Can we see from space if a vineyard needs water?

BigDataGrapes is about how the Grapevine industry can be supported with new Big Data technologies to generate new insights and allow informed decisions.

In the Farm Management Pilot Abaco together with Geocledian, KUL and CNR are developing a Dashboard on Water Availability and Irrigation Recommendations to help farmers decide on these critical business decisions:

  • How to increase productivity and grape quality?
  • How to reduce the use of emergency irrigation?
  • How to detect the moment in which the stress level is critical for the plant?
  • Which part of the vineyards really needs irrigation?

The first question: “How to increase productivity and grape quality?” represents a fundamental business philosophy. The balance between grape quality and quantity is perceived in a different way form farmer to farmer with respect also to the grape variety. Related to the first question, is also the third one: “How to detect the moment in which the stress level is critical for the plant?”. This is a philosophical question too. There is a very thin ridge that distinguishes a stress condition that can lead to an increase in grape quality and a stress condition that can damage the plant and of course the grape productivity and quality.

The Dashboard is fed with diverse data sets from the BigDataGrapes platform and utilizes new technologies to combine, analyse and convey the message to the user in a transparent and user-centred way: “what he should do”.

The Dashboard is based on a classical reservoir water balance modelling that is fed with meteorological data, soil data, and user entered irrigation amounts. The water balance model calculates the current water availability to the plants based on a simple formulation that balances the water input (precipitation and irrigation) with the soil water storage capacity and the output (evapotranspiration and percolation or runoff) which mainly depends on the soil type, vegetation status and weather conditions. This method is widely used and has proven valuable for millions of farmers around the world if used properly. But uncertainties can be introduced by the user if he enters e.g. wrong irrigation amounts or if assumptions about the soil properties or the vegetation status are inaccurate. In particular the plant phenological status is strictly related to the real water consumption (evapotranspiration).Therefore a new uncertainty aware and user centred method is currently developed to combine this approach with multispectral Earth observation data from the Copernicus Sentinel-2 satellite constellation that was started in 2015 and has since disrupted the remote sensing community around the world. This satellite provides high quality data in an unprecedented temporal frequency and resolution. Sensing in the visible, near infrared and shortwave infrared part of the electromagnetic spectrum it allows to measure the plant’s Chlorophyll absorption and near infrared light reflection by the leaf structure which is typically used to determine the vegetation status and health situation. In the shortwave infrared part of the spectrum the measurements are correlated with the vegetation water content which is used here. The advantage of using this type of data is that the satellite delivers an additional, independent measurement from space that allows to gain insights about vegetation status and the vegetation water content every few days. This can help to reduce the uncertainties for the user. Moreover the remote sensing data can provide “signals” of plant stress before the human eye can detect it. In this way decisions on best practice activities can be taken promptly and plant damages reduced.

In the next weeks an end user evaluation will take place to clarify with experts whether the newly developed system provides the expected benefits to the user.

Image: