Fig 1.
Geographic centroids of agrarian reform settlement areas.
Settlements distributed throughout Brazilian Legal Amazonia were considered in terms of (A) the overall natural vegetation conversion analysis (N = 1911 settlements) and (B) settlements for which pre- and post-settlement deforestation data were available up to 7 years before and 11 years after the onset of agrarian settlements (N = 300). Sizes of circles are proportional to the log-transformed polygon areas of agrarian settlement projects, and colors indicate the proportion of land area that had been converted as of 2011 (darker colours indicate higher conversion rates).
Table 1.
Explanatory variables used in deforestation and land-use conversion models in this study.
Table 2.
Summary of land cover, fire incidence and rural population size for the entire Legal Amazon administrative region, distinguishing areas within and outside INCRA agrarian reform settlements.
Fig 2.
Settlement growth since the 1960s.
Cumulative growth over time in both the total area allocated to INCRA settlement areas (shaded area) and the number of settler families (solid line) occupying those settlements throughout the Brazilian Legal Amazon administrative region.
Fig 3.
Agrarian reform settlement areas within Brazilian Legal Amazonia.
Main map showing (a) land cover as of 2011 and state administrative boundaries (topographic map background from [52]). Smaller maps show (b) the level of traffic intensity along major roads, expressed in terms of the number of heavy cargo and passenger vehicles per day, (c) the spatial distribution of fires (hot pixels) in 2011, and (d) major classes of agricultural soil fertility throughout the region.
Fig 4.
Proportion of vegetation loss inside and outside settlements.
Municipal county scale proportion of natural vegetation cover lost both inside and outside the boundaries of INCRA agrarian settlements for the 568 counties of Legal Amazonia containing settlements. Sizes of circles are proportional to the size (log10 x) of counties. All circles above the dashed diagonal line indicate counties that have lost a higher proportion of forest cover within settlement polygons than areas outside.
Fig 5.
Extractive production before and after settlements.
Area-weighted average extractive production of (A) raw timber (roundlogs), (B) firewood, and (C) charcoal for counties containing agrarian settlements, estimated on the basis of municipal county scale data overlapping all settlements for which annual offtake data from IBGE (2000–2010) were available (N = 424) for up to 8 years before (yr‒8) and 12 years after (yr+12) the formal establishment of settlements. Error bars indicate standard errors.
Fig 6.
Deforestation and fire incidence before and after settlements.
Variation in annual deforestation and fire incidence rates within agrarian settlements within Brazilian Amazonia. Relative annual deforestation rate is expressed as the proportion of additional loss in the forest cover remaining from the previous year. Solid and shaded circles represent mean deforestation rates and mean fire incidence calculated from 300 and 1397 settlement areas, respectively, for which reliable data are available. Error bars indicate standard errors.
Fig 7.
Forest cover retention before and after settlements.
Mean annual deforestation within 300 agrarian settlement areas across Brazilian Amazonia, up to 7 years before (yr‒7) and 11 years after (yr+11) their formal establishment. Vertical dashed line indicates the official recognition of any given settlement at year zero (yr0). Forest loss prior to year zero was often associated with agricultural activity by land squatters. Error bars indicate standard errors.
Table 3.
GLM model results (slope coefficients and associated ± SE) of predictors of cumulative vegetation (forest and cerrado) conversion rate as of 2011 within agrarian settlement areas across the Brazilian Legal Amazon region (N = 1,911); and mean difference in annual deforestation rate before (until year–1) and after (since year+1) the creation of settlements (N = 300).
Fig 8.
Human population density vs land conversion.
Relationship between human population density (HPD), expressed in terms of the cubic root of households/km2, both (A) within and (B) outside INCRA settlement project areas and the proportion of natural forest, cerrado and grasslands in these areas that has been converted to any agropastoral land-use. R2-values, which represent best-fit lines from 3-parameter sigmoidal functions, indicate that landscape-scale HPD has played a greater role in driving historical vegetation conversion rates within settlement areas than the density of families of small farmers who were actually relocated into those settlements.
Fig 9.
Proportion of forest cover and human population density.
Pre- and post-settlement (a) forest cover and (b) mean post-settlement deforestation rates within agrarian reform settlement areas across the Brazilian Amazonia. Circle size represents the landscape-scale human population density (log10 households km-2) within a 10-km buffer area outside settlements.
Fig 10.
Annual deforestation polygons in selected settlements.
Examples of land cover change within three INCRA settlements of the Brazilian Amazon. The Rio Juma and Acari Settlements (top) were created in 1981 and 1982, respectively, far from the southern Amazonian agricultural frontier, whereas the more recent Juruena Settlement (created in 1997) lies in a vibrant deforestation hotspot of the northern state of Mato Grosso. These settlement polygons had experienced very little loss in forest cover prior to the arrival of settlers, as indicated by the color-coded deforestation chronosequence (where darker colors indicate more recent deforestation).