A comprehensive China soil dataset of high-resolution microstructure, topographical characteristics and physical properties

A comprehensive China soil dataset of high-resolution microstructure, topographical characteristics and physical properties 1 Abstract Soil is the domain where hydro-bio-geo-chemical processes are closely coupled in the ecosystem. Soil microstructure, the arrangement of soil particles and aggregates into a porous structure, regulates the flow of resources. Soil microstructure data includes both the 3D structures of the soil and the morphological parameters characterizing their topological and physical properties, all of which are essential for understanding solute transport dynamics such as oxygen diffusion, water flow, organic matter accessibility, and nutrient availability. However, it is challenging to access high-resolution soil microstructure data, for the heterogeneity and opacity of soils, as well as the costs associated with advanced imaging techniques. In this study, we present a novel 3D soil microstructure dataset that integrates high-resolution imaging data of soils across China (CHARM3D) with comprehensive measurements of topological and physical properties to quantitatively characterize soil morphological parameters. Thirty sampling sites were selected within the Chinese Ecosystem Research Network (CERN), including six major ecosystems: forests, grasslands, deserts, agricultural lands, wetlands, and lakes. A total of 120 soil samples representing diverse soil orders were collected, reconstructed using high-resolution X-ray computed tomography (XCT), and analyzed with advanced image processing techniques via ImageJ and Avizo software. Morphological parameters are consisted of topological characteristics that describe the geometry of pore networks, and physical parameters associated with soil functions, particularly permeability and connectivity. Topological features were extracted using ImageJ combined with pore network analysis, whereas permeability-related connectivity was simulated through pore-scale modeling. A case study analyzing representative soil samples from the six ecosystems demonstrates the usefulness of the dataset in providing key parameters for reactive transport modeling in soils, such as pore connectivity, permeability, porosity distribution, as well as in characterizing the impact of soil aggregation on flow transport. This dataset provides valuable insights into soil structure variation across ecosystems and serves as a critical resource for calibrating soil hydraulic models in regional studies. 2 Keywords Soil; Soil microstructure; Soil topographical characteristics; Soil physical properties; China; X-ray computed tomography; Network analysis; Particulate Organic Matter; 3 Overview of the dataset The high-resolution 3D soil microstructure dataset in this study is consist of 120 in-situ soil core samples collected across 30 sites in China, representing 6 distinct ecosystem types: agricultural, forest, grassland, desert, lake, and wetland ecosystems. This dataset provides a comprehensive soil microstructure archive, which promotes multi-scale modeling of eco-hydrological processes in subsurface systems. Future studies can use this high-resolution dataset for comparative analyses, improving soil property predictions and land management strategies. As shown in Table 1, each dataset follows the naming convention Soil_xx, where “xx” represents the numerical order of the soil sample. The dataset consists of multiple file types and can be categorized into 3D soil microstructure data, morphological parameter files, and soil physical property files. Each soil sample is structured into three primary levels: 1. 3D soil microstructure Data This top-level category contains the original grayscale and binary images of the soil microstructure. Each sample is stored in TIFF format under the filename Soil_xx_xxxx.TIFF. Additionally, binary segmentation images and 3D renderings of the pore structure are included to enhance visualization and structural analysis. 