Research · spectral index

Shrimp Index (SI)

Remote sensing had no index for shrimp farming. This work derives one — a number that quantifies what is happening inside a shrimp pond, from orbit.

Role: Spectral analysis
& index derivation
Sensor: Sentinel-2 MSI L2A
12 bands · ~5-day revisit
Method: Multi-temporal &
mono-temporal analysis
Bands: RE3 (783 nm)
SWIR (1610 nm)
Ground truth: Field reports
via WhatsApp & Metabase

The problem

There is no remote sensing index related to shrimp farming. Without one, satellite imagery cannot really be used to monitor shrimp ponds — you can see the water, but you cannot measure the culture. The proposal was a Shrimp Index (SI) that captures the distinctive spectral response of shrimp ponds and supports monitoring and management at scale.

My role

Studying the spectral signatures of aquaculture ponds across different geographic regions, and comparing them against other water bodies.

Research objectives

Multi-temporal spectral analysis at pond level, to understand the phenology of different ponds.
Mono-temporal spectral response analysis of ponds at different phenological stages, to understand response across wavelengths of the electromagnetic spectrum.

Data

Source	What it gives
Sentinel-2 AWS & Planetary Computer	Imagery across 12 spectral bands at roughly a five-day revisit, enabling continuous observation of the phenological and environmental changes associated with shrimp farming.
Ground truth WhatsApp & Metabase	Field information from farms, used to validate and improve the insights drawn from geospatial analysis.

Sample spectral signature plot of shrimp ponds across Sentinel-2 bands — **Fig. 01**Sample spectral signature of shrimp ponds — the curve that the index is derived from.

What the signatures showed

Detailed multi-temporal and mono-temporal analysis was carried out at pond level across several farms, alongside other water bodies. Three findings held consistently:

A significant increase in reflectance in the Red Edge band across all ponds during culture — evidence of phytoplankton activity.
Following the red edge response, strong reflectance in the NIR band. In water bodies, higher NIR reflectance indicates increased suspended material; in an aquaculture pond that is a strong indicator of shrimp and feed.
The spectral responses of water bodies other than shrimp ponds were completely different — which is what makes a species-specific index possible in the first place.

Further plots and analysis: spectral analysis deck ↗

The index

SI is a numerical value that quantifies the properties of ponds under shrimp culture. It is used to measure shrimp health and other dynamics of the culture pond using remotely sensed data.

Shrimp Index SI = (RE3 − SWIR) / (RE3 + SWIR)

RE3 is reflectance in the Red Edge band; SWIR is reflectance in the shortwave infrared band.

Value	Interpretation
High SI	Healthy shrimp, with active photosynthesis aerating the pond.
Low SI	Distressed shrimp, or no culture in the pond at all.

SI over time

The value on its own matters less than its trajectory. Tracked across a culture cycle, the index separates healthy ponds from failing ones well before the difference is visible on the ground.

Time series plot of Shrimp Index for a diseased pond, showing a downward trend — **Fig. 02**SI response for a diseased / unhealthy pond — a sustained downward trend.

Second time series plot of Shrimp Index for an unhealthy pond — **Fig. 03**A second unhealthy pond, same pattern, different farm.

Time series plot of Shrimp Index for a healthy pond, showing an upward trend — **Fig. 04**SI response for a healthy pond — the index climbs with the culture.

Conclusion

The SI pattern was studied across different geographic regions. It showed a downward trend in the majority of diseased or unhealthy ponds, and an upward trend in healthy ponds — a consistent enough signal to build an alert system on, which became SIRAS / EWIS.

← Water Quality Estimation Next: SIRAS / EWIS →