Spectral-Aware Text-to-Time Series Generation with Billion-Scale Multimodal Meteorological Data

TL;DR

This paper introduces a new large-scale meteorological dataset and a spectral-aware diffusion model for text-guided weather time-series generation, achieving state-of-the-art results and strong semantic control.

Contribution

The work presents MeteoCap-3B, a billion-scale weather dataset with expert captions, and MTransformer, a spectral-aware diffusion model for precise text-to-weather time-series synthesis.

Findings

State-of-the-art generation quality on real-world benchmarks

Accurate cross-modal alignment between text and weather signals

Enhanced semantic controllability and improved forecasting in data-sparse scenarios

Abstract

Text-to-time-series generation is particularly important in meteorology, where natural language offers intuitive control over complex, multi-scale atmospheric dynamics. Existing approaches are constrained by the lack of large-scale, physically grounded multimodal datasets and by architectures that overlook the spectral-temporal structure of weather signals. We address these challenges with a unified framework for text-guided meteorological time-series generation. First, we introduce MeteoCap-3B, a billion-scale weather dataset paired with expert-level captions constructed via a Multi-agent Collaborative Captioning (MACC) pipeline, yielding information-dense and physically consistent annotations. Building on this dataset, we propose MTransformer, a diffusion-based model that enables precise semantic control by mapping textual descriptions into multi-band spectral priors through a Spectral Prompt Generator, which guides generation via frequency-aware attention. Extensive experiments on real-world benchmarks demonstrate state-of-the-art generation quality, accurate cross-modal alignment, strong semantic controllability, and substantial gains in downstream forecasting under data-sparse and zero-shot settings. Additional results on general time-series benchmarks indicate that the proposed framework generalizes beyond meteorology.