Satellite intelligence
NDMI vs NDVI: how to know when to irrigate with satellite imagery
NDVI tells you whether the crop is green. NDMI tells you whether it has water. What each index measures, how to read them on your field, and why satellite data alone can't decide your irrigation.
In short
- NDVI measures vigor (how much green vegetation there is). NDMI measures water in the leaf.
- NDMI usually shows water shortage earlier than NDVI, which reacts once the crop has already lost color.
- Sentinel-2 passes roughly every 5 days and clouds can hide the field: satellite tells you where to look, not when to irrigate.
- The irrigation decision is best made with a soil moisture sensor in the zone the satellite flagged.
What is NDVI and what does it measure?#
NDVI (Normalized Difference Vegetation Index) compares how much red light the plant absorbs with how much near-infrared light it reflects. A healthy leaf absorbs a lot of red for photosynthesis and reflects a lot of infrared because of its cell structure. The higher the value, the more green, active vegetation in the pixel.
It is the most widely used index in agriculture because it is simple and sharp: on Sentinel-2 it is computed from 10-meter bands. Its limit is that it reacts late to water stress. By the time NDVI drops for lack of water, the crop is already losing leaf or color.
What is NDMI and why does it matter for irrigation?#
NDMI (Normalized Difference Moisture Index) swaps the red band for short-wave infrared (SWIR). Water inside the leaf absorbs that light: the more water in the canopy, the less SWIR it reflects and the higher the index.
That is why NDMI works as an early warning. A zone can still look green on NDVI and already show less water on NDMI. The trade-off is resolution: Sentinel-2's B11 band is 20 meters per pixel.
NDVI · Sentinel-2
(B8 − B4) / (B8 + B4)NDMI · Sentinel-2
(B8 − B11) / (B8 + B11)NDMI vs NDVI side by side#
| Criterion | NDVI | NDMI |
|---|---|---|
| Measures | Vigor and green cover | Water in the canopy |
| Bands | Red (B4) + near infrared (B8) | Near infrared (B8) + SWIR (B11) |
| Resolution | 10 m | 20 m |
| Response to water stress | Late, once damage is visible | Earlier, as leaf water drops |
| Best use | Spot vigor differences and plan scouting | Find dry zones and prioritize where to check irrigation |
| Main limit | Saturates in very dense canopy | With little foliage, mostly reflects the soil |
How to read the values on your field#
Both indices range from −1 to 1. The most common mistake is to take a single number and compare it with a table found online. Values change with the crop, the growth stage and how much foliage there is. A freshly pruned vineyard and one in full canopy can't be judged by the same rule.
Compare zones, not absolute numbers#
The most useful signal is the difference between zones of the same field on the same date. If three zones sit at 0.20 and one at 0.08, that zone has a problem worth checking, whatever the crop's “ideal” value is.
Follow the trend over time#
An NDMI drop between two cloud-free images, with no rain in between and no change in management, is a stronger signal than any single reading.
Why satellite data alone can't time your irrigation#
NDMI is an excellent tool to find where there is a problem, but it has three limits for deciding when to irrigate:
- Frequency: Sentinel-2 passes roughly every 5 days. In summer, a field can go from fine to stressed between two images.
- Clouds: a cloudy image is discarded. In the rainy season weeks can pass without a usable reading.
- It measures the plant, not the soil: when NDMI drops, the crop is already feeling the shortage. A soil sensor sees the problem before it reaches the leaf.
An NDMI drop can also come from disease, frost, pruning or an image artifact. Before changing your irrigation, confirm it in the field or with a sensor.
How to combine satellite and soil sensors#
The efficient approach is to use each source for what it does best. Satellite covers the whole field with no installation cost; a sensor gives a continuous reading, every few minutes, at one point. Together they answer both questions:
Satellite flags the zone. The NDMI zone map shows which part of the field is drifting from the rest.
The sensor confirms it in the soil. A moisture sensor in that zone tells you whether it is a water shortage or something else.
Line pressure finds the fault. If the soil is dry and irrigation is scheduled, a pressure sensor shows whether there is a leak or a clogged emitter.
You irrigate by data, not by calendar. With continuous soil moisture, irrigation runs when the soil needs it.
That is how Bloomiot Workspace works: it combines Satellite Intelligence with your LoRaWAN or Wi-Fi/MQTT sensors, even third-party ones, and tells you what to do in plain language. See the details in how it works.
Frequently asked questions
What is the difference between NDMI and NDVI?
NDVI measures vigor and the amount of green vegetation using the red and near-infrared bands. NDMI measures the water content of the canopy using near infrared and short-wave infrared. NDMI usually detects water shortage earlier than NDVI.
Can I decide when to irrigate from satellite imagery alone?
It is not advisable. Sentinel-2 passes roughly every 5 days, clouds can hide the field and the index measures the plant, not the soil. Satellite tells you where to look; the decision of when to irrigate is best made with a soil moisture sensor in the flagged zone.
What resolution does Sentinel-2 NDMI have?
Sentinel-2's short-wave infrared band (B11) is 20 meters per pixel, so NDMI has that resolution. NDVI, built from 10-meter bands, shows more detail.
Does NDMI work on freshly pruned crops or sparse foliage?
With little vegetation cover the pixel mixes plant and soil, and the value mostly reflects the soil. At those stages compare the index between zones of the same field rather than against absolute values.
Sources
- NDMI
- NDVI
- Sentinel-2
- Irrigation
- Soil sensors
About the author
Elio Fagúndez
Founder of Bloomiot. Mechatronics engineer specializing in hardware development, embedded systems and IoT; he designed the platform that turns satellite and field data into irrigation decisions.