Browsing by Subject "AOD"

Filter results by typing the first few letters
Now showing 1 - 2 of 2
  • Thumbnail Image
    ItemOpen Access
    Seasonal and Interannual Variations of Carbonaceous Aerosols over the Amazon
    (2020) Hu, Allen

    This study examines the seasonal and interannual variabilities of carbonaceous aerosols, including black carbon (BC) and organic carbon (OC), over the years of 2005-2016 by using outputs from the NASA GISS ModelE simulations and observations from the OMI instrument aboard Aura, AERONET stations in Amazon region, and the GoAmazon aircraft campaigns.

    Simulated seasonal variations and spatial distributions of surface concentrations of BC and OC in Amazon agree well with those of biomass burning emissions. The concentrations are the highest in the dry season (July-September) and lowest in the wet season (February-May), and the locations of high concentrations follow those of high emissions. ModelE is found to underestimate concentrations of OC and BC. Comparisons of the vertical profiles of OC from ModelE with GoAmazon observations in 2014 show that ModelE underestimates OC at all altitudes. In the dry season, when biomass burning dominates, ModelE captures 42%-86% of OMI AAOD in Amazon over 2005-2016, suggesting a low bias in simulated BC concentrations. Simulated seasonal variations in AOD and AAOD in ModelE differ from OMI observations; simulated AOD (AAOD) values are the highest in the dry season, while OMI observed AOD (AAOD) values are the highest in October-January.

    Interannual variations in BC and OC are quantified by relative deviation from the mean (RDEVM). Interannual variations of BC and OC in dry season are much higher than those in wet season. RDEVM values are in the range of -63.2% to 127.2% (-70.8% to 143.8%) for BC (OC) in dry season and in the range of -17.8% to 32.7% (-26.3% to 53.4%) for BC (OC) in wet season. Simulated OC concentrations have larger interannual variability than simulated BC for both the dry and wet seasons. We also found that, compared with OMI observations, ModelE overestimates the interannual variability of AOD and AAOD in the Amazon region for both the dry and wet seasons.

    Results from this study contribute to the understanding of aerosol distributions in the Amazon and have implications for the impact of carbonaceous aerosols on climate on an interannual timescale.

  • Thumbnail Image
    ItemOpen Access
    Unmasking the sky: high-resolution PM2.5 prediction in Texas using machine learning techniques.
    (Journal of exposure science & environmental epidemiology, 2024-04) Zhang, Kai; Lin, Jeffrey; Li, Yuanfei; Sun, Yue; Tong, Weitian; Li, Fangyu; Chien, Lung-Chang; Yang, Yiping; Su, Wei-Chung; Tian, Hezhong; Fu, Peng; Qiao, Fengxiang; Romeiko, Xiaobo Xue; Lin, Shao; Luo, Sheng; Craft, Elena

    Background

    Although PM2.5 (fine particulate matter with an aerodynamic diameter less than 2.5 µm) is an air pollutant of great concern in Texas, limited regulatory monitors pose a significant challenge for decision-making and environmental studies.

    Objective

    This study aimed to predict PM2.5 concentrations at a fine spatial scale on a daily basis by using novel machine learning approaches and incorporating satellite-derived Aerosol Optical Depth (AOD) and a variety of weather and land use variables.

    Methods

    We compiled a comprehensive dataset in Texas from 2013 to 2017, including ground-level PM2.5 concentrations from regulatory monitors; AOD values at 1-km resolution based on images retrieved from the MODIS satellite; and weather, land-use, population density, among others. We built predictive models for each year separately to estimate PM2.5 concentrations using two machine learning approaches called gradient boosted trees and random forest. We evaluated the model prediction performance using in-sample and out-of-sample validations.

    Results

    Our predictive models demonstrate excellent in-sample model performance, as indicated by high R2 values generated from the gradient boosting models (0.94-0.97) and random forest models (0.81-0.90). However, the out-of-sample R2 values fall within a range of 0.52-0.75 for gradient boosting models and 0.44-0.69 for random forest models. Model performance varies slightly across years. A generally decreasing trend in predicted PM2.5 concentrations over time is observed in Eastern Texas.

    Impact statement

    We utilized machine learning approaches to predict PM2.5 levels in Texas. Both gradient boosting and random forest models perform well. Gradient boosting models perform slightly better than random forest models. Our models showed excellent in-sample prediction performance (R2 > 0.9).