A Ground-Penetrating Radar-Based Study of the Structure and Moisture Content of Complex Reconfigured Soils

Yunlan He, Lulu Fang, Suping Peng, Wenyu Liu, Changhao CuiWater 2024 · DOI

TL;DR

This study leverages Ground-Penetrating Radar (GPR) with advanced signal processing techniques (envelope detection and AEA method) to accurately map the structure and moisture content of complex, reconstructed soils. The research demonstrates the effectiveness of these methods in identifying soil layers and inverting moisture content up to 3 meters deep, offering a non-destructive and efficient alternative to traditional soil analysis.

What Problem Does This Paper Address?

Reclaimed and reconstructed soils, particularly in mining areas, often exhibit complex and heterogeneous structures that are challenging to characterize using conventional methods. Traditional soil analysis techniques, such as on-site profile excavation, are time-consuming, labor-intensive, and destructive, making it difficult to obtain a comprehensive understanding of soil structure and moisture distribution across a region. Existing GPR studies have primarily focused on natural soils, leaving a gap in understanding its application for these complex, engineered soil environments. This paper aims to address this by developing and validating GPR-based methods for accurately assessing both the structural layering and moisture content of these challenging reconstituted soils.

Methods

The core of this research lies in the application and refinement of GPR data processing techniques. The authors employed a 400 MHz GPR system to collect data. Key methodological steps include:

  • Signal Preprocessing: Standard techniques like background denoising, gain processing, bandpass filtering, and sliding average were used to enhance the signal-to-noise ratio and extract clearer waveform characteristics.
  • Envelope Detection: This technique was used to obtain the amplitude envelope of the GPR signals. By applying a Hilbert transform, upper and lower envelopes were generated.
  • Instantaneous Phase Analysis: Least squares fitting was applied to the instantaneous phase of the envelopes. The inflection points on these fitted curves were identified as demarcation points for soil structure and layer boundaries. This allowed for the calculation of soil layer thicknesses.
  • Average Envelope Amplitude (AEA) Method: This novel approach establishes a correlation between the early signal amplitude envelope of the GPR and the soil's dielectric constant. This correlation is then used to invert for soil moisture content. The authors specifically investigated the relationship between the radar amplitude envelope and the soil's dielectric constant, and subsequently, the moisture content.
  • Hilbert-Huang Transform: This was used to analyze the frequency components of the GPR signals, further aiding in the identification of structural characteristics and anomalous reflections.

The novelty lies in the combined application of these advanced signal processing techniques, particularly the AEA method, for the specific context of complex, multilayered, and mixed reconstituted soils.

Datasets

The study was conducted in a specific reconstructed soil area within the Jungar Banner Junneng Mining Area, Ordos City, Inner Mongolia Province. This area was characterized by:

  • Zone A: A mixed soil area composed of red clay, sandy loam, and loess.
  • Zone B: A three-layered reconstructed soil area with distinct layers of red clay (bottom), sandy loam (middle), and loess (top).

Soil samples were collected via boreholes to a depth of 3 meters at various points along the GPR survey lines. These samples were used for laboratory drying to determine the actual moisture content, serving as ground truth for validating the GPR-based inversion results. No new benchmark datasets were introduced; rather, the study utilized a real-world field site with known soil types and structures.

Key Insights and Ideas

  • Envelope detection is a powerful tool for soil stratification: The study clearly demonstrates that envelope detection, combined with Hilbert transforms and phase analysis, can effectively delineate soil layer boundaries, even in complex reconstituted soil profiles. This is a significant improvement over raw GPR data for structural analysis.
  • AEA method shows promise for moisture inversion in challenging soils: The AEA method proved to be a reliable technique for estimating soil moisture content up to 3 meters deep in these engineered soils, achieving relatively low error rates. This expands the applicability of GPR for hydrological studies in such environments.
  • Layered soils offer better GPR performance than mixed soils: The stratified soil structure in Zone B yielded more accurate and consistent results for both structure identification and moisture content inversion compared to the mixed soil structure in Zone A. This highlights the impact of soil homogeneity on GPR signal interpretation.
  • Moisture content increases with depth, but the pattern differs: While both zones showed an increasing trend in moisture content with depth, the layered structure of Zone B exhibited more pronounced, stepwise increases, reflecting the distinct hydrological properties of each layer. This provides valuable insights into water movement and retention in these engineered systems.

