Dr. Jiabi Du sits at a supercomputer screen at Texas A&M University at Galveston, watching 400 processors hum through a one-year ocean simulation in just 24 hours—a digital window into the hidden architecture of the Gulf of Mexico. For years, Du, an assistant professor of marine and coastal environmental science, has been building high-resolution 3D ocean models that can replicate water levels, currents, waves, temperature, and salinity with stunning accuracy. These models are transforming how scientists understand a landscape that shapes everything from marine life to disaster recovery.
The Gulf is a dynamic, uncontrolled system where weather, seasons, and geography never stop shifting. Understanding how water actually moves through this complexity matters enormously—it touches marine biology, coastal science, environmental policy, and crisis response. Yet real water movement has always been maddeningly difficult to predict. Du's approach sidesteps the problem by creating customizable digital twins of the Gulf's physics, tools that can isolate the impact of individual forces like tides, wind, or hurricanes and measure how they ripple through the whole system.
The practical payoff has already arrived. During Hurricane Harvey's extreme precipitation, Du's simulations revealed that the prolonged drop in salinity afterward may have killed every oyster on affected reefs—a 100 percent mortality rate that might have remained invisible without the model. More recently, Du used his framework to track microplastics produced in Texas as they distribute through the Gulf, discovering they can travel all the way to Mexican shorelines, while microplastics from Mexico have far less chance of reaching Texas. Seeing particle transport at this scale would be nearly impossible without visualization.
His work caught the attention of the Texas Water Development Board, which asked Du to help develop a hydrodynamic model for the entire Texas coast. The result is BAYCAST, delivered in 2025 and now publicly accessible, providing a continuous five-day forecast of water conditions including currents, salinity, level, and temperature. The model will support the Texas General Land Office's oil spill response by predicting how spilled oil moves through coastal waters. Du is already planning the next phase, incorporating water quality, chemical processes, and sediments—work he describes as a specialty only his group can deliver for government.
But Du is not stopping there. He is leveraging more than 20 years of hydrodynamic simulation data to develop a coupled physical-biological model focused on how oyster larvae move through Galveston Bay. That work will guide oyster restoration, aquaculture decisions, and farm planning across the region. Simultaneously, recognizing that even the best models have blind spots, Du is building an autonomous robot that can venture into storm conditions and collect live readings of floodwaters—measuring velocity, salinity, and temperature when digital simulation alone cannot capture ground truth. By blending supercomputer models with real-world data, he is chasing a clearer, more honest picture of how the Gulf actually moves.