Volume 58

Real-time Estimation of Surface Downwelling Longwave Radiation from Satellite Imagery for Sky Radiative Cooling Applications Yuying XIE, Mengying LI

Abstract

Passive sky radiative cooling systems, which utilize the universe as a natural heat sink, have emerged as a critical technology for providing low-carbon solutions to urban cooling. Surface downwelling longwave radiation (SDLR), originating from the atmosphere, significantly influences the cooling potential of such systems. Therefore, accurately estimating SDLR is pivotal for the design and performance evaluation of these cooling systems. However, real-time SDLR data is generally scarce or lacks accuracy, primarily due to the complex interactions between atmospheric radiation and the inherent variability of clouds. To address this challenge, this work proposes a novel K-means-multilayer perceptron (MLP) model to estimate cloud optical properties and SDLR using high-resolution (5-min, 2-km) geostationary satellite imagery combined with an enhanced two-stream, spectrally resolved radiative model. When validated against one year (2019) of SDLR measurements from the Surface Radiation Budget Network (SURFRAD) across diverse climatic regions in the contiguous United States, the proposed model achieves root mean square error values ranging from 20-25 W/m² across all stations. These findings highlight the capability of this data-driven model to deliver low-latency, accurate estimations of cloud optical properties and SDLR using real-time satellite imagery. The advancement provides promising tools and atmospheric data sources to support the development of advanced sustainable energy systems in urban environment.

Keywords sky radiative cooling, surface downwelling longwave radiation, cloud optical properties, radiative transfer model, remote sensing, deep learning