Alfalfa yield estimation using the combination of Sentinel-2 and meteorological data

dc.contributor.authorGámez Guzmán, Angie Lorena
dc.contributor.authorSegarra, Joel
dc.contributor.authorVatter, Thomas
dc.contributor.authorSantesteban García, Gonzaga
dc.contributor.authorAraus, José Luis
dc.contributor.authorAranjuelo Michelena, Iker
dc.contributor.departmentAgronomía, Biotecnología y Alimentaciónes_ES
dc.contributor.departmentAgronomia, Bioteknologia eta Elikaduraeu
dc.contributor.departmentInstitute for Multidisciplinary Research in Applied Biology - IMABen
dc.contributor.funderGobierno de Navarra / Nafarroako Gobernua
dc.date.accessioned2025-06-11T10:18:48Z
dc.date.available2025-06-11T10:18:48Z
dc.date.issued2025-03-19
dc.date.updated2025-06-11T10:15:15Z
dc.description.abstractContext: Alfalfa (Medicago sativa L.) is one of the world's most important forages for livestock feeding. Timely yield estimates could provide information to guide management decisions to improve production. Since alfalfa crops typically undergo multiple harvests in a year and demonstrate rapid regrowth, satellite remote sensing techniques present a promising solution for alfalfa monitoring. Objective: To generate alfalfa yield estimation models at three phenological stages (early vegetative, late vegetative, and budding stages) using vegetation indices (VIs) derived from satellite Sentinel-2 images and their combination with meteorological data. Methods: We analyzed fields located in Navarre (northern Spain) over two consecutive seasons (2020 and 2021). To generate the yield estimation models, we applied a conventional multilinear regression and two machine learning algorithms (Least Absolute Shrinkage and Selection Operator - LASSO and Random Forest - RF). Results: Regardless of the statistical approach, the three phenological stages were not optimal when either VIs or meteorological data were used singularly as the predictor. However, the combination of VIs and meteorological data significantly improved the yield estimations, and in the case of LASSO model reached percentages of variance explained (R2) and normalized root mean square error (nRMSE) of R2= 0.61, nRMSE= 0.16 at the budding stage, but RF reached a R2= 0.44, nRMSE= 0.22 at the late vegetative stage, and R2= 0.36, nRMSE= 0.24 at the early vegetative stage. The most suitable variables identified were the minimum temperature, accumulated precipitation, the renormalized difference vegetation index (RDVI) and the normalized difference water index (NDWI). The RF model achieved more accurate yield estimations in early and late vegetative stages, but LASSO at bud stage. Conclusion: These models could be used for alfalfa yield estimations at the three phenological stages prior to harvest. The results provide an approach to remotely monitor alfalfa fields and can guide effective management strategies from the early development stages.en
dc.description.sponsorshipThis study was supported by equal financial contributions from the Government of Navarre and NAFOSA S.L. through the PhD grant provided to Angie L. Gámez. The authors would also like to thank José María Osés and Bernardo Monreal for overseeing management of the alfalfa crop during the course of the study. J. Segarra is a recipient of a POP postdoctoral fellowship from the Ministerio de Ciencia e Innovación, Spain (grant: PRE2020- 091907).
dc.format.mimetypeapplication/pdfen
dc.format.mimetypeapplication/msworden
dc.identifier.citationGámez, A. L., Segarra, J., Vatter, T., Santesteban, L. G., Araus, J. L., Aranjuelo, I. (2025). Alfalfa yield estimation using the combination of Sentinel-2 and meteorological data. Field Crops Research, 326, 1-11. https://doi.org/10.1016/j.fcr.2025.109857.
dc.identifier.doi10.1016/j.fcr.2025.109857
dc.identifier.issn0378-4290
dc.identifier.urihttps://academica-e.unavarra.es/handle/2454/54217
dc.language.isoeng
dc.publisherElsevier
dc.relation.ispartofField Crops Research (2025), vol. 326, 109857
dc.relation.publisherversionhttps://doi.org/10.1016/j.fcr.2025.109857
dc.rights© 2025 The Authors. Published by Elsevier B.V. This is an open access article under the CC BY license.
dc.rights.accessRightsinfo:eu-repo/semantics/openAccess
dc.rights.urihttp://creativecommons.org/licenses/by/4.0/
dc.subjectAlfalfaen
dc.subjectMeteorological variablesen
dc.subjectRandom foresten
dc.subjectSatelliteen
dc.subjectYield estimationen
dc.titleAlfalfa yield estimation using the combination of Sentinel-2 and meteorological dataen
dc.typeinfo:eu-repo/semantics/article
dc.type.versioninfo:eu-repo/semantics/publishedVersion
dspace.entity.typePublication
relation.isAuthorOfPublication8afeb09d-4593-4d40-8219-c00a6da4ef90
relation.isAuthorOfPublication70e95546-7fe8-4555-b366-74356bdb746e
relation.isAuthorOfPublicationb8dd84ae-83ed-4e3f-873e-b0023505b3df
relation.isAuthorOfPublication.latestForDiscovery70e95546-7fe8-4555-b366-74356bdb746e

Files

Original bundle
Now showing 1 - 2 of 2
No Thumbnail Available
Name:
Gamez_Alfalfa.pdf
Size:
6.58 MB
Format:
Adobe Portable Document Format
No Thumbnail Available
Name:
Gamez_AlfalfaYield_MatCompl.docx
Size:
337.37 KB
Format:
Microsoft Word XML
License bundle
Now showing 1 - 1 of 1
No Thumbnail Available
Name:
license.txt
Size:
1.71 KB
Format:
Item-specific license agreed to upon submission
Description: