Close Window

CD-16 Abstract

Scaling Forest Carbon Flux Measurements from Sites to Landscapes: A Hyperspectral Remote Sensing Analysis of FLUXNET Research Sites Along a Breal to Tropical Gradient

Scott Ollinger — University of New Hampshire (US-PI)
Humberto Rocha — USP (Universidade de Sao Paulo) (SA-PI)

Cycles of carbon and nitrogen in forest ecosystems

are tightly linked through a shared set of biological processes.  At the

leaf level, rates of net photosynthesis have been related to nitrogen

concentrations among species from diverse biomes around the world.  At

the canopy level, nitrogen concentrations have been related to rates of forest

production, soil C:N ratios and rates of N cycling.  In aggregate, this

suggests that remote sensing of foliar nitrogen chemistry can offer a useful

approach to scaling point-based carbon flux measurements to local landscapes

or regions.  In one recent study, data from NASA’s AVIRIS sensor was

successfully used to map aboveground NPP and soil carbon to nitrogen ratios

across a large and heterogeneous forest landscape. 



Here, we propose to extend this approach across wider

range of forest ecosystems that span a boreal to tropical gradient, where

carbon flux measurements are available from eddy covariance towers and

extensive plot-based measurements.  The sites to be analyzed are part of

the AmeriFlux or FLUXNET CO2 exchange research programs and all are located in

Brazil and the USA.  The Brazilian sites include: LBA sites km 67 and km

83 in the Tapajos National Forest near Santarem, and LBA sites km 14 and km 34

in the Amazonas region.  The US sites include Howland, ME; Harvard

Forest, MA; Duke Forest, NC and Gainsville, FL. 



In building our field and remote sensing campaign

around AmeriFlux and FLUXNET research sites, the proposed work will bring the

added benefit of allowing eddy covariance CO2 exchange measurements to be

interpreted for the first time with respect to patterns of remotely-sensed

canopy chemistry within the surrounding landscapes.  We will examine

several approaches to undertaking this, including qualitative analyses of CO2

exchange with respect to varying wind direction and associated canopy nitrogen

levels as well as spatially-explicit application of a forest ecosystem model.



Research Plan:



For the 2003 field season, we plan to visit the 4

tower sites that are part of our investigation during June of 2003 with the

following objectives:  At each site we will coordinate with other

cooperating investigators to supplement existing field data with new leaf

nutrient measurements to build complete data sets for foliar chemistry,

biomass production and soil C:N ratios in order to examine relations among

these variables both within and across sites. 



Hyperspectral remote sensing data will be acquired

from NASA’s AVIRIS (4 m pixel resolution) and Hyperion (30 m resolution)

instruments for development of data planes for foliar chemistry and associated

carbon fluxes using relationships observed within the field data.  This

will provide insight into the nature and generality of forest carbon-nitrogen

relations across ecosystems and climate regimes and could open new

possibilities for mapping fine-scale patterns of growth in other regions

around the world.



After completion of the field campaign, field and

remote sensing data will be used to modify and run a forest ecosystem model

with the aim of modeling canopy carbon fluxes for comparison with tower

measurements.  Refinement of canopy physiology algorithms (for Amax, SLA

and light attenuation) for model application to Amazonian forests will be

achieved through involvement of a Brazilian graduate student.  Use of a

common modeling platform across all sites will provide a means of examining

variation across sites (i.e. with respect to differences in climate and canopy

physiology) as well as spatial variation within sites.  

Close Window