The ability to accurately measure the geometric characteristics of soil wetted bodies (SWBs) is very important for conserving water in agriculture. However, measurements of SWBs obtained using conventional methods have a number of defects. Ground penetrating radar (GPR) is a promising technique for detecting buried features. In this paper, we used GPR to nondestructively investigate SWBs formed under drip irrigation. First, numerical simulations were performed to obtain the theoretical radar-gram of SWBs. Then, controlled irrigation experiments were performed to obtain radar scan datasets in a laboratory. The GPR image was interpreted according to the numerical simulation results, and the SWB thickness detection accuracy was estimated. Finally, GPR detection was performed in the field with different irrigation volumes, and the detection effect was assessed. The GPR reflections in the laboratory and field measurements were more complex than in the numerical simulation images, but the location and thickness of SWBs were still clear; the accuracy of the measured thickness was high, and the accuracy decreased with an increase in irrigation volume. The radar image resolution and thickness accuracy measured in the field were slightly less than the values measured in the laboratory. Thus, GPR is able to quickly and accurately characterize SWBs formed by drip irrigation based on the thickness and relative position in the soil. Furthermore, the F-K offset transformation was an effective GPR data processing method for focusing reflections from SWBs to obtain their true position and physical extent.

The accurate detection of the volume of the soil wetted body (SWB) is the theoretical basis for the optimal design of a drip irrigation system. In view of the shortcomings of traditional detection methods, such as low efficiency, high cost, and poor in situ detection, this paper explores the rapid, in situ, and nondestructive estimation techniques of SWB volume using Ground Penetrating Radar (GPR). Through drip irrigation of different water volumes and geotechnical experiments, the SWBs with different sizes were formed in air-dried soil and natural soil, respectively. The GPR scan lines on the ground were arranged to obtain the SWB radargram, and the profile shape of the wetted body was restored by the F-K transformation process (frequency-wave domain migration algorithm) that can transform the wave equation in the time-space domain into the frequency-wavenumber domain to achieve offset homing. On this basis, the SWB volume estimation model was constructed by three methods: regression analysis, geometric calculation, and profile triple integration (profile area accumulation). Additionally, the accuracies of the models were verified by the measured data of the SWB excavation. For each of these methodologies, we illustrate the quality of the output with the measured data and discuss the possibilities and limitations. The results show that the profile triple integration model is not accurate. The regression analysis model has the greatest accuracy. However, owing to the difference in soil characteristics, the model input parameters must be changed. Relatively, the geometric calculation model is the best model to detect the SWB volume. From the results, we can draw a conclusion that the method combining F-K transformation process and the geometric calculation model can accurately measure the volume of SWB, and it has better adaptability in different soils. We demonstrate the GPR as a rapid geophysical tool that can be used successfully to explore the volume of the SWB formed under drip irrigation. This study expands the agricultural application of GPR and provides a new method for in situ nondestructive rapid detections of the SWB volume.

How homologous straw and biochar alter soil carbon fractions and associated microbial processes in coastal saline soils remains unclear. A 120-day microcosm experiment compared reed (Phragmites australis) straw and its biochar at equal mass rates of 1%, 5%, and 9% of dry soil mass. Both amendments increased the soil organic carbon (SOC) and reduced electrical conductivity. At 9%, straw increased the particulate organic carbon share of SOC to 78.2%, decreased the bacterial oligotroph/copiotroph ratio from 1.82 to 0.86, and raised potential hydrolytic enzyme activities to 4.3–9.1 times the control levels. Across straw treatments, cumulative CO₂ emissions were 375.0–1645.9% higher than in the unamended control. Biochar maintained operationally defined mineral-associated organic carbon dominance (62.1–68.0% of SOC), with weaker hydrolytic responses and cumulative CO₂ emissions 58.0–102.2% above the control levels. Exploratory structural equation models showed that lower bacterial oligotroph/copiotroph ratios were associated with higher hydrolytic enzyme activity under straw. Under biochar, higher fungal Basidiomycota/Ascomycota ratios were associated with higher polyphenol oxidase activity. Both enzyme indicators were positively associated with cumulative CO₂ emissions in their respective models. These findings demonstrate contrasting carbon fraction distributions and respiratory responses to homologous amendments, with implications for carbon management in coastal saline soils.

Introduction: Land-use/land-cover (LULC) change in Cambodia is relevant to forest conservation and land management in the Lower Mekong region. Methods: This study describes mapped change between three national LULC snapshots (2000, 2010, and 2020) using Landsat TM/ETM+/OLI imagery, supervised support vector machine (SVM) classification, post-classification transition matrices, and random forest (RF) predictor-importance analysis. Results: Mapped cropland increased from 57,972.40 km ² in 2000 to 71,281.39 km ² in 2020, whereas woodland/forest declined from 103,756.88 to 89,428.94 km ² . Woodland/forest to cropland was the largest mapped 2000-2020 pathway (13,426.25 km ² ). Water and built-up/impervious transitions were excluded from substantive interpretation because both classes had F1 = 0 in the available 2020 validation matrix. Human footprint, terrain, and ecosystem productivity ranked highest in the RF models, indicating model-specific predictor associations rather than causal effects. Discussion: The study provides an empirical synthesis of mapped transitions, predictor rankings, and class-specific accuracy constraints at the national scale. The three snapshots do not resolve continuous land-system trajectories, and the uncalibrated suitability surface should be interpreted as an exploratory visualization rather than a validated basis for management zoning.