Welcome to the Global Spectra-Trait Initiative (GSTI).
This project aims to generate spectra trait models using reflectance data to predict leaf traits associated with the photosynthesis capacity of leaves. This includes the maximum carboxylation rate of rubisco (Vcmax), the maximum electron transport rate (Jmax), the dark respiration, as well as the prediction of leaf nitrogen, leaf mass per area (LMA), and leaf water content (LWC). We aim to gather datasets from multiple species and biomes worldwide to build generalizable spectra trait models.
If you want to participate, please email us or submit an issue in this GitHub repository. We welcome raw A-Ci data, dark-adapted respiration data as well as other structural and chemical leaf traits in a free format, and of course leaf reflectance data (ideally full range from 500 nm to 2400 nm). An overview of the data curation workflow that we follow is presented here: Curation workflow. It also details the data that is required to participate in the project.
More details on the processing chain to import, homogenize, and produce a standardized database from multiple datasets are given in this guide: Guide.
The project utilizes the data and metadata formatting recommendations presented in the Leaf-level gas exchange data and metadata reporting format (Ely et al, 2021).
The A-Ci fitting is based on the FvCB model of photosynthesis (Farquhar et al. 1980), as implemented and parametrized in CLM4.5, and detailed here: FvCB equations and parametrization.
Only free use data (CC BY 4) accepted. We request that users cite the Zenodo DOI for this repository, and strongly encourage them to (i) cite all dataset primary publications, and (ii) involve data contributors as co-authors when possible.
All data contributors will be included in an introductory paper planned for 2024.
GSTI is not designed to be, and should not be treated as, a permanent data repository. It is a community resource of standardized spectra-trait datasets to facilitate a living set of algorithms that can be used by researchers to predict a host of leaf traits using spectral measurements. It is not an institutionally-backed repository like Figshare, DataONE, ESS-DIVE, etc. We recommend (but not require) depositing your data in one of these first, and providing its DOI in your dataset metadata.
The raw datasets are stored in individual folders Datasets.
The curated GSTI database is available here Database.
Examples of generalized spectra trait models for the prediction of Vcmax25, Jmax25, and Rdark25 using the GSTI database are presented here: Spectra trait models. These models were obtained using partial least squares regressions following Burnett et al.'s 2021 best practices.
More to come
Burnett, A.C., Anderson, J., Davidson, K.J., Ely, K.S., Lamour, J., Li, Q., Morrison, B.D., Yang, D., Rogers, A. and Serbin, S.P., 2021. A best-practice guide to predicting plant traits from leaf-level hyperspectral data using partial least squares regression. Journal of Experimental Botany, 72(18), pp.6175-6189.
Ely KS, Rogers A, Agarwal DA, Ainsworth EA, Albert LP, Ali A, et al. A reporting format for leaf-level gas exchange data and metadata. Ecol Inform. 2021;61: 101232. https:doi.org/10.1016/j.ecoinf.2021.101232
Farquhar, G.D., von Caemmerer, S. & Berry, J.A. A biochemical model of photosynthetic CO2 assimilation in leaves of C3 species. Planta 149, 78–90 (1980). https://doi.org/10.1007/BF00386231