From Satellite Change to the First Walk

The useful endpoint of orchard monitoring is not a coloured map; it is a better first decision. A grower or field lead usually does not need the satellite to name a disease from orbit. They need to know which block changed, how strong the evidence is, how much of the block is involved, and where a person should walk first. That is a narrower promise than diagnosis, but it is also a promise that can be tested against what the crew finds on the ground.

A field-first workflow can begin with a legal or confirmed field boundary, a clean Sentinel-2 observation window, and a comparison with the block’s recent history. It can then combine changes in vegetation and moisture indices with spatial concentration, observation coverage, and confidence. A result might say that a portion of the block has moved away from its recent baseline and deserves inspection, while also stating that the imagery cannot identify whether the cause is water, nutrition, disease, canopy damage, or management. That language is not a weakness. It prevents a remote-sensing signal from becoming an unsupported agronomic claim.

The next step is the part that turns monitoring into a working system: the field visit should record what was actually found, where the observation was made, and whether the satellite priority was useful. Over time, those confirmations can improve thresholds and local reference curves without pretending that a short initial window contains a complete history. For Farmbit, the product boundary should remain practical: show what changed, show how credible the signal is, and show where to walk first. Sentinel-2 is well suited to block-level change and scouting priority; it should not be presented as leaf-level disease detection, single-tree diagnosis, or proof of causal irrigation failure.