Research ArticleENVIRONMENTAL SCIENCES

Types and rates of forest disturbance in Brazilian Legal Amazon, 2000–2013

See allHide authors and affiliations

Science Advances  12 Apr 2017:
Vol. 3, no. 4, e1601047
DOI: 10.1126/sciadv.1601047
  • Fig. 1 Sample-based estimates of the total 2000–2013 tree cover loss area in BLA.

    Estimates are disaggregated by predisturbance forest type and disturbance type. Selective logging and fire categories do not represent complete tree cover loss but rather the area affected by these processes. See table S1 for SEs of the estimates.

  • Fig. 2 Sample-based estimates of annual tree cover loss area in BLA.

    Estimates are disaggregated by (A) disturbance type and (B) predisturbance forest type and disturbance type group. Selective logging and fire categories do not represent complete tree cover loss but rather the area affected by these processes. See tables S2 and S3 for SEs of the estimates.

  • Fig. 3 The 2000–2013 state-level tree cover loss area estimates.

    Estimates are disaggregated by disturbance type in (A) all forests, (B) primary forests, (C) natural woodlands, and (D) secondary forests, woodlands, and plantations. See table S4 for SEs of the estimates.

  • Fig. 4 Study area—BLA.
  • Fig. 5 Annual human forest clearing by state.

    (A) In all forests, (B) in primary forests, (C) in natural (primary) woodlands, and (D) in secondary forests, woodlands, and plantations. See table S5 for SEs of the estimates.

  • Fig. 6 Estimated annual percent of gross AGC loss from human clearing of primary forests versus other forest disturbances.

    Other disturbances include human clearing of woodlands and secondary forests, fires, and selective logging. Uncertainty is based on the range of mean AGC estimates per forest type from Table 1.

  • Fig. 7 Comparison of sample- and map-based annual deforestation estimates.

    Three-year averages of sample-based annual tree cover loss estimates by disturbance type (stand-replacement disturbances, selective logging, and fire) and forest type (primary forests and other forests and woodlands) compared with 3-year averages of annual map-based deforestation estimates from PRODES and Souza et al. (20) and tree cover loss estimates from UMD map.

  • Fig. 8 Comparison of forest degradation estimates.

    Sample-based fire and selective logging estimates are compared with DEGRAD map within sampling region and Souza et al. (20) degradation estimate. Error bars represent ±SE.

  • Fig. 9 PRODES forest mask and 2001–2013 forest cover loss and UMD 2001–2013 tree cover loss within BLA.
  • Fig. 10 Population from which the simple random sample of 10,000 pixels was selected.
  • Table 1 Mean AGC density in predisturbance forest types (MgC/ha).

    For carbon data source description, see Materials and Methods.

    Sample size (n)Predisturbance (year 2000) AGC density (MgC/ha)
    Baccini et al. (48)Saatchi et al. (50)Avitabile et al. (51)Range
    Primary forests270299.394.977.477.4–99.3
    Natural (primary) woodlands38727.528.418.918.9–28.4
    Secondary forests, woodlands,
    and plantations
    81948.448.344.844.8–48.4
  • Table 2 Comparison between annual deforestation estimates.

    (A) Current study (human clearing of primary forests), (B) PRODES, and (C) Souza et al. (20). Total difference between (A) and (C), and (B) and (C) is calculated only for 2001–2010 because of the absence of Souza et al. (20) estimates for 2011–2013.

    Area of deforestation (Mha)
    2001200220032004200520062007200820092010201120122013Total
    (A) Sample1.512.302.772.592.331.521.381.240.730.560.650.530.6318.72
    (B) PRODES1.822.172.542.781.901.431.171.290.750.700.640.460.5918.22
    (C) Souza et al. (20)1.722.332.222.442.221.601.381.241.200.5516.91
    Difference between estimates (%)
    2001200220032004200520062007200820092010201120122013Total
    Sample versus PRODES
    (A − B)/A × 100%
    −20.56.08.2−7.418.36.315.3−4.4−2.3−25.21.413.56.32.7
    Sample versus Souza
    (A − C)/A × 100%
    −14.2−1.419.65.54.5−4.90.1−0.3−64.21.70.04
    PRODES versus Souza
    (B − C)/B × 100%
    5.3−7.812.512.0−17.0−11.9−18.03.9−60.521.5−2.3
  • Table 3 Monthly distribution of sample pixels identified as fire disturbance, 2000–2013.

    “End of year—uncertain date” indicates that the fire scar was observed in the first 16-day composite of the year and there were no cloud-free 16-day composites at the end of the previous year; in this case, fire was attributed to the end of the previous year.

    JanFebMarAprMayJuneJulyAugSepOctNovDecEnd of year—uncertain date
    Number of pixels225351645931831915
  • Table 4 Comparison between types of human clearing in primary forests (2001–2013) identified from the sample and postdeforestation land-use types from TerraClass.

