Reviewed August 2026 against NASA Earthdata, the USDA Economic Research Service, and the Alabama Cooperative Extension System.

Try it: Run your own numbers →

NDVI in Precision Agriculture: How EVI and SAVI Compare

NDVI (Normalized Difference Vegetation Index) is the ratio (NIR − Red) / (NIR + Red) computed from satellite or drone imagery, and it is the standard way precision-agriculture tools score how green and vigorous a crop canopy is, on a scale of -1 to 1. EVI (Enhanced Vegetation Index) is a variant that adds a blue band and three correction coefficients so it stays accurate in dense canopies and hazy air, while SAVI (Soil-Adjusted Vegetation Index) adds a soil-brightness term so it does not overstate vegetation where bare ground still dominates the pixel. What follows are the exact formulas, the value ranges that separate a healthy stand from a stressed one, the satellites that actually produce this data, and a calculator that runs your own reflectance numbers through all three indices.

NDVI, EVI, and SAVI: Full Forms and Quick Definitions

• NDVI full form: Normalized Difference Vegetation Index. Formula: (NIR − Red) / (NIR + Red). Range: -1 to 1. Per the Alabama Cooperative Extension System (published September 23, 2025), a healthy crop canopy reads 0.6–0.9; below 0.4 signals stress or sparse cover.
• EVI full form: Enhanced Vegetation Index. Formula: 2.5 × (NIR − Red) / (NIR + 6 × Red − 7.5 × Blue + 1). The same Alabama Extension guidance puts a healthy EVI at 0.2–0.8, with values under 0.15 flagged as problematic.
• SAVI full form: Soil-Adjusted Vegetation Index. Formula: [(NIR − Red) / (NIR + Red + L)] × (1 + L), where L is a soil-brightness correction, set to 0.5 by default per the Landscape Toolbox remote-sensing methods reference.

These three numbers are not academic. The USDA Economic Research Service reports that in 2023, autosteering guidance systems — which rely on the same imagery pipelines that generate NDVI — were used on 52% of midsize U.S. crop farms and 70% of large-scale crop farms, while decision-support technologies built on yield monitors, yield maps, and soil maps reached 68% of large-scale crop farms. Adoption rises sharply with farm size and falls off fastest among small, low-revenue operations. The chart below shows the 2023 split; ERS re-publishes farm-size adoption figures periodically inside its “America’s Farms and Ranches at a Glance” series, so check the linked chart for the current year’s numbers before citing a figure in a grant application or board memo.

NDVI: The Normalized Difference Vegetation Index Explained

NDVI works because chlorophyll absorbs red light for photosynthesis and a healthy leaf’s cell structure scatters near-infrared (NIR) light back out. Compare the two bands and you get a number: NDVI = (NIR − Red) / (NIR + Red), always between -1 and 1. Per NASA Earthdata, low values correspond to bare rock, sand, or snow, while higher values indicate denser, greener vegetation such as forests, cropland, and wetlands. NASA has run this calculation continuously for more than 20 years using the MODIS instruments on the Terra and Aqua satellites, and now extends that record with the VIIRS instrument on Suomi NPP, which observes the Earth daily.

On the ground, that translates into four concrete jobs: flagging drought stress before it is visible to a scout, classifying land cover across a whole county, spotting yield-limiting zones inside a single field, and confirming whether a cover crop actually established after planting. NDVI’s weakness is the same mechanism that makes it useful — it responds to whatever the sensor sees, including bare soil and thin cloud haze, which is why a field at 20% emergence and a field of gravel can produce a confusingly similar low NDVI reading. That gap is what EVI and SAVI were each built to close, from two different directions: EVI toward dense canopies and atmospheric noise, SAVI toward exposed soil.

EVI: Enhanced Vegetation Index for Dense Canopies

EVI keeps NDVI’s core logic but adds a blue band and three fixed coefficients to correct for atmosphere and soil background. The formula used for the standard MODIS Vegetation Index product, documented in the MODIS Vegetation Index User’s Guide hosted by the University of Hawaii’s College of Tropical Agriculture, is:

EVI Calculation Formula

EVI = G × (NIR − Red) / (NIR + C1 × Red − C2 × Blue + L)
where G = 2.5, C1 = 6, C2 = 7.5, and L = 1 — the coefficients NASA fixed for the operational MODIS EVI product.

