Scholarlink

Context <p>The spectral variation hypothesis (SVH) suggests that remotely sensed spectral variation correlates with biodiversity. However, conflicting evidence has been presented regarding the SVH, it has not been widely tested in high-latitude landscapes, and its assessments mainly employ species richness, omitting other aspects of biodiversity.</p> Objectives <p>We examine plant diversity, spectral variation and the SVH across a northern boreal landscape utilizing high-spatial-resolution satellite imagery. We incorporate taxonomic, phylogenetic, functional, and indicator-based plant diversity metrics and visualize spatial differences between proposed plant diversity proxies.</p> Methods <p>We surveyed plant community composition in peatlands, forests, and tundra in northwest Finland and calculated 14 diversity metrics for&#xa0;355 sites. We compared observed diversity with (1) spatial, (2) temporal, (3) multivariate spatiotemporal spectral variation, and (4) multivariate machine learning. We used statistical relationships and machine learning models to create continuous maps of plant diversity.</p> Results <p>While forests were generally more biodiverse than peatlands or tundra, each of these ecosystems obtained the highest values in some metrics. Spectral variation had positive correlations (0.14–0.72) across most dimensions of plant diversity, but relationships varied between ecosystems and were sometimes insignificant or negative. For example, plant diversity in forests was largely uncorrelated with spatial spectral variation. Temporal spectral variation was more strongly correlated with plant diversity than spatial spectral variation in a majority (41 of 56) of cases.</p> Conclusions <p>While spectral variation was usually correlated with plant diversity, correlations varied among peatlands, forests, and tundra. Thus, methods for estimating plant diversity should be tailored to different biodiversity metrics and ecosystem types.</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

Mapping plant diversity in a northern boreal landscape using remotely sensed spectral variation

  • 期刊论文

  • 出版模式：

    开放获取

  • 发表日期：

    2025年10月31日

查看全文
  • Pauli Putkiranta,
  • Aleksi Räsänen,
  • Miska Luoto,
  • Heidi Mod,
  • Tarmo Virtanen

摘要

Context

The spectral variation hypothesis (SVH) suggests that remotely sensed spectral variation correlates with biodiversity. However, conflicting evidence has been presented regarding the SVH, it has not been widely tested in high-latitude landscapes, and its assessments mainly employ species richness, omitting other aspects of biodiversity.

Objectives

We examine plant diversity, spectral variation and the SVH across a northern boreal landscape utilizing high-spatial-resolution satellite imagery. We incorporate taxonomic, phylogenetic, functional, and indicator-based plant diversity metrics and visualize spatial differences between proposed plant diversity proxies.

Methods

We surveyed plant community composition in peatlands, forests, and tundra in northwest Finland and calculated 14 diversity metrics for 355 sites. We compared observed diversity with (1) spatial, (2) temporal, (3) multivariate spatiotemporal spectral variation, and (4) multivariate machine learning. We used statistical relationships and machine learning models to create continuous maps of plant diversity.

Results

While forests were generally more biodiverse than peatlands or tundra, each of these ecosystems obtained the highest values in some metrics. Spectral variation had positive correlations (0.14–0.72) across most dimensions of plant diversity, but relationships varied between ecosystems and were sometimes insignificant or negative. For example, plant diversity in forests was largely uncorrelated with spatial spectral variation. Temporal spectral variation was more strongly correlated with plant diversity than spatial spectral variation in a majority (41 of 56) of cases.

Conclusions

While spectral variation was usually correlated with plant diversity, correlations varied among peatlands, forests, and tundra. Thus, methods for estimating plant diversity should be tailored to different biodiversity metrics and ecosystem types.