Answers are grounded in this paper and cite their sources.
Overlapping clouds (OLCs) play a critical role in Earth’s radiation budget and climate change. The accurate characterization of their properties is essential for cloud radiative effect (CRE) estimates. The existing physical retrieval methods can analytically derive properties of OLCs with optically thin ice clouds overlying water clouds, but require rigorous radiative transfer solutions, which are computationally expensive. This study proposes a retrieval framework integrating a radiative transfer model (RTM) and machine learning (ML) to facilitate the efficient and multiparameter simultaneous retrieval of OLCs with optically thin ice clouds overlying water clouds, including cloud base height (CBH), cloud top height (CTH), cloud effective radius (CER), and cloud optical thickness (COT). The visible-to-infrared observations (0.86, 1.6, 2.25, 8.6, 11, and 12 μm) from visible infrared imaging radiometer suite are used as predictors. CTH and CBH from the active sensor and RTM-derived COT and CER are utilized as reference values. Based on the data from 2017 to 2019, the physics-guided ML method is demonstrated to effectively retrieve both ice and water cloud properties in OLCs, improving predictive performance and physical consistency. Specifically, Pearson’s correlation coefficients for CBH and CTH exceed 0.85, while those for CER and COT exceed 0.90. Validated against active observations, the bias and root-mean-square error of predicted CTH (CBH) are −0.36 ± 1.79 (−0.30 ± 1.75) km for upper level thin ice clouds and −0.18 ± 1.74 (−0.27 ± 1.65) km for lower level water clouds. Moreover, the spatial and seasonal distribution characteristics of COT and CER are revealed. Overall, this approach offers efficient and reliable property retrieval of OLCs, showing potential for improving their CRE estimations.
No takes yet. Share an insight, caveat, or question.
Yang et al. (2026) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: