Young-Min Yang

Artificial intelligence has advanced global weather forecasting, outperforming traditional numerical models in both accuracy and computational efficiency. Nevertheless, extending predictions beyond subseasonal timescales requires the development of deep learning (DL)–based ocean-atmosphere coupled models that can realistically simulate complex oceanic responses to atmospheric forcing. This study presents KIST-Ocean, a DL-based global three-dimensional ocean general circulation model. Comprehensive evaluations demonstrate the model’s robust ocean simulation skill and efficiency. Moreover, it reproduces ocean responses, such as Kelvin and Rossby wave propagation, and vertical motions induced by wind stress curl, demonstrating its ability to represent key atmospherically forced ocean dynamics underlying climate phenomena, including the El Niño–Southern Oscillation. These findings reinforce confidence in DL-based global weather and climate models by demonstrating their capacity to capture essential ocean-atmosphere relationships. Building on this foundation, the present study paves the way for extending DL-based modeling frameworks toward integrated Earth system simulations, thereby offering substantial potential for advancing long-range climate prediction capabilities.

Labor is a fundamental driver of economic productivity but is increasingly threatened by rising heat exposure under climate change. While mitigation policies are often framed around avoided damages and health co-benefits, the persistence of labor productivity losses under negative CO₂ emission pathways remains poorly understood. Here, we use the Community Earth System Model version 1.2 (CESM1.2) large ensemble simulations to investigate hysteresis in Wet Bulb Globe Temperature (WBGT), labor productivity, and associated economic impacts under a CO₂ overshoot scenario. Our results show that midday heat exposure produces the most severe productivity reductions, with WBGT recovery lagging behind surface temperature due to humidity-driven hysteresis. Even after atmospheric CO₂ return to present climate levels, global labor losses remain above 100 billion hours annually, with South Asia, Central Africa, and the Middle East experiencing the strongest irreversibility. These persistent damages account for more than 60% of total climate-related economic losses. We provide the global assessment of hysteresis in labor productivity under overshoot pathways. The findings demonstrate that mitigation alone cannot fully restore labor capacity and highlight the necessity of complementary adaptation strategies—including heat-resilient infrastructure, work-rest scheduling, and legal protections for outdoor workers. Our study emphasizes the importance of incorporating hysteresis effects into benefit–cost assessments of climate policies to more accurately capture long-term economic and social risks, particularly for vulnerable populations in tropical and low-income regions.

Understanding the relationship between fire activity and climate variability is a major concern for the scientific community and is essential for reducing economic losses and life-threatening fire hazards. However, the drivers of fire activity and the influence of climate variability remain uncertain. Here, we show that the Madden-Julian Oscillation (MJO)—a dominant tropical subseasonal variability—influences fire activity by modulating local fire-supporting weather through atmospheric teleconnections. Our results show that midlatitude fire emissions exhibit significant subseasonal variability, with MJO-related weather influencing the fire intensity and contributing to large fire events. MJO-related fire events account for about 10%–20% of total midlatitude fire events, suggesting that if MJO teleconnections strengthen in the future, fire emissions and associated economic losses could worsen.