What question did this study set out to answer?

This research aims to improve the prediction of particle number size distributions (PNSD) by utilizing deep learning models that incorporate historical aerosol data.

June 10, 2026npj Climate and Atmospheric Science0 citationsOpen Access

Reconstructing particle number size distributions via aerosol history–informed multiple deep learning models

YCYuhan ChengWestlake University XXX XuZhejiang International Studies University LWLiwen WangWestlake University

Key Points

This research aims to improve the prediction of particle number size distributions (PNSD) by utilizing deep learning models that incorporate historical aerosol data.
Developed a framework combining air-parcel location history with aerosol and meteorological data.
Trained two Long Short-Term Memory (LSTM) models and one Bidirectional LSTM (BiLSTM) model over 10 years of data from Cape Verde Atmospheric Observatory.
Evaluated model transferability on Ascension Island dataset.
Achieved mean fractional errors (MFEs) < 0.17 during training and < 0.16 during transferability testing.
Demonstrated potential for projecting PNSD in remote marine regions.
SHAP analysis indicated varied feature attributions across models, reflecting challenges in interpretability.

Abstract

Particle number size distribution (PNSD) is fundamental for characterizing aerosols and quantifying aerosol–cloud interactions, while simulations remain uncertain due to strong spatiotemporal variability during air-parcel transport. We developed a framework that combines air-parcel location history with co-located aerosol, cloud, meteorological, and gas variables into deep learning (DL) models to reconstruct PNSD. Two Long Short-Term Memory (LSTM) models and one Bidirectional LSTM (BiLSTM) model were trained and evaluated on 10 years of measurements at the Cape Verde Atmospheric Observatory (CVAO) in the central Atlantic, achieving mean fractional errors (MFEs) < 0.17. Applying the three well-trained models to the Ascension Island (ASI) dataset in the South Atlantic achieved MFE < 0.16, demonstrating strong transferability and enabling PNSD projection in remote marine regions. SHapley Additive exPlanations (SHAP) analysis showed inconsistent feature attributions across three models, suggesting limited interpretability of mechanisms during air-parcel transportation. Thus, multi-model cross-validation is recommended for interpreting DL models.

Ask AI

Helpful

Bookmark

View Full Paper