    Cell entries of the confusion matrix denote the number of sample pixels in each category (a mixed loss pixel was recorded as 0.5). The 113.5 sample pixels with TerraClass showing later loss date than the current analysis (for example, 2004 instead of 2001–2003) were excluded from the analysis and are not displayed in the table.

    Human clearing
    of primary forests
    (current study)
    TerraClass
    PastureAnnual
    agriculture
    (cropland)
    Mosaic of land
    uses
    Secondary
    regrowth and
    reforestation
    ForestNonforested
    areas
    WaterNo
    data
    MiningUrban
    areas
    Total
    Agroindustrial
    clearing
    Pasture94411.535129250561.580.5001507.5
    Crops5286061017.504.500176
    Trees431822000020
    Small-scale clearing73.5013.543.51301037.500281
    ConstructionRoads5.500.52.515.53000027
    Other2.5010.510.500005.5
    Dam construction30004200009
    Mining20000.5000002.5
    Total1086.5100.551189.5413914.592.5002028.5
  • Table 5 Comparison between types of human clearing in natural woodlands (2001–2012) identified from the sample and 2013 land use according to TerraClass Cerrado.

    Cell entries of the confusion matrix denote the number of sample pixels (1 and 0.5 loss) in each category.

    Human clearing of
    natural woodlands
    (current study)
    TerraClass Cerrado
    PastureAgriculture
    (annual and
    perennial)
    Mosaic of
    land uses
    ForestryNatural
    vegetation
    WaterNo dataTotal
    Agroindustrial
    clearing
    Pasture1153014100160
    Crops25.573.5039.501112.5
    Trees20021005
    Small-scale clearing3.50003.5007
    ConstructionRoads00.5104005.5
    Other00101002
    Dam construction00001405
    Mining00000000
    Total14677266141297
  • Table 6 Comparison between types of human clearing in all forest types (2001–2013) identified from the sample and land cover/land use according to MapBiomas.

    Cell entries of the confusion matrix denote the number of sample pixels (1 and 0.5 loss) in each category.

    Human clearing of
    all forest types
    (current study)
    MapBiomas
    PastureAgricultureForestPlanted
    forest
    Coastal
    forest
    WaterOtherNo dataTotal
    Agroindustrial
    clearing
    Pasture997.573.553600171702325
    Crops87.5132.528.50001010349.5
    Trees313000020054
    Small-scale clearing61.50271.50001211455
    ConstructionRoads9.511500013038.5
    Other40100010015
    Dam construction00000184022
    Mining1.500001608.5
    Total1164.5208882002099213267.5
  • Table 7 Sample size (number of pixels) and area of target region by state in BLA.
    StateSample size, nArea of target region, Atot (Mha)
    Acre3102.74
    Amapá1511.29
    Amazonas8777.15
    Maranhão1,27811.50
    Mato Grosso2,55022.75
    Pará3,03026.37
    Rondônia9097.81
    Roraima2101.88
    Tocantins6855.88
    BLA total10,00087.36
  • Table 8 Types of forest disturbance.

    Images are subsets of pre- and postdisturbance (top and bottom, respectively) for annual Landsat composites (band combination, 5-4-3). Small red rectangles represent sampled pixels.


    Embedded Image
  • Table 9 Predisturbance forest types.

    Images are subsets of pre- and postdisturbance (top and bottom, respectively) for annual Landsat composites on the left (band combination, 5-4-3) and Google Earth imagery on the right. Small red rectangles represent sampled pixels.


    Embedded Image

Supplementary Materials

  • Supplementary material for this article is available at http://advances.sciencemag.org/cgi/content/full/3/4/e1601047/DC1

    table S1. Total 2001–2013 forest cover loss in BLA by disturbance type and forest type (Mha ± SE).

    table S2A. Annual forest cover loss in BLA by disturbance type in all forests (Mha ± SE).

    table S2B. Annual tree cover loss in BLA by forest type (Mha ± SE), all disturbance types.

    table S3. Annual tree cover loss in BLA by major disturbance types and types of forest cover (Mha ± SE).

    table S4. Disturbance types by state and forest type (Mha ± SE), corresponding to Fig. 3.

    table S5. Annual human forest clearing by state and forest type (Mha ± SE), corresponding to Fig. 5.

  • Supplementary Materials

    This PDF file includes:

    • table S1. Total 2001–2013 forest cover loss in BLA by disturbance type and forest type (Mha ± SE).
    • table S2A. Annual forest cover loss in BLA by disturbance type in all forests (Mha ± SE).
    • table S2B. Annual tree cover loss in BLA by forest type (Mha ± SE), all disturbance types.
    • table S3. Annual tree cover loss in BLA by major disturbance types and types of forest cover (Mha ± SE).
    • table S4. Disturbance types by state and forest type (Mha ± SE), corresponding to Fig. 3.
    • table S5. Annual human forest clearing by state and forest type (Mha ± SE), corresponding to Fig. 5.

    Download PDF

    Files in this Data Supplement:

Navigate This Article