The importance of reducing water consumption is paramount, especially as agriculture is estimated to account for over 70% of global water use. With food demands only rising, water use is expected to increase an additional 15% to meet this demand.

Technology companies, along with growers, have risen to the challenge of solving this complex issue through precision farming methods and visibility tools. One area of technology that is making strides towards increasing water efficiency in the field as well as in the greenhouse is artificial intelligence (AI).

Emerging technologies, devices and platforms enable us to collect and leverage unprecedented amounts of data from across multiple sources: historic rainfall patterns, aerial imagery, yield records, on-field sensors, etc. In return, the aggregated data can be processed and combined alongside forecast data (from market demand to weather) to help us make “intelligent” decisions based on the most accurate predictions we’ve ever had access to.

Identifying areas that are being overwatered or underwatered is key. A daily task, that can be tough to assess for any farmer or agronomist, is determining the right amount of water to get the optimal yield and quality. Depending on the type of plant, overwatering also carries risks. For example, over-irrigated cotton crops will lead to the growth of more leaves, instead of the cotton flowers that hold the value of the crop.

Farmers aim to create an optimal irrigation schedule for their crops that will optimize yield and quality, while keeping costs in check. Evapotranspiration has been a key metric to create an irrigation system that is tailored to the needs of a plant. It represents the sum of evaporation from the land surface plus transpiration from plants. Modern satellite imagery and weather predictions help farmers improve the assessment of evapotranspiration. However, breakthroughs in internet of things (IoT) sensor technology help inform much more incisive irrigation decisions by measuring the plants’ behaviour instead of (or in addition to) the soil and the weather.

Powerful AI engines are able to process and analyze data feeds from satellite, plane or drone imagery. Machine learning, and in particular deep-learning algorithms, can help us interpret data from images and identify patterns that spotlight irrigation issues (as well as other issues such as pests). If imagery is combined with soil and plant-based sensors, data can give us an extremely accurate read of the irrigation needs in real time – as well as alert us about potential issues.

Wasting water, especially in areas in which it is a scarce resource, is a huge headache (and expense) for farmers and food growers worldwide. While at the one end of the spectrum, there are technologies such as drip irrigation and sophisticated controlled environments such as soilless greenhouses, they involve technologies and systems that are costly and therefore not suitable for extensive agriculture (or lower value crops). One area that can be hugely improved upon is the discovery of malfunctions such as leaks on irrigation systems.

In the past, it might have taken a personal inspection to find a broken piece of equipment or identify a leak. IoT devices mean that the software itself can alert when something is wrong or suspicious – and this is only available when devices are interconnected. This way, an irrigation sensor can detect an irregularity and link it to the root issue or variable – especially if it is connected to other data points such as weather data so it can rule out other potential causes.

One grower I have worked with manages 2,630 hectares of farmland in east central Idaho, within a region where temperatures can fluctuate drastically as much as 25°C in two days. Controlling irrigation is their biggest challenge. With 80 irrigation pivots to turn on when it heats up, any issue can become a big problem as the ground dries up quickly. Using AI-based tools such as Valley Insights, they are able to access aerial visuals and other data on the field, including thermal imaging of each plant.

The captured imagery and AI field analysis is able to provide accurate alerts that pinpoint the exact problematic spots, determining irrigation issues that require immediate attention. This means they are able to tackle issues such as pivot-related leaks, which are difficult to detect with the naked eye. The power of AI goes beyond spotlighting an issue. It provides insights into how to rectify an irrigation irregularity.

In a similar way in which the introduction of autonomous cars is bound to change driving as we know it, agriculture and farming will be redefined within a decade with the adoption of AI-driven autonomous tools. While today the function of AI and predictive analytics is mostly to inform farmers’ decision-making processes, in a not-so-distant future machines will be able to operate autonomously.

Autonomous machines in agriculture won’t only take into account crop requirements. They will have the “intelligence” to take into consideration factors such as yield quality and financial considerations associated with energy costs, as well as other parameters. While irrigation and water consumption in general is an important place to start, this technology will also become a cornerstone for other agronomic processes including fertilization and crop protection.