Sample report · generated unedited by the Sequential Pipeline in GISFORUS Spatial Research Suite (Land Mask → Allen Coral Atlas → Coral Health).Get it for Mac  ·  Windows  ·  gisforus.com

Sequential Pipeline - Spatial Analysis Report

Satellite Remote-Sensing Analysis · Spatial Research Suite

Generated: 2026-08-24 10:38  |  Study area: 354.4 km²  |  Date range: 2024-06-01 to 2024-07-01  |  Compositing: Median

Summary

This report screens Coral Health (Landsat, long-record) across a 354.4 km² study area from satellite imagery composited over 2024-06-01 to 2024-07-01.

The Allen Coral Atlas maps 72.457 km² (71.4%) of the area as coral/algae habitat, over which the coral, landscape and trend results were computed.

These are satellite-derived screening results, not field-validated; see Limitations.

composite map
Final composite = Coral Health (Landsat, long-record), shown as 5 equal-interval classes whose colours match the Area-by-class table, clipped to ACA-confirmed coral (deep water & land excluded). Chain: Land Mask (exclude ocean) -> Coral Habitat (Allen Coral Atlas) -> Coral Health (Landsat, long-record).

What the map shows

How this map was built. The pipeline ran these steps in order, each handing its region/mask to the next:

  1. Land Mask (exclude ocean) — Land-only mask published (ocean excluded via the SRTM land footprint). Dry-land indicators placed after this are now restricted to land.
  2. Coral Habitat (Allen Coral Atlas) — Allen Coral Atlas: Coral/Algae covers 72.457 km2 (71.4%). Its coral mask is now published — any Coral Health / Trend / index step AFTER this runs over ACA coral only.
  3. Coral Health (Landsat, long-record) — Coral Health on Landsat (30 m): ln(green) - ln(blue) on optically-shallow reef; oriented so HIGHER = more coral/algae cover (darker bottom), LOWER = bright sand/rubble; stretched 0.22 to 2.46. Built from the harmonized Landsat 5/7/8/9 archive (1984+) so the record spans the FULL timeline, incl. pre-2018. Coarser than the Sentinel-2 version and NOT sun-glint corrected -> treat as a screening long-record index, and note small cross-sensor (TM/ETM+/OLI) offsets.

What you are looking at. The colours show Coral Health (Landsat, long-record). On the legend, values run from about 0.22 to 2.46:

Higher = more coral-like cover, lower = more bare substrate. Note the index cannot separate live coral from macroalgae (both are dark) — a high value can mean healthy coral OR algal overgrowth, so pair it with turbidity/chlorophyll before concluding. Deep open water has been removed, so every coloured pixel is genuine shallow reef/seabed.

Source & caveat. USGS Landsat Collection 2 Level-2 Science Products (LC08 / LC09 / LE07 / LT05). This is a satellite-derived screening index — treat it as a guide and confirm on the ground before drawing firm conclusions.

Layer gallery

Every layer the pipeline computed, in sequence. Colours are each layer's own display stretch.

Area by class

ClassValue rangeArea (km2)Area (ha)Share (%)
Very Low0.22 to 0.6731.8163181.670
Low0.67 to 1.128.862886.219.5
Medium1.12 to 1.572.249224.94.9
High1.57 to 2.021.047104.72.3
Very High2.02 to 2.461.469146.93.2

Landscape metrics

Landscape diversity
Richness (classes present)5Shannon diversity (SHDI)0.915
Shannon evenness (SHEI)0.568Simpson diversity0.468
Dominant class share70.0%
Across these value classes: SHDI rises with more and more-even classes (0 = a single class dominates). SHEI is 0-1 (1 = perfectly even split). Simpson = the chance two random pixels belong to different classes.
Landscape configuration (patch structure)
Number of patches1,013Patch density (/100 ha)22.29
Mean patch size (ha)4.49Area-weighted mean patch (ha)123.97
Largest Patch Index (%)4.6Edge density (m/ha)66.5
A patch is a contiguous block of one class (8-connected). Higher patch & edge density with lower mean patch size = a more fragmented map; a high Largest Patch Index means one class forms a single big block. Approximate — computed on a coarsened grid (~3× the native pixel) to stay within Earth Engine memory, and single-patch pixel counts are capped at ~256 px, so very large patches (and LPI) can read low.

Coral habitat (Allen Coral Atlas)

Benthic classArea (km2)Area (ha)Share (%)
Coral/Algae72.4577245.771.4
Rubble19.7171971.719.4
Sand7.319731.97.2
Rock1.617161.71.6
Seagrass0.34734.70.3
Microalgal Mats0.0686.80.1

Allen Coral Atlas mapped 72.457 km² (71.4%) of this area as Coral/Algae. Because the ACA step ran before Coral Health, the Coral Health index, landscape metrics and trend above were all computed over this ACA-confirmed coral only.

Processing chain

StepLayerRoleStatusDetail
1Land Mask (exclude ocean)maskcomputedLand-only mask published (ocean excluded via the SRTM land footprint). Dry-land indicators placed after this are now restricted to land.
2Coral Habitat (Allen Coral Atlas)maskcomputedAllen Coral Atlas: Coral/Algae covers 72.457 km2 (71.4%). Its coral mask is now published — any Coral Health / Trend / index step AFTER this runs over ACA coral only.
3Coral Health (Landsat, long-record)indexcomputedCoral Health on Landsat (30 m): ln(green) - ln(blue) on optically-shallow reef; oriented so HIGHER = more coral/algae cover (darker bottom), LOWER = bright sand/rubble; stretched 0.22 to 2.46. Built from the harmonized Landsat 5/7/8/9 archive (1984+) so the record spans the FULL timeline, incl. pre-2018. Coarser than the Sentinel-2 version and NOT sun-glint corrected -> treat as a screening long-record index, and note small cross-sensor (TM/ETM+/OLI) offsets.

