Center-fixing of tropical cyclones using uncertainty-aware deep learning applied to high-temporal-resolution geostationary satellite imagery

2024-09-24

arXiv.org

AI Summary

Low tropical cyclone (TC) center localization accuracy and reliance on sparse microwave/scatterometer observations hinder operational forecast timeliness and robustness. This paper proposes GeoCenter, a deep learning model that requires only high-temporal-resolution (10-minute cadence) and low-latency (<10 minutes) geostationary infrared imagery. GeoCenter introduces the first uncertainty-aware end-to-end framework, taking as input a 9-channel × 4-hour infrared time-series animation and ATCF scalar priors, and employs a hybrid 3D CNN–LSTM architecture enabling real-time rolling updates. On an independent test set, it achieves mean/median/RMS localization errors of 26.6/22.2/32.4 km across all TCs, improving to 14.6/12.5/17.3 km for major hurricanes (Saffir–Simpson Categories 2–5)—significantly outperforming the infrared-only ARCHER-2 baseline. Notably, GeoCenter attains accuracy comparable to microwave-assisted methods under infrared-only conditions and outputs a calibrated 150-member ensemble to quantify localization uncertainty.

Technology Category

• Machine Learning: Calibration & Uncertainty Quantification
• Computer Vision: Remote Sensing / Geospatial AI
• Knowledge Representation and Reasoning: Geometric, Spatial, and Temporal Reasoning

Application Category

• Systems and Infrastructure for Web, Mobile and WoT: Location- and context-aware Web and WoT applications and services
• Graph Algorithms and Modeling for the Web: Graph neural networks and deep learning approaches for Web-related graphs
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking

Abstract

Determining the location of a tropical cyclone's (TC) surface circulation center --"center-fixing"-- is a critical first step in the TC-forecasting process, affecting current and future estimates of track, intensity, and structure. Despite a recent increase in automated center-fixing methods, only one such method (ARCHER-2) is operational, and its best performance is achieved when using microwave or scatterometer data, which are not available at every forecast cycle. We develop a deep-learning algorithm called GeoCenter; besides a few scalars in the operational ATCF, it relies only on geostationary IR satellite imagery, which is available for all TC basins at high frequency (10 min) and low latency (<10 min) during both day and night. GeoCenter ingests an animation (time series) of IR images, including 9 channels at lag times up to 4 hours. The animation is centered at a"first guess"location, offset from the true TC-center location by 48 km on average and sometimes>100 km; GeoCenter is tasked with correcting this offset. On an independent testing dataset, GeoCenter achieves a mean/median/RMS (root mean square) error of 26.6/22.2/32.4 km for all systems, 24.7/20.8/30.0 km for tropical systems, and 14.6/12.5/17.3 km for category-2--5 hurricanes. These values are similar to ARCHER-2 errors with microwave or scatterometer data, and better than ARCHER-2 errors when only IR data are available. GeoCenter also performs skillful uncertainty quantification, producing a well calibrated ensemble of 150 TC-center locations. Furthermore, all predictors used by GeoCenter are available in real time, which would make GeoCenter easy to implement operationally every 10 min.

Problem

• Automating tropical cyclone center-fixing using deep learning
• Improving accuracy with geostationary IR satellite imagery
• Providing real-time, uncertainty-aware TC location estimates

Innovation

• Deep learning for tropical cyclone center-fixing
• Uses geostationary IR satellite imagery
• Real-time uncertainty-aware ensemble predictions