One click on the map and Naturametrics draws five circles, pulls forty years of MapBiomas, estimates the age of the vegetation, counts the fragments, measures biomass and shows what Brazil's environmental agency (IBAMA) has recorded nearby. This walkthrough opens a real plot of the National Forest Inventory in the order the app presents it, with where each number comes from and where it stops.
Fetching the example point from Earth Engine…
measured computed from satellite imagery or a public service. estimated a modelled product, with a known error margin. missing a layer that did not respond this time.
Naturametrics needs no polygon: any click on the map of Brazil becomes a study point. The search box takes a postcode, an address, a coordinate (even one pasted from Google Maps) or a municipality, but it only moves the map — the click chooses the point.
In the app: the Search box and a click on the map, or a direct link with
/?lat=…&lon=… (or /?ifn= for an inventory plot).
Brazil's National Forest Inventory (IFN) measures trees on the ground at a grid of plot clusters across the country. The app carries the 17,479 published points, filterable by region, state, municipality and biome. Clicking a plot uses its published coordinates, not the exact spot you clicked: 40 m away is already a different question.
In the app: the IFN, IBAMA and uploaded data group. Hovering a plot previews its land cover; clicking starts the analysis. A switch turns on multiple selection.
Every analysis runs in five circles, from 0.5 to 10 km in radius. They are full discs, not rings: the 5 km disc contains the 2 km one. The question is "what is the landscape like up to this distance", so the larger radius always includes the smaller.
In the app: the 0.5 km / 1 km / 2 km / 5 km / 10 km buttons above each chart. The square buffers switch swaps the circle for a square whose side is the diameter.
Each column is one MapBiomas year, with the disc's hectares in each land-cover class. The map shows the latest year's land cover clipped to the largest disc, as the app does as soon as a point is chosen.
In the app: the Land-use history panel, with the radius selector and the % switch; beside it, the latest year's main classes.
The same series, now as the share of each disc that is still native vegetation, one line per radius. When the small disc falls before the large one, conversion started near the point and spread; when it falls later, it came from outside.
In the app: the Land-use history panel with the % switch, changing radius. Here the map shows the first year of the series.
The discs are averages. The 30 m pixel right under the point shows what the classifier saw there, year after year, with nothing to smooth it. It is the most direct record and also the noisiest: one misclassified year shows up as a change that never happened.
In the app: the point's series goes into every export, always with the warning that a single pixel is noisy.
The change mask compares native vegetation in 2008, the Forest Code's cut-off, with the latest year. Red is what was native and no longer is; green is what came back. The sum is the same at every radius.
In the app: the Biomass and forest change group, Change since 2008 layer.
MapBiomas' Deforestation and Secondary Vegetation product records when each pixel was cleared and when it grew back. From that comes the age of today's vegetation. Vegetation never cleared since the record began has no known age: all we know is that it is at least as old as the series.
That is why the app never gives a mean age: it shows the median of the dated part and, next to it, how much has no date.
In the app: the Vegetation age tab, with the point's series and the distribution by radius.
A landscape has a shape too. The app counts fragments — continuous patches of one class — and measures edge density and class diversity at each radius. In the same disc, more fragments and more edge mean a more broken-up landscape.
In the app: the Metrics tab. Fragments are vectorised inside each disc with no size limit; a fragment cut by the disc's edge counts only the part inside.
Above-ground biomass comes from ESA CCI, in 100 m pixels for ten years between 2007 and 2022. The product publishes each pixel's standard deviation alongside, and the app carries it into the chart: the shaded band is the product's own uncertainty.
In the app: the Biomass tab, in Mg/ha by radius.
IBGE's vegetation map is drawn as polygons at 1:250,000; MapBiomas classifies every 30 m pixel every year. Where clearing is recent, IBGE still draws forest.
In the app: the IBGE tab, with the matrix by radius, and the IBGE layer in the reference group.
Embargoed areas and infraction notices come straight from IBAMA's public services, for the map's extent. They show that enforcement happened, not what happened in each pixel: an embargo can cover more than the clearing, and a notice's coordinate is where it was recorded, not necessarily where the infraction was.
In the app: the IFN, IBAMA and uploaded data group, Embargoes and Infraction notices layers.
The Download data button writes an ODS spreadsheet with one sheet per table: history by radius, the point's pixel, age, metrics, biomass and the provenance of each. With multiple selection the same spreadsheet covers hundreds of plots at once, up to a whole biome.
In the app: the Download data button at the top, and the point's PDF report.
More examples:
São Félix do Xingu (PA), Amazon ·
Água Boa (MT), Cerrado ·
Cáceres (MT), Pantanal ·
Colniza (MT), this page's example.
To read another point, change the plot (?ifn=) or pass a coordinate (?lat=…&lon=…).
The numbers come from the app's own services: this page's example is stored with it, and any other point is computed
on request. Versão em português.
Built with Reflex