© 2003 Heron Publishing—Victoria, Canada
Partitioning interannual variability in net ecosystem exchange between climatic variability and functional change
Dafeng Hui (1, 2), Yiqi Luo (1) and Gabriel Katul (3)
1. Department of Botany and Microbiology, University of Oklahoma, Norman, OK 73019, USA / 2. Author to whom correspondence should be addressed (dafeng@ou.edu) / 3. Nicholas School of the Environment and Earth Sciences, Duke University, Durham, NC 27708, USA / Received June 3, 2002; accepted October 28, 2002; published online April 1, 2003
Summary
Interannual variability (IAV) in net ecosystem exchange of carbon (NEE) is a critical factor in projections of future ecosystem
changes. However, our understanding of IAV is limited because of the difficulty in isolating its numerous causes. We proposed
that IAV in NEE is primarily caused by climatic variability, through its direct effects on photosynthesis and respiration
and through its indirect effects on carbon fluxes (i.e., the parameters that govern photosynthesis and respiration), hereafter
called functional change. We employed a homogeneity-of-slopes model to identify the functional change contributing to IAV
in NEE and nighttime ecosystem respiration (RE). The model uses multiple regression analysis to relate NEE and RE with climatic variables for individual years and for all years. If the use of different slopes for each year significantly
improves the model fitting compared to the use of one slope for all years, we consider that functional change exists, at least
on annual time scales. With the functional change detected, we then partition the observed variation in NEE or RE to four components, namely, the functional change, the direct effect of interannual climatic variability, the direct effect
of seasonal climatic variation, and random error. Application of this approach to a data set collected at the Duke Forest
AmeriFlux site from August 1997 to December 2001 indicated that functional change, interannual climatic variability, seasonal
climatic variation and random error explained 9.9, 8.9, 59.9 and 21.3%, respectively, of the observed variation in NEE and
13.1, 5.0, 38.1 and 43.8%, respectively, of the observed variation in RE.
Keywords:
CO2 flux, ecosystem respiration, eddy-covariance measurement, homogeneity-of-slopes model.
