An Orchard Is Not a Single Image

A single satellite image can show that a block is green, dry, bright, or darker than the block next to it. It cannot reliably explain whether that difference belongs to the crop, the soil, a recent irrigation event, a shadow, or a cloud that survived the first round of masking. This is why orchard mapping has gradually moved away from the idea that one well-timed image is enough. A fruit tree is not only an object in space; it is a plant with a seasonal rhythm, and that rhythm is often more distinctive than its colour on any one day.

Sentinel-2 makes that rhythm observable at a useful regional scale. Its surface-reflectance products combine visible, near-infrared, red-edge, and short-wave infrared information with a revisit interval that can be close to five days under favourable conditions, although clouds and shadows reduce the number of observations that can actually be used. When those observations are arranged through the year, an orchard begins to look like a curve: it greens up at a particular time, reaches a particular level, holds that level for a while, and then declines or remains evergreen. Research on orchards, vineyards, olives, apples, and hazelnuts repeatedly finds that these changes over time help separate crops with similar single-date spectra. See Copernicus Sentinel-2 documentation, Abubakar et al., and Sengul et al..

The practical lesson is modest but important: a satellite time series should first be used to compare the behaviour of blocks, not to pretend that it has inspected every tree. A useful orchard map starts with clean observations, a known field boundary, and enough history to tell a persistent pattern from a noisy day. It can then help a team decide where a closer walk is worth the time. The map becomes more valuable when it makes the next field action clearer, while leaving diagnosis to the evidence collected on the ground.