Vegetation Indices

SAVI

Soil Adjusted Vegetation Index. A vegetation index that minimizes soil brightness influences by incorporating a soil brightness correction factor (L). Particularly useful in areas with sparse vegetation where exposed soil affects NDVI readings.

Formula

SAVI = ((NIR - Red) / (NIR + Red + L)) × (1 + L)

Overview

The Soil Adjusted Vegetation Index (SAVI) minimizes the influence of soil brightness on vegetation measurements by introducing a soil correction factor L. Developed by Alfredo Huete in 1988, it addresses NDVI's unreliability in areas where exposed soil significantly affects readings.

How It Works

SAVI = ((NIR − Red) / (NIR + Red + L)) × (1 + L). The L factor shifts the origin of the soil line in NIR-Red spectral space. L = 0 equals NDVI; L = 0.5 (default) provides good compromise for moderate cover; L = 1 applies maximum soil suppression.

Key Facts

  • Developed by Alfredo Huete in 1988.
  • L = 0.5 is the widely used default that minimizes soil brightness variations.
  • Best suited for 15–50% vegetation cover where soil influence is strongest.
  • When cover exceeds ~70%, SAVI converges with NDVI.

Applications

Arid Vegetation Mapping

Monitoring sparse shrubland and dryland crops where exposed soil makes NDVI unreliable.

Rangeland Assessment

Estimating vegetation cover on grazing lands with varying soil backgrounds.

Desertification Monitoring

Tracking vegetation loss more accurately than NDVI by removing soil artifacts.

Early-Season Crop Monitoring

Assessing canopy development when significant soil is visible between rows.

Limitations & Considerations

Requires user to select L value, creating circular dependency since vegetation density is what's being estimated. This led to MSAVI's self-adjusting approach. Saturates at high LAI values like NDVI. Remains sensitive to atmospheric effects.

History & Background

Published in 1988 in Remote Sensing of Environment, demonstrating that soil variations could shift NDVI by 0.2 units for identical vegetation. Spawned TSAVI, OSAVI, and MSAVI.

Analyze SAVI data with LYRASENSE

Use our agentic notebook environment to work with satellite data and compute indices like SAVI — no setup required.