Preface to the Special Issue: AI Applications in Atmospheric and Oceanic Science: Pioneering the Future (Part I)

As AI continues to establish itself as a cornerstone technology across various industries and scientific disciplines, its profound impact on atmospheric and oceanic science is becoming increasingly apparent.The advantages of AI in surmounting obstacles within our field are undeniable, as evidenced by breakthroughs in weather forecasting (e.g., Bi et al., 2023), climate prediction (e.g., Ham et al., 2019), AI-based parameterization schemes (e.g., Rasp et al., 2018;Wang and Tan, 2023), and beyond.Recognizing the transformative potential of AI in atmospheric and oceanic science, this special issue endeavors to explore the extensive applications of AI in our domain.Covering a broad spectrum of subjects, from weather and climate to oceanic science, this issue aims to enrich our understanding of how AI is transforming our approaches.With articles focusing on enhancing forecasts, optimizing numerical models, and conducting mechanistic research through explainable analysis, this collection provides valuable insights into the current state and future potential of AI in atmospheric and oceanic science.This special issue (Part I) highlights recent advancements in AI applications in atmospheric and oceanic science, comprising a total of 14 articles that span a broad spectrum of research topics.Huang et al. ( 2024) commence by succinctly addressing the ongoing debate surrounding the physical consistency of AI weather models, elucidating the significance of physical constraints and delineating methodologies for their integration into AI models.Additionally, they underscore the adoption of online coupled AI-physical models as a viable approach.Lyu et al. (2024) and Wang et al. (2024) employ supervised and unsupervised learning techniques to undertake ENSO predictions and correct ENSO simulations within climate models, respectively.Meanwhile, Lin et al. (2024) and Wang et al. (2024) conduct mechanistic analyses using AI models, exploring the impacts of fires on hail events and investigating the origins of the low predictability of El Niño.At weather and subseasonal scales, Li et al. (2024) investigate the integration of deep learning and physical models for soil moisture prediction, while Zhou et al. (2024) propose a physics-informed deep learning method for typhoon intensity prediction.Additionally, Liu et al. (2024) utilize deep learning for thunderstorm wind forecasting, and Song et al. (2024a) construct advanced deep learning models for predicting Arctic sea-ice concentration one month in advance.Yuan et al. (2024) employ deep learning models for predicting ocean sound velocity.Addressing the issue of ensemble forecast integration, Song et al. (2024b) develop a novel non-cross-quantile regression neural network method for calibrating numerical forecast ensembles.Furthermore, Zhuang et al. (2024) apply symbolic classification algorithms to the monitoring of turbulence anomalies, bearing significance for aviation safety.Moreover, Du and Zhang (2024) advocate for a deep learning scheme called Representing Wind Stress Anomalies over the Tropical Pacific, integrated with an intermediate coupled model.Collectively, these contributions underscore the significant progress made at the confluence of AI and atmospheric and oceanic science, offering promising prospects for enhanced understanding and predictive capabilities in our ongoing efforts to comprehend and mitigate environmental challenges.