Area and Age of Secondary Forests in Brazilian Amazonia 1978–2002: An Empirical Estimate

Abstract

In quantifying the carbon budget of the Amazon region, temporal estimates of the extent and age of regenerating tropical forests are fundamental. However, retrieving such information from remote-sensing data is difficult, largely because of spectral similarities between different successional stages and variations in the reflectance of forests following different pathways of regeneration. In this study, secondary-forest dynamics in Brazilian Amazonia were modeled for the 1978–2002 period to determine area and age on a grid basis. We modeled the area, age, and age class distribution of secondary forests using empirical relationships with the percentage of remaining primary forest, as determined from large-area remote-sensing campaigns (the Pathfinder and Prodes projects). The statistical models were calibrated using detailed maps of secondary-forest age generated for seven sites in the Brazilian Legal Amazon. The area–age distribution was then specified from mean age by a distribution assumption. Over the period 1978–2002, secondary-forest area was shown to have increased from 29,000 to 161,000 km2 (that is, by a factor of 5). The mean age increased from 4.4 to 4.8 years. We generated a time series of secondary-forest area fractions and successional stages that provides wall-to-wall coverage of the Brazilian Amazon at a spatial resolution of 0.1 decimal degrees (approximately 11 km). Validation against reference data yielded root mean squared errors of 8% of the total area for estimate of secondary-forest area and 2.4 years for mean secondary-forest age. Using this approach, we provide the first published update on the state of secondary forests in Amazonia since the early 1990s and a time series of secondary-forest area over the 25-year period.