• Blue band correction: C1 and C2 use the blue band to strip out aerosol scattering from the red band before the ratio is taken — the step NDVI skips entirely.
• Healthy range: 0.2–0.8, per the Alabama Cooperative Extension System guidance cited above; below 0.15 is flagged as low vigor.
• Where it wins: forest canopies, plantations, and any field past full canopy closure, where NDVI’s ratio starts to saturate and stops distinguishing “very green” from “extremely green.”

When the blue band is unavailable or noisy — over bright snow, ice, or cloud edges — MODIS falls back to a two-band EVI: 2.5 × (NIR − Red) / (NIR + Red + 1), per the same University of Hawaii guide. That backup formula drops the atmospheric correction but keeps EVI’s wider dynamic range over dense canopy.

EVI vs. NDVI: Five Practical Differences

1. Band count: NDVI uses two bands (Red, NIR); EVI uses three (Red, NIR, Blue) plus fixed correction coefficients.
2. Atmospheric resistance: EVI actively corrects for aerosol scattering using the blue band; NDVI has no such term, so haze and thin cloud shift its values.
3. Saturation behavior: NDVI compresses toward 1 as biomass climbs past full canopy closure; EVI stays more linear at high biomass, which is why it is the preferred index for forestry and dense plantation monitoring.
4. Healthy-value range: NDVI’s healthy band is 0.6–0.9; EVI’s is lower, 0.2–0.8, because its formula scales the ratio differently — the two are not interchangeable on the same threshold.
5. Soil sensitivity: both are affected by exposed soil, but EVI’s canopy-background term reduces that influence more than NDVI’s plain ratio does.

For the full derivation of each of these five points, including worked examples, see our dedicated breakdown: EVI vs NDVI: 5 Key Differences for Precision Farming.

SAVI: Soil-Adjusted Vegetation Index for Sparse Cover

SAVI in remote sensing exists to fix one specific failure mode: over sparsely vegetated ground, exposed soil brightness pushes NDVI around in ways that have nothing to do with plant health. The fix, developed by Huete (1988) and described by the Landscape Toolbox remote-sensing methods reference, inserts a soil-brightness correction factor, L, into the NDVI denominator:

SAVI Calculation Formula

SAVI = [(NIR − Red) / (NIR + Red + L)] × (1 + L)
L = 0.5 is the default that “works well in most situations,” per the Landscape Toolbox reference; L ranges from 0 (very high vegetation cover) to 1 (bare ground).

• Decision rule: the same source states SAVI is intended for areas with vegetation cover under 40% — young stands, recently harvested or tilled fields, semi-arid rangeland — where soil brightness would otherwise distort NDVI.
• Trade-off: SAVI is less sensitive to real changes in vegetation amount than NDVI, and more sensitive to atmospheric differences, in exchange for that soil correction.
• Worked example the source gives: in a rocky canyon test case, NDVI overstated vegetation cover because of bright exposed rock, while SAVI produced a closer match to ground-truthed cover.

For sustainable-agriculture context around soil condition monitoring more broadly, see Soil Health: 7 Ways to Boost Sustainable Agriculture.

SAVI vs. NDVI: Which One Fits Your Field

The decision comes down to canopy cover, not preference. If a field is past 40% canopy cover — most row crops from mid-vegetative stage onward — NDVI’s simpler two-band ratio is accurate enough and cheaper to compute at scale. Below that threshold — pre-emergence, drought-stunted stands, recently disked ground, arid rangeland — SAVI’s soil correction keeps a bare-dirt field from reading as if it had a stand of crop on it. A practical routine many agronomists use: run NDVI as the default weekly layer, then switch the same field to SAVI immediately after planting and again after any tillage or harvest pass, until the canopy closes back over 40%.

