EGU General Assembly 2022
© Author(s) 2022. This work is distributed under
the Creative Commons Attribution 4.0 License.

A comparison of different NDVI based methods for grapevines Kcb

Benoit Huet2,1, Michel Le Page2, David Tous3, Pascal Fanise2, Sylvie Duthoit4, and Joaquim Bellvert5
Benoit Huet et al.
  • 1ONERA, Toulouse, France
  • 2IRD, CESBIO, Toulouse, France
  • 3Safsampling, Verdu, Spain
  • 4TerraNIS, Toulouse, France
  • 5IRTA, Lleida, Spain

The basal crop coefficient (Kcb) is the ratio of crop evapotranspiration that primarily corresponds to transpiration. A comparison of different approaches using remote sensing observations for the estimation of Kcb of grapevines is carried out. The study is done over a cv.Tempranillo vineyard located in Catalunya, Spain from March to July 2021.

The different approaches tested are the following: 1- a linear relation between NDVI and Kcb(Campos et al., 2010), 2 and 3 - the two different approaches proposed in (Allen and Pereira, 2009), where Kcb is estimated thanks to a density factor. The first version lies on an exponential of Leaf Area Index (LAI), the second version lies on the tree height and the fraction of effective exposed area. 4 and 5- an intercepted photosynthetic radiation (fiPAR) model (Oyarzun et al., 2007) using inferred crop height and width is related to Kcb through the (Lebon et al., 2003) and (Picón-Toro et al., 2012) proposals. 6- The generic tabulated approach proposed by (Allen et al., 1998) is also used to compare to a reference, however it must be remembered that those tabulations are only indicative.

The different approaches are compared to the actual Kcb retrieved from a flux tower and the reference evapotranspiration of a nearby weather station. The resulting Kcb are injected into a water budget and daily evapotranspirations are finally compared to actual measurements.

The simple linear method did not transfer well on this particular vineyard. The "Allen& Pereira LAI" and the "Oyarzun/Picon" had the best performance with a respective r2, RMSE of 0.57, 0.60 and 0.54 0.62 on evapotranspiration estimates. However, the former approach through LAI does not seem really operational. The later method only needs to be parameterized with some easy to retrieve descriptions of the plot and plantation. In any case, a drawback of these NDVI based methods is the possible appearance of adventices, or the presence of a inter-row crop. Those must be withdrawn from the NDVI signal of the grapevine.

How to cite: Huet, B., Le Page, M., Tous, D., Fanise, P., Duthoit, S., and Bellvert, J.: A comparison of different NDVI based methods for grapevines Kcb, EGU General Assembly 2022, Vienna, Austria, 23–27 May 2022, EGU22-10231,, 2022.


Display file

Comments on the display

to access the discussion