In recent years, air pollution has become an important public health concern. The high concentration of fine particulate matter with diameter less than 2.5 µm (PM_2.5) is known to be associated with lung cancer, cardiovascular disease, respiratory disease, and metabolic disease. Predicting PM_2.5 concentrations can help governments warn people at high risk, thus mitigating the complications. Although attempts have been made to predict PM_2.5 concentrations, the factors influencing PM_2.5 prediction have not been investigated. In this work, we study feature importance for PM_2.5 prediction in Tehran’s urban area, implementing random forest, extreme gradient boosting, and deep learning machine learning (ML) approaches. We use 23 features, including satellite and meteorological data, ground-measured PM_2.5, and geographical data, in the modeling. The best model performance obtained was R² = 0.81 (R = 0.9), MAE = 9.93 µg/m³, and RMSE = 13.58 µg/m³ using the XGBoost approach, incorporating elimination of unimportant features. However, all three ML methods performed similarly and R² varied from 0.63 to 0.67, when Aerosol Optical Depth (AOD) at 3 km resolution was included, and 0.77 to 0.81, when AOD at 3 km resolution was excluded. Contrary to the PM_2.5 lag data, satellite-derived AODs did not improve model performance.

Monitoring global vegetation dynamics is of great importance for many environmental applications. The vegetation optical depth (VOD), derived from passive microwave observation, is sensitive to the water content in all aboveground vegetation and could serve as complementary information to optical observations for global vegetation monitoring. The microwave vegetation index (MVI), which is originally derived from the zero-order model, is a potential approach to derive VOD and vegetation water content (VWC), however, it has limited application at dense vegetation in the global scale. In this study, we preferred to use a more complex vegetation model, the Tor Vergata model, which takes into account multi-scattering effects inside the vegetation and between the vegetation and soil layer. Validation with ground-based measurements proved this model is an efficient tool to describe the microwave emissions of corn and wheat. The MVI has been derived through two methods: (i) polarization independent ( ${\mathrm{MVI}}_{\mathrm{B}}^{\mathrm{P}}$ ) and (ii) time invariant ( ${\mathrm{MVI}}_{\mathrm{B}}^{\mathrm{T}}$ ), based on model simulations at the L band. Results show that the ${\mathrm{MVI}}_{\mathrm{B}}^{\mathrm{T}}$ has a stronger sensitivity to vegetation properties compared with ${\mathrm{MVI}}_{\mathrm{B}}^{\mathrm{P}}$ . ${\mathrm{MVI}}_{\mathrm{B}}^{\mathrm{T}}$ is used to retrieve VOD and VWC, and the results were compared to physical VOD and measured VWC. Comparisons indicated that ${\mathrm{MVI}}_{\mathrm{B}}^{\mathrm{T}}$ has a great potential to retrieve VOD and VWC. By using L band time-series information, the performance of MVIs could be enhanced and its application in a global scale could be improved while paying attention to vegetation structure and saturation effects.