Emerging Topics
Tipping Points, Unprecedented Events and Irreversibilities in the Climate System
High resolution paleoclimatic records, supported by Earth system modelling studies, have demonstrated that perturbations in the Earth system can lead to abrupt, non-linear, often irreversible changes and impacts at local, regional, and global scales. From a local point of view, even small changes in climate conditions may result in high impacts that threaten the livelihoods of people and the ecosystems they depend on. Such changes are referred to as surprises, tipping points, or high impact-low likelihood (HILL) events. They encompass regime changes across scales in atmospheric and ocean circulation, the hydrological cycle, extreme weather statistics, large-scale vegetation, monsoon systems, and ice masses in Greenland and Antarctica. Knowledge on impacts of future climate change is based on scenario simulations using the current generation of comprehensive models. However, two limitations are evident: (i) current models have difficulties simulating amplitudes and patterns of HILL events; and (ii) scenarios do not include surprises, nor out-of-scale extreme events. These limitations may be overcome by enhancing instruments capable on delivering insights into the consequences of HILL events. Improved model resolution, a more impact-oriented model analysis, and bolstered connections between the scientific and service sectors can ensure more efficient global to local adaptation initiatives and support National Meteorological and Hydrological Services (NMHSs) in protecting people, assets, and infrastructure.
Big Data for Addressing Global Environmental Challenges
Data volumes are rapidly increasing due to technological advancements in in-situ observations, satellite remote sensing, and Earth system models. The resulting big data is crucial for addressing interlinked global challenges such as climate change, air quality, and water, food, and energy supplies. However, exploding data volumes are posing challenges in the stocking, manoeuvring, and tracing of data, highlighting the need for energy-efficient and innovative cyberinfrastructures, technologies and methods for the collection, generation, storage, distribution, and use of big data. Further challenges are related to insufficient data documentation, misunderstood user needs and a lack of capacity to use big data effectively. However, global open access platforms and effective cloud/data-sharing frameworks are providing opportunities to improve the quality and traceability of data as well as bridge global data inequalities through enhanced access to data.
AI for Seamless Numerical Weather-Climate Prediction
Artificial Intelligence (Al) offers an innovative approach to numerical weather climate prediction (NWCP) by training models on historical data, as opposed to solving physical equations, resulting in rapid computations compared to traditional numerical models. Advancing Al NWCP tools may present challenges and opportunities for National Meteorological and Hydrological Services (NMHSs), including impacts on the workforce and influencing future investments. However, training the Al models requires significant data and computation, frequent updates to account for climate variability and improved physical understanding to design and produce the new training datasets. It is still unknown how effectively Al NWCP models forecast extreme events and questions remain on how to best integrate Al NWCP and traditional physics based NWCP models into operational production systems that harness the best of both methods. Nevertheless, Al NWCP could revolutionize operational systems resulting in democratized access to forecast information and insights, benefitting NMHSs that previously lacked resources for in-house NWCP models. Additionally, this may result in new players entering the market, notably the private sector, including big-tech and impact-based weather startups, which may impact the role of NMHSs and the quality and reliability of weather and climate services in an expanding market. Other Al opportunities across the value cycle, such as natural language generation, computer vision or machine learning, may enhance the transformation of NWCP outputs into useful and actionable information for various users and sectors.