Forests
What does tree-cover loss mean?
Tree-cover loss is the removal of tree canopy in a mapped pixel, detected from Landsat imagery. It counts canopy disappearing for any reason — logging, fire, disease, storm, or a plantation being harvested on schedule. Its publisher states plainly that it is not equivalent to deforestation, and every figure depends on a canopy-density threshold that is a choice.
What the satellite sees
It sees canopy present in one year and absent in the next. It does not see why, and it does not see what came back.
The Hansen / University of Maryland dataset, published through Global Forest Watch, compares Landsat imagery year on year at 30-metre resolution and flags pixels where tree canopy was removed. It is a measurement, not a model — but what it is a measurement of is canopy, not forest.
So a eucalyptus plantation felled on a seven-year rotation registers as loss, and registers again every seven years. A patch of primary forest cleared for a mine registers once. Both are "tree cover loss" and only one is deforestation.
The threshold is the other trap. GFW's headline figures use 30 per cent canopy density in the year 2000 as the definition of "tree cover". A 10 per cent threshold gives a much larger area, and the thresholds are cumulative nested subsets — the 10 per cent figure includes the 30 per cent figure. Summing across thresholds double-counts and gives roughly eight times the real total.
- Canopy density threshold
- The minimum crown cover in 2000 for a pixel to count as tree cover. A choice, not a fact, and it must be printed wherever the number is.
- Gross loss
- Canopy removed. It is not netted against regrowth, so it is not a net change in forest area.
- Primary forest
- Mature natural forest. GFW reports loss within it separately, and that subset is much closer to what people mean by deforestation.
The numbers, and the choice behind them
India's cumulative tree-cover loss, as published, with the threshold that defines it.
Canopy density threshold used
30 % canopy in 2000
Global Forest Watch headline convention
Cumulative tree-cover loss
2.43 million hectares
Hansen / UMD, published by Global Forest Watch
Same loss, in square kilometres
24,257 km²
Hansen / UMD, published by Global Forest Watch
Period covered
2,001 – 2,025
Hansen / UMD annual series
Same loss at a 10 per cent threshold
26,19,067.05 ha
Hansen / UMD, same dataset, different threshold
The last two rows are the same country over the same period. The only difference is where the threshold was set. That is why a tree-cover-loss figure quoted without its threshold cannot be checked and should not be used.
How to read a loss figure
Four checks. A figure that survives them is usable.
What threshold? Without it the number has no definition.
Gross or net? This series is gross loss. It does not subtract gain, and gain is tracked separately and less well.
All tree cover or primary forest? The primary-forest subset is much smaller and much closer to "deforestation".
Which years? The series runs annually from 2001. Methodology improved over that period, and the publisher advises against reading short-run changes as trend.
What this number cannot tell you
- It cannot distinguish a harvest from a clearance. Both are canopy removed.
- It cannot see regrowth in the same figure, so it is not a net change in forest area.
- It cannot be compared with India's own forest-cover figures without accounting for both definitions — see forest cover or tree cover.
- Its thresholds cannot be summed. They are nested subsets, and adding them multiplies the answer.
Where this comes from
Every figure above is read out of one of these, at the address printed beside it.
- Global Forest Watch — tree cover loss World Resources Institute, using Hansen / University of Maryland data · The dataset and its own caveat that loss is not deforestation
- GFW Data API World Resources Institute · The authoritative interface; a key is required
Reuse freely — CC BY 4.0. The grant covers this page. Each source above keeps its own terms, which is why every one of them is named.
Next
The reading this explains, the explanations next to it, and the one place to stand in it.