Hierarchical AI-Meteorologist: LLM-Agent System for Multi-Scale and Explainable Weather Forecast Reporting

2025-11-28

AI Summary

This study addresses three key limitations of large language models (LLMs) in weather report generation: poor interpretability, weak multi-scale temporal consistency, and insufficient factual accuracy. To this end, we propose a hierarchical LLM-agent meteorological system that integrates hourly, 6-hourly, and daily time-series analysis to establish a multi-granularity reasoning architecture. We further introduce a keyword-guided semantic anchoring verification mechanism to explicitly enforce temporal coherence and factual alignment. The system combines structured meteorological data processing (from OpenWeather and Meteostat), multi-scale temporal modeling, and controllable natural language generation. Experimental results demonstrate significant improvements in factual accuracy (+12.7%) and cross-scale consistency. Moreover, we present the first reproducible, semantics-aware evaluation framework tailored to meteorological reporting—laying foundational groundwork for trustworthy, automated weather forecasting and dissemination.

Technology Category

• Machine Learning: Large Multimodal Models (LMMs)
• Natural Language Processing: (Large) Language Models
• Planning, Routing, and Scheduling: Planning with Language Models

Application Category

• Semantics and Knowledge: Data modeling to support human-machine intelligence, including LLMs agents, intelligent system behavior, explanations, and user-friendly interactions
• Search and Retrieval-Augmented AI: Large language models for search
• User Modeling, Personalization and Recommendation: Large Language Models (LLM) for user modeling and recommendation

Abstract

We present the Hierarchical AI-Meteorologist, an LLM-agent system that generates explainable weather reports using a hierarchical forecast reasoning and weather keyword generation. Unlike standard approaches that treat forecasts as flat time series, our framework performs multi-scale reasoning across hourly, 6-hour, and daily aggregations to capture both short-term dynamics and long-term trends. Its core reasoning agent converts structured meteorological inputs into coherent narratives while simultaneously extracting a few keywords effectively summarizing the dominant meteorological events. These keywords serve as semantic anchors for validating consistency, temporal coherence and factual alignment of the generated reports. Using OpenWeather and Meteostat data, we demonstrate that hierarchical context and keyword-based validation substantially improve interpretability and robustness of LLM-generated weather narratives, offering a reproducible framework for semantic evaluation of automated meteorological reporting and advancing agent-based scientific reasoning.

Problem

• Generates explainable weather reports using hierarchical reasoning
• Converts meteorological data into coherent narratives with keywords
• Improves interpretability of automated weather forecasting systems

Innovation

• Hierarchical reasoning across multiple time scales
• Generating keywords to summarize meteorological events
• Validating reports using keyword-based semantic anchors