Scholarlink

<p>Methane, a potent greenhouse gas, significantly contributes to climate change, with agriculture accounting for ~50% of anthropogenic emissions. This study investigates the impact of agriculture on methane emissions and surface temperature variation in the Loiret region, France, a mixed urban-rural landscape dominated by crops like Wheat, Maize, Barley and Rapeseed. Globally, it addresses how methane emissions from agricultural and urban sources interact to influence local temperature dynamics in mixed landscapes, and whether high-resolution data can enhance these insights. Using Sentinel-5P TROPOMI XCH₄ data statistically downscaled from its native ~7 km footprint (resampled to 1 km in Google Earth Engine) to 100 m grids via Random Forest machine learning. XCH<sub>4</sub> integrated with Landsat-derived indices (Land Surface Temperature, Normalized Difference Vegetation Index, Normalized Difference Built-up Index, Land Use Land Cover), methane dynamics were analyzed from 2019 to 2024. Results reveal seasonal peaks of 1900.1 ppb in Autumn (post-harvest) and 1899.8 ppb in Winter (manure management), with a diffuse production pattern confirmed by wind back-trajectory analysis. A strong methane-temperature correlation (R² = 0.81) and urban-rural feedback loop highlight temperature-driven methanogenesis, amplified by agricultural and urban sources. Annual analysis shows a yearly increase in the concentration of methane between 2019 and 2024. Validation against field data yielded an R² of 0.85 and RMSE of 9.3 ppb, affirming the downscaled data’s reliability. The study demonstrates that machine learning enhances methane mapping resolution, urban and peri-urban emissions interact with agriculture to shape local temperatures. These findings provide a high-resolution, replicable, and relatively simple framework for monitoring methane-temperature interactions, offering insights for sustainable land management in temperate regions.</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

Assessing Agricultural Methane Emissions and Temperature Feedbacks in the Loiret Region, France: A High-Resolution Sentinel-5P and Machine Learning Approach

  • 期刊论文

  • 发表日期：

    2026年1月30日

查看全文
  • Naji El Beyrouthy,
  • Mario Al Sayah,
  • Rita Der Sarkissian,
  • Rachid Nedjai

摘要

Methane, a potent greenhouse gas, significantly contributes to climate change, with agriculture accounting for ~50% of anthropogenic emissions. This study investigates the impact of agriculture on methane emissions and surface temperature variation in the Loiret region, France, a mixed urban-rural landscape dominated by crops like Wheat, Maize, Barley and Rapeseed. Globally, it addresses how methane emissions from agricultural and urban sources interact to influence local temperature dynamics in mixed landscapes, and whether high-resolution data can enhance these insights. Using Sentinel-5P TROPOMI XCH₄ data statistically downscaled from its native ~7 km footprint (resampled to 1 km in Google Earth Engine) to 100 m grids via Random Forest machine learning. XCH4 integrated with Landsat-derived indices (Land Surface Temperature, Normalized Difference Vegetation Index, Normalized Difference Built-up Index, Land Use Land Cover), methane dynamics were analyzed from 2019 to 2024. Results reveal seasonal peaks of 1900.1 ppb in Autumn (post-harvest) and 1899.8 ppb in Winter (manure management), with a diffuse production pattern confirmed by wind back-trajectory analysis. A strong methane-temperature correlation (R² = 0.81) and urban-rural feedback loop highlight temperature-driven methanogenesis, amplified by agricultural and urban sources. Annual analysis shows a yearly increase in the concentration of methane between 2019 and 2024. Validation against field data yielded an R² of 0.85 and RMSE of 9.3 ppb, affirming the downscaled data’s reliability. The study demonstrates that machine learning enhances methane mapping resolution, urban and peri-urban emissions interact with agriculture to shape local temperatures. These findings provide a high-resolution, replicable, and relatively simple framework for monitoring methane-temperature interactions, offering insights for sustainable land management in temperate regions.