ANALYSIS: Battered US cornfields open new front for disruptive tech

With the US still facing huge uncertainty over the state of its corn plantings, there is an array of technology that is manoeuvring to unseat the traditional survey-based approach that is the hallmark of the USDA.

This harvest could be the watershed moment that propels disruptive tech into the vanguard of crop progress and planting reports.

And yet, for all the wonders of satellite imagery and digital technology, agricultural markets will still have to wait for at least another two or three weeks before a clear picture of the state and size of the US corn crop emerges.

On July 11 the USDA confounded analysts by again revising upwards its estimate of the corn acreage to 91.7 million acres – the same estimate it made in late June, a survey that had already been widely discounted.

The domino effect of that estimate has spawned phantom bushels, which  pushed next year’s ending stocks to 51 million mt, a huge 8.5 million mt clip higher than what analysts had expected.

With futures rising again on Monday to hit a five-year high in the face of what should have been a bearish Wasde report, it raises the question as to how faulty the USDA model of estimating crops is.

Highly sceptical

“The market is highly sceptical of the USDA’s increase in acreage and production,” said futures brokerage ED&F Man in a report to clients, adding that the market simply doesn’t believe the ending stocks figure.

In theory, it should be an opportune time for a slew of companies that rely on remote sensing and satellites to showcase their technology and deploy it to estimate the size and quality of the crop to front-run the more traditional market-moving Wasde report.

But even they have struggled to pin down a figure after weeks of rain that has seen unprecedented delays in planting the world’s biggest corn crop across the US.

Indigo Atlas Insights, which markets itself as delivering a “living map of the world’s food system” called for corn plantings on an 81-89 million acre area, with beans at 89-95 million acres in its June report, at a time when Rabobank was calling for a precise 86.7 million acres.

Indigo didn’t want to comment on why that range was so wide.

But subsequently, it narrowed its findings to a single, outright figure for yield and area in its July update – 85.4 million acres, at 159.4 bushels/acre.

Meanwhile other companies, such as Maxar, say more time will be needed before it can give an accurate picture of the crop.

“We are not yet able to use satellite imagery to estimate how much land is planted versus unplanted. With some soybean planting still to be finished and much of the corn crop still too short to be identified by satellite, it is just too early to use imagery,” Kyle Tapley, a senior agricultural meteorologist at Maxar told Agricensus.

“The satellite data is getting more popular… acreage is of course a huge deal, but I see more interest in the weather-related yield models,” Kelly Herrick of Advance Trading said, highlighting that the development in seed technology makes the results hard to gauge.

“The genetic curve is very difficult to stay ahead of right now; we seem to be able to test the agronomic limits of a plant each year and are still surprised,” he said.

Alongside that, the USDA earlier this month confirmed US farmers can plant cover crops on land claimed under prevent planting and still be eligible for a bailout fund, known as the Market Facilitation Program.

With so many moving parts, the reality is that the market will have to wait for another few weeks before getting a handle on what yield impact the heavy rain and sodden fields followed by dry, warm weather has had on corn.

Getting there

In four-weeks’ time, it is likely the USDA will republish the June acreage report after surveying farmers across 18 states in a bid to provide more transparency on likely production.

For companies such as Maxar, Indigo and Geosys – the latter which provides remote-sensing solution tools by reviewing satellite data daily – the goal will be to provide market-moving forecasts ahead of that survey.

“We can start evaluating crop health as soon as plants have sprouted… (and) use the data to build NDVI (normalized difference vegetation index) curves. As we watch the curve for the current year grow, we compare that to curves in our extensive historical database,” Emily Negrin of Geosys told Agricensus.

“If the current year growth curve is similar to a historical year, we can anticipate similar yields to that reference year,” she added.

However, others highlight that – in precedent terms – this season may be out on its own.

“Given the record slow planting pace, we are in uncharted territory this growing season. The uncertainty this has created will certainly make imagery-based analyses of acreage much more relevant to the ag market this year,” Tapley said.

“I’m sure the technology is getting there, but with all the variances and if looking at how ‘green’ a pixel of satellite data is – the ability to identify an acre of corn, soybean, cotton, sorghum this year might be very difficult,” Herrick said.