Scholarlink

<p>The Lower Mekong Region (LMR), which includes Vietnam, Thailand, Laos, Cambodia, and Myanmar, faces growing energy demand driven by urbanization and economic development. Wind energy offers a promising alternative, yet comprehensive wind suitability assessments for this region remain limited. This study bridges this gap by using machine learning (ML) models like Support Vector Machine (SVM), Random Forest (RF), K-Nearest Neighbor (KNN), and XGBoost to assess wind power potential across the LMR. These models were trained on 11 spatial variables, including wind speed, elevation, and proximity to infrastructure, to classify land suitability. The ensemble model, combining outputs from all classifiers, demonstrated a 23.17% area suitability for high wind potential, concentrated in coastal and elevated terrains, particularly in Vietnam and southern Laos. All models achieved ROC-AUC scores above 0.90, indicating strong predictive reliability. Variable importance analysis revealed that wind speed and distance to roads were the most influential factors in determining suitability. Additionally, areas with very high potential accounted for 11.52% of the total land area, covering 117,351&#xa0;km². The moderate suitability class accounted for 15.44% of the total area, highlighting significant wind potential in mid-suitability zones. This research innovatively integrates multiple ML models to provide an accurate, spatially explicit assessment of wind energy potential in the LMR. The ensemble model not only improves mapping precision but also enhances spatial coherence, offering a useful framework for future wind farm development and energy planning. These findings are critical for informing policy decisions and supporting sustainable energy strategies in the region.</p> Graphical Abstract <p></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

Wind power potential assessment in the lower Mekong countries: a data-driven approach

  • 期刊论文

  • 发表日期：

    2026年4月24日

查看全文
  • Harekrishna Manna,
  • Malay Pramanik,
  • Rizwan Ahamed,
  • Sanjit Sarkar

摘要

The Lower Mekong Region (LMR), which includes Vietnam, Thailand, Laos, Cambodia, and Myanmar, faces growing energy demand driven by urbanization and economic development. Wind energy offers a promising alternative, yet comprehensive wind suitability assessments for this region remain limited. This study bridges this gap by using machine learning (ML) models like Support Vector Machine (SVM), Random Forest (RF), K-Nearest Neighbor (KNN), and XGBoost to assess wind power potential across the LMR. These models were trained on 11 spatial variables, including wind speed, elevation, and proximity to infrastructure, to classify land suitability. The ensemble model, combining outputs from all classifiers, demonstrated a 23.17% area suitability for high wind potential, concentrated in coastal and elevated terrains, particularly in Vietnam and southern Laos. All models achieved ROC-AUC scores above 0.90, indicating strong predictive reliability. Variable importance analysis revealed that wind speed and distance to roads were the most influential factors in determining suitability. Additionally, areas with very high potential accounted for 11.52% of the total land area, covering 117,351 km². The moderate suitability class accounted for 15.44% of the total area, highlighting significant wind potential in mid-suitability zones. This research innovatively integrates multiple ML models to provide an accurate, spatially explicit assessment of wind energy potential in the LMR. The ensemble model not only improves mapping precision but also enhances spatial coherence, offering a useful framework for future wind farm development and energy planning. These findings are critical for informing policy decisions and supporting sustainable energy strategies in the region.

Graphical Abstract