Methods

All layers were computed on Google Earth Engine over the study boundary. Point-in-time layers were composited over 2024-06-01 to 2024-07-01 using the Median reducer at an adaptive spatial scale matched to the boundary size (Sentinel-2 SR Harmonized with per-pixel SCL cloud masking; Landsat Collection-2 for land indices / pre-2015). Marine optical layers used a blue-green bottom index, ln(green)−ln(blue) (ln B3 − ln B2), over optically-shallow reef (deep water excluded), oriented so higher values indicate darker, light-absorbing bottoms (coral / algae / seagrass) and lower values bright sand / rubble; it reflects benthic cover/brightness and cannot on its own distinguish live coral from macroalgae. The Sentinel-2 index applies sun-glint correction (Hedley et al., 2005); the Landsat long-record index (30 m, 1984+) is not sun-glint corrected and carries small cross-sensor (TM/ETM+/OLI) offsets, so it is treated as a screening long-record index. Processing chain: Land Mask (exclude ocean) → Coral Habitat (Allen Coral Atlas) → Coral Health (Landsat, long-record).

Sensors & datasets

Sensor / platformDataset (Earth Engine ID)Temporal coverageNative resolutionBands usedRole in this study
Landsat 5/7/8/9 (Collection 2, Level-2 SR)LANDSAT/LC09|LC08|LE07|LT05/C02/T1_L21984 → present (multi-mission)30 m (optical); 100 m thermal, resampled 30 mSR blue/green/red/NIR/SWIR1/SWIR2 + ST thermalUrban Sprawl (NDBI), Urban Heat / LST, long-record indices
NASA SRTM digital elevationUSGS/SRTMGL1_003Static (acquired Feb 2000)30 melevationLand Mask (land/ocean footprint), terrain layers
Allen Coral Atlas — benthic habitatACA/reef_habitat/v2_0Static ~2018–2020 baseline~5 m mapped classesbenthic class (Coral/Algae, Sand, Rubble, Rock, Seagrass, Microalgal mats)Coral-habitat mask — restricts marine steps to mapped coral

Reproducibility

ParameterValue
Boundary area (km²)354.365
Date range2024-06-01 to 2024-07-01
Compositing reducerMedian
Adaptive spatial scale (m)30
Composite thumbnail (px)500
Processing chainLand Mask (exclude ocean) → Coral Habitat (Allen Coral Atlas) → Coral Health (Landsat, long-record)
DatasetsNASA SRTM Digital Elevation 30m (USGS/SRTMGL1_003); Allen Coral Atlas (Earth Engine: ACA/reef_habitat/v2_0).; USGS Landsat Collection 2 Level-2 Science Products (LC08 / LC09 / LE07 / LT05)
Cloud handlingSentinel-2: per-pixel SCL mask; image-level filter relaxed (<=90%); clear pixels reduced by the reducer above
Coordinate reference systemEPSG:4326 (WGS84)
Analysis parametersShallow-water mask: green reflectance > 0.015; display stretch: 2-98th percentile; OISST anomaly raw x0.01 °C vs 1971-2000 climatology; Sentinel-2 image cloud filter <=90% + per-pixel SCL; benthic composite max cloud <=90%.
Bounding box (WGS84)99.5979 W, 6.4934 S, 99.7141 E, 6.7417 N
Software versionsR 4.6.0; rgee 1.1.8; earthengine-api 1.6.11; app db7d10a
PlatformGoogle Earth Engine (rgee) - Spatial Research Suite pipeline
Generated2026-08-24 10:38 IST

Data sources & references

  1. NASA SRTM Digital Elevation 30m (USGS/SRTMGL1_003)
  2. Allen Coral Atlas (Earth Engine: ACA/reef_habitat/v2_0).
  3. USGS Landsat Collection 2 Level-2 Science Products (LC08 / LC09 / LE07 / LT05)
  4. Olofsson, P. et al. (2014). Good practices for estimating area and assessing accuracy of land change. Remote Sensing of Environment, 148, 42-57.
  5. Pontius, R.G. & Millones, M. (2011). Death to Kappa. International Journal of Remote Sensing, 32(15), 4407-4429.
  6. Hedley, J.D., Harborne, A.R. & Mumby, P.J. (2005). Simple and robust removal of sun glint for mapping shallow-water benthos. International Journal of Remote Sensing, 26(10), 2107-2112.
  7. Lyzenga, D.R. (1981). Remote sensing of bottom reflectance and water attenuation parameters in shallow water. International Journal of Remote Sensing, 2(1), 71-82.
  8. Lyons, M.B. et al. (2020). Mapping the world's coral reefs using a global multiscale earth observation framework. Remote Sensing in Ecology and Conservation, 6(4), 557-568. https://doi.org/10.1002/rse2.157
  9. Farr, T.G. et al. (2007). The Shuttle Radar Topography Mission. Reviews of Geophysics, 45, RG2004. https://doi.org/10.1029/2005RG000183

Limitations

Generated by the Spatial Research Suite. Figures and statistics are reproducible from the Methods above.