://lnkd.in/gMxW5sgY Slides and Summary: https://lnkd.in/gPzWBSea Abstract: While AI has been disrupting conventional weather forecasting, we are only beginning to witness the impact of AI on long-term climate simulations. The fidelity and reliability of climate models has been limited by computing capabilities. These limitations lead to inaccurate representations of key processes such as convection, cloud, or mixing or restrict the ensemble size of climate predictions. Therefore, these issues are a significant hurdle in enhancing climate simulations and their predictions. Here, I will discuss a new generation of climate models with AI representations of unresolved ocean physics, learned from high-fidelity simulations, and their impact on reducing biases in climate simulations. The simulations are performed with operational ocean model components. I will further demonstrate the potential of AI to accelerate climate predictions and increase their reliability through the generation of fully AI-driven emulators, which can reproduce decades of climate model output in seconds with high accuracy. Bio: Laure Zanna is a physical oceanographer and climate physicist in the Department of Mathematics at the Courant Institute and the Center for Data Science, NYU. She holds the Joseph B. Keller and Herbert B. Keller Professorship in Applied Mathematics. Her research focuses on understanding, simulating and predicting the role of the ocean in climate on local and global scales. She combines theory, numerical simulations, statistics, and machine learning to tackle a wide range of problems in fluid dynamics and climate, including turbulence, multiscale modeling, ocean heat and carbon uptake, and sea level. Since 2020, she is leading M²LInES, an international collaboration sponsored by Schmidt Sciences dedicated to improving climate models using scientific machine learning. In 2020, Prof Zanna received the Nicholas P. Fofonoff Award from the American Meteorological Society “for exceptional creativity in the development and application of new concepts in ocean and climate dynamics”, and was the 2022 WHOI Geophysical Fluid Dynamics principal lecturer. #modeling #simulation #ai #ml #research #datascience #climatechange #climatemodeling #oceanmodeling #weathermodels #neuralemulator

AI is redesigning how we explore the oceans. Literally! MIT CSAIL Alliances Alumni Diplomat Program (in collaboration with University of Wisconsin-Madison) is reshaping the future of autonomous underwater gliders, and it's all powered by AI. They developed an AI‑driven design pipeline that generates new, efficient shapes for autonomous underwater gliders. Started with conventional 3D models (submarines, manta rays, whales, etc.), enclosed them in “deformation cages,” and used physics simulations to generate and evaluate thousands of variants. A neural network predicted lift-to-drag ratios at different angles, guiding iterative refinement. 𝗧𝗵𝗲 𝗿𝗲𝘀𝘂𝗹𝘁: Two innovative gliders were 3D printed one with wings, another with fins. 𝗪𝗵𝘆 𝗶𝘁 𝗺𝗮𝘁𝘁𝗲𝗿𝘀: These next-gen gliders will allow researchers to gather richer data like temperature, salinity, and currents across larger distances. A game-changer for climate science, ocean mapping, and marine research. Full story from MIT CSAIL: https://lnkd.in/df4zAvxs 𝗛𝗼𝘄 𝗱𝗼 𝘆𝗼𝘂 𝘀𝗲𝗲 𝗔𝗜 𝗶𝗺𝗽𝗮𝗰𝘁𝗶𝗻𝗴 𝘆𝗼𝘂𝗿 𝗶𝗻𝗱𝘂𝘀𝘁𝗿𝘆’𝘀 𝗱𝗲𝘀𝗶𝗴𝗻 𝗼𝗿 𝗿𝗲𝘀𝗲𝗮𝗿𝗰𝗵 𝘄𝗼𝗿𝗸𝗳𝗹𝗼𝘄𝘀? 📌 𝗣𝗦: Follow for more information and latest tech insights. Let’s connect https://lnkd.in/d7FR8yK2 and discuss how we can collaborate! #AIinScience #OceanTech #MITCSAIL #SustainableInnovation #ArtificialIntelligence

Clouds block satellites. AI fills in the gaps. And, suddenly, we’re better at predicting typhoons. A new system called PARAN is helping scientists see the ocean more clearly—literally. When clouds get in the way of satellite readings, this AI model steps in to reconstruct sea surface temperature data in real time. That means better insight into how heat moves between the ocean and atmosphere… and way better forecasting of heatwaves, storms, and marine disasters. It’s a smart blend of AI and physics, and a huge leap for climate resilience. Because when you can’t see the problem, you definitely can’t solve it. Science, meet sharp vision. #ClimateTech #AI