The Study on Landslide Hazards Based on Multi-Source Data and GMLCM Approach DOI Creative Commons
Zhifang Zhao,

Z. Y. Li,

Ping Lv

et al.

Remote Sensing, Journal Year: 2025, Volume and Issue: 17(9), P. 1634 - 1634

Published: May 5, 2025

The southwest region of China is characterized by numerous rugged mountains and valleys, which create favorable conditions for landslide disasters. landslide-influencing factors show different sensitivities regionally, induces the occurrence disasters to degrees, especially in small sample areas. This study constructs a framework identification, analysis, evaluation hazards complex mountainous regions within utilizes baseline subset interferometric synthetic aperture radar (SBAS-InSAR) technology high-resolution optical imagery comprehensive interpretation identify hazards. A geodetector employed analyze disaster-inducing factors, machine-learning models such as random forest (RF), gradient boosting decision tree (GBDT), categorical (CatBoost), logistic regression (LR), stacking ensemble strategies (Stacking) are applied sensitivity evaluation. GMLCM stands geodetector–machine-learning-coupled modeling. results indicate following: (1) 172 were identified, primarily concentrated along banks Lancang River. (2) analysis shows that key landslides include digital elevation model (DEM) (1321–1857 m), rainfall (1181–1290 mm/a), distance from roads (0–1285 geological rock formation (soft formation). (3) Based on application K-means clustering algorithm Bayesian optimization algorithm, GD-CatBoost excellent performance. High-sensitivity zones predominantly River, accounting 24.2% area. method identifying small-sample can provide guidance insights monitoring harnessing similar environments.

Language: Английский

Evolution of urban vitality drivers from 2014 to 2022: a case study of Kunming, China DOI
Yinrui Xie, C. Shang, Xin Deng

et al.

International Journal of Environmental Science and Technology, Journal Year: 2025, Volume and Issue: unknown

Published: Feb. 23, 2025

Language: Английский

Citations

1
Multi-Source Data-Driven Spatiotemporal Study on Integrated Ecosystem Service Value for Sustainable Ecosystem Management in Lake Dianchi Basin DOI Open Access
Tian Bai,

Junming Yang,

Xinyu Wang

et al.

Sustainability, Journal Year: 2025, Volume and Issue: 17(9), P. 3832 - 3832

Published: April 24, 2025

Ecosystem services are pivotal in assessing environmental health and societal well-being. Focusing on Lake Dianchi Basin (LDB), China, our research evaluated the IESV (Integrated Service Value) from 2000 to 2020, utilizing remote sensing multiple statistical datasets. The analysis incorporates LSV (Landscape Value), CSV (Carbon Sequestration NPPV (Net Primary Productivity Value). results show that exhibited an expansion of low-yield zones near urban areas, contrasted by NPPV’s growth high-yield outskirt areas. LSV’s normal distribution indicates stability, while CSV’s bimodal structure points partial integration systemic divergence. pronounced clustering both low- regions, with congregating centers dispersed along basin’s periphery. Despite overall downward trajectory IESV, augmentation suggested underlying resilience. A southeastward shift IESV’s focus was driven patterns expansion. Finally, we produced projections CA-MC (Cellular Automata–Markov Chain) model analyze ongoing areas around Kunming. By 2030, aggregate value is expected modestly diminish, ascension mitigating declines CSV. In essence, fluctuations within LDB intricately linked development.

Language: Английский

Citations

0
The Study on Landslide Hazards Based on Multi-Source Data and GMLCM Approach DOI Creative Commons
Zhifang Zhao,

Z. Y. Li,

Ping Lv

et al.

Remote Sensing, Journal Year: 2025, Volume and Issue: 17(9), P. 1634 - 1634

Published: May 5, 2025

The southwest region of China is characterized by numerous rugged mountains and valleys, which create favorable conditions for landslide disasters. landslide-influencing factors show different sensitivities regionally, induces the occurrence disasters to degrees, especially in small sample areas. This study constructs a framework identification, analysis, evaluation hazards complex mountainous regions within utilizes baseline subset interferometric synthetic aperture radar (SBAS-InSAR) technology high-resolution optical imagery comprehensive interpretation identify hazards. A geodetector employed analyze disaster-inducing factors, machine-learning models such as random forest (RF), gradient boosting decision tree (GBDT), categorical (CatBoost), logistic regression (LR), stacking ensemble strategies (Stacking) are applied sensitivity evaluation. GMLCM stands geodetector–machine-learning-coupled modeling. results indicate following: (1) 172 were identified, primarily concentrated along banks Lancang River. (2) analysis shows that key landslides include digital elevation model (DEM) (1321–1857 m), rainfall (1181–1290 mm/a), distance from roads (0–1285 geological rock formation (soft formation). (3) Based on application K-means clustering algorithm Bayesian optimization algorithm, GD-CatBoost excellent performance. High-sensitivity zones predominantly River, accounting 24.2% area. method identifying small-sample can provide guidance insights monitoring harnessing similar environments.

Language: Английский

Citations

0