Novel Ecological Suitability Evaluation Perspective for Large-Scale Wind Power Construction Zoning and Geographical Potential Prediction:An Explainable Artificial Intelligence-Based Approach DOI
Liting Wang, Ruijia Zhang, Bingran Ma

et al.

Published: Jan. 1, 2025

The framework development for large-scale energy construction zoning and the consideration of geographical potential are critical to achieving carbon neutrality goals. However, previous wind power studies have mostly focused on micro-site selection, rarely quantitatively analyzed response relationship between ecosystem from a macroscopic perspective. This study introduced new explainable artificial intelligence ecological suitability prediction framework, which can serve as link mitigation adaptation measures in global climate change. Twenty indicator layers were incorporated, quantitative evaluation method was used generate twenty raster data subsequent models' input layers. results showed that accuracies XGBoost, LightGBM, CatBoost CNN 93.36%, 96.50%, 97.17% 97.46% respectively. performed consistently across all categories, while more distinctive very unsuitable suitable categories. analysis found sensitivity desertification distance roads contributed most results, influences socio-economic indicators minimal. Applying Qinghai-Tibetan Plateau revealed 71.4% area farm construction, but total high 1.90×106 GW. It is recommended 11.32% areas with should be developed soon possible future after detailed assessment, order achieve nationally determined contributions target slow down warming.

Language: Английский

Novel Ecological Suitability Evaluation Perspective for Large-Scale Wind Power Construction Zoning and Geographical Potential Prediction:An Explainable Artificial Intelligence-Based Approach DOI
Liting Wang, Ruijia Zhang, Bingran Ma

et al.

Published: Jan. 1, 2025

The framework development for large-scale energy construction zoning and the consideration of geographical potential are critical to achieving carbon neutrality goals. However, previous wind power studies have mostly focused on micro-site selection, rarely quantitatively analyzed response relationship between ecosystem from a macroscopic perspective. This study introduced new explainable artificial intelligence ecological suitability prediction framework, which can serve as link mitigation adaptation measures in global climate change. Twenty indicator layers were incorporated, quantitative evaluation method was used generate twenty raster data subsequent models' input layers. results showed that accuracies XGBoost, LightGBM, CatBoost CNN 93.36%, 96.50%, 97.17% 97.46% respectively. performed consistently across all categories, while more distinctive very unsuitable suitable categories. analysis found sensitivity desertification distance roads contributed most results, influences socio-economic indicators minimal. Applying Qinghai-Tibetan Plateau revealed 71.4% area farm construction, but total high 1.90×106 GW. It is recommended 11.32% areas with should be developed soon possible future after detailed assessment, order achieve nationally determined contributions target slow down warming.

Language: Английский

Citations

0