Joint Retrieval of Cloud properties using Attention-based Deep Learning Models

2025-04-04

AI Summary

Joint retrieval of cloud optical thickness (COT) and cloud effective radius (CER) is hindered by limitations of the independent pixel approximation (IPA), including neglect of 3D radiative effects, edge-induced errors, and failure to model overlapping or heterogeneous cloud fields. Existing deep learning approaches suffer from high memory overhead, single-attribute output, or insufficient joint retrieval accuracy. To address these challenges, we propose CloudUNet with Attention Module (CAM): a lightweight UNet-based architecture incorporating spatial self-attention to enhance contextual modeling, coupled with a physics-informed joint loss function for end-to-end multi-parameter co-retrieval. Evaluated on large-eddy simulation (LES) synthetic data, CAM reduces mean absolute error (MAE) for COT and CER by 34% and 42%, respectively, compared to state-of-the-art deep learning methods, and by 76% and 86% relative to IPA. The method significantly improves retrieval accuracy over thick clouds, overlapping cloud systems, and cloud edges.

Technology Category

• Machine Learning: Calibration & Uncertainty Quantification
• Computer Vision: Multi-modal Vision
• Search and Optimization: Learning to Search

Application Category

• Search and Retrieval-Augmented AI: Retrieval-Augmented Generation (RAG) and multi-modal RAG
• Web Mining and Content Analysis: Large pretrained models with web data
• Systems and Infrastructure for Web, Mobile and WoT: Cloud, edge and content delivery systems for the Web

Abstract

Accurate cloud property retrieval is vital for understanding cloud behavior and its impact on climate, including applications in weather forecasting, climate modeling, and estimating Earth's radiation balance. The Independent Pixel Approximation (IPA), a widely used physics-based approach, simplifies radiative transfer calculations by assuming each pixel is independent of its neighbors. While computationally efficient, IPA has significant limitations, such as inaccuracies from 3D radiative effects, errors at cloud edges, and ineffectiveness for overlapping or heterogeneous cloud fields. Recent AI/ML-based deep learning models have improved retrieval accuracy by leveraging spatial relationships across pixels. However, these models are often memory-intensive, retrieve only a single cloud property, or struggle with joint property retrievals. To overcome these challenges, we introduce CloudUNet with Attention Module (CAM), a compact UNet-based model that employs attention mechanisms to reduce errors in thick, overlapping cloud regions and a specialized loss function for joint retrieval of Cloud Optical Thickness (COT) and Cloud Effective Radius (CER). Experiments on a Large Eddy Simulation (LES) dataset show that our CAM model outperforms state-of-the-art deep learning methods, reducing mean absolute errors (MAE) by 34% for COT and 42% for CER, and achieving 76% and 86% lower MAE for COT and CER retrievals compared to the IPA method.

Problem

• Retrieve accurate cloud properties for climate studies
• Overcome IPA limitations in 3D radiative effects
• Improve joint retrieval of COT and CER efficiently

Innovation

• Attention-based UNet model for cloud retrieval
• Joint retrieval of COT and CER properties
• Specialized loss function for accuracy improvement