State-Space Approach to Understand Soil-Plant-Atmosphere Relationships

This chapter presents two different state-space approaches to evaluate the relation between soil and plant properties using examples of sugarcane, coffee and forage. These state-space approaches take into account sampling positions and allow a better interpretation of the data in relation to the field. Concepts of autocorrelation and crosscorrelation functions are first introduced, followed by theoretical aspects of both state-space approaches. More emphasis is given to the last one based on the Bayesian formulation, which gives more attention to the evolution of the estimated observations. It is concluded that the use of these dynamic regression models improve data analyses, being therefore recommended for several studies involving time and space data series, related to the performance of a given soil-plant-atmosphere system.