NDVI, EVI, and SAVI: Full Comparison Table

Which Satellites Actually Produce This Data

Every one of these indices is only as good as the imagery feeding it. Three public satellite programs supply nearly all of the red/NIR/blue reflectance data used in U.S. agriculture:

• MODIS and VIIRS (NASA): the instruments behind NASA’s operational NDVI and EVI products, observing the Earth on a near-daily basis and holding a continuous record of more than 20 years, per NASA Earthdata.
• Landsat 8 and 9 (USGS/NASA): 30-meter resolution, collecting up to 750 scenes a day; Landsat 9 alone revisits a given location every 16 days, and the Landsat 8/9 pair together cut that to 8 days, according to NASA Science’s Landsat 9 mission page.
• Sentinel-2 (ESA/Copernicus): four bands, including Red and NIR, at 10-meter resolution; a single Sentinel-2 satellite revisits every 10 days, and the two-satellite constellation cuts that to 5 days at the equator, per the Copernicus SentiWiki Sentinel-2 mission page.

Field-scale platforms, including Farmonaut, typically blend these sources: Sentinel-2’s 10-meter bands for in-field detail, Landsat’s longer archive for year-over-year comparison, and MODIS/VIIRS for near-daily large-area screening when cloud cover blocks the higher-resolution passes.

Calculate Your Own NDVI, EVI, and SAVI

Enter reflectance values pulled from a Sentinel-2, Landsat, or drone band export below to see NDVI, EVI, and SAVI side by side, and which one to trust for that field’s canopy cover.

Assumptions: uses the standard two-band NDVI/SAVI ratio (L = 0.5) and the fixed MODIS EVI coefficients (G = 2.5, C1 = 6, C2 = 7.5, L = 1). It does not correct for atmosphere, sensor calibration, or mixed pixels the way a full satellite processing pipeline does — treat the output as a teaching approximation, not a certified index value for a specific sensor product.

Beyond the Crop Row: Mining, Carbon, and Lending

The same red/NIR math extends past crop scouting into three adjacent use cases worth knowing about even if your primary interest is farming:

• Mine-site restoration: NDVI and SAVI track vegetation regrowth after disturbance and support compliance reporting; see Mining Site Restoration: Top 5 Strategies.
• Carbon and climate reporting: vegetation-index time series feed carbon-sequestration estimates and regenerative-agriculture tracking; Farmonaut’s carbon footprinting tool uses this data to help landowners document climate impact.
• Crop loans and insurance: lenders and insurers increasingly reference vegetation-index history as one input for risk assessment and claims verification; see crop loan and insurance verification.
• Supply-chain traceability: satellite-derived growth records get logged against product traceability systems to document where and how a crop was grown.

How Farmonaut Applies These Indices

Farmonaut runs NDVI, EVI, and SAVI calculations against satellite imagery and surfaces the results through a web app, Android and iOS apps, and a developer API, so the choice between indices does not have to be made by hand for every field:

• API and developer access: pull raw vegetation-index values into your own tools via the Farmonaut satellite API and its developer documentation.
• Farm-scale administration: the agro admin app applies these indices across multiple fields and operators for larger operations.
• Equipment and logistics: fleet management ties variable-rate input decisions, flagged by vegetation-index maps, to the equipment that executes them.
• AI advisory: Farmonaut’s Jeevn AI system interprets NDVI, EVI, and SAVI streams and turns them into a plain-language recommendation for the next field action.

FAQ

Conclusion

NDVI, EVI, and SAVI answer the same underlying question — how much healthy vegetation is in this pixel — with three different corrections for three different failure modes: none for NDVI, atmosphere and canopy saturation for EVI, and soil brightness for SAVI. The formulas do not change; what changes is which satellite feeds them (MODIS/VIIRS daily, Sentinel-2 every 5–10 days, Landsat every 8–16 days) and where your field sits on the canopy-cover scale. Use the 40% cover threshold from the SAVI section above as your decision rule, use the comparison table as your reference sheet, and re-check the USDA ERS adoption chart and satellite mission pages linked throughout whenever you need a number fresher than August 2026.

Run your own reflectance numbers through the calculator above, or open Farmonaut’s satellite platform to pull live NDVI, EVI, and SAVI for your fields.