Key Results

  • Soil Structure: The envelope detection method successfully identified stratification boundaries in Zone B, with calculated layer thicknesses closely matching actual measurements (e.g., upper layer 1.47 m, middle 0.88 m, bottom 0.65 m). Zone A, with its mixed structure, showed less distinct stratification.
  • Soil Moisture Content: The AEA method achieved an average relative error ranging from 2.81% to 7.41% for moisture content inversion up to 3 meters deep.
  • The upper layers (0-1.4 m) showed the most accurate inversion (MRE 2.81-2.92%).
  • The middle layers (1.4-2.2 m) had slightly higher errors (MRE 3.78-5.01%).
  • The bottom layers exhibited the highest errors (MRE 6.64-7.41%), suggesting potential challenges with deeper penetration or signal attenuation.
  • Comparison: The AEA method's inversion results showed a strong correlation (R² as high as 0.99) with measured moisture content from the drying method. Layered soil structures (Zone B) generally yielded better inversion results than mixed soil structures (Zone A).

Limitations and Future Work

  • Signal Attenuation: The paper acknowledges that signal attenuation in high-conductivity soils remains a challenge, potentially limiting the effective detection depth. This is a common limitation of GPR.
  • Mixed Soil Complexity: The mixed soil structure in Zone A presented more difficulties for clear stratification and moisture inversion compared to the layered structure in Zone B.
  • Future Research Directions:
    • Wider Applicability: Validate these methods across a broader range of soil types and environmental conditions.
  • Enhanced Accuracy: Explore integrating GPR with other geophysical techniques or advanced numerical models to improve accuracy and detection depth.
  • Refined Signal Processing: Further research into refining signal processing techniques could mitigate issues like signal attenuation.

Summary updated

Abstract

To increase the detection accuracy of soil structure and moisture content in reconstituted soils under complex conditions, this study utilizes a 400 MHz ground-penetrating radar (GPR) to examine a study area consisting of loess, sandy loam, red clay, and mixed soil. The research involves analyzing the single-channel waveforms and two-dimensional images of GPR, preprocessing the data, obtaining envelope information via amplitude envelope detection, and performing a Hilbert transformation. This study employs a least squares fitting approach to the instantaneous phase envelope to ascertain the thickness of various soil layers. By utilizing the average envelope amplitude (AEA) method, a correlation between the radar’s early signal amplitude envelope and the soil’s shallow dielectric constant is established to invert the moisture content of the soil. The analysis integrates soil structure and moisture distribution data to investigate soil structure characteristics and moisture content performance under diverse soil properties and depths. The findings indicate that the envelope detection method effectively identifies stratification boundaries across different soil types; the AEA method is particularly efficacious in inverting the moisture content of reconstituted soils up to 3 m deep, with an average relative error ranging from 2.81% to 7.41%. Notably, moisture content variations in stratified reconstituted soils are more pronounced than those in mixed soil areas, displaying a marked stepwise increase with depth. The moisture content trends in the vertical direction of the same soil profile are generally consistent. This research offers a novel approach to studying reconstituted soils under complex conditions, confirming the viability of the envelope detection and AEA methods for intricate soil investigations and broadening the application spectrum of GPR in soil studies.
BibTeX
@inproceedings{he2024ground,
  author = {Yunlan He and Lulu Fang and Suping Peng and Wenyu Liu and Changhao Cui},
  title = {{A Ground-Penetrating Radar-Based Study of the Structure and Moisture Content of Complex Reconfigured Soils}},
  booktitle = {Water},
  year = {2024},
  doi = {10.3390/w16162332},
}
Figure