Satellite vegetation indices
Draw the field on the map and see the condition of the vegetation as recorded by Sentinel-2. The satellite passes roughly every five days and sees at 10 metre resolution.
The classic biomass index. Works excellently in arable crops with continuous cover.
Press the polygon or the rectangle icon, top left on the map, and draw around the field. With the polygon, click each corner and click the first point again to close it. With the rectangle, simply drag diagonally.
What each result means
How the number arises
A leaf does not reflect light uniformly. Chlorophyll absorbs red strongly, around 665 nm, because that is where photosynthesis is driven. In the near infrared, around 842 nm, the leaf has no use for the energy and reflects almost all of it — that reflection comes from the spongy structure of the mesophyll and from the air spaces between the cells.
A healthy, well hydrated leaf therefore shows a large difference between these two values. A leaf that yellows, wilts or senesces loses chlorophyll and its mesophyll structure collapses: absorption in the red falls, reflection in the infrared falls too, and the difference narrows. All vegetation indices measure that difference, normalised so as not to be affected by sun angle or the brightness of the day.
Sentinel-2 records 13 spectral bands — four at 10 metres, six at 20 and three at 60 — with a 290 kilometre swath and a revisit of five days from the pair of satellites. The four indices on this page use four of them.
| Index | Bands | Formula | What it measures | Resolution |
|---|---|---|---|---|
| NDVI | B08 (842 nm) B04 (665 nm) |
(B08−B04) (B08+B04) |
Density and vigour of green biomass | 10 m |
| NDRE | B08 (842 nm) B05 (705 nm) |
(B08−B05) (B08+B05) |
Chlorophyll content, indirectly nitrogen | 20 m |
| NDMI | B08 (842 nm) B11 (1610 nm) |
(B08−B11) (B08+B11) |
Water content of the canopy | 20 m |
| SAVI | B08, B04 + factor L |
(B08−B04)·1.5 (B08+B04+0.5) |
Vegetation, corrected for bare soil | 10 m |
NDVI — the biomass index
The oldest and most widespread. It takes values from −1 to +1, but in practice the numbers met in the field are far narrower.
| NDVI value | What it usually means |
|---|---|
| below 0 | Water, wet soil, paved surface |
| 0 – 0.15 | Bare soil, rock, freshly ploughed field |
| 0.15 – 0.30 | Sparse or just emerging vegetation, dry grass |
| 0.30 – 0.50 | Moderate cover, young crop, sparse orchard |
| 0.50 – 0.70 | Good, active vegetation in full growth |
| above 0.70 | Dense, vigorous canopy with full ground cover |
Once leaf area index exceeds about 3, that is when the canopy becomes dense and leaves start shading one another, NDVI stops rising. A maize crop with normal nutrition and an over-fertilised one beside it will give almost the same NDVI, around 0.85. The index has "filled up" and no longer discriminates. From mid-season onwards NDVI loses much of its discriminating power — and that is exactly where NDRE comes in.
NDRE — the chlorophyll index
Band B05, at 705 nm, lies in the so-called red edge: the narrow region of the spectrum where leaf reflectance rises abruptly from the low level of the red to the high level of the infrared. The position and slope of this "step" shift with the chlorophyll concentration in the leaf.
The practical difference from NDVI is twofold:
- It does not saturate. Red edge radiation is absorbed less strongly, so it penetrates deeper into the canopy and brings information from the lower leaves as well. The index keeps discriminating even in a dense canopy.
- It relates to nitrogen. Because most of the leaf nitrogen sits in Rubisco and in the chlorophylls, chlorophyll content is an indirect indicator of nitrogen nutrition. That is why NDRE is used internationally for variable rate fertilisation maps.
In fruit trees it is the right choice. An orchard has a dense canopy in which NDVI saturates early, while NDRE continues to show differences between trees. Note though that its resolution is 20 metres, four times coarser in area than the 10 of NDVI.
NDMI — the water status
It uses band B11 at 1610 nm, in the shortwave infrared. There, water absorbs strongly: the more water the leaf cells hold, the less radiation returns to the satellite. The index falls when the canopy starts to lose water.
Its value is that it precedes the visible symptom. The drop in water content appears in the spectrum before the leaf wilts or changes colour. Combined with NDRE it separates two states that look alike: a fall in NDRE with NDMI steady points to nutrient deficiency or senescence, while a simultaneous fall in both points to water stress.
SAVI — when the soil shows
NDVI tacitly assumes the soil plays no part. When cover is low, however — a young planting, a sparse orchard, the start of the season — bare soil occupies a large part of the pixel and its reflectance distorts the index. Dry pale soil and wet dark soil give a different NDVI for exactly the same vegetation.
SAVI introduces the soil adjustment factor L, with an established value of 0.5 for intermediate cover, and multiplies by (1+L) so that the range of values stays comparable with NDVI. The result is less sensitive to the background. Use it in the first months of the season and in young plantings; once cover closes, NDVI and SAVI almost coincide.
How to read the time series table
| Column | What it is | How to use it |
|---|---|---|
| Mean | The average of the index over all clear pixels of the field | The general condition. It is meaningful mainly as a comparison with earlier dates, not as an absolute number |
| Deviation | The standard deviation, that is how much the points differ from one another | The most underrated number. A large deviation means an uneven field — something is happening in one part and not across the whole |
| Clear | The percentage of the field the satellite saw without cloud or shadow | Below 70% the measurement is unsafe. Below 40%, disregard it |
Two fields with a mean NDVI of 0.55 are not the same thing. If the first has a deviation of 0.04, it is uniform and healthy. If the second has 0.18, then zones of 0.35 and zones of 0.75 coexist within it — and the mean hides exactly the problem being looked for. When the deviation rises while the mean stays steady, something is starting locally. That is when a visit to the field is worth it.
What the image cannot say
A satellite image shows that something is happening and where. It never shows why. A low index in one zone may be due to nutrient deficiency, nematodes, soil compaction, a blocked emitter, salinity, a root infection, or simply to that part of the field being sandier. The image indicates where to dig. The diagnosis is made on the ground.
Two further limitations that must be clear:
- The mixed pixel. At 10 metres, one pixel in an orchard contains trees, bare soil, grass and shade. The number is the weighted average of all of them. It is not the condition of the tree.
- The atmosphere. The data are level 2A, that is atmospherically corrected, but thin translucent cloud and haze are not always caught by the mask. A sudden, unexplained drop on a single date is often cloud, not an event in the field.
The right way to use it
- Compare within the same field, on the same day. Relative differences between zones are reliable. The absolute number depends on the crop, the growth stage and the background.
- Follow the curve, not the point. A single measurement says nothing. The course of three or four consecutive passes does.
- Compare the same date against another year. It is the most reliable way to see whether the season is going better or worse.
- Confirm in the field. Every zone that stands out in the image deserves a visit before a decision is taken.
- In fruit trees, NDRE. In arable crops, NDVI. Early in the season, SAVI. In a heatwave, NDMI.
Data source. Copernicus Sentinel data 2026, through the Copernicus Data Space Ecosystem of the European Union and the European Space Agency. The service is provided free of charge by the EU.
Clouds. Where the satellite saw cloud or shadow, the image comes out grey and translucent and those points are excluded from the calculations. If the whole image is grey, choose a wider date range.
This is not an agronomic recommendation. A satellite image shows where to look, not what to do. Every intervention follows an on-site assessment.