Non-intrusive soil carbon content quantification methods using machine learning algorithms:A comparison of microwave and millimeter wave radar sensors

Agricultural and forestry biomass can be converted to biochar through pyrolysis gasification,making it a significant carbon source for soil.Applying biochar to soil is a carbon-negative process that helps combat climate change,sustain soil biodiversity,and regulate water cycling.However,quantifying soil carbon content conventionally is time-consuming,labor-intensive,imprecise,and expensive,making it difficult to accurately measure in-field soil carbon’s effect on storage water and nutrients.To address this challenge,this paper for the first time,reports on extensive lab tests demonstrating non-intrusive methods for sensing soil carbon and related smart biochar applications,such as differentiating between biochar types from various biomass feedstock species,monitoring soil moisture,and biochar water retention capacity using portable microwave and millimeter wave sensors,and machine learning.These methods can be scaled up by deploying the sensor in-field on a mobility platform,either ground or aerial.The paper provides details on the materials,methods,machine learning workflow,and results of our investigations.The significance of this work lays the foundation for assessing carbon-negative technology applications,such as soil carbon content accounting.We validated our quantification method using supervised machine learning algorithms by collecting real soil mixed with known biochar contents in the field.The results show that the millimeter wave sensor achieves high sensing accuracy(up to 100%)with proper classifiers selected and outperforms the microwave sensor by approximately 10%–15%accuracy in sensing soil carbon content.

A flexible dual-function capacitive sensor enhanced by looppatterned fibrous electrode and doped dielectric pillars for spatial perception

The integrated perception capable of detecting and monitoring varieties of activities is one of the ultimate purposes of wearable electronics and intelligent robots.Limited by the space occupation,it lacks practical feasibility to stack multiple types of single sensors on each other.Herein,a high-sensitivity dual-function capacitive sensor with proximity sensing and pressure sensing is proposed.The fringing electric field can be confined in the proximity-sensitive area by fibrous loop-patterned electrode,leading to more stolen charges when object approaching and thus a high proximity sensitivity.The high-permittivity doped structured dielectric layer reduces the compressive stiffness and enhances the rate of compression-caused increase in the equivalent relative permittivity of the dielectric layer,resulting in a larger increase in capacitance and thus a high pressure sensitivity.The electrodes and dielectric layer together compose the capacitor and act as the sensor without taking up additional space.The decoupling of proximity-sensing and pressure-sensing modes can be achieved by decrease or increase in capacitance.Combined with array distribution and sequential scanning,the sensors can be used for detection of motion trajectory,contour recognition,pressure distribution.

In-Situ Differentiation of Acidic and Non-Acidic Tundra via Portable X-ray Fluorescence (PXRF) Spectrometry

Frozen soils or those with permafrost cover large areas of the earth's surface and support unique vegetative ecosystems. Plants growing in such harsh conditions have adapted to small niches, which allow them to survive. In northern Alaska, USA, both moist acidic and non-acidic tundra occur, yet determination of frozen soil p Hs currently requires thawing of the soil so that electrometric pH methods can be utilized. Contrariwise, a portable X-ray fluorescence(PXRF) spectrometer was used in this study to assess elemental abundances and relate those characteristics to soil pH through predictive multiple linear regressions. Two operational modes, Soil Mode and Geochem Mode, were utilized to scan frozen soils in-situ and under laboratory conditions, respectively, after soil samples were dried and ground. Results showed that lab scanning produced optimal results with adjusted coefficient of determination(R^2) of 0.88 and 0.33 and root mean squared errors(RMSEs) of 0.87 and 0.34 between elemental data and lab-determined pH for Soil Mode and Geochem Mode, respectively. Even though the presence of ice attenuated fluoresced radiation under field conditions, adjusted R^2 and RMSEs between the datasets still provided reasonable model generalization(e.g., 0.73 and 0.49 for field Geochem Mode). Principal component analysis qualitatively separated multiple sampling sites based on elemