Mapping (by District)

Why Ground-truthed?

Ground-truthed atlas of Mauritius is a superior scientific record because it bypasses the inherent technical limitations of remote sensing. While satellite and airborne imagery provide a broad "bird's-eye view," they often struggle with the biological and physical complexities of the Mauritian coastline. Extracting mangrove data from these imageries is difficult due to several factors:

1. Spectral Confusion ("The Green-on-Green" Problem)

The most significant challenge for automated software is that mangroves look very similar to other vegetation from above.

Signature Overlap: Mangroves, coastal shrubs (like Veloutier), and nearby sugar cane fields all have high chlorophyll content. This causes them to reflect light in the Near-Infrared (NIR) spectrum almost identically.

Lack of Differentiation: Standard multi-spectral sensors (like Sentinel-2 or Landsat) often cannot distinguish between a mangrove tree and a terrestrial tree at the edge of the forest without manual ground-truth validation.

2. Tidal Inundation and Masking

Because mangroves exist in the intertidal zone, the timing of the imagery is a major "noise" factor.

Submerged Features: If an image is taken during high tide, younger mangroves, seedlings, and complex root systems (pneumatophores) are submerged. The satellite sensor then registers a "mixed signature" of water and leaves, which often causes the software to misclassify the area as "shallow water" or "mudflat."

Water Attenuation: Water absorbs infrared light. Even a shallow layer of water over mangrove roots can significantly alter the total reflectance value, making the forest appear less dense or "unhealthy" in the data.

3. The "Mixed Pixel" Problem

In Mauritius, many mangrove stands are narrow, fringing belts rather than massive forests.

Resolution Limits: A standard free satellite pixel (like Landsat) is 30m x 30m. If a mangrove belt at a place like Le Morne is only 10 meters wide, that single pixel will contain 30% mangrove, 40% ocean water, and 30% sand or soil.

Averaged Data: The computer averages these colors into a single "dirty" value, often resulting in the fringe being omitted from the map entirely.

4. Fragmented and Narrow Distributions

Unlike the massive delta mangroves in other parts of the world, Mauritian mangroves are often:

Fragmented: Small clusters around river estuaries or islets like Île d'Ambre.

Linear: Long, thin lines along the coast. Airborne sensors often "blur" these fine-scale features due to atmospheric haze or sensor limitations, whereas your ground-truthed maps capture every individual cluster regardless of size.

5. Persistent Tropical Cloud Cover

Mauritius’s tropical climate poses a literal physical barrier to optical satellites.

Cloud Contamination: Finding a perfectly clear day where the tide is also at its lowest point is a statistical rarity. Researchers often have to use "composite" images from different dates, which can introduce errors if the mangrove extent changed between those dates (due to planting or coastal development).

6. Species Differentiation

Mauritius primarily hosts two species: Rhizophora mucronata and Bruguiera gymnorrhiza.

• Similarity: These species often grow in mixed stands. Standard satellite imagery cannot distinguish between them.

• Mangrove Associates: Other coastal plants (like the "Veloutier") often grow right up to the mangrove edge. Without "ground-truthing," a satellite cannot tell where the mangrove ends and the coastal scrub begins.