What have we learnt?
What have we learnt?
Over the past decade, our research on forest modelling has advanced our understanding of growth dynamics, management practices, and biomass potential across diverse regions. Combining field trials, permanent plots, and modeling approaches, these studies inform how silviculture, climate, and site factors interact to shape yields, wood quality, and carbon stocks for timber and bioenergy.
Silviculture and Stand Management. In Brazil, pruning Eucalyptus grandis × urophylla clones up to 70% height at age 1.3 years optimizes clear-wood production without reducing growth (Ferraz-Filho et al., 2016), while thinning trials reveal that early, intensive thinning boosts tree size, whereas no thinning maximizes stand basal area depending on product goals (Ferraz-Filho et al., 2018). Coppice-with-standards systems offer dual outputs of biomass and high-quality logs (Ferraz-Filho et al., 2014).
Growth and Yield Modeling. Height–diameter allometry and mixed-effects models enable accurate volume estimates in Brazilian eucalyptus and Swedish poplars (Ferraz-Filho et al., 2018b; Hjelm et al., 2015). Bayesian and distribution-based models allow for reliable diameter predictions in data-poor pine plantations (Tian et al., 2020), while climate-sensitive models for Spain and the Nordics forecast future site productivity and forest dynamics (Antón-Fernández et al., 2016; Trasobares et al., 2022).
Biomass and Carbon Assessments. Life-cycle analysis of Chilean Eucalyptus globulus shows harvesting drives most environmental impacts, stressing the role of input efficiency (Morales et al., 2015). In Panama, allometric analysis of mangrove plots refines national carbon stock estimates and supports remote-sensing calibration (Hoyos-Santillan et al., 2025). González-García et al. (2014) performed cradle-to-gate LCAs of twelve European forestry systems—covering maritime pine, Norway spruce, willow, poplar, and Douglas-fir in Sweden, Germany, France, Italy, and Portugal—with 1 m³ yr⁻¹ roundwood as the functional unit. By harmonizing inventory data and methodological assumptions, they revealed that harvesting and forwarding dominate both environmental impacts and energy use (hotspots driven by diesel fuel), with fertilization, thinning, and weed control also contributing significantly. Variations in productivity, management intensity, and regional conditions drove wide differences in GWP, acidification, eutrophication, and photochemical‐oxidant potentials, underscoring that biomass choice should balance yield gains against the environmental costs of silvicultural practices.
Remote‐Sensing Imputation. Valbuena et al. (2017) evaluated how fusing airborne LIDAR height metrics with NDVI‐derived multispectral descriptors enhances most‐similar‐neighbor (MSN) imputation of key forest attributes. Across Scots pine plots in Spain, they computed three predictor sets—LIDAR only, NDVI only, and combined—and applied MSN to estimate diameter‐dependent (QMD, BA), height‐dependent (HL, V), density (N, SDI), and structural‐heterogeneity (Gini, BALM) variables. Results showed that while LIDAR alone already predicted height‐based attributes well, the addition of NDVI metrics notably improved estimates of diameter‐related and heterogeneity indicators, though benefits were attribute‐specific and required case-by-case assessment.