Scholarlink

<p>The invasive fall webworm (<i>Hyphantria cunea</i>) threatens China’s ecosystems and agroforestry, yet accurate prediction of its spread remains challenging. Here, we developed an ensemble model (GWO-RF-CART-ANN) integrating climatic (temperature, precipitation) and anthropogenic factors (nighttime light, road density) to predict <i>H. cunea</i> distribution across China. Using county-level occurrence data (2004–2022), we show that the optimized ensemble model reduced prediction errors (RMSE: 13.21 km<sup>2</sup>, MAE: 8.97 km<sup>2</sup>) by 24.7–31.1% compared to single models. Minimum temperature and human activity intensity were the dominant drivers of infestation patterns. Our results identify high-risk zones (Liaoning, Hebei, Shandong) and demonstrate that ensemble models significantly improve ecological risk assessment. This study provides a scalable framework for invasive species management, supporting evidence-based policy decisions.</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

An ensemble model for predicting Hyphantria cunea invasion risks: implications for ecological monitoring and management

  • 期刊论文

  • 发表日期：

    2026年1月27日

查看全文
  • Hongwei Zhou,
  • Yunbo Yan,
  • Yantao Zhou,
  • Yifan Chen,
  • Haochang Hu,
  • Shuo Li,
  • Yue Wang,
  • Zihan Xu,
  • Di Cui,
  • Yumo Chen,
  • Jun Yang,
  • Jun Chen

摘要

The invasive fall webworm (Hyphantria cunea) threatens China’s ecosystems and agroforestry, yet accurate prediction of its spread remains challenging. Here, we developed an ensemble model (GWO-RF-CART-ANN) integrating climatic (temperature, precipitation) and anthropogenic factors (nighttime light, road density) to predict H. cunea distribution across China. Using county-level occurrence data (2004–2022), we show that the optimized ensemble model reduced prediction errors (RMSE: 13.21 km2, MAE: 8.97 km2) by 24.7–31.1% compared to single models. Minimum temperature and human activity intensity were the dominant drivers of infestation patterns. Our results identify high-risk zones (Liaoning, Hebei, Shandong) and demonstrate that ensemble models significantly improve ecological risk assessment. This study provides a scalable framework for invasive species management, supporting evidence-based policy decisions.