Monitoring Selective Logging in Tropical Evergreen Forests Using Landsat: Multitemporal Regional Analyses in Mato Grosso, Brazil

Abstract Selective logging degrades tropical forests. Logging operations vary in timing, location, and intensity. Evidence of this land use is rapidly obscured by forest regeneration and ongoing deforestation. A detailed study of selective logging operations was conducted near Sinop, State of Mato Grosso, Brazil, one of the key Amazonian logging centers. An 11-yr series of annual Lansdat images (1992–2002) was used to detect and track logged forests across the landscape. A semiautomated method was applied and compared to both visual interpretation and field data. Although visual detection provided precise delineation of some logged areas, it missed many areas. The semiautomated technique provided the best estimates of logging extent that are largely independent of potential user bias. Multitemporal analyses allowed the authors to analyze the annual variations in logging and deforestation, as well as the interaction between them. It is shown that, because of both rapid regrowth and deforestation, evidence of logging activities often disappeared within 1–3 yr. During the 1992–2002 interval, a total of 11 449 km2 of forest was selectively logged. Around 17% of these logged forests had been deforested by 2002. An intra-annual analysis was also conducted using four images spread over a single year. Nearly 3% of logged forests were rapidly deforested during the year in which logging occurred, indicating that even annual monitoring will underestimate logging extent. Great care will need to be taken when inferring logging rates from observations greater than a year apart because of the partial detection of previous years of logging activity.