Fig 1.
Map of the study area and the land-use history of the active shifting cultivation landscape retrieved by remote sensing.
(a) Map of the Amazon Basin indicating the location of the study area (green diamond). Landsat 8 image (September 29th 2014; band combination 6/5/2) of the study area (buffer zone) showing the location of riverine communities and respective number of households. The red square indicates the zoomed area represented in maps c and d. Maps show the land-use history of agricultural fields during the time period of 1985 until 2015. Segments represent pixels that have similar land-use history; (b) Map of the study area showing all the segments belonging to the swidden-fallow fields, i.e. which had at least one clear-cut event detected over the 1984–2015 period. Segments are coloured according to the date of the first clear-cut event detected by the BFAST analysis; (c) Map of the number of clear-cut events (slash and burn events) detected at each segment during the monitoring period; (d) Map of the age of secondary-forest-fallows in 2014 for each segment.
Fig 2.
Secondary forest age and land-use intensity of the shifting cultivation landscape in 2014.
(a) Cumulative area with secondary-forest fallows with different ages in 2014. (b) Cumulative area of fields that experienced different number of clear-cut events (swidden-fallow cycles). Dashed lines indicate the age of secondary forests (a) and the number of clearcut events in 50% and 80% of the fields.
Fig 3.
Expansion of shifting cultivation fields over old-growth forest between 1987 and 2014.
a) Total area cultivated per year, i.e. total area where a clear cut occurred regardless if the previous land cover was an old-growth forest or a secondary-forest fallow (in black), and the area where clear cuts occurred in old-growth forests (in grey). b) The percentage of clear cut area in old-growth forests relative to the total clear cut area each year.
Table 1.
Results of the mixed-effects model for age of secondary forests and number of past clear-cut events.
In both models, communities are included as random effects. The values of the slope (β) and its standard error (SE), the degrees of freedom (DF), t-value and P-value for each variable are provided. For the dummy variable community size, isolated households are taken by the model as the reference category.
Fig 4.
Kaplan-Meier survival curves for fallows fields at two time-periods: 1987–2000 and 2001–2014.
Cumulative survival probability of fallows refers to the probability of fallows not been cut, i.e. of continuing regrowth.
Fig 5.
Frequency distribution of fallow ages in two time-periods: 1987–2000 and 2001–2014, in the riverine settlements of the region of the middle-Amazonas river, Brazil.
Grey dashed line and the number indicate the average fallow age in each time-period. Note the different scales in y axes of the two graphs.