This datasets archives the processed data supporting the research article titled “A new dimension of the rhizosphere: Microscale spatial organization of soil invertebrates around roots.” The data originates from an innovative edaphic monitoring method based on high-resolution in-situ imaging.

The experimental setup utilized 12 localized scanner-based devices configured to capture A4-sized images at a resolution of 1200 dpi. These sensors operated at a six-hour temporal resolution, generating a high-density longitudinal image series. The data were acquired at the DIAMS experimental agroforestry site in Montpellier, France, between March 2024 and November 2025. For the purposes of this study, a curated subset of this image bank was selected and analyzed.

The raw imaging signals were processed through two specialized analytical pipelines to generate the provided datasets. The first workflow employed image segmentation to convert raw optical signals into binary surface masks, which are archived in the root_mask_data.tar compressed file. The corresponding quantitative data is available in root_pixels_count.tab, which provides a synthesized database of pixel counts and structural metrics derived from these binary images.

The second workflow targeted faunal soil organisms, translating raw imaging signals into a structured taxonomic occurrence count. This dataset, found in fauna_data.tab, details the abundance and classification of soil animal taxa identified throughout the whole monitoring period.

All subsequent statistical modeling was conducted within the R programming environment. The complete source code, including the specific scripts used for data cleaning and processing, is maintained and publicly accessible at the following GitHub repository: https://github.com/emmabelaud/rhizosphere.