Mangrove Restoration Suitability & Eligibility Screening
Two views of the same skill set. First, a generalized methodology overview for a confidential client engagement. No client data, sites, or results are shown. Below that, a full real implementation I built myself: a live VM0033 mangrove restoration eligibility screen across all 28 coastal provinces of Vietnam, computed on real public Google Earth Engine data, with an interactive map you can try right on this page.
VM0033 Screening Map
Real VM0033 mangrove restoration eligibility results across all 28 coastal provinces of Vietnam, computed on Google Earth Engine. Click into any province and manually step through its six screening layers: historical mask, raw restorable area, and each of the four exclusion filters. Live preview below, or open it on its own page. The full story (methodology, tools, and takeaways) is below the map.
How a coastline becomes a shortlist
A generalized version of a confidential client engagement: the approach, not the deliverable.
Methodology Summary
- Objective: identify candidate coastal sites for mangrove restoration and blue-carbon project development, replacing ad hoc site selection with a defensible, criteria-based process.
- Approach: integrate coastal land-cover, hydrology, and historical degradation layers in QGIS, screening sites against ecological suitability criteria and combining them into a multi-criteria score per site.
- My role: designed the screening workflow and criteria weighting, and produced ranked outputs to support restoration-site prioritization. Generalized here to illustrate methodology, not an actual client deliverable.
Sample Work Notice
This section illustrates methodology and approach for a confidential client engagement. It does not disclose actual client data, site locations, deliverables, or results.
VM0033 Mangrove Restoration Screen
A self-directed extension of the same methodology, applied for real: VM0033 mangrove restoration eligibility screening across all 28 coastal provinces of Vietnam, built end to end, from authenticating Google Earth Engine to shipping the interactive map below. Not a client engagement: every number and polygon here comes from open data (FAO GAUL boundaries, Global Mangrove Watch, GLC-FCS30D, JRC Global Surface Water, Allen Coral Atlas, ESRI Land Cover, SRTM, and CGMD-Extent30), computed live for this project.
Objective
- Apply the VM0033 mangrove restoration eligibility methodology to a real country, not a mock-up.
- Identify where historically mangrove land has been lost and could realistically be restored.
- Make the result explorable: not just a number, a real map you can click into.
Approach
- Union GLC-FCS30D (2000) and Global Mangrove Watch (2007–2020) into a historical mangrove baseline.
- Subtract confirmed 2023 mangrove extent (CGMD-Extent30) to isolate the loss zone.
- Exclude permanent water, reef/shallow habitat, built-up land, and elevation ≥ 50 m, in that order.
- Repeat per province across all 28 coastal provinces, not just one example site.
Tools
- Google Earth Engine (Python API) for all raster/vector processing and analysis.
- FAO GAUL administrative boundaries for the 28-province coastal AOI.
- Vanilla JS/SVG front end: no mapping library, no framework, no external tile requests.
My Role
- Authenticated and ran the entire Earth Engine pipeline myself, province by province.
- Caught and fixed real bugs along the way (a band-name mismatch in the GMW union step, a wrong filtering approach for a tiled dataset, an incorrect asset ID) rather than just using what worked on the first try.
- Designed and built the interactive front end: click-to-zoom map, manual per-layer toggles, ranked province table.
Key Takeaways
- Cà Mau alone accounts for roughly 40% of Vietnam's national restorable mangrove area, so the shortlist is not evenly spread.
- Which exclusion filter matters most is regional: reef/shallow-habitat exclusion is near-irrelevant in the Mekong Delta but material in reef-adjacent provinces, while elevation exclusion barely matters anywhere on this flat coastline.
- "Confirmed current extent" is a genuinely hard input to source well: the standard Global Mangrove Watch vector series stops at 2020, so getting a real 2023 baseline required finding and verifying a second, less well-known dataset.
Data Notice
This section is real output from public datasets, not a schematic or sample dataset, but the province boundaries reflect Vietnam's pre-2025 administrative structure (63 provinces), since that's what the underlying FAO GAUL boundary dataset provides. Vietnam merged to 34 provinces in July 2025; this project has not yet been updated to match.