Effects of tillage intensities on spatial soil variability and site-specific management in early growth of Eucalyptus grandis

• We compared three tillage intensities in early growth of Eucalyptus.
• Models that include soil spatial variability information as a covariate improve efficiency.
• Moderate intensive tillage is preferred when soil conditions are not restrictive.
• The use of soil spatial variability is useful for delineating site-specific management.

Soil tillage is one of the most common and important site preparation managements in forestry. However, in highly variable soils, uniform management practices might not be the best alternative. Site-specific management on the other hand, allows an optimal resource management as well as decreased environmental impact. However, the choice of a suitable strategy to manage areas with high soil variability is still a challenge. Our goal was to compare strategies that use soil characteristics to improve the comparison of tillage managements on Eucalyptus grandis growth. Specifically, we aimed to: compare strategies that incorporate soil characteristics into the models to compare tillage treatments; to determine the most useful soil characteristics for zone management delineation; and to compare tillage methods for site-specific management. We compared tillage intensities in contrasting soil types in a randomized complete block design with four and five replications. Tillage treatments included pit-planting, disc harrowing, and subsoiler. Experimental units consisted of three rows of fifteen trees each. Soil characteristics as well as plant height and diameter were evaluated periodically during the first 30 month after implantation. Intra-plot variability was described with multivariate geostatistical models. Using soil properties as covariates in the model to compare tillage treatments improved model fit. When root development is limited by soil conditions and electroconductivity is high, tillage intensity makes a difference in plant growth; subsoiler is the best treatment when electroconductivity is high, while disc harrowing is the best when electroconductivity is low. However, when root development is not limited by soil conditions, no differences were found between subsoiler and disc harrowing. We show how the use of soil characterization is a tool that provides better comparisons among treatments when high intra-plot variability is present. Additionally, the use of soil characterization either directly into the model or to determine zones provides useful information for site-specific management. Site-specific management could therefore easily be implemented to decrease the environmental impact of soil tillage as well as to increase wood production in forestry.