Search

• We analysed changes in APAR and LUE along a complete rotation on 16 eucalypt genotypes. • APAR was estimated using the MAESTRA tridimensional model precisely parameterized. • Light absorption was a major limiting factor only over the first year of growth. • Differences in productivity and LUE among genotypes remained large throughout the rotation. • LUEs over the second half of the rotation was the major driver of final stemwood biomass. Stemwood productivity in forest ecosystems depends on the amount of light absorbed by the trees (APAR) and on the Light Use Efficiency (LUE), i.e. the amount of stemwood produced per amount of absorbed light. In fertilized Eucalyptus plantations of Brazil, growth is expected to be strongly limited by light absorption in the first years after planting, when trees can benefit from high soil water stocks, recharged after clearcutting the previous stand. Other limiting factors, such as water or nutrient shortage are thought to increase in importance after canopy closure, and changes in allocation patterns are expected, affecting the LUE. Studying changes in APAR and LUE along a complete rotation is paramount for gaining insight into the mechanisms that drive the inter- and intra-genotype variabilities of productivity and stemwood biomass at the time of harvest. Here, we present a 6-year survey of productivity, APAR and LUE of 16 Eucalyptus genotypes of several species used in commercial plantations and planted in 10 randomized replications in the São Paulo Region, Brazil. APAR was estimated using the MAESTRA tridimensional model parameterized at tree scale for each tree in each plot (a total of 16,000 trees) using local measurements of leaf and canopy properties. Stand growth was estimated based on allometric relationships established through successive destructive biomass measurements at the study site. Allometric relationships predicting biomass of tree components, leaf surface, crown dimension and leaf inclination angle distribution throughout the rotation for the 16 productive genotypes are shown. Results at stand scale showed that (1) LUE increased with stand age for all genotypes, from 0.15 at age 1 yr to 1.70 g MJ−1 at age 6 yrs on average; (2) light absorption was a major limiting factor over the first year of growth (R2 between APAR and stand biomass ranging from 0.5 to 0.95), explaining most of the inter- and intra-genotype growth variability; (3) at rotation scale, the variability of final stemwood biomass among genotypes was in general attributable to other factors than average APAR; (4) differences in stemwood productions among genotypes remained large throughout the rotation; (5) LUEs over the second half of the rotation, rather than initial growth or APAR, was the major driver of stemwood biomass at the time of harvest. [ABSTRACT FROM AUTHOR]

• G'Day model simulates carbon and water budgets of Eucalyptus plantations. • We followed a multi-site calibration approach based on 4 datasets of different precision. • Model prediction of key biomass variables closely matched measurements. • Model captures 89% of stem biomass variability on 27,500 ha of Eucalyptus plantation. • Calibrating the model for each genotype was critical for stem biomass prediction at that scale. Carbon and water budgets of forest plantations are spatially and temporally variable and hardly empirically predictable. We applied G'DAY, a process-based ecophysiological model, to simulate carbon and water budgets and stem biomass production of Eucalyptus plantations in São Paulo State, Brazil. Our main objective was to assess the drivers of spatial variability in plantation production at regional scale. We followed a multi-site calibration approach: the model was first parameterized using a detailed experimental dataset. Then a subset of the parameters were re-calibrated on two independent experimental datasets. An additional genotype-specific calibration of a subset of parameters was performed. Model predictions of key carbon-related variables (e.g., gross primary production, leaf area index and stem biomass) and key water-related variables (e.g., plant available water and evapotranspiration) agreed closely with measurements. Application of the model across ca. 27,500 ha of forests planted with different genotypes of Eucalyptus indicated that the model was able to capture 89% of stem biomass variability measured at different ages. Several factors controlling Eucalyptus production variability in time and space were grouped in three categories: soil, climate, and the planted genotype. Modelling analysis showed that calibrating the model for genotypic differences was critical for stem biomass prediction at regional scale, but that taking into account climate and soil variability significantly improved the results. We conclude that application of process-based models at regional scale can be used for accurate predictions of Eucalyptus production, provided that an accurate calibration of the model for key genotype-specific parameters is conducted. [ABSTRACT FROM AUTHOR]