Abstract:Atmospheric ozone plays a pivotal role in Earth's climate system, influencing solar radiation absorption in the stratosphere and regulating ultraviolet light reaching the surface. Accurate monitoring of ozone concentration is crucial for environmental assessments, air quality monitoring, and climate change studies. The Ozone Monitoring Instrument (OMI) and Sentinel-5 Precursor TROPOspheric Monitoring Instrument (TROPOMI) provide valuable data for such monitoring. While OMI offers a long data record since 2004, but its effectiveness is hindered by its limitations in spatial resolution and signal-to-noise ratio, stemming from satellite hardware and retrieval algorithms. Sentinel-5P TROPOMI provides higher spatial resolution and improved signal-to-noise ratio, nevertheless, data record from it is rather short. Harmonizing these two datasets by taking the best use of their specific advantages is essential for creating a comprehensive and accurate atmospheric ozone concentration dataset. To maximize the advantages of these multi-source data products, our method utilizes a neural network to learn the mapping relationship between OMI and Sentinel-5P TROPOMI ozone column concentration products, constructing a harmonized model that optimizes the spatial and temporal sequence of historical OMI ozone column concentrations while considering topographic factors. The reconstructed ozone column concentration product is a long time series with the high spatial resolution and accuracy characteristics of Sentinel-5P TROPOMI. This research leverages powerful nonlinear modeling and spatial feature mapping capabilities based on deep learning networks to create a harmonized dataset of atmospheric ozone column concentrations, offering a comprehensive understanding of ozone distribution across the Tibetan Plateau. This dataset not only improves accuracy and precision in ozone concentration measurements but also facilitates in-depth analysis of local ozone variations, providing reliable dataset for scientific investigations into the atmospheric environment. The complete dataset is openly accessible at https://doi.org/10.5281/zenodo.10430751.

DeepSAT4D: Deep learning empowers four-dimensional atmospheric chemical concentration and emission retrieval from satellite

High Spatial-Resolution Classification of Urban Surfaces Using a Deep Learning Method

Deep Learning with Pretrained Framework Unleashes the Power of Satellite-Based Global Fine-Mode Aerosol Retrieval

A Deep Learning Approach to Increase the Value of Satellite Data for PM2.5 Monitoring in China

Combining physical mechanisms and deep learning models for hourly surface ozone retrieval in China

Spatio-temporal modeling of satellite-observed CO 2 columns in China using deep learning

Wide and Deep Learning Model for Satellite-Based Real-Time Aerosol Retrievals in China

Satellite Data Science: A Case Study for Smog Disaster Prediction from Multiple Satellite Observations.

A Physically Constrained Deep-Learning Fusion Method for Estimating Surface NO2 Concentration from Satellite and Ground Monitors

Physical informed neural network improving the WRF-CHEM results of air pollution using satellite-based remote sensing data

First estimation of hourly full-coverage ground-level ozone from Fengyun-4A Satellite using machine learning

Tree-based ensemble deep learning model for spatiotemporal surface ozone (O3) prediction and interpretation

Harmonizing Atmospheric Ozone Column Concentrations over the Tibetan Plateau from 2005 to 2022 Using OMI and Sentinel-5P TROPOMI: A Deep Learning Approach

Real-time and Seamless Monitoring of Ground-level PM2.5 Using Satellite Remote Sensing

Estimating regional ground-level PM2.5 directly from satellite top-of-atmosphere reflectance using deep learning

Improved Consistency of Satellite XCO2 Retrievals Based on Machine Learning

Diurnal Carbon Monoxide Retrieval from FY-4B/GIIRS Using a Novel Machine Learning Method

Estimating Monthly Surface Ozone Using Multi-Source Satellite Products in China Based on Deep Forest Model

Improve Observation-Based Ground-Level Ozone Spatial Distribution by Compositing Satellite and Surface Observations: A Simulation Experiment

Near Real-time CO$_2$ Emissions Based on Carbon Satellite and Artificial Intelligence

Mimicking atmospheric photochemical modeling with a deep neural network