Answers are grounded in this paper and cite their sources.
Precise characterization of the spatiotemporal dynamics of soil moisture is of profound significance for optimizing water resource allocation, preventing and mitigating drought and flood disasters, and safeguarding food security. However, complex farmland environmental conditions tend to degrade the laboratory-calibrated accuracy of soil moisture sensors after long-term field deployment, thereby impairing in-situ monitoring performance. A consecutive three-year in-situ field experiment was conducted at the National Precision Agriculture Research and Demonstration Base (Changping, Beijing, China). Six widely used frequency domain reflectometry (FDR) soil moisture sensors, including four probe-type sensors (P1–P4) and two tube-type sensors (T1–T2), were selected to systematically evaluate their long-term monitoring performance, and the random forest (RF) algorithm was employed to identify the key environmental factors affecting their monitoring performance. The results indicated that the monitoring values of all six sensors could effectively capture the spatiotemporal variation characteristics of soil moisture consistent with those measured by the standard oven-drying method; however, due to the influence of environmental variations, their root mean square error (RMSE) ranged from 0.026 to 0.085 m³/m³. Different sensors exhibited distinct performance variations under varying monitoring periods and soil depth conditions. Specifically, probe-type sensors demonstrated overall superior accuracy in long-term field monitoring compared with tube-type counterparts, with the P4 sensor achieving the optimal overall performance (RMSE = 0.028 m³/m³). In terms of seasonal variation, all sensors achieved the highest monitoring accuracy in autumn (RMSE = 0.044 m³/m³), while monitoring errors were relatively higher in spring (RMSE = 0.054 m³/m³) and summer (RMSE = 0.052 m³/m³). Moreover, the monitoring error of the sensors increased with increasing soil depth, with the RMSE in the deep soil layer (60–80 cm) rising by 8.16% compared with that in the shallow soil layer (0–20 cm). Environmental dependence analysis revealed that monitoring errors increased under conditions of high humidity (≥80%), low or high temperature (≤5 °C or ≥25 °C), and high wind speed. Further, feature importance ranking via the RF algorithm identified soil temperature as the dominant environmental factor affecting sensor monitoring accuracy. The six sensors failed to consistently achieve the nominal accuracy specified by the manufacturers across all soil layers during long-term field operation. Post-factory secondary calibration tailored to actual soil types and climatic conditions constitutes a crucial approach to enhancing their measurement accuracy. In this study, sensor-specific secondary calibration equations were established based on data obtained from in-situ soil sampling and the oven-drying method, reducing the mean absolute error (MAE) of the sensors by 4.3–58.7%. This research provides data and methodological support for ensuring the accuracy of soil moisture sensors during long-term field operation.
No takes yet. Share an insight, caveat, or question.
Shi et al. (2026) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: