in Brazil

Deforestation in Brazil’s Amazon hits highest rate for ten years, according to Brazil’s government. It’s way worse according to Global Forest Watch

Data released by the Brazilian government last week reveals that deforestation in Brazil’s Amazon has reached its highest rate since 2008. In the period August 2017 to July 2018, an area of 7,900 square kilometres of forest was cleared. That’s an increase of 13.7% compared to the previous 12 months.

Edson Duarte, Brazil’s outgoing environment minister, said in a statement that illegal logging was behind the increase in deforestation. He blamed “an upsurge in organised crime”, and called for the government to increase its policing of the Amazon rainforest.

Between them, the states of Pará, Rondônia, and Mato Grosso accounted for almost three-quarters of the deforestation. The state of Acre, which since 2010 has been the site of the world’s first jurisdictional REDD programme, saw deforestation increase by 84% compared to the previous 12 months.

Norway’s US$1 billion REDD deal. For what, exactly?

Norway’s US$1 billion REDD deal with Brazil was announced in December 2008 (that’s the dotted line on the graphic below). Before that deal, deforestation came down dramatically. But between 2009 and 2014, when most of the Norwegian money flowed to the Amazon Fund, deforestation remained pretty much stable.

Since then, it’s been going back up:

One of the reasons for the increase was that in 2013 President Dilma Rousseff approved a new forest code, that gave small landholders amnesty for previous deforestation.

Deforestation in Brazil is likely to increase dramatically under the government of President-elect Jair Bolsonaro. During his election campaign, Bolsonaro committed to limiting fines for illegal logging, allowing mining in protected areas, weakening indigenous peoples’ rights, and weakening the environment agency.

BBC News reports one of Bolsonaro’s aide’s as saying that the agriculture and environment ministries will be merged under Bolsonaro’s government.

Marina Silva, a former environment minister, said,

“This disastrous decision will bring serious damage to Brazil and will pass on to consumers abroad the idea that all Brazilian agribusiness survives thanks to the destruction of forests.”

Deforestation is even worse than the government’s figures

The actual area of deforestation in Brazil is even worse than the figures released by the Brazilian government. Data released by the University of Maryland on Global Forest Watch shows far higher rates of tree loss than the official Brazilian government forest monitoring system, PRODES.

In 2017, for example, PRODES reported 690,000 hectares of forest loss in the Brazilian Amazon. Global Forest Watch’s figure for 2017 was 3.47 million hectares.

The differences in the deforestation figures are accounted for by several factors. Global Forest Watch measures forest loss from January to December. PRODES uses August to July.

PRODES only measures new deforestation in primary forest. Global Forest Watch includes forest loss in secondary forest, and forest degradation from fire.

PRODES only includes areas of forest lost that are larger than 6.5 hectares. Global Forest Watch records tree loss on an area as small as 0.1 hectares.

Despite the fact that Brazil’s reduction in deforestation came before REDD payments started, Brazil is often reported to be a REDD success story.

The discrepancy between the Brazilian government’s deforestation figures and those from Global Forest Watch is important. The figures for deforestation of secondary forests, forest degradation from fires, and areas of deforestation covering areas less than 6.5 hectares must be included in Brazil’s deforestation statistics.

Meanwhile, the increased rate of deforestation in Brazil is terrifying. As Marcio Astrini, Greenpeace Brasil’s public policy coordinator, points out the tipping point at which the Amazon rainforest will disappear is approaching rapidly. Astrini told The Guardian that,

“A moment will arrive in which the accumulation of this deforestation will cause an effect in which the forest will stop being a forest. The scientists calculate this is between 20-30%. We are very close to the 20%.”


Leave a Reply

  1. This comparison is utter nonsense!! PRODES has been operating continuously for 30 years, and has a clear and well defined methodology. It has been the basis of Brazil’s policies. Despite ups and downs in climate policy in Brazil, PRODES is trusted by government, NGOs, media, and scientists. There are over 200 peer-reviewed papers that use PRODES data. You cannot criticise PRODES for not measuring what it does not aim to measure. As a former director of INPE (Brazil’s National Institute for Space Research), the institute that does PRODES, I can attest to how difficult it has been to keep the integrity and independence of the system, despite enormous political pressure. Furthermore, there is nothing like PRODES anywhere. Please name any other national forest monitoring system where all the information is high quality and is openly available. PRODES has a measured accuracy of 95%. It cannot be compared with a GFW whose accuracy has not been verified independently. Those who criticise PRODES should consider the counterfactual. It the system did not exist, do you think the Brazilian government would trust the GFW estimates, which has not been verified independently? Would the Brazilian forest be better protected?

    Gilberto Camara, director of the GEO Secretariat, former director of INPE (2006-2012).

  2. @Gilberto Camara – Thanks for this comment. This post was not intended to be a criticism of the accuracy of PRODES measurements of deforestation, or that PRODES exists. It’s great that PRODES has existed for 30 years. Brazil’s data on its forests is excellent. The counterfactual (PRODES not existing) is way, way worse.

    However, PRODES only measures new deforestation in primary forests. That means PRODES is only telling part of the story of deforestation in Brazil. Deforestation in secondary forests, forest degradation from fire, and deforestation of areas smaller than 6.5 hectares cannot just be ignored in any country’s deforestation statistics. Can they?

    My post links to a blog post by Global Forest Watch that looks into the differences between PRODES and GFW figures for deforestation. Here are two extracts from GFW’s post:

    PRODES results demonstrate that while the battle is far from over, Brazil has had impressive results in reducing the deforestation of primary forest since the early 2000s. The University of Maryland data shows an increase in other tree cover losses, indicating new battles to be fought for the ecological integrity of the Amazon….

