Scholarlink

<p>Accurate estimation of ground-level particulate matter (PM) using satellite-derived data is critical for air quality monitoring and public health assessments<i>.</i> SEMARA is a high-resolution Aerosol Optical Depth (AOD) retrieval algorithm that integrates two distinct approaches: the Simplified and Robust Surface Reflectance Estimation Method (SREM) and the Simplified Aerosol Retrieval Algorithm (SARA). This study evaluates the performance of the novel SEMARA AOD product compared to the widely used MAIAC AOD in modeling concentrations of fine (PM<sub>2.5</sub>) and coarse (PM<sub>10</sub>) particulate matter. OD values were normalized using planetary boundary layer height to enhance model performance, and PM concentrations were corrected using relative humidity data. Meteorological and land-use variables served as auxiliary inputs. Four machine learning algorithms were employed, with Random Forest (RF) and Gradient Boosting (GB) yielding the highest accuracy. Overall, SEMARA-based models outperformed MAIAC AOD models. Among the pollutants, PM<sub>10</sub> estimation achieved better results (R<sup>2</sup> = 0.72, MAPE = 20%) compared to PM<sub>2.5</sub> (R<sup>2</sup> = 0.68, MAPE = 26%). The findings of this research highlight SEMARA’s potential for improving satellite-based PM estimation.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

High-Resolution Modeling of Particulate Matter: Comparing SEMARA AOD and MAIAC AOD Models

  • 期刊论文

  • 发表日期：

    2025年11月8日

查看全文
  • Mozhgan Bagherinia,
  • Siamak Boudaghpour

摘要

Accurate estimation of ground-level particulate matter (PM) using satellite-derived data is critical for air quality monitoring and public health assessments. SEMARA is a high-resolution Aerosol Optical Depth (AOD) retrieval algorithm that integrates two distinct approaches: the Simplified and Robust Surface Reflectance Estimation Method (SREM) and the Simplified Aerosol Retrieval Algorithm (SARA). This study evaluates the performance of the novel SEMARA AOD product compared to the widely used MAIAC AOD in modeling concentrations of fine (PM2.5) and coarse (PM10) particulate matter. OD values were normalized using planetary boundary layer height to enhance model performance, and PM concentrations were corrected using relative humidity data. Meteorological and land-use variables served as auxiliary inputs. Four machine learning algorithms were employed, with Random Forest (RF) and Gradient Boosting (GB) yielding the highest accuracy. Overall, SEMARA-based models outperformed MAIAC AOD models. Among the pollutants, PM10 estimation achieved better results (R2 = 0.72, MAPE = 20%) compared to PM2.5 (R2 = 0.68, MAPE = 26%). The findings of this research highlight SEMARA’s potential for improving satellite-based PM estimation.