Imagens microtomográficas e simulação computacional na análise do sistema poroso do solo

Abstract

Soils represent complex and dynamic systems in their natural or undisturbed state, such as those found in areas of native forest or secondary forest. When destined for agricultural purposes, soils undergo intensive and non-intensive preparation practices, in addition to being subject to climatic processes and continuous irrigation, which cause wetting and drying cycles (W-D). The different preparation systems and application of W-D cycles cause changes mainly in the micrometric scale, and thus fractal and multifractal approaches can be used to understand complex systems. Initially, this study used a computational algorithm based on the mercury intrusion porosimetry (MIP) technique to obtain pore size distribution curves (PSD) by microtomographic image analysis (CT) of a Dystrophic Red Latosol submitted to different preparation practices (conventional preparation - CT, minimum preparation - MT and no-preparation - NT) and area under secondary forest (F) as reference. The PSDs were also compared and complemented with those obtained by the traditional method based on the water retention curve (RC) using the first derivative of the mathematical model of RC interpolation. The same algorithm was also used in the reverse process as an alternative methodology for RCs acquisition. The results showed that CT and MT presented significant changes in PSDs compared to F and NT, and in both methods the CT and NT porosities suffered the most changes, mainly by the increase in transmission pores (<500 μm). The RCs determined by the traditional method and those based on the computational algorithm were concordant, revealing that the latter can be used for representative measurements based on images of water retention properties by porous systems. The results of the analysis of the effects of W-D cycles on the PSDs, also obtained by the MIP-based algorithm, revealed that such cycles altered the porosity and functionality of the pores by distributing their sizes for all preparations. MT was the preparation that suffered the most significant changes in the functionality of its pores, in which a large portion of the transmission pores became fissures. The main result obtained in the comparison of preparation indicates that according to the preparation adopted, the soil will react differently to the W-D cycles. Finally, fractal and multifractal approaches such as lacunarities, generalized fractal dimensions (Dqs), multifractal spectra, and entropy in three dimensions (3D) together with the geometric parameters of pore connectivity and tortuosity, allowed to show that the preparation practices were divided into two groups, which showed great self-similarity or proximity in their porous characteristics (F with NT and MT with CT). High linear relationships were found between 3D lacunarities, Dqs, and entropies with the porosities. The 3D lacunarities, Dqs and degrees of asymmetries arising from the multifractal spectra confirmed spatial variabilities of the pore size distributions of the preparation and F. In this way, the potential and sensitivity of the 3D multifractal approach for the characterization and differentiation of dynamic and complex porous spaces, such as the soil, became evident.