    While the additional changes captured by the University of Maryland data may not be caused by clear-cutting of pristine forests, the loss still has an impact on ecosystem services, biodiversity, and climate. Studies show that degradation from fire is now counteracting the reduction in emissions from reducing deforestation in Brazil. Thus, both PRODES and the University of Maryland data have a critical role in monitoring one of the most important forests in the world.

  3. Dear Chris,

    After 30 years of work monitoring the Amazon deforestation, I never cease to be amazed and disturbed on how few so-called experts on forest monitoring from satellites have yet to learn the lessons of PRODES.

    I would like to take your points one by one:

    1. Minimal mapping area

    Any consistent, long-term area survey using remote sensing has to establish, when it begins, its minimal mapping area (MMA). Back in 1988, when PRODES started, there was no internet, computers had limited graphical capabilities, and interpretation had to be done in analogue images (printed). At that time, the PRODES team set the MMA of 6,25 ha as a practical threshold that would ensure a reliable and reproducible result. To maintain the coherence of the PRODES time series, INPE has keep this threshold since.

    We have found that, on the long term, this MMA has not been a limiting factor for PRODES. This has to do with the nature of Amazon deforestation. Almost all deforestation is motivated by large-scale farming and/or large pasture for big cattle herds. Almost, if not all, areas less than the PRODES MMA are accounted for in subsequent years, because of the way farming works in the Amazon.

    2. Emissions from fires

    INPE has been monitoring fires in Brasil since the 1980s. We have one of the world’s longest running fire monitoring systems. Since its inception, we have tried to establish strong correlations between fires and actual burnt areas. The results are not consistent enough to be able to be used together with the PRODES clear-cut areas. Roughly speaking, there are two situations: (a) forest fires that lead to clear-cut areas; (b) forest fires that degrade the forest, but do not lead to clear cuts. In case (a), sooner or later these areas will be accounted for by PRODES. In case (b), one would have to develop a very reliable estimate of the carbon emissions resulting from fire, which to my understanding, does not yet exist.

    One thing is to write a scientific paper that estimates emissions from forest fires in small, well-controlled areas. Another is to have a method that is robust for an area the size of Europe, and works well year after year. We are not there yet.

    3. Secondary vegetation

    If you and the GFW team had ever read the PRODES methodology, they would know that PRODES does not account for secondary forests. In other words, once a natural forest area is clear-cut, all of the emissions resulting from this cut are accounted at once. And we don’t look back. Once cleared, always cleared. This practice is conservative in best sense of the word. It avoids the need for stocktaking in secondary forests. Otherwise, one would need to measure carefully the carbon sequestration resulting from forest regrowth in secondary forest areas, and then estimate the carbon emission when the area is cut again.

    The PRODES forest area is a strictly monotonic time series. The PRODES forest area, which captures how much native forest is there in Amazonia, only decreases, This is based on a strong (and correct) assumption about the time it takes for a tropical forest to reach the same level of biodiversity, once it is clear-cut. Current estimates range from 50 to 100 years for forest patches that are left undisturbed after clear-cuts. Thus, PRODES is currently the best guide for the loss of biodiversity rich areas in Amazonia.

    To ensure that previously clear-cut areas are not double counted, the PRODES classification methods include a temporal dependency. PRODES has a mask composed of all areas that have been identified in previous years as clear-cut. These areas are off-limits to the interpreters. They only identify newly clear cut areas. In statistical terms, the interpretations of different years are dependent, thus ensuring a strong temporal coherence.

    When GFW adds secondary forest emission to the PRODES emissions at a given year, this is so unwise that it is not even wrong, as Wolfgang Pauli would say. Since all of the areas of secondary forest in the Amazon have been clear-cut in the past (and thus already accounted for), what GFW is doing is double-counting, which is plain nonsense!!

    Accounting for secondary forests is not an easy task. The PRODES database shows that secondary forests in Amazonia are highly dynamic. The cycle of cut-regrowth-cut depends on many factors, including the economic gains to be accrued from the land. There are cases when clear-cut areas are left without use for some years, leading to forest regrowth. Then, in times of economic growth, these areas may be cut again.

    One of my strong criticisms of most of those involved in REDD+ is their failure to learn from PRODES. Most REDD+ algorithms (including those of GFW) treat images from different years as independent data. In so doing, they lose temporal coherence. I know of some tropical countries involved in REDD+ that they report forest gain, as well as forest loss. Again, this is not even wrong!! This implies that areas of secondary forest can be counted together with areas of natural forest. From the point of view of emissions, biodiversity and carbon accounting, this is nonsense!!

    This lack of awareness is quite serious. It has endangered REDD+ and reduced its effectiveness and global impact. As REDD+ moves into the umbrella of the GCF, if the REDD+ communities continues to praise PRODES, but ignore why it works the way it does, they have only themselves to blame.

    I think the matter is serious enough to warrant a guest blog post at you site, that I would be glad to write.

    Gilberto Camara
    Secretariat Director, Group on Earth Observation
    Former director, INPE (National Institute for Space Research).

    P.S. Do yourself and the REDD community a favour: read the PRODES methodology, which is available at I know, it’s in Portuguese. Use google translate.

  4. @Gilberto Camara – Thanks very much for this detailed response (and for your patience!). I agree with you that this is an important issue. I’d like to accept your offer of writing a guest post – I’ll be in touch by email.