Remote Sensing Estimation of Plant Diversity in Sandy Ecosystem Based on Sentinel-2 Data

Abstract

Plant diversity is a key indicator of ecosystem structure, function, and restoration status, yet its rapid assessment remains challenging in sandy ecosystems where vegetation is sparse, spatially heterogeneous, and strongly affected by exposed soil backgrounds. In such environments, conventional greenness-based spectral indices may not adequately capture species-level variation because plant communities are controlled not only by photosynthetic biomass but also by soil moisture, micro-topography, and dune-related habitat heterogeneity. This study evaluated the potential of Sentinel-2-derived spectral indices for estimating plant α-diversity in the Hunshandak Sandland, northern China. Based on field observations from 888 plots collected during 2017–2024, four α-diversity metrics—species richness, Shannon–Wiener index, Simpson index, and Pielou evenness index—were calculated and compared with 21 spectral indices using correlation analysis, partial least squares regression (PLSR), and random forest (RF) models. The results showed that model performance varied substantially among diversity metrics. Species richness was estimated with the highest accuracy, whereas Shannon–Wiener, Simpson, and Pielou indices showed weaker predictability, indicating that remotely sensed spectral indices were more sensitive to species number than to abundance distribution and evenness. Moisture- and soil-background-sensitive indices, including the Normalized Difference Water Index (NDWI), Modified Normalized Difference Water Index (MNDWI), Bare Soil Index (BSI/BRI), and Chlorophyll Absorption Ratio Index (CARI), showed relatively stable relationships with plant diversity across different vegetation gradients. Although the overall explanatory power was moderate rather than high, the results demonstrate the practical value of Sentinel-2 spectral indices for regional screening of plant diversity patterns in sandy ecosystems. This study provides empirical evidence for biodiversity monitoring and ecological restoration assessment in semi-arid sandy landscapes and highlights the need to integrate environmental covariates, multi-source remote sensing, and phenological information in future studies.

Plant phenotyping relevance

Sentinel-2のスペクトル指標と回帰・機械学習により、圃場プロットの植物α多様性を推定・検証する測定ワークフローが研究の中心であり、単なる生物学的実験のルーチン測定ではない。

Code and data availability

The supplied article blocks describe field diversity data (888 plots, 2017–2024), Sentinel-2 composites, and PLSR/RF analyses, but contain no data availability statement, no public repository deposit, and no author code URL. No paper-specific public asset is identified.

No evidence-backed public reproduction asset is currently recorded.