Scholarlink

<p>Fine-scale, species-level monitoring of mangrove dynamics is critical yet challenging, especially in urban coastal wetlands where native and exotic species interact. To address this, we developed an integrated remote sensing workflow that combines Object-Based Image Analysis (OBIA) with a Convolutional Neural Network (CNN). This OBIA-CNN approach was applied to high-resolution WorldView-3 imagery to map decadal changes (2015–2025) in mangrove cover and species composition for five key species in Shenzhen’s Futian Mangrove Nature Reserve. Our results revealed a substantial 43% expansion in total mangrove area (from 96.65 to 137.71&#xa0;ha). However, species-level analysis uncovered strikingly divergent trajectories: the exotic <i>Sonneratia</i> species experienced initial rapid expansion, followed by a sharp decline primarily driven by targeted management interventions (i.e., manual removal of exotic trees), with natural succession playing only a secondary role. In contrast, the native <i>Kandelia obovata</i> exhibited steady expansion and increased landscape consolidation, ultimately re-establishing dominance. Analysis of landscape metrics and species replacement pathways highlighted the competitive interplay between species, identifying mangrove-mudflat edges as critical hotspots of change. During its expansion phase, the exotic <i>Sonneratia</i> competitively suppressed native mangrove growth through canopy shading; however, the subsequent replacement by <i>K. obovata</i> was facilitated by management-driven removal rather than natural competitive displacement. This study provides not only a robust and transferable methodology for fine-scale, long-term mangrove monitoring but also critical insights for managing exotic species and conserving biodiversity in urban coastal wetlands.</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

Tracking the decadal dynamics of native and exotic mangroves in an urban coastal wetland: an integrated OBIA-CNN approach

  • 期刊论文

  • 发表日期：

    2026年5月8日

查看全文
  • Jianfeng Pan,
  • Zijian Huang,
  • Shixiao Yu,
  • Qiong Yang,
  • Ping Hu

摘要

Fine-scale, species-level monitoring of mangrove dynamics is critical yet challenging, especially in urban coastal wetlands where native and exotic species interact. To address this, we developed an integrated remote sensing workflow that combines Object-Based Image Analysis (OBIA) with a Convolutional Neural Network (CNN). This OBIA-CNN approach was applied to high-resolution WorldView-3 imagery to map decadal changes (2015–2025) in mangrove cover and species composition for five key species in Shenzhen’s Futian Mangrove Nature Reserve. Our results revealed a substantial 43% expansion in total mangrove area (from 96.65 to 137.71 ha). However, species-level analysis uncovered strikingly divergent trajectories: the exotic Sonneratia species experienced initial rapid expansion, followed by a sharp decline primarily driven by targeted management interventions (i.e., manual removal of exotic trees), with natural succession playing only a secondary role. In contrast, the native Kandelia obovata exhibited steady expansion and increased landscape consolidation, ultimately re-establishing dominance. Analysis of landscape metrics and species replacement pathways highlighted the competitive interplay between species, identifying mangrove-mudflat edges as critical hotspots of change. During its expansion phase, the exotic Sonneratia competitively suppressed native mangrove growth through canopy shading; however, the subsequent replacement by K. obovata was facilitated by management-driven removal rather than natural competitive displacement. This study provides not only a robust and transferable methodology for fine-scale, long-term mangrove monitoring but also critical insights for managing exotic species and conserving biodiversity in urban coastal wetlands.