en 483 concentration. The application of our mapping efforts to all the scenes used from DESIS and EMIT are 484 available at Harvard Dataverse. 485 486 Data, Materials, and Software Availability 487 488 Forest inventory data were obtained from the USDA Forest Service’s FIA Program and are available 489 through the FIA DataMart (https://apps.fs.usda.gov/fia/datamart/datamart.html). However, as noted in the 490 Methods, we used a federally protected version of the FIA database to access actual plot locations for our 491 analyses (for more information on federally protected FIA data, see 492 https://research.fs.usda.gov/programs/fia/sds). All code associated with this research is available on 4
Open resource ↗FIA DataMart · pdf-raw-page:14 lines:1-100Unverified paper record
Mapping multiple dimensions of forest diversity using spaceborne spectroscopy
31 Aug 2025 · 10.32942/x24h1d
Abstract
Observing biodiversity across space and time is essential for advancing and verifying conservation efforts toward global biodiversity and sustainability goals. Spaceborne imaging spectroscopy has emerged as a revolutionary tool for quantifying and tracking forest diversity, yet its application at large spatial scales remains a central challenge. We develop a framework to map multiple dimensions of forest community composition and diversity by integrating imaging spectroscopy from two spaceborne sensors (DESIS and EMIT) with taxonomic, phylogenetic, and functional trait datasets, and 43,155 forest inventory plots across the Eastern United States. We find that spectral dissimilarity among forest communities is positively correlated with β-diversity matrices of compositional dissimilarity. We then show that imaging spectroscopy can be used to predict ordination axes of β-diversity and to map multiple dimensions of forest diversity at high spatial resolution (30 or 60 m). Predicted β-diversity axes can be used to model forest attributes, including forest types, plant lineages, and community plant traits. On average, β-diversity axes explain more than 48% of the variance—outperforming climatic and topographic predictors—and enable accurate mapping of 95 forest attributes. Our framework shows that spaceborne imaging spectroscopy, when combined with inventory data, allows indirect yet comprehensive observation of forest diversity attributes across broad spatial extents. This integrative approach sets the stage for scalable forest monitoring in support of global biodiversity conservation and forthcoming satellite missions.
Plant phenotyping relevance
宇宙空間イメージング分光と在庫データを統合し、森林群集の多様性や植物形質を推定・マッピングする枠組みが研究の中心であり、植物状態の大規模な表現型推定に該当する。
abstractWe develop a framework to map multiple dimensions of forest community composition and diversity by integrating imaging spectroscopy from two spaceborne sensors (DESIS and EMIT) with taxonomic, phylogenetic, and functional trait datasets, and 43,155 forest inventory plots across the Eastern United States.
abstractWe then show that imaging spectroscopy can be used to predict ordination axes of β-diversity and to map multiple dimensions of forest diversity at high spatial resolution (30 or 60 m).
abstractPredicted β-diversity axes can be used to model forest attributes, including forest types, plant lineages, and community plant traits.
Code and data availability
The paper's plant-phenotyping/community-composition analysis relies on FIA forest inventory data (public via FIA DataMart), author analysis code publicly hosted on GitHub, and paper-specific spaceborne data products (Level 3/4 maps of β-diversity and forest attributes) released via Harvard Dataverse. The SDS link only
). However, as noted in the 490 Methods, we used a federally protected version of the FIA database to access actual plot locations for our 491 analyses (for more information on federally protected FIA data, see 492 https://research.fs.usda.gov/programs/fia/sds). All code associated with this research is available on 493 GitHub (https://github.com/Antguz/mapping-communities), and will be archived in Zenodo under version 494 1.0 upon publication. Data that do not compromise federally protected information are being prepared for 495 release in the Harvard Dataverse. The spaceborne data products developed in this research are also 496 available through the Harvard Dataverse 497 (https://datavers
Open resource ↗GitHub · Antguz/mapping-communities · pdf-raw-page:14 lines:1-100z/mapping-communities), and will be archived in Zenodo under version 494 1.0 upon publication. Data that do not compromise federally protected information are being prepared for 495 release in the Harvard Dataverse. The spaceborne data products developed in this research are also 496 available through the Harvard Dataverse 497 (https://dataverse.harvard.edu/previewurl.xhtml?token=cfb44b92-ec7f-4cd8-9c7c-c2b4b43612d6).498 499 Acknowledgments 500 501
Open resource ↗Harvard Dataverse · pdf-raw-page:14 lines:1-100This is an automatically classified, unverified record. Curator approval is required before any resource enters the Catalog.