The Payoffs of Precision Agriculture

by

By Peter Keen

Precision agriculture (PA) is a farming management concept that will have the same degree of impact on farming as the internet has had on retailing. The technology components break down into three main areas of functionality:

Capturing usable data: One of the main limitations of farming methods has been that they treat crops and fields as homogenous, whereas in reality, every plant, row, or patch is a little — or even a lot — different from its neighbor. PA uses technology to sharpen focus, zero in on, and narrow down, for example, land mapping, soil sampling, fertilization, pest and disease control, and weather alerts. It will pick out a puddle in a field or a change in gradient within a tea field for treatment.

IBM estimates that PA generates 500,000 data points per farm each day. The Farmer’s Edge service collects data points from sensors every 20 feet in 100,000 fields amounting to 100 million acres. It maintains a database of 25 billion records. Five years ago, that would have been big news. Today, it’s ho-hum. Farming has been constrained by what the grower can see. PA is able to “see” more.

Making it useful: PA analyzes and processes captured data to produce alerts, insights, and recommendations. The main areas of value to farmers are “hyperlocalized” weather forecasting, field mapping with advice on planting schedules, and yield forecasting. There’s a high degree of customization needed to meet the needs, environments, farmer, farm, and crop profiles. One instance of specialization is a smartphone app for fish-pond farmers in Iran that makes real-time recommendations for optimal feeding schedules; this has cut fish deaths by 30-40% within a year,

Translating information: In many ways, PA is a map creator for farmers to navigate by. It typically delivers visual displays, schematics, and graphics to farmers’ cell phones. This is acritical counterbalance to the massive increase in data PA creates; it must be made meaningful, communicative, and assimilable, especially among the many farming communities with varying levels of literacy and comfort with technology.

The economic advantages of PA

Just one striking area of PA technology innovation shows the extensive payoff. Machine vision breaks all the limits on “seeing.” Drones, sensors, and satellites break the limits of how much useable information can be captured. This creates the ability to apply fertilizers, pesticides, seeds, and water to a single targeted plant, delivered from a drone or an internet connected tractor. The drone is a flying farmhand that so accurately focuses spraying that it cuts water use by 30-50% and chemicals by 90%.

In its moves towards a “reforestation economy” in tea farming, Greenpop uses drones in areas of Malawi selected by satellite imaging to shoot tree seeds into the ground, with nutrients, at a rate 150 times faster than by hand: 40,000 a day. Planting costs are decreased by as much as 90% and weeding time is sped up by a factor of 5 with labor savings cut by a factor of 10.

In Tanzania, fertigation —precision fertilizer and water feeding —increased the average yield for three-year-old bushes by 25%; four-year old plants by 48%; and five-year old plants by 52%. Labor savings were 48% and power reduction 16% versus sprinklers. Fertilizer costs were lower in all instances. Less weeding was required. Shoots regenerated faster and bush mortality was lowered.

Heat imaging now spots leaf disease weeks before it is visible to human view. AI machine learning models routinely achieve levels of 96% plus in any area where PA data can be “seen” at the level of invisible light, molecular structure or genetic coding (DNA “fingerprints”), thereby transforming authentication and supply chain integrity,

The hitch: some assembly required

These are just a few examples of PA in operational use. But all the shiny drones and cameras, the compact sensors, and the telecommunications connections come with the equivalent of that existential and chilling footnote equivalent of your online Amazon purchase: “some assembly required.” The frustrating element of PA is that the tools are all ready to go but need infrastructures and support services, many of which pose extra burdens and challenges for the tea industry:

The comprehensive technology platform to have the tools work together. Think of these as all the invisible communication links, hardware, software, and data management systems behind your smart phone app.

Agencies to fund capital investment and operating costs. PA has high payoff for a relatively low financial investment. “Relatively” is very different for the large-scale farming units of a consumer foods firm and a cash-strapped smallholder operating on thin margins in a market of fierce global competitive and price pressures.

A services base to support PA deployment and effective use. The very purpose of PA is to change business processes and farming practices. That makes PA as much a matter of education, community, coalition-building, and trust as technology implementation supporting networks of hardware, software, and telecoms.

Platforms: the big tech players are in the game

The main components of the PA platform include dozens of types on the ground sensors, robotics, and equipment configured to be IoT-linked — the internet of things —plus above-the-ground telecoms, anywhere in the “cloud” data and software, and in the hand smart phones, laptops, and tractor controls. How they work together defines the PA technology platform.

Here, scale and experience dominate, and the leaders are all major players in technology that are building experience in farming. They have increasingly made small farms a new priority. Microsoft’s Farm Beat, for example, has added tools to specifically address the major problem of lack of high-speed internet telecommunications in rural regions — no one can cable the Himalayas. This includes its tethered eye arrays of liquid helium balloons floating 100 feet above the ground.

IBM’s Farm Trust, Amazon’s AWS, and China’s VE Chain are among the major platform providers with tea applications. VE Chain is very active in using blockchain and cryptocurrency as the government-endorsed, and even mandated, base for regional development. Its tool chain is tracking 200,000 bricks of expensive puehr tea to reduce counterfeiting.

Funding

This is a never-ending blockage to capital investment in the tea industry. While the financial returns are high and unit prices of individual PA technology tools attractively low, the total costs are just not affordable for many farmers. The situation seems unlikely to improve after the inevitable cash crunches and drying up of grants, international agency aids, and credit availability of the post-Covid virus pandemic environment.

The main costs are for hardware, around 70% of the total outlay. IoT sensors average around $350 according to Farm Beats. (Its platform cleverly cuts this down for limited soil data collection by altering the chipset circuit board of a standard Android smart phone to measure the time a signal takes to penetrate a given depth. That drops costs to $25.)

A basic copter drone costs as little as $1,500 with a camera for visual imaging just $200. Most drone configurations, though, are in the $8-15,000 range. Hyperspectral cameras that capture near-infrared invisible light are $500-5,000. A top-of-the-line LIDAR (light imaging and detection ranging) sensor/camera costs between $60,000-150,000 and an autonomous robot tractor $250,000 or more. The costs of a fertigation system are about $4,000 for a smallholder farm, but commercial scale automatic machines are $6,000.

The main sources of financing are from existing farm equipment lenders and government and international NGO development agencies, such as the UN International Fund for Agricultural Development (IFAD). The irony of PA today is that the highest returns on investment are from smallholder adoption, but the highest barriers are also for smallholder investment.

Services

Precision Agriculture demands new expertise both on the ground and from remote talent pools. It includes the management and dissemination of data, technical, and operational support for equipment, training, and research. Currently, service organizations constitute a small and slow growing fraction of PA investment. IBM estimates it will be just $140 million out of its forecast for 2025 of a $2.6 billion industry sector. Sensors will comprise 18%: about half a billion dollars.

Many services are provided by public sector organizations, especially research-based centers. Many of these are regionally focused.

The path to value creation

The value potential of PA is very clear. The challenge, of course, is how to turn potential into action. There are few established strategies to use as guides. There appear to be a few principles worth following:

Begin by recognizing the specific and proven opportunities relevant to your own business that offer high proven payoffs. Focus on projects that are larger than pilots but do not involve heavy investment, project management, and change management. Soil, weather, irrigation, and field management are the most likely candidates.

Ensure three barriers are recognized and leapt over: platform needs, funding, and services. Get a lift up from alliances and development agencies; do not try to hop over them from a standing start.