LC-21 Abstract

Regional Effects of Selective Logging on Canopy Damage and Nutrient Dynamics in Amazonia: Linking Landsat ETM+ and Field Biogeochemical Studies

Gregory Paul Asner, Carnegie Institution (US-PI)
Mercedes M.C. Bustamante, UnB (SA-PI)

Selective logging is a dominant form of land use in the Brazilian Amazon.  Several LBA studies are quantifying the effects of logging on carbon dynamics, including the role of fire as an agent of continued change following timber harvest.  Despite ongoing efforts, surprisingly little is known about the extent, intensity or biogeochemical effects of logging in Amazon forests.  At the scale of the entire Basin, the extent of logging is hotly contested, with current estimates ranging from 2,000-15,000 km2 yr-1.  Even less is known about the intensity of logging, where intensity is defined here as canopy structural damage caused by the harvest operation.  Moreover, there exists almost no information on changes in nutrient stocks and dynamics in selectively logged forests.  The fate of nutrients could be central to determining rates of regrowth and the long-term sustainability of timber harvesting in Amazonia.

During the past two years, we have tested a method to quantify both the extent and intensity of selective logging using Landsat ETM+ data with a Monte Carlo spectral mixture model.  The method produces coverages of fractional canopy, bare soil and surface necromass (slash) cover, along with statistical uncertainty maps for each cover fraction.  The canopy fractional cover results have proven highly correlated with field-measured forest gap fraction, which in turn, is spatially correlated with the volume (and biomass) of wood removed from the forest and the coarse woody debris remaining in the harvest sites.  This approach opens the door to regional-scale studies of logging extent and intensity as well as the resulting changes in carbon and nutrient stocks.  Before considering the method viable for basin-wide logging studies, additional testing is needed across a wider range of forest structural types, logging regimes and in areas subjected to fire following harvest.  Additional studies are also needed to determine the functional, quantitative linkages between satellite-observed changes in forest canopy cover and exposed surface slash, carbon stock changes in vegetation and soils, and nutrient cycling.

We propose to extend our remote sensing approach to a much larger region of the Amazon, and to further develop linkages between our satellite analyses and measured carbon and nutrient changes.  The satellite studies will take place primarily in Para and Mato Grosso states, and will cover the period 1999-2004.  We will use the fractional cover maps to direct field studies and continued canopy gap fraction validation efforts across a range of forest types.  The field sites will include low- and high-damage logging blocks with and without fire, and additional areas containing forest structural variation under study by groups led by Dan Nepstad, Foster Brown, Emilio Moran, and other LBA collaborators.  At a subset of sites within the imagery, we will quantify changes in vegetation and soil nitrogen, phosphorus, and base cation stocks.  Using established logging chronosequences, we will measure changes in nutrient pools, mineralization rates and other key nutrient cycling processes to determine the short-term and the potential long-term effects of harvest on nutrient availability.  We will use the canopy damage information derived from Landsat ETM+ to spatially integrate nutrient and carbon data, and to predict regional-scale changes in nutrient stocks and their partitioning in vegetation and soils.  Our results will be compared to studies of intact forest underway by other LBA teams.  We will also use the Landsat analyses of logging extent and intensity to assess linkages between logging practices and the spatial and temporal patterns of fire occurrence over large regions of the eastern Amazon.  Fire occurrence data have been and will be provided by the Tropical Rainfall Mapping Mission (TRMM) satellite, and will be validated during our (and Nepstad’s) proposed field studies.

The project will produce a set of tangible products to be freely shared by LBA investigators and Brazilian agencies such as EMBRAPA and INPE.  These include: (1) regional maps of selective logging extent and canopy damage for 1999-2004; (2) nutrient stock and flux data partitioned by forest and soil types, and by landscape units such as logging decks, roads, skid trails, and tree-falls; (3) spatially integrated nutrient budgets showing how nutrients are altered by logging; and (4) regional maps of canopy damage from logging in relation to fire occurrence from field and TRMM satellite data.  This project will have a strong educational component, including: (a) training a Brazilian and an American post-doc, including exchange between the co-PIs labs; (b) training students from the Universidade de Brasilia in remote sensing studies and field methods; (c) training of EMBRAPA personnel in the use and validation of the satellite results; and (d) outreach to Brazilian land managers interested in the use of satellite data for logging operations (technology transfer).

The project will directly address LBA questions LC-Q3, CD-Q3b, CD-Q3c, and ND-Q1.   In doing so, we will contribute to answering the overall LBA question: “How do tropical forest conversion, re-growth and selective logging influence carbon storage, nutrient dynamics, trace gas fluxes, and the prospect for sustainable land use in Amazonia?”