Scholarlink

<p>Upward-looking ground-based microwave radiometers provide continuous sounding of atmospheric temperature and humidity. This study assesses a ground-based microwave radiometer (MWR) installed in Ahmedabad by comparing the measured brightness temperatures and retrieved atmospheric profiles with simulated brightness temperatures using the Radiative Transfer for TOVS (RTTOV) model adapted for ground-based microwave radiometers (RTTOV-gb) and atmospheric soundings from radiosondes. The analysis revealed a mean bias of approximately − 5&#xa0;K and − 2.5&#xa0;K in the brightness temperatures measured by humidity and temperature-sensitive channels, respectively, relative to RTTOV simulations. Furthermore, the retrieved profiles exhibited substantial errors compared to radiosonde measurements. The maximum bias in MWR-retrieved temperature and humidity profiles was ~ 6&#xa0;K and 6.5&#xa0;g/m³, respectively, when compared to radiosonde data. The MWR-retrieved near-surface temperature had a maximum RMSD and mean bias of 1.9&#xa0;K and − 1.5&#xa0;K, respectively, while the MWR-retrieved near-surface absolute humidity had a maximum RMSD and mean bias of 3.2&#xa0;g/m³ and − 2&#xa0;g/m³, respectively. Additionally, a comparison of Random Forest (RF) and Artificial Neural Network (ANN)-based machine learning retrieval methods was conducted to retrieve atmospheric temperature and humidity profiles from ground-based microwave radiometer measurements. The mean square error for the retrieved temperature and humidity using these machine learning-based methods was approximately 1.1&#xa0;K and 1.5&#xa0;g/m³ for the RF method and 1.8&#xa0;K and 1.6&#xa0;g/m³ for the ANN method. The study also demonstrates the diurnal evolution of the boundary layer for two contrasting days based on the ML-retrieved profiles.</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

Evaluation of Ground-Based Upward-Looking Microwave Radiometer Brightness Temperature Observations and Assessment of Machine Learning-Based Methods for Retrieving Atmospheric Profiles and Their Application for Studying the Evolution of the Atmospheric Boundary Layer

  • 期刊论文

  • 发表日期：

    2025年3月5日

查看全文
  • Manoj Kumar Mishra,
  • Pradeep Kumar Thapliyal

摘要

Upward-looking ground-based microwave radiometers provide continuous sounding of atmospheric temperature and humidity. This study assesses a ground-based microwave radiometer (MWR) installed in Ahmedabad by comparing the measured brightness temperatures and retrieved atmospheric profiles with simulated brightness temperatures using the Radiative Transfer for TOVS (RTTOV) model adapted for ground-based microwave radiometers (RTTOV-gb) and atmospheric soundings from radiosondes. The analysis revealed a mean bias of approximately − 5 K and − 2.5 K in the brightness temperatures measured by humidity and temperature-sensitive channels, respectively, relative to RTTOV simulations. Furthermore, the retrieved profiles exhibited substantial errors compared to radiosonde measurements. The maximum bias in MWR-retrieved temperature and humidity profiles was ~ 6 K and 6.5 g/m³, respectively, when compared to radiosonde data. The MWR-retrieved near-surface temperature had a maximum RMSD and mean bias of 1.9 K and − 1.5 K, respectively, while the MWR-retrieved near-surface absolute humidity had a maximum RMSD and mean bias of 3.2 g/m³ and − 2 g/m³, respectively. Additionally, a comparison of Random Forest (RF) and Artificial Neural Network (ANN)-based machine learning retrieval methods was conducted to retrieve atmospheric temperature and humidity profiles from ground-based microwave radiometer measurements. The mean square error for the retrieved temperature and humidity using these machine learning-based methods was approximately 1.1 K and 1.5 g/m³ for the RF method and 1.8 K and 1.6 g/m³ for the ANN method. The study also demonstrates the diurnal evolution of the boundary layer for two contrasting days based on the ML-retrieved profiles.