Radiative transfer simulations and remote sensing studies fundamentally require accurate and efficient computation of the optical properties of non-spherical particles.This paper proposes a deep learning(DL)scheme in conjunction with an optical property database to achieve this goal.Deep neural network(DNN)architectures were obtained from a dataset of the optical properties of super-spheroids with extensive shape parameters,size parameters,and refractive indices.The dataset was computed through the invariant imbedding T-matrix method.Four separate DNN architectures were created to compute the extinction efficiency factor,single-scattering albedo,asymmetry factor,and phase matrix.The criterion for designing these neural networks was the achievement of the highest prediction accuracy with minimal DNN parameters.The numerical results demonstrate that the determination coefficients are greater than 0.999 between the prediction values from the neural networks and the truth values from the database,which indicates that the DNN can reproduce the optical properties in the dataset with high accuracy.In addition,the DNN model can robustly predict the optical properties of particles with high accuracy for shape parameters or refractive indices that are unavailable in the database.Importantly,the ratio of the database size(～127 GB)to that of the DNN parameters(～20 MB)is approximately 6810,implying that the DNN model can be treated as a highly compressed database that can be used as an alternative to the original database for real-time computing of the optical properties of non-spherical particles in radiative transfer and atmospheric models.

Jinhe YU;Lei BI;Wei HAN;Xiaoye ZHANG

Key Laboratory of Geoscience Big Data and Deep Resource of Zhejiang Province,School of Earth Sciences,Zhejiang University,Hangzhou 310027,China;Center for Earth System Modeling and Prediction,China Meteorological Administration,Beijing 100081,China;Numerical Weather Prediction Center,China Meteorological Administration,Beijing 100081,China;Key Laboratory of Atmospheric Chemistry of China Meteorological Administration,Institute of Atmospheric Composition,Chinese Academy of Meteorological Sciences,Beijing 100081,China

大气科学进展（英文版）

[1]Jinhe YU;Lei BI;Wei HAN;Xiaoye ZHANG-.Application of a Neural Network to Store and Compute the Optical Properties of Non-Spherical Particles)[J].大气科学进展（英文版）,2022(12):2024-2039

imbedding

Application,Neural,Network,Store,Compute,Optical,Properties,Non,Spherical,Particles,Radiative,transfer,simulations,remote,sensing,studies,fundamentally,require,accurate,computation,optical,properties,spherical,particles,This,paper,proposes,deep,learning,DL,scheme,conjunction,property,database,this,goal,Deep,neural,DNN,architectures,were,obtained,from,dataset,super,spheroids,extensive,shape,parameters,size,refractive,indices,was,computed,through,invariant,matrix,method,Four,separate,created,extinction,efficiency,single,scattering,albedo,asymmetry,phase,criterion,designing,these,networks,achievement,highest,prediction,accuracy,minimal,numerical,results,demonstrate,that,determination,coefficients,are,greater,than,between,values,truth,which,indicates,can,reproduce,In,addition,robustly,unavailable,Importantly,ratio,MB,approximately,implying,treated,highly,compressed,used,alternative,original,real,computing,radiative,atmospheric,models

0.539399