Soil burrow characterization by 3D image analysis: Prediction of macroinvertebrate groups from biopore size distribution parameters

Autor: Laura Gargiulo, Fabio Terribile, Gilda Buscemi, Giacomo Mele
Přispěvatelé: Mele, G., Buscemi, G., Gargiulo, L., Terribile, F.
Rok vydání: 2021
Předmět:
Zdroj: Geoderma (Amst.) 404 (2021): 1–11. doi:10.1016/j.geoderma.2021.115292
info:cnr-pdr/source/autori:Giacomo Mele, Gilda Buscemi, Laura Gargiulo*, Fabio Terribile/titolo:Soil burrow characterization by 3D image analysis: Prediction of macroinvertebrate groups from biopore size distribution parameters/doi:10.1016%2Fj.geoderma.2021.115292/rivista:Geoderma (Amst.)/anno:2021/pagina_da:1/pagina_a:11/intervallo_pagine:1–11/volume:404
DOI: 10.1016/j.geoderma.2021.115292
Popis: Although scientific literature focuses mainly on earthworms, all soil macroinvertebrates play an important role in modifying the architecture of the soil pore space and, in turn, in soil functions. Notwithstanding the fact that non-invasive technology, such as X-ray tomography, has long been used to differentiate non-biopores from biopores, it is still difficult to distinguish the specific contribution different macroinvertebrates make to the soil biopore system. Unlike the object-based image analysis approach, when applied to a soil pore system, mathematical morphology permits the user to obtain a very accurate pore size distribution consistent with the physical principle of water retention. The aim of this work was to evaluate the potential of the parameters of this kind of biopore size distribution to differentiate between the burrows of five different macroinvertebrate groups, namely Earthworms, Millipedes juliform, Centipedes, Campodeiform larvae and Elateriform larvae, inoculated into repacked soil mesocosms and incubated (14 days) in the field from where the soil animals were originally collected. A two-fold approach was proposed in this work so as to obtain parameters by both pore size population distributions and Weibull modelling of the cumulative distributions. Then a predictive discriminant analysis was performed on selected parameters by using macroinvertebrate groups as grouping variables and a very good prediction was obtained in both cases. The most useful parameters were the skewness and FFT indices in the first case and the shape parameter α of the Weibull model along with its RMSE in the second one. In addition, topological characterization was performed on gallery-shaped biopores. Vertical deviation was the only parameter that was independent of the individual body size and showed the statistically significant lowest value for the earthworms. The experiment and analyses performed in this work to explore the connection between macroinvertebrate groups and the corresponding biopore size distributions may represent a suitable methodological approach to performing a general investigation into the relationships between soil management and its impact on the system of soil macropores.
Databáze: OpenAIRE