What Free Satellites Can (and Cannot) See in a Rice Field

A rice field, seen from orbit, is a surprisingly rich document. The European Union's Sentinel-2 satellites pass over most of the world every few days and give the imagery away for free. The question for anyone building agricultural intelligence is not whether the data exists. It is what that data can honestly tell you — and what it cannot.

What the free bands can see

Sentinel-2 reads the landscape in around a dozen spectral bands at up to ten-metre resolution, including the red-edge and near-infrared light that vegetation reflects most tellingly. From that you can track how vigorous a canopy is, spot standing water and flooding, watch a crop green up and fade across a season, and flag stress across a district long before it is obvious from the roadside. For rice, tea, and coconut grown at scale, that is a genuine early-warning layer.

What it cannot

It cannot see a single struggling plant. It cannot look under the canopy, name a pest, or tell you why a patch is stressed — only that it is. And it does not see on demand: revisit gaps and cloud cover, common over monsoon Asia, mean the clearest image is sometimes a week old. Treating a ten-metre pixel as ground truth is the fastest way to give a farmer confident, wrong advice.

The useful middle

The value is in the honest combination — free satellite imagery for scale and early signal, ground data and local knowledge for detail, and models that know the difference. Our crop-stress work treats Sentinel-2 as one instrument among several, not an oracle. A first pilot is targeted within six months [TARGET]. Naming what the satellite cannot see is the only way to trust what it can.