Data and method
The layers the tool reads, the licence each one carries, every threshold in the score, and the things it does not attempt.
What it does
An early screening tool for coastal development. You place a proposed site, say what is being built, and the tool measures the site against the mangrove stands, drainage channels, shoreline, roads and built-up land mapped around it. It returns a score out of 100, where a higher score means less mapped environmental conflict, and it lists every deduction with the distance that caused it.
It started as BlueBalance, and we renamed it because Guyana’s coast is not blue.
What it does not do
It does not stand in for an engineer, the Environmental Protection Agency, planning approval, an environmental impact assessment, a flood model, land-ownership research or legal advice. A good score is not permission to build and a bad score is not a refusal.
Coverage
The Region 3 coast between Vreed-en-Hoop and Uitvlugt. The layers were cut to a box running 6.74 to 6.98 north and 58.45 to 58.13 west. Put a site outside that box and the tool refuses it rather than screening against layers it does not hold.
A site goes on the map either by tapping it or by typing a latitude and longitude into the two fields on the check page. Both routes end up at the same point and are screened the same way.
Layers and licences
- MangrovesGlobal Mangrove Watch v3.0, the 2020 epoch, CC BY 4.0. 48 shapes, cut out of the published global file. OpenStreetMap has one mangrove way in this region and seven in the whole country, so it was no use here.
- WaterwaysOpenStreetMap through the Overpass API, ODbL 1.0, pulled 28 August 2026. 682 lines, mostly the drainage canals rather than rivers.
- RoadsOpenStreetMap, ODbL 1.0. 4296 lines. Used to work out whether a site can be reached without cutting a new road, so residential streets count.
- Built-up landOpenStreetMap, ODbL 1.0. 41 areas already in residential, industrial, commercial or retail use.
- The coastlineOpenStreetMap, ODbL 1.0. 29 lines.
Protected areas
There are none on this stretch. Guyana’s only coastal protected area is Shell Beach, which is in Region 1, so there is nothing here to score against and no protected-area deduction in the rules.
What it measures
Five things, in this order: how close the site is to a mapped mangrove stand, to a drainage channel, to the shoreline, to the nearest road, and to land that is already built on. The first three are about what the site damages. The last two are about what it costs to reach, since building where access already exists avoids cutting a new road through something else.
Land and water
A housing scheme cannot be screened in the sea, so the tool refuses a site that is not on land. The land shape is built from the same OpenStreetMap coastline the map draws: the 29 segments join end to end into one line across the whole area, and that line is closed off inland to make a polygon. The site is tested against it with a point-in-polygon check.
We checked it against the other layers rather than trusting it. Of 614 road points sampled, 612 fall on the land side, as do 40 of the 41 built-up areas. Of 15 points placed just out to sea from the coastline, none do. It is the real coastline rather than a guess at where the water starts, though it is only as good as OpenStreetMap’s mapping of it.
The score
Every site starts at 100 and loses points, and each deduction shows on the result with the distance that produced it. The thresholds are ours, picked as a screening aid. They are not drawn from any agency’s standard and nobody has signed them off. Where the two heaviest bands came from is set out under the table.
| Factor | Distance | Deduction |
|---|---|---|
| Mangroves | Inside a stand | 40 |
| Within 50 m | 25 | |
| Within 150 m | 12 | |
| Within 300 m | 5 | |
| Waterways | Within 30 m | 15 |
| Within 80 m | 8 | |
| Shoreline | Within 100 m | 15 |
| Within 250 m | 8 | |
| Road access | Beyond 500 m from a road | 15 |
| Beyond 200 m | 7 | |
| Existing development | Beyond 1 km from built-up land | 10 |
| Beyond 400 m | 5 |
The mangrove, waterway and shoreline deductions are multiplied by what is being built: housing 1, small commercial 1.15, light industrial 1.35. Road access and existing development are not multiplied, since those are about siting rather than the weight of the impact.
Why the mangrove bands are set where they are
Mangroves carry the heaviest deduction because building inside a stand is the hardest thing on this list to do lawfully. Guyana has no single mangrove statute, but the species is protected and the Guyana Forestry Commission has to issue a permit before any mangrove tree is destroyed, with clauses in the Environmental Protection Act, the Sea Defence Act, the Fisheries Act and the Forest Acts all bearing on it. A site inside a stand is not a site with a planning problem, it is a site that needs a permit that may not come.
The 50 m band is the width at which coastal greenbelt policies start. A 2025 review of wave attenuation across mangrove forests found that greenbelt rules worldwide prescribe widths of 50 m to 200 m, and that the median reduction in wave height across the first 100 m of forest is 62 per cent. Our 50 m and 150 m bands sit inside that range. We are not claiming the review produced our numbers. It tells you the numbers are the right order of size, and that a stand thinner than 50 m is doing measurably less work.
Why the shoreline bands are set where they are
The Sea Defence Act reaches further inland than most people expect. Section 2 puts all land fifty feet landward of the centre of a sea or river dam inside the definition of a sea defence, and the same fifty feet landward of the crest of any reef, bank or other natural feature doing the same job. Section 20 makes it an offence to remove earth, sand, shell or vegetation from a sea defence or the land along it without the Board’s consent.
Fifty feet is about 15 m. Our first shoreline band is 100 m, so the tool flags sites well before the statute does. That is deliberate for a screening aid: the legal reserve is the point at which you are already in trouble, not the point at which it is worth thinking again.
The road access and existing development deductions have no equivalent source. They are ours, and they are about what a site costs to reach rather than what it damages.
Finding a better spot
The search steps out from the site on sixteen bearings at 250 m, 500 m and 1 km, drops any candidate outside the box or sitting in a mangrove stand or on a drainage channel, scores the rest by the same rules, and shows the best three that beat the original. It is a fixed grid, so the same site always returns the same suggestions.
Missing data
When nothing is found within about 3 km the distance reads “over 3 km”, because the search window stops there. A layer that fails to load reads “not assessed” and the score is marked incomplete rather than counted as clear. Not assessed means we did not look, not that the site is clear.
Sources
Three different things end up in one number. The distances are measurement. The sources below are evidence, published law and published research, and they set the order of size the bands sit at. The point values are judgement, and they are ours.
- Sea Defence Act, Chapter 64:02Sections 2 and 20. The fifty foot reserve landward of a sea defence or natural protective feature, and the offence of removing vegetation from it. faolex.fao.org
- Mangroves as a protected speciesEnvironmental Protection Agency Guyana, public education note. Permit required from the Guyana Forestry Commission before any mangrove tree is destroyed, with the relevant clauses spread across four Acts. epaguyana.org
- Greenbelt widths and wave attenuationvan Wesenbeeck, van Zelst, Antolinez and de Boer, Quantifying uncertainty in wave attenuation by mangroves to inform coastal green belt policies, Communications Earth and Environment 6, 258, 2025. Prescribed widths of 50 m to 200 m, and a median 62 per cent wave height reduction over the first 100 m. nature.com
These explain why the bands sit where they do. None of them approves the point values, and nobody has reviewed the deductions themselves.
Limitations
- Flood risk. We found no public flood map for this coast.
- Sea defence reserve boundaries. Those are not published as data anywhere we could find.
- Land ownership. Nothing here knows who owns the ground.
- Anything outside the box above, where we hold no layers.
The mangrove and shoreline data is a snapshot, not a live feed. A stand cleared last month still shows as standing until the source is republished. The point values would be the first thing to calibrate if a coastal engineer or the EPA were willing to look at them.