Application of remote sensing technology to estimate productivity and assess phylogenetic heritability

Abstract

Premise Measuring plant productivity is critical to understanding complex community interactions. Many traditional methods for estimating productivity, such as direct measurements of biomass and cover, are resource intensive, and remote sensing techniques are emerging as viable alternatives. Methods We explore drone-based remote sensing tools to estimate productivity in a tallgrass prairie restoration experiment and evaluate their ability to predict direct measures of productivity. We apply these various productivity measures to trace the evolution of plant productivity and the traits underlying it. Results The correlation between remote sensing data and direct measurements of productivity varies depending on vegetation diversity, but the volume of vegetation estimated from drone-based photogrammetry is among the best predictors of biomass and cover regardless of community composition. The commonly used normalized difference vegetation index (NDVI) is a less accurate predictor of biomass and cover than other equally accessible vegetation indices. We found that the traits most strongly correlated with productivity have lower phylogenetic signal, reflecting the fact that high productivity is convergent across the phylogeny of prairie species. This history of trait convergence connects phylogenetic diversity to plant community assembly and succession. Discussion Our study demonstrates (1) the importance of considering phylogenetic diversity when setting management goals in a threatened North American grassland ecosystem and (2) the utility of remote sensing as a complement to ground measurements of grassland productivity for both applied and fundamental questions.

Plant phenotyping relevance

ドローン遠隔センシングとフォトグラメトリで植物群落の生産性・バイオマス・被覆を推定し、地上測定との予測性能を比較検証しており、植物形質取得手法が中心的です。

Code and data availability

The authors publicly deposited the scripts and data used in this drone-based prairie phenotyping study (biomass, cover, vegetation index measurements, trait data) on GitHub and archived on Zenodo, with explicit availability statements and URLs in the article.