Multifractal analysis on CT soil images: Fluctuation analysis versus mass distribution
Xiaoqin Sun,Dongli She,Ernesto Sanz,Juan J. Martín-Sotoca,Ana M. Tarquis,Lei Gao
DOI: https://doi.org/10.1016/j.chaos.2023.114080
2023-10-01
Abstract:The combined action of physical, chemical and biological processes at different scales creates a highly complex soil architecture. At the same time, this soil pores' complexity is crucial in maintaining soil biogeochemical and biophysical processes. The development of digital image processing and a multiscaling analysis allow a better study to quantify pore complexity and evaluate pore spatial variability. This study took saline soils from coastal reclamation areas as an example and aimed to compare the multifractality of porosity series using different methods on sample cubes of 512 pixels in length extracted at different soil depths. Two measures were selected to represent the porosity series: the average grey values (AVG) of each slice and the CT-porosity (CTP) of binarized slices. The mass distribution (MD), detrended fluctuation analysis (DFA), and the moments function (K) were the methods applied in the multiscaling study. Shuffled and surrogated series were used to analyze the two possible multifractality sources, the probability distribution density function (PDF) and long-range correlations, respectively. The results showed that MD gave the minimum multifractal spectrum, which appeared as a point in AGV and a narrow concave parabola in CTP. On the other hand, both AGV and CTP showed a higher complexity, a clear and significant multifractal behaviour when DFA and K were used. The shuffled and surrogated series pointed out that the long-range correlations were the main factor of the multifractal sources of the porosity series. Both long-range correlations and PDF influenced more heavily on the multifractality of porosity in topsoil (0– 20 cm) than in the deeper soil (20– 60 cm), implying that the porosity series in topsoil had stronger multifractality. In addition, the DFA and K, combined with shuffled and surrogated series analysis, can distinguish between different factors in the multiscaling behaviour of different soil textures and time reclamation.
mathematics, interdisciplinary applications,physics, mathematical, multidisciplinary