Production ecology of clonal Eucalyptus plantations in northeastern Brazil

Millions of hectares of Eucalyptus are managed for wood production in Brazil but little is known about the physiological processes that control growth. I used one gradient study and two manipulative studies with clonal E. grandis x urophylla in northeastern Brazil to: (1) identify environmental factors controlling growth and the associated resource use; (2) quantify water and nutrient supply effects on carbon gain-allocation; and (3) evaluate the ability of a process-based model to simulate Eucalyptus growth. For the first objective, fourteen stands with aboveground net primary production (ANPP) spanning from 9--39 Mg ha-1 yr-1 where characterized in their use of resources. Water was the most limiting resource, and ANPP increased by 2.3 Mg ha-1 yr-1 for each 100 mm yr-1 increase in rainfall. Indirectly, water supply increased light and nitrogen use by increasing their capture-efficiency. The most productive sites were the most efficient in using water (3.21 kg of ANPP per m3 transpired water), light (1.14 kg per GJ APAR) and nitrogen (381 kg per kg N taken up). Lower VPD, soil water stress and root-to-aboveground ratio were associated with these high-efficiency stands. For the second objective, an irrigation-fertilization design was used in a 3-year-old plantation. Carbon budgets were obtained for two years, including ANPP, total belowground C allocation (TBCA) and gross primary production (GPP). Irrigation increased GPP from 4.8 to 6.7 kg C m-2 yr -1 primarily by increasing apparent canopy quantum efficiency from 0.034 to 0.052 mol C mol-1 PAR and secondarily by increasing LAI. Irrigation decreased the fraction of GPP allocated belowground from 34% to 28%, increasing ANPP from 1.5 to 2.3 kg C m-2 yr-1. Wood and coarse root increments were the dominant components of C sequestration, with no change in soil C. The final objective was met by calibrating the 3-PG model with two tropical Eucalyptus trials, and validating it with independent data from forty plots monitored for 2 years. The model responded well to soil fertility and climate conditions during validation, and was particularly sensitive to LAI estimates. Actual wood production ranged from 2 to 51 Mg ha-1 yr -1, compared with model estimates of 10 to 42 Mg ha-1 yr-1 (r2 = 0.78). Overall, our study indicates that for fertilized Eucalyptus tropical plantations water is the most influential resource controlling C gain, allocation, resource use and the efficiency in using these resources; and process-based models can play an important role in improving the management of these almost-agricultural forests.