This blog post is based on the 23rd Practice Abstract from the second batch of Farmtopia Practice Abstracts.
Water scarcity, irregular rainfall and heat events are making irrigation decisions increasingly important for almond farmers in Spain. At the same time, farms can already have access to large amounts of useful information – from soil moisture sensors and weather forecasts to satellite data, irrigation records and field observations. The challenge is turning all of this information into a clear picture of what is happening in the orchard and what action may be needed. Through the GAIA Open Call Sustainable Innovation Pilot, Farmtopia explored how artificial intelligence can help bring these different data sources together and support more informed irrigation and field-management decisions.
Bringing farm data together
GAIA, developed by Xilbi Sistemas de Informacion SL, was tested on an almond farm in Salamanca, Spain, together with farmer Gregorio Garcia Sanchez.
Instead of requiring farmers to manually compare information from different systems, GAIA brings several data sources into one irrigation-oriented workflow. Its AI analyses information including:
- soil moisture;
- weather conditions and forecasts;
- expected rainfall;
- field context;
- Earth Observation indicators; and
- irrigation information.
Based on these inputs, the platform generates parcel-level recommendations and explains the main factors behind each recommendation. Importantly, the farmer remains in control. Recommendations can be confirmed, adjusted or rejected rather than being applied automatically. A conversational interface also allows farmers to request field summaries and irrigation guidance using plain language.
From data to practical irrigation decisions
The GAIA pilot was configured around the farm’s existing sensors and irrigation practices rather than requiring a completely new digital infrastructure. This is an important aspect of the approach. For farmers already collecting data, the challenge is often not the absence of information but the time and effort required to interpret it. By combining different sources, GAIA aims to give farmers faster access to relevant information and clearer priorities for individual parcels.
The pilot reported the following changes:
- Water use decreased by 21.6%, from 5,000 to 3,920 m³/ha.
- Operational costs decreased by 16.7%, from €1,650 to €1,375/ha.
- The fertiliser-loss-risk proxy decreased by 32.9%, from 170 to 114 kg N/ha.
These figures reflect the results observed during the pilot. They should be interpreted in the context of the specific farm, season and management conditions rather than as results that can automatically be expected on every almond farm.
Keeping the farmer in control
One of the practical principles behind GAIA is that AI should support the farmer’s decision-making rather than replace it. Before acting on a recommendation, farmers should consider the recommendation rationale alongside current weather conditions, field observations and their own knowledge of the orchard.
For reliable day-to-day use, the pilot also highlights several basic requirements:
- Check sensors and meters regularly. Soil sensors and water meters should be correctly installed, mapped to the appropriate parcels and periodically checked or calibrated.
- Keep farm records up to date. Consistent recording of irrigation events and field observations helps provide better information for decision support.
- Review the reasoning behind recommendations. AI-generated guidance should be considered together with current field conditions before an irrigation decision is made.
- Use AI as decision support. The farmer should retain the final decision on whether and how to act.
A more manageable way to use existing information
For almond farms facing increasing pressure on water resources, tools such as GAIA show how AI can help make existing farm data more useful. Rather than adding another isolated source of information, the platform brings sensor data, weather information, Earth Observation indicators and farm records into a single workflow. The result is intended to be a more timely and understandable basis for irrigation decisions. The Salamanca pilot demonstrates the potential of this approach to support more efficient water use, lower operational costs and clearer parcel-level decision-making, while keeping farmers at the centre of the process.
Based on Farmtopia Practice Abstract 23
Official title: An AI-Powered Decision Support Tool for Sustainable Irrigation and Field Management in Almond Farming (GAIA OC SIP)
Publisher: Xilbi Sistemas de Informacion SL
Author: Pedro Branco
Contact: Xilbi Sistemas de Informacion SL contact pag