Scholarlink

<p>Global climate change and rapid urbanization are transforming land use and thermal environments, particularly in developing megacities, impacting regional climate and sustainable development. In cities like Dhaka, Bangladesh, urbanization has significantly altered land use and land cover (LULC), directly affecting urban climate and land surface temperature (LST).This study investigates the impacts of rapid urbanization on LULC changes and LST in Dhaka, Bangladesh, using multi-temporal satellite imagery from Landsat 5, 7, and 8 from 2009 to 2023. The classification analysis was conducted using Support Vector Machine classification and Random Forest (RF) modeling in Google Earth Engine to predict future LST. The classification achieved high accuracy, with kappa values over 80%. Results found that, due to the Dhaka Metropolitan Development Plan (DMDP) urban settlements expanded by 139.52 km², and vegetation and water bodies declined by 16.71% and 51.71% respectively. The study also found a 4 °C increase in LST (from 34 °C in 2009 to 38 °C in 2023), with predictions indicating further increases up to 41 °C by 2030. Statistical analysis revealed strong correlations between LST and LULC indices, with R² values of 0.42 and − 0.68 for NDVI and NDWI (negative correlations), and 0.04 and 0.26 for NDBI (positive correlation). The RF model, with an R² of 0.953 between observed and predicted values, further predicts a 3 °C rise in LST over the next decade. Spatial analysis revealed the highest urban expansion occurred in the northeastern and southeastern regions of the city. This study demonstrates the utility of integrating multi-temporal satellite data, machine learning, and spatial modeling to quantify urban growth patterns, associated land cover changes, and thermal impacts. The findings highlight the need for climate-adaptive urban planning in rapidly developing megacities to mitigate rising urban temperatures and associated environmental and health risks. The modeling approach presented can support evidence-based policymaking for sustainable urban development and climate change adaptation in Dhaka and similar urban contexts globally.</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

Spatiotemporal analysis of urban expansion, land use dynamics, and thermal characteristics in a rapidly growing megacity using remote sensing and machine learning techniques

  • 期刊论文

  • 发表日期：

    2025年1月10日

查看全文
  • M. Shahriar Sonet,
  • Md. Yeasir Hasan,
  • Abdulla Al Kafy,
  • Nobonita Shobnom

摘要

Global climate change and rapid urbanization are transforming land use and thermal environments, particularly in developing megacities, impacting regional climate and sustainable development. In cities like Dhaka, Bangladesh, urbanization has significantly altered land use and land cover (LULC), directly affecting urban climate and land surface temperature (LST).This study investigates the impacts of rapid urbanization on LULC changes and LST in Dhaka, Bangladesh, using multi-temporal satellite imagery from Landsat 5, 7, and 8 from 2009 to 2023. The classification analysis was conducted using Support Vector Machine classification and Random Forest (RF) modeling in Google Earth Engine to predict future LST. The classification achieved high accuracy, with kappa values over 80%. Results found that, due to the Dhaka Metropolitan Development Plan (DMDP) urban settlements expanded by 139.52 km², and vegetation and water bodies declined by 16.71% and 51.71% respectively. The study also found a 4 °C increase in LST (from 34 °C in 2009 to 38 °C in 2023), with predictions indicating further increases up to 41 °C by 2030. Statistical analysis revealed strong correlations between LST and LULC indices, with R² values of 0.42 and − 0.68 for NDVI and NDWI (negative correlations), and 0.04 and 0.26 for NDBI (positive correlation). The RF model, with an R² of 0.953 between observed and predicted values, further predicts a 3 °C rise in LST over the next decade. Spatial analysis revealed the highest urban expansion occurred in the northeastern and southeastern regions of the city. This study demonstrates the utility of integrating multi-temporal satellite data, machine learning, and spatial modeling to quantify urban growth patterns, associated land cover changes, and thermal impacts. The findings highlight the need for climate-adaptive urban planning in rapidly developing megacities to mitigate rising urban temperatures and associated environmental and health risks. The modeling approach presented can support evidence-based policymaking for sustainable urban development and climate change adaptation in Dhaka and similar urban contexts globally.