Our plots · Advice, yield and harvest
Advice, yield and water
Weekly advice, a yield comparison with last year, and five other satellite checks. Early results are marked experimental.
Job 2
Weekly advice
Each week FarmBot will write a short note for each plot. It will bring together the crop's stage seen from space, the pest check, rain against drying by sun and wind, and the newest official farm advice for Koraput. Every line will name its source. Anything that is an estimate will say so.
Crop stage
From the greenness line now, and from radar later. Shown on each plot page.
Water
Rain minus the water pulled out by sun and wind, for the past and coming week Estimate. Also the pond's water area Seen by satellite.
Pests and official advice
The pest check, with links to the matching official advice. The note won't give pesticide doses.
Job 3
Is this year's crop better or worse than last year's? Experimental
Satellites can't weigh grain. They can show how green and thick the crop was compared with other seasons and plots. FarmBot will use each season's greenest point, and its greenness added up over the whole season, as a yield comparison score. It is a comparison, not tonnes. The numbers below use only the 10 m squares inside the exact outlines. Technical name: relative yield index
| Plot | Middle greenness, 1 Oct–23 Oct 2025 | Same dates, 2026 | Score: 2026 divided by 2025 |
|---|---|---|---|
| Nilabadi Fishing Pond fields (A + B) | 0.84 (0.75–0.90, 6 photos) | 0.76 (0.59–0.91, 10 photos) | 0.90 |
| · Pond-area field A (north-east, by the road) | 0.85 (0.74–0.90, 6 photos) | 0.76 (0.60–0.92, 10 photos) | 0.90 |
| · Pond-area field B (south-west) | 0.82 (0.76–0.88, 6 photos) | 0.71 (0.56–0.89, 10 photos) | 0.87 |
| Agnigangamma Temple plot | 0.79 (0.71–0.85, 6 photos) | 0.87 (0.68–0.95, 8 photos) | 1.10 |
How to read this: the last column is the score. 1.00 means as green as last year. Above 1 is greener, below 1 is less green. The small numbers in brackets are the lowest and highest photo, and how many photos were used.
What this does and doesn't show
- The photos start on 1 Oct 2025. So the only dates both seasons share are 1–23 October, near the greenest point. Only those dates are compared here, not the whole season.
- Inside the exact outlines, the pond fields were less green this October than last (score 0.90). The temple plot was greener (1.10). The earlier rough squares had given 0.98 for the pond and 1.07 for the temple, but the pond square held almost none of these fields.
- The pond fields' crop is later this year, not necessarily weaker. Their greenness was still rising on 23 Oct (0.79, smoothed). The last two photos were close to 2025's values. The 2026 dates also include a few hazy, low photos (the brackets show the spread).
- The 2026 season isn't finished, so its greenest point and season total aren't known yet. Read these as early signs only.
Also planned · A
When will it be ready to harvest? Experimental
FarmBot watches how each plot's smoothed greenness changes each week. When it starts falling steadily after the greenest point, the crop is ripening. In 2025, a fall below 0.7 and then below 0.5 marked the harvest weeks.
Nilabadi Fishing Pond fields
In 2025 the smoothed greenness fell below 0.7 on 12 Nov and below 0.5 on 27 Nov. Latest smoothed value: 0.79 on 23 Oct, a change of +0.05 over the past week. Still getting greener. Not ripening yet. Seen by satellite
Agnigangamma Temple plot
In 2025 the smoothed greenness fell below 0.7 on 2 Nov and below 0.5 on 22 Nov. Latest smoothed value: 0.87 on 23 Oct, a change of +0.00 over the past week. At its greenest. Not falling yet. Seen by satellite
Last year's dates are a guide, not a forecast. The rice variety and the transplanting date change the timing. November cloud can also delay the signal.
Also planned · B
Weak patches inside a plot Experimental
For each 10 m square fully inside the outline, FarmBot compares its greenness with the field's middle value on the same day. It uses the cloud-free photos of 1–23 October in 2025 and 2026. A square in the lowest 20% in both years could be a weak patch. About 4% of squares would land there by chance.
What it shows
No weak patch is claimed in any plot. The squares that were low in 2025 were not the ones low in 2026. The match scores were -0.37 in pond field A, -0.04 in field B and -0.32 in the temple plot. A score near 1 would mean the same squares were low both times. Fewer squares were low in both years than chance would give.
Pond fields: only 85 and 37 squares are fully inside. So these satellite photos can't map patches within them. Drone or phone photos would be needed.
The earlier weak strip (old snapshot): the strip north of the pond in the old snapshot from 31 Oct 2026 (kept, not updated) came from the 600 m square around the pond. It lies outside both fields: none of its 127 squares are inside, and they are about 249 m away on average. So it says nothing about this land.
Where to put sensors: with no steady weak patch, the first soil-wetness sensors can go near the centre of each field, one per plot.
Also planned · C
Radar that sees through clouds
A radar satellite can see through monsoon cloud. Standing water on a paddy field makes the radar echo dark. The echo gets brighter as the crop grows. That shows when fields were flooded or waterlogged, and roughly when transplanting happened, in July and August when the photo satellites saw only cloud. This will feed the crop-stage and humidity parts of the pest warnings. It is a plan only: no radar data has been used yet.
Also planned · D
Is there enough pond water for a winter crop? Experimental
Could the pond water a low-water winter crop such as green gram, black gram or mustard? This compares the water in the pond with what a crop would need after the little winter rain.
1.30 ha
Middle pond water area in November 2025 (10 cloud-free photos, water score). It was 0.99 ha in March–May 2026. Seen by satellite
471 − 20 mm
Water pulled out by sun and wind, minus rain, from 1 Nov 2025 to 28 Feb 2026. That leaves a shortfall of 451 mm that a crop would need from the pond. Estimate From Open-Meteo.
≈ 2.9 ha per metre
Each metre of usable depth over 1.30 ha holds about 13,000 cubic metres of water. At that drying rate, it could water about 2.9 ha of crop.
Big unknowns
- Nobody knows how deep the pond is. The answer grows with the usable depth, so it has to be measured, for example with a water-level sensor or a marked pole.
- The figure assumes the crop uses water at the standard drying rate. It doesn't yet include each crop's real needs, losses in the field, water lost from the pond itself by drying and soaking away, or what the fish need. These will lower the hectares.
- Whether the farmers want a winter crop, and whether there are other water sources, is still to be confirmed.
Also planned · E · later
Compare with neighbouring fields
Comparing a plot only with itself can't tell a bad season from a weak field. Later, FarmBot will compare each plot's greenest point and season total with nearby Nilabadi fields in the same season. That turns the yield comparison into "better or worse than the neighbours". This comes with the village-wide view.