Scholarlink

Key message <p>Using monthly Harmonized Landsat–Sentinel imagery and an adapted LandTrendr-based workflow, we achieved reliable month-resolved estimation of coppiced-stand age for <i>Eucalyptus</i> spp. plantations in Luzhai County, Guangxi, China. The workflow achieved 83.2% overall accuracy and a mean absolute timing error of 1.58&#xa0;months, enabling reliable month-scale harvest dating for short-rotation plantations.</p> Context <p>Short-rotation <i>Eucalyptus</i> plantations are frequently harvested, and annual composites often fail to identify harvest timing precisely. Month-resolved stand age is therefore important for plantation management and carbon assessment.</p> Aims <p>We developed and validated a monthly stand age estimation framework by combining Harmonized Landsat–Sentinel time series with a LandTrendr-based workflow and multi-index fusion.</p> Methods <p>We used monthly Harmonized Landsat–Sentinel composites from 2019 to 2024 to derive four spectral indices. After index screening with confirmed harvest samples, we applied LandTrendr to each index time series and used multi-index consensus fusion with a ± 1-month tolerance rule to refine stand age estimates. Performance was then evaluated against reference samples using accuracy metrics and mean absolute error.</p> Results <p>Monthly inputs improved harvest detection compared with annual inputs, increasing overall accuracy from 82.2% to 89.4%. Multi-index fusion further improved stand age estimation accuracy from 76.8% to 83.2% and reduced mean absolute error from 2.29 to 1.58&#xa0;months.</p> Conclusion <p>Monthly Harmonized Landsat–Sentinel time series combined with LandTrendr enabled accurate and cost-effective month-resolved stand age assessment for <i>Eucalyptus</i> plantations. The workflow can support harvest scheduling, plantation inventory updating, and carbon-related assessments using free satellite data and cloud-based processing.</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

Estimating the age of Eucalyptus plantations using monthly Harmonized Landsat–Sentinel imagery and a LandTrendr-based workflow

  • 期刊论文

  • 出版模式：

    开放获取

  • 发表日期：

    2026年5月9日

查看全文
  • Juncheng Huang,
  • Huazhou Wei,
  • Jirong Ding,
  • Yehua Liang,
  • Zhiyong Wu,
  • Jianjun Chen,
  • Haotian You

摘要

Key message

Using monthly Harmonized Landsat–Sentinel imagery and an adapted LandTrendr-based workflow, we achieved reliable month-resolved estimation of coppiced-stand age for Eucalyptus spp. plantations in Luzhai County, Guangxi, China. The workflow achieved 83.2% overall accuracy and a mean absolute timing error of 1.58 months, enabling reliable month-scale harvest dating for short-rotation plantations.

Context

Short-rotation Eucalyptus plantations are frequently harvested, and annual composites often fail to identify harvest timing precisely. Month-resolved stand age is therefore important for plantation management and carbon assessment.

Aims

We developed and validated a monthly stand age estimation framework by combining Harmonized Landsat–Sentinel time series with a LandTrendr-based workflow and multi-index fusion.

Methods

We used monthly Harmonized Landsat–Sentinel composites from 2019 to 2024 to derive four spectral indices. After index screening with confirmed harvest samples, we applied LandTrendr to each index time series and used multi-index consensus fusion with a ± 1-month tolerance rule to refine stand age estimates. Performance was then evaluated against reference samples using accuracy metrics and mean absolute error.

Results

Monthly inputs improved harvest detection compared with annual inputs, increasing overall accuracy from 82.2% to 89.4%. Multi-index fusion further improved stand age estimation accuracy from 76.8% to 83.2% and reduced mean absolute error from 2.29 to 1.58 months.

Conclusion

Monthly Harmonized Landsat–Sentinel time series combined with LandTrendr enabled accurate and cost-effective month-resolved stand age assessment for Eucalyptus plantations. The workflow can support harvest scheduling, plantation inventory updating, and carbon-related assessments using free satellite data and cloud-based processing.