2. Morphological parameter files This level contains quantitative morphological characterizations stored in S oil_xx_Minkowski.csv, Soil_xx_pore_analysis.csv, and Soil_xx_network_analysis.c sv. Soil_xx_Minkowski.csv include key topological parameters such as pore volume fraction (𝜙), surface area (𝑆𝐴), circularity (𝜋), Euler characteristics (𝜒), perimeter (𝑃), perimeter density (𝜌𝑃), convex area (𝐴(𝑥)), and convexity (𝜘). These parameters are derived using Minkowski functionals and mathematical morphology frameworks (Legland et al., 2016; Ohser & Schladitz, 2009). Soil_xx_pore_analysis.csv contain throat radius (𝑇𝑅 ) and coordination number (𝐶𝑁), which describe the connectivity and structural characteristics of the pore network. These parameters are essential for understanding fluid transport properties and were computed following methodologies from Domander et al. (2021) and Doube et al. (2010). Soil_xx_network_analysis.csv includes graph-based metrics such as clustering coefficient (𝐶 ), average distance (𝐷 ), average degree (𝐾 ), and node betweenness centrality (𝑁𝐵𝐶). These parameters quantify the complexity and connectivity of the pore network and were calculated using methods from Wang et al. (2023). 3. Soil physical property data The particulate organic matter (POM) data includes the Soil_xx_POM.csv files and corresponding visual representations of soil organic matter. These images are stored in .png format within the dataset, providing a spatial representation of organic matter distribution within the soil matrix. This information is essential for evaluating soil fertility and carbon storage dynamics (Li et al., 2024). The Soil_xx_permeability.csv files contain permeability (K) values and the porosity-permeability relationship, calculated using JHNY-DPM V3 software. This dataset is structured in two components: (1) COMSOL simulation files, which include velocity and pressure field data, and (2) visualizations of the simulation results, stored in .csv files. The Soil_xx_Seepage_simulation.mph is the example file of seepage simulation through COMSOL, and the Soil_xx_Pore_structure.stl is the grid file involved. Each file in the dataset is structured for seamless integration into computational analysis workflows. By offering detailed spatial characterizations of soil microstructures, the dataset supports advanced research in soil physics, hydrology, and biogeochemistry. Table 1. Summary of the file systems of the CHARM3D dataset. All soil microstructure data are stored in .TIFF format, ensuring compatibility with various visualization software. The morphological parameters of the 3D soil microstructure are provided in .CSV files. File name Key Definition Reference Soil_xx_xxxx.TIFF Grayscale images Binary images − − Soil_xx.png Rendering images Pore structure images Pore network model images POM distribution Images Velocity field and pressure field Images − − Soil_xx_Pore_struct ure.stl Grid file of pore structure − − Soil_xx_Seepage_si mulation.mph x-component of flow velocity y-component of flow velocity z-component of flow velocity (flow direction) Pressure field − − Soil_xx_Minkowski. csv Pore volume fraction (𝜙) 𝜙=𝑀0/𝑉 Legland et al., 2016; 2007 Surface area (𝑆𝐴) 𝑆𝐴=∫𝑑𝑆 Circularity (𝜋) 𝜋=1 2∫𝜅𝑑𝑆 Euler characteristics (𝜒) 𝜒∝∫1 𝑟11 𝑟2𝑑𝑆 Perimeter (𝑃) Crofton formula Ohser and Schladitz, 2009; Legland et al., 2007 Perimeter density (𝜌𝑃) − Convex Area (𝐴(𝑥)) − Convexity (𝜘) 𝜘= 𝐴(𝑋) 𝐴(ConvHull⁡(𝑋)) Soil_xx_Pore_analy sis.csv Throat Radius(𝑇𝑅) The throat is a narrow channel between the connected pores Domander et al., 2021; Doube et al., 2010; Parkinson and Fazzalari, 2000; Harrigan and Mann, 1984 Coordination Number (𝐶𝑁) Describe the number of throats connected to a single individual pore in the pore space Soil_xx_Network_a nalysis.csv Clustering coefficient 𝐶 =∑ 𝑁 𝑖=1 2𝐸𝑖 𝑘𝑖(𝑘𝑖−1) 𝑁P Wang et al., 2023 Average distance (𝐷) 𝐷 =2 𝑁P(𝑁P−1)∑ 𝑖≠𝑗∈𝑉 𝑑𝑖𝑗 Average degree 𝐾=2𝑁t 𝑁p NBC characterizes 𝑁𝐵𝐶(𝑖)=∑ 𝑗≠𝑘 𝑔𝑗𝑘(𝑖) 𝑔𝑗𝑘 Soil_xx_POM.csv Ratio of POM POM (particulate organic matter) refers to the amount of organic matter in the soil Li et al., 2024 Soil_xx_Permeabilit y.csv Permeability (𝐾) Darcy’s Law Bear, 1972 Note: “xx” represents the numerical order of the soil sample. 4 Time range July 1, 2023 -- December 1, 2024