NVIDIA collaborates with the University of California's climate simulation research team to prevent future flood disasters.

NVIDIA's press release reveals the latest academic research indicating that the probability of sea level rise due to climate change is gradually increasing. In the next 30 years, the likelihood of flooding in coastal areas of the United States has reached 26%. In the face of increasingly severe risks from extreme weather events, Professor Michael Beck's research team at the University of California, Santa Cruz, is using NVIDIA to develop more advanced simulation tools to predict and mitigate the impacts of flooding on coastal communities. Professor Beck and his team have created detailed visualizations of coastal flooding to help government agencies, non-governmental organizations, and Financial Institutions better understand the risks posed by flooding, while also showcasing solutions to mitigate onshore damage.

CUDA-X Accelerated Flood Simulation Operations

This new study is led by Professor Michael Beck, director of the “Coastal Climate Resilience Center” at the University of California, Santa Cruz, focusing on simulating the benefits of natural resources such as coral reefs and mangroves in flood prevention. By utilizing CUDA-X software and RTX GPU accelerators, the team accelerated the acquisition of flood simulation operation reports, enhancing model computation efficiency and accuracy with the NVIDIA RTX 6000 Ada. To expedite the simulation, the center employs NVIDIA CUDA-X software, including the cuPyNumeric library and nvfortran compiler.

Data-driven videos present the real situation under extreme weather conditions.

Beck stated that visualizing the research report is a key way to explain the crisis of flooding to the public. Many solutions are costly, and only by allowing the public to intuitively see the impact and benefits of floods can actual action be promoted. In practice, the research team used the SPhinx simulation tool and Unreal Engine 5 rendering engine to create realistic data-driven animations, presenting scenarios such as “a once-in-a-century storm hitting Santa Cruz, California.” These simulated scenarios not only provide references for the government and NGOs but also offer decision-making basis for the insurance industry and investment institutions. Senior modeler David Gutiérrez indicated that high-performance GPU computing allows the team to run multiple simulation models simultaneously, conduct more detailed sensitivity analyses, and reduce errors in model assumptions and parameter settings, with workload computing speed increased by 3 to 4 times, and some even faster.

Insure coral reefs and treat nature as an asset for protection.

The research team is also promoting an innovative concept of viewing coral reefs as “natural infrastructure” and protecting them through insurance mechanisms. This new insurance scheme has been implemented in the southern part of Cancun, Mexico, where local governments and institutions such as the World Bank have initiated compensation for hurricane wind speeds exceeding 100 knots, aimed at restoring damaged coral reefs and ensuring coastal communities can recover quickly after wind disasters. This new type of insurance is not only a tool for risk transfer but also an investment in post-disaster recovery and reconstruction. After flooding, it is necessary to replant damaged corals and clear beach debris and waste, all of which require financial support.

CoSMoS Project: Establishing Decision-Making Tools for Coastal Areas Across the United States

CoSMoS collaborates with the United States Geological Survey (USGS) to enhance the Coastal Storm Modeling System CoSMoS. This system not only predicts risks but also evaluates the effectiveness of various disaster mitigation strategies, including dune construction, seawall structures, and nature-based solutions. Simulating and assessing the effectiveness of these mitigation plans requires substantial computing resources, and GPUs are an essential key.

The team is currently expanding the application of this technology to produce high-resolution flood risk maps for global small island developing states (SIDS) including Tonga, Trinidad and Tobago, which are expected to be presented at the COP30 climate conference in November this year. Beck stated that the team aims not only to present climate issues but also to simulate the feasibility and effectiveness of solutions.

Introduction to NVIDIA CUDA-X:

Developers, researchers, and inventors from various industries use GPU programming to accelerate applications, which require a powerful programming environment that includes highly optimized, domain-specific microservices and libraries. NVIDIA CUDA-X is built on CUDA®, integrating microservices, libraries, tools, and technologies for building applications, data processing, AI, and high-performance computing (HPC).

This article discusses NVIDIA's collaboration with the University of California's climate simulation research team to prevent future flood disasters, first published in Chain News ABMedia.