Stand-level growth and yield model system for clonal eucalypt plantations in Brazil that accounts for water availability

Abstract

Growth and yield (G &Y) model systems aim at forecasting forest productivity. The lack of environmental variables to account for how water availability constrains eucalyptus production in Brazil, however, is argued to be a major drawback of these model systems. Thus, this study aimed to develop a stand-level G & Y model system that accounts for water availability (G & Y with SWD), highlighting its usefulness when applied for clonal eucalypt stands under drier climatic conditions. The dataset is composed of remeasurement information of sixteen research sites that span all climatic regions in Brazil. A total of eleven eucalypt clones were planted in single block plots at each site, and extra replications under the rainfall exclusion system were also installed for these eleven clones in fourteen sites. Linear algebra techniques were used to simultaneously fit a compatible set of prediction and projection basal area equations. A stand-level volume equation was also developed. These equations were validated through the use of an independent dataset composed of the rainfall exclusion plots. Finally, the accuracy and usefulness of a conventional G & Y model system applied to clonal eucalypt stands in Brazil was compared to the new proposed G & Y model system, which accounts for the impact of water availability in eucalyptus productivity. The prediction and projection basal area equations accounting for water availability displayed estimates in the order of 5% more accurate compared to the conventional basal area modeling. Stand-level volume estimates were 40% and 74% less biased through the use of the new G & Y model system. This result highlighted how useful and powerful the newly developed approach is, since the model system was capable to provide accurate estimates through the use of the rainfall exclusion plots. The new G & Y model system is a powerful alternative to estimate forest afforestation yield and is fully capable to accurately update forest inventories. The model system can also be used for projecting how forest growth may be impacted by short-term climate variation.