Abstract

Process-based models for forest plantations simulate differences in genetics, soil, climate, and management across stands. We tested a calibration protocol using the APSIM model to predict 2.5 years of growth for 16 Eucalyptus genotypes (E. badjensis, E. globulus, E. nitens, and hybrids of E. camaldulensis × globulus and E. globulus × nitens) with four contrasting growing conditions in Chile: north and south sites, each under high and low irrigation. Two APSIM pre-calibrated Australian genotypes were also included. After validating simulated aboveground net primary production (NPP) and stem volume, the model was used to examine relationships between NPP, light use efficiency attained (LUEA), and water use efficiency (WUE) across site, irrigation, and genotype. Chilean genotype specifications included potential light use efficiency (LUEP), wood density, specific leaf area (SLA), and bark thickness; other parameters were adopted from pre-calibrated Australian E. globulus or E. nitens profiles. LUEP was adjusted to fit predictions at the southern site with high irrigation and reused for the other growing conditions. LUEP was significantly correlated with volume rankings at south-high irrigation, which aligned with rankings at the northern site for both irrigation treatments. However, genotype volume rankings were poorly predicted under the least favorable conditions (south-low irrigation). Simulated LUEA ranged from 1.76–1.98 g MJ?¹ (high irrigation) and 0.92–1.12 g MJ?¹ (low irrigation). Irrigation at the northern site moderately improved WUE (+28 %), while the southern site showed a dramatic WUE gain (+283 %), enhancing total and stem NPP. Overall, the protocol enabled adequate prediction of tree growth and may support yield forecasting across diverse genotypes and growing conditions. © 2025 Elsevier B.V.