Nutrient prediction for tef (Eragrostis tef) plant and grain with hyperspectral data and partial least squares regression: Replicating methods and results across environments

Authors: Flynn, Kyle; FRAZIER, AMY - Arizona State University; ADMAS, SINTAYEHU - Ethiopian Biodiversity Institute

Submitted to: Remote Sensing
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: 8/30/2020
Publication Date: 9/2/2020
Citation: Flynn, K.C., Frazier, A.E., Admas, S. 2020. Nutrient prediction for tef (Eragrostis tef) plant and grain with hyperspectral data and partial least squares regression: Replicating methods and results across environments. Remote Sensing. 12(18). Article 2867. https://doi.org/10.3390/rs12182867.
DOI: https://doi.org/10.3390/rs12182867

Interpretive Summary: The advancement of science using the scientific method is fully dependent on the reproducibility and replication (R&R) of the methods utilized. R&R has become a key focus for many fields of science, but less attention has been given within the field of remote sensing science, especially when applying hyperspectral data. Hyperspectral data consist of electromagnetic spectrum reflectance values from an object, in this case, plant canopy. Given the extremely local environments in which many hyperspectral studies are conducted (e.g., agricultural field plots), purposeful attention to the repeatability of findings across study locales is necessary to ensure methods are generalizable. This study undertakes an investigation of the nutrient content of tef (Eragrostis tef), an understudied plant that is growing in importance worldwide due to both food and forage benefits, but does so within the context of the replicability of methods and findings across two study sites situated in different international and environmental contexts. The aims are to determine (1) whether calcium, magnesium, and protein of both the plant and grain can be predicted using hyperspectral data with partial least squares (PLS) regression with waveband selection, and (2) whether the findings are replicable across differing environments. Results suggest the method can produce high nutrient prediction accuracy for both the plant and grain in individual environments, but selection of wavebands for nutrient prediction was not successfully replicated across study areas. The findings suggest that when incorporating hyperspectral data in precision agriculture methods, the correlations should be found for the location of interest, specifically. Our findings highlight the need for greater attention to methods and results replication in remote sensing, specifically hyperspectral analyses, in order for scientific findings to be repeatable beyond the plot level.

Technical Abstract: Achieving reproducibility and replication (R&R) of scientific results is tantamount for science to progress, and it is also necessary for ensuring the self-correcting mechanism of the scientific method. Topics of R&R have sailed to the forefront of research agenda in many fields recently but have received less attention in remote sensing in general and specifically for studies utilizing hyperspectral data. Given the extremely local environments in which many hyperspectral studies are conducted (e.g., agricultural field plots), purposeful attention to the repeatability of findings across study locales is necessary to ensure methods are generalizable. This study undertakes an investigation of the nutrient content of tef (Eragrostis tef), an understudied plant that is growing in importance worldwide due to both food and forage benefits, but does so within the context of the replicability of methods and findings across two study sites situated in different international and environmental contexts. The aims are to determine (1) whether calcium, magnesium, and protein of both the plant and grain can be predicted using hyperspectral data with partial least squares (PLS) regression with waveband selection, and (2) whether the findings are replicable across differing environments. Results suggest the method can produce high nutrient prediction accuracy for both the plant and grain in individual environments, but selection of wavebands for nutrient prediction was not successfully replicated across study areas. The findings suggest that the method must be calibrated in each location, thereby reducing the potential to extrapolate methods to different areas. Our findings highlight the need for greater attention to methods and results replication in remote sensing, specifically hyperspectral analyses, in order for scientific findings to be repeatable beyond the plot level.