Amazonas Brazil


Forest Carbon
30,770 MtC
(using average carbon stock 207 tC/ha)
IPCC Measurement Methodology
Deforestation vs. Degradation

Reference Levels and Targets

Average Deforestation Rate
726 km²/yr
1996 - 2012
Deforestation Reduction Goal for State/Province and REDD Program
60 % Reduction[1]
for the State/Province and for REDD+ program by 2020
Tons of CO₂e Avoided Target
158 MtCO₂e
by 2012
Needs Identified to Improve Baseline Definition

-Calculation of a state-wide baseline for emissions and state reduction targets that specifically discuss carbon from forest (REDD-specific) and non-forest sources

-Continuous improvements to baseline model over time

Deforestation Dynamics Monitoring

Are current deforestation rates known?
Deforestation Rate Target
350 km²/yr[2]
Deforestation Rates

Expansion of ranching, agriculture and illegal land occupation has increased pressure on forests, particularly in southern Amazonas. In the southreastern counties of Apuí, Manicoré, and Novo Aripuanã, small family agriculture is increasingly being replaced by cattle ranching in large INCRA settlement projects. At the border with Acre and Rondônia, in the counties of Canutama, Lábrea and Boca do Acre, immigration from the neighboring states via the BR-364 and BR-317 highways is accompanied by the expansion of cattle ranching and logging. Forests in other counties, such as Manicoré, Humaitá, Canutama, and Lábrea, are under pressure from expanding industrial crop production, driven by increased access to financial resources and improved technology.

Amazonas has access to PRODES, DETER, CCal, and SAD to assist in monitoring deforestation and forest degradation.  

INPE (National Institute for Space Research): PRODES (Deforestation Monitoring Project for the Legal Amazon—releases annual deforestation data) and DETER (Project to Detect Deforested Areas in Real Time—alert system to support anti-deforestation law enforcement)

LANDSAT satellite images (30-m) are chosen for analysis based on 1) level of cloud cover and 2) proximity to date of reference (August 1);

The images are geo-referenced using SPRING (Sistema de Processamento de Informacoes Geo-referenciadas);

Linear spectral mixing model (MLME) is used to extract soil, vegetation, and shade fractions from the images;

Regions are identified based on similarity and minimum size, and analyzed by an algorithm;

The daily deforestation rate is calculated first and yearly rates are extrapolated from these, taking into account yearly climate differences that may impact detection of deforestation.


Daily MODIS (250-m) images are filtered for cloud cover and spectral resolution bands are fused to increase the detail of images;

NPV (non-photosynthetic components) are estimated and NDFI (normalized difference fraction index) is calculated: values range from -1 (soil) to 1 (forest);

Deforestation is detected by comparing two consecutive monthly NDFIs: a change of -200 to -50 reflects probable deforestation;

The resolution of SAD (250-m) is larger than for the satellites used by INPE (CBERS (20-m) and LANDSAT (30-m)), but in a comparison, 80-90% of deforestation reported by SAD was confirmed

Carbon Calculator (CCal) is a data management platform that can integrate data from SAD, DETER, and PRODES to a user-friendly interface allowing the state to compare data and monitor deforestation and associated emissions.

-Capacity for the State to process its own data and have a mapping unit of its own for faster turn-around for day-to-day detection of deforestation infractions (INPE PRODES data (6.25-ha resolution) is published annually and DETER (25-ha resolution) is published monthly); this increased capacity will require new sources of funding, as the costs of such monitoring are often prohibitive

-Development of a methodology that allows for partial analysis of individual satellite scenes (to eliminate portions under cloud cover) and the merging of these cloud-free segments, in order to generate a more complete and uniform product

Forest Degradation Dynamics Monitoring

Are current degradation rates known?
Forest Degradation Rates

- No information


-Low degradation: predominance of green pixels with some small purple pixels found in low density and frequency

-Medium degradation: dominance of green pixels with slightly larger purple pixels in a mid-level density and frequency

-High degradation: dominance of purple pixels, or smooth green ones, with spots of forest

-LANDSAT and CBERS images (minimum scale of 6.25 hectares) are enhanced for contrast

-Images classified in fairly arbitrary way described above

-86% of degradation/deforestation detected was confirmed, but false negatives (non-detection) is common at low and medium degradation


-Only one level of degradation

-Characterization of degradation dynamics (direct and underlying causes and drivers of forest degradation)

-Direct: selective logging, small-scale wood harvesting and agriculture, population growth/relocation (Manaus)

-Indirect: displacement of activities and people from other parts of Amazonia, due to heavier land use (e.g., commercial agriculture) there, roads

-NDFI is calculated and compared monthly, as described above in deforestation monitoring methodology

-A change between -20 and -49 reflect probable degradation

-Increased image resolution (spatial and spectral); Current INPE DEGRAD method uses high resolution CBERS sensor (2.7 m resolution, 27 km scene width)

-Increase internal processing capacity

Forest Carbon Stocks Quantification

Are forest carbon stocks known?
Forest Classes
-- Forest Types Represented by
-- Plots
Above Ground Carbon Stock
151.2 ± -- tC/ha
Below Ground Carbon Stock
55.7 ± -- tC/ha

Indirect method – continuous forest inventory;

Direct method (allometric method) - cutting and weighing the parts of trees

Development of allometric equations according to the vegetation types studied

Uncertainty index of less than 10% (IPCC indicates below 20%)

-More frequent monitoring of forest inventory plots, to obtain better estimates of carbon sequestration

-Add forest inventory plots in southwestern Amazonas to achieve more complete coverage of the state


Reference Levels and Targets
1. Decreto n.º 2.055, de 19 de dezembro de 2013.
Deforestation Dynamics Monitoring
2. MMA, 2012. REDD+ nos estados da Amazônia: Mapeamento de iniciativas e desafios para integração com a estratégia brasileira, Brasilia,.
3. Prodes / INPE.
4. Prodes / INPE.
5. Prodes / INPE.
6. Prodes.
Forest Degradation Dynamics Monitoring
7. INPE, DEGRAD Mapeamento da degradação florestal na Amazônia brasileira, accessed on 21 June 2013.
8. INPE, DEGRAD Mapeamento da degradação florestal na Amazônia brasileira, accessed on 21 June 2013.
Forest Carbon Stocks Quantification
10. IGBE, 2003. Geoestatísticas de Recursos Naturais da Amazônia Legal. Serie Estudos e Pesquisas Informação Geográfica no. 8.