‹ volver al glosario de SIG

SAVI (Soil-Adjusted Vegetation Index)

Definition

The Soil-Adjusted Vegetation Index (SAVI) is a vegetation index developed to minimize the influence of soil brightness on vegetation signals, particularly in areas with sparse vegetation cover where the underlying soil significantly affects reflectance. It introduces a constant soil adjustment factor (L) to the NDVI formula to account for first-order soil-vegetation interactions.

Application

SAVI is widely used in arid, semi-arid, and agricultural regions where vegetation cover is low to moderate and the soil background is bright and highly variable. Applications include monitoring rangeland health, assessing early crop emergence and growth, studying desertification, and evaluating vegetation recovery in degraded lands. It is especially useful in the early growing season when canopy cover is incomplete.

FAQ

What is the formula for SAVI and what is the 'L' factor?

The SAVI formula is: SAVI = ((NIR - Red) / (NIR + Red + L)) * (1 + L). The 'L' factor is a soil brightness correction factor, typically set to 0.5 for most applications with intermediate vegetation cover. This factor accounts for the differential NIR and red reflectance of the soil and helps isolate the vegetation signal.

How does SAVI improve upon NDVI in sparse vegetation areas?

In sparse vegetation, NDVI values are highly sensitive to the color and brightness of the underlying soil, leading to unreliable vegetation assessments. SAVI's L factor reduces this soil background effect, providing a more stable and accurate measure of the actual green vegetation amount, independent of whether the soil is dark, bright, wet, or dry.

How do I choose the appropriate value for the L factor?

The optimal L value varies with vegetation density:

For very low vegetation cover (≤15%), use L=1.0.

For intermediate cover (~30-50%), L=0.5 is standard.

For dense vegetation (≥80%), the soil influence is minimal, and L can approach 0, making SAVI essentially equivalent to NDVI. Many applications use the default L=0.5 as a good compromise.

What are the limitations of SAVI?

Limitations include:

1) The need to estimate or assume an appropriate L value, which may not be constant across a scene;

2) It is less effective than more complex indices (like MSAVI) in completely eliminating soil noise across all conditions;

3) Like NDVI, it can still saturate in very high biomass conditions;

4) It does not explicitly correct for atmospheric effects.