The Journal of Geo-information Science published a review article by researcher Hengcai Zhang (State Key Laboratory of Resources and Environmental Information System, Institute of Geographic Sciences and Natural Resources Research, Chinese Academy of Sciences). They reviewed the historical development of GeoAI-driven spatiotemporal forecasting, providing an overview of prediction models based on statistical learning, deep learning, and generative large models. They presented general computational operators for modeling temporal, spatial, and spatiotemporal relationships, and explained the intrinsic structure and functional interdependencies of complex models. They also examined how these models embed spatiotemporal dependencies and decoupled several computational operators. In addtion, the open-source spatial-temporal prediction model were provoided to enhance the flexibility and reproducibility of the modeling process, streamlining workflows and improving overall efficiency. Standardizing operator design allows researchers to build and adjust models more efficiently without starting from scratch. Then, they summarized and generalized five main challenges in this field: sparse labeled data, lack of interpretability, insufficient generalization ability, inadequate model compression and lightweighting, and insufficient high reliability of prediction models. Finally, they highlighted the practical value and pointed out four future research directions in spatial-temporal intelligent prediction : a generalized spatial intelligent prediction platform incorporating multiple operators, generative prediction models integrating multimodal knowledge, prior-guided deep learning-based intelligent prediction models, and the expansion of geospatial intelligent prediction models into deep predictive applications for earth system analysis. For more details, please refer to the original article:
GeoAI-driven Spatiotemporal Prediction: Progress and Prospects. http://doi.org/10.12082/dqxxkx.2025.240718
https://www.sciengine.com/JGIS/doi/10.12082/dqxxkx.2025.240718（If you want to see the English version of the full text, please click on the in the article page.）
