Current remote sensing-based aerosol optical depth (AOD) products have coarse spatial resolutions, which are useful for studies at continental and global scales, but unsatisfactory for local scale applications, such as urban air pollution monitoring. In this study, we investigated the possibility of using Landsat imagery to develop high-resolution AOD estimations at 30 m based on machine learning algorithms. We assessed the performance of six machine learning algorithms, including Extreme Gradient Boosting, Random Forest, Cascade Random Forest, Gradient Boosted Decision Trees, Extremely Randomized Trees, and Multiple Linear Regression. To obtain accurate AOD estimations, we used prior knowledge from multiple sources as inputs to the machine learning models, including the Global Land Surface Satellite (GLASS) albedo, the 1-km AOD product from MODIS data using the Multi-Angle Implementation of Atmospheric Correction (MAIAC) algorithm, and meteorological and surface elevation data. A total of 13,624 AOD measurements from Aerosol Robotic Network (AERONET) sites were used for model training and validation. We found that all six algorithms exhibited good performance, with R² values ranging from 0.73 to 0.78 and AOD root-mean-square errors (RMSE) ranging from 0.089 to 0.098. The extremely randomized trees algorithm, however, demonstrated marginally superior performance as compared to the other algorithms; hence, it was used to produce AOD estimates at a 30 m resolution for one Landsat scene coving Beijing in 2013–2019. Through a comparison with overlapping AERONET observations, a high level of accuracy was achieved, with an R² = 0.889 and an RMSE = 0.156. Our method can be potentially used to generate a global high-resolution AOD dataset based on Landsat imagery.

Retrieving aerosol optical depth (AOD) over land remains challenging because the relatively weak aerosol signal in top-of-atmosphere (TOA) observations must be separated from strong and spatially heterogeneous surface reflectance. Here, we develop a physically informed random forest framework for global 1 km land AOD retrieval from MODIS. The framework integrates multispectral TOA reflectance, surface properties, observation geometry, meteorological conditions, topography, and physically informed aerosol–surface features. Long-term Aerosol Robotic Network (AERONET) observations from 2001 to 2017 were collocated with MODIS and ancillary datasets for model development and evaluation. Two physically informed features were introduced to improve retrieval robustness across diverse aerosol and surface conditions, including minimum AOD derived from long-term AERONET observations and time-series clear-sky reflectance (TSCR) in the blue, red, and shortwave-infrared bands derived using the 6S radiative-transfer model. Independent retrieval evaluation for 2013–2014 showed good agreement with AERONET observations, with R = 0.81, MAE = 0.063, RMSE = 0.096, and 74.93% of matched samples falling within the MODIS land expected-error envelope, although increasing underestimation was observed at high aerosol loading (AOD > 1). The proposed retrievals also showed better agreement with AERONET than the MOD04 Dark Target and Deep Blue products. These results demonstrate the value of incorporating physically interpretable aerosol-background and surface-reflectance information into data-driven retrievals for AOD over land surfaces.