526 Forest Science • October 2020 For. Sci. 66(5):526–536 doi:10.1093/forsci/fxaa007 RESEARCH ARTICLE biometrics Spatio-Temporal Modeling and Competition Dynamics in Forest Tillage Experiments on Early Growth of Eucalyptus grandis L. Pablo González Barrios † , Alejandra Borges † , José Terra, Mario Pérez Bidegain, and Lucía Gutiérrez Forest tillage experiments regularly use long-term evaluations of large plots creating temporal and/or spatial correlations among observations. Not modeling these correlations could compromise treatment comparisons. The aim of this study was to evaluate the effect of modeling spatio-temporal (ST) variability in forest tillage experiments. We used different strategies that incorporate spatial and/or temporal correlations in the evaluation of tillage intensity effect in initial Eucalyptus growth as well as evaluate the effect of intraplot mortality and competition dynamics. Three tillage intensities in two contrasting soil conditions were compared for tree height and wood volume. Additionally, we compared the use of three individual growth curves for plant height to evaluate the time needed to reach 2 m in height (T 2m ). We modeled the spatial correlation of T 2m using mixed models. In both sites, ST models were superior for plant height and wood volume per hectare, whereas for individual-tree wood volume, temporal models were superior. Pit planting always had a lower performance than disk harrowing and subsoiler, which behaved similarly. The competition dynamics within the plot because of tree mortality was affected by treatments and site. Modeling ST variability is key to improving treatment comparisons in forest experiments. Study Implications: The management of tillage practices is a key aspect in the productivity of Eucalyptus plantations. The intensity of tillage directly affects the soil physical properties generating changes in the growth and mortality dynamics of trees. Our results indicate that pit planting showed the lower performance in growth variables for all evaluation moments and in both sites. Disk harrowing improves the growth conditions in the frst year after planting compared to subsoiler under limiting soil conditions. However, at 30 months after planting, there were no signifcant differences between disk harrowing and subsoiler treatments. Because disk harrowing is a practice with lower operating costs and environmental impact, it appears to be the superior alternative in terms of productivity and sustainability of forest systems. Keywords: site preparation, tillage intensity, spatio-temporal variability, growth curves C ommercial plantations of Eucalyptus species play an im- portant role in worldwide forest production systems (Grattapaglia and Kirst 2008), representing the largest source of cellulose for the paper industry (Overbeek et al. 2012). Te main advantages of these species are high wood yield, adapta- bility to a wide range of environmental conditions, and diversifed areas of fnal products (Verhaegen and Plomion 1996). Te increasing demand for Eucalyptus products has caused an expansion of aforestation areas in several regions of the world, leading to strong changes in production systems and land use (Kröger 2012). Te expansion of new forested areas is mainly located on marginal or low-productivity soils, where, in many cases, a strong spatial var- iability structure is present. Furthermore, most forest experiments require large experimental units, evaluated for long periods of time, with increased spatial and temporal variability that challenge their power and precision. Manuscript received July 30, 2019; accepted January 30, 2020; published online March 28, 2020. Affliations: Pablo González Barrios (pgonzalez6@wisc.edu), Alejandra Borges (aborges@fagro.edu.uy), and Lucía Gutiérrez (gutierrezcha@wisc.edu), Statistics Department, Facultad de Agronomía, Universidad de la Republica, Uruguay. José Terra (jterra@tyt.inia.org.uy), National Agricultural Research Institute (INIA), Uruguay. Mario Pérez Bidegain (mperezb@fagro.edu.uy), Soils Department, Facultad de Agronomía, Universidad de la Republica, Uruguay. Lucía Gutiérrez (gutierrezcha@wisc.edu), Department of Agronomy, University of Wisconsin, Madison. † The first two authors made equal contributions to the manuscript. Acknowledgments: Te research that provided the results presented in this study received funding from the Agencia Nacional de Innovación e Investigación (ANII) under the code POS_NAC_2012_1_8747, through a fellowship to PGB, Comisión Sectorial de Investigación Científca (CSIC) through an internship to AB and PGB and through an industry partnership research project with Forestal Oriental S.A. (FOSA). Te authors would like to thank two anonymous reviewers and the Associate Editor for substantial revision that improved the manuscript. Declaration on conficts of interest: Te authors declare that they have no confict of interest. Copyright © 2020 Society of American Foresters Downloaded from https://academic.oup.com/forestscience/article/66/5/526/5813133 by guest on 17 November 2022