Generative Artificial Intelligence Driven Metrological Insight Engine with Reasoning


Climate and weather phenomena critically impact sectors such as agriculture, energy, and disaster management. Despite advances in existing meteorological frameworks, these systems often lack interpretability, effective integration of heterogeneous meteorology domain multimodal data, and robust cause–and–effect reasoning across meteorological and related domains. To address these limitations, this study proposes a semantically enriched multi-agent Visual Question Answering (VQA) framework that integrates multimodal meteorological data, domain ontologies, and foundation models. Semantic grounding enables cross-domain reasoning and interpretability, while specialized agents support causal inference. Through both graphical and conversational interfaces, the framework expects to support user-centric interaction, enhance interpretability, and empower informed decision-making across multiple domains.