What is NDMI? (Normalized Difference Moisture Index)
What is NDMI?
NDMI, the Normalized Difference Moisture Index, measures how much water is inside your crop’s leaves. Where NDVI tells you how green and dense the canopy is, NDMI tells you how well-watered it is — and it does so before drought stress becomes visible to the eye or to NDVI. It is the single most important index for drought monitoring, irrigation scheduling, and wildfire risk assessment.
The reason NDMI beats NDVI for drought: leaf water content drops before chlorophyll breaks down. A crop can be running low on water and still look green. NDMI catches this; NDVI only catches it once the stress is severe enough to start yellowing.
How it’s calculated
When vegetation is well-watered, its internal leaf structure strongly reflects near-infrared and its water content absorbs shortwave-infrared (SWIR). As leaves dry out, SWIR reflectance rises sharply — the gap between NIR and SWIR shrinks. NDMI turns that gap into a number:
NDMI = (NIR − SWIR1) / (NIR + SWIR1)On Sentinel-2 these are Band 8 (NIR) and Band 11 (SWIR1, ~1.6 µm). The ratio structure cancels out raw brightness, so values are comparable across scenes and dates.
Typical value ranges
| NDMI range | Meaning | Fire risk |
|---|---|---|
| > 0.3 | High moisture, lush vegetation | Very low |
| 0.1 – 0.3 | Moderate moisture, healthy crop | Low |
| −0.1 – 0.1 | Transitional, water stress beginning | Moderate |
| −0.3 – −0.1 | Significant stress, drying fast | Elevated |
| Below −0.3 | Extreme dryness, dead or dormant vegetation | Critical |
These thresholds are the basis for fire-danger classification. Most fire-impact monitoring systems weight NDMI heavily — at AgroReports, the pre-fire risk score is 60% NDMI, 25% NBR, and 15% NDVI.
When to use it
Use NDMI for drought early-warning, irrigation scheduling, and wildfire risk assessment. Over a season, NDMI follows the crop’s water budget: high after rainfall or irrigation, declining steadily during dry spells, crashing during drought. On an irrigated field, the curve shows your watering cycle as a sawtooth. On a rain-fed field, it mirrors the rainfall pattern. A divergence where NDVI stays flat but NDMI falls is the earliest sign of hidden drought stress.
NDMI gives you days to weeks of lead time over NDVI for drought detection: water stress begins, leaf moisture drops, NDMI falls — but NDVI still looks fine. By the time NDVI declines, the damage is already underway.
Comparison with other indices
NDMI answers “is the canopy well-watered?” while NDVI answers “is the canopy green and dense?” The two diverge during drought onset, with NDMI providing the earlier warning. NDWI (McFeeters) is often confused with NDMI but measures surface water, not leaf moisture — use NDWI for flooding and drainage, NDMI for crop moisture stress. NBR pairs with NDMI for fire risk assessment: NDMI covers fuel moisture, NBR covers fuel condition, and together they form the core of any satellite-based fire danger score.
Try it free: Get NDMI values for your field — free, no signup. Works on any land worldwide. Check my field’s moisture →
Frequently asked questions
What is a good NDMI value?
For a healthy, well-watered crop, NDMI above 0.2 is typical at peak growth. Values between 0.1 and 0.2 indicate adequate but declining moisture. Below 0.0, the crop is under measurable water stress. As always, compare against your field’s own seasonal history rather than a single absolute threshold.
Is NDMI better than NDVI for drought?
Yes — for the onset of drought. Leaf water content drops before chlorophyll degrades, so NDMI falls earlier than NDVI. NDVI catches stress once the canopy starts losing greenness; NDMI catches it days to weeks sooner. For irrigation decisions, NDMI is the better early signal.
Can NDMI predict wildfire risk?
Yes. NDMI is the primary input to most satellite-based fire-danger scores because it directly reflects fuel moisture. Low NDMI (below −0.1) indicates dry, flammable vegetation; below −0.3 indicates critically dry fuel. Fire agencies use NDMI trends alongside weather data to anticipate danger windows.
Why is NDMI at 20 m resolution when NDVI is 10 m?
Because Sentinel-2’s SWIR1 band (Band 11) is collected at 20 m resolution, while the Red and NIR bands used by NDVI are at 10 m. NDMI must use SWIR, so it inherits the coarser resolution. This is a sensor limitation, not a software choice — it is fine for field-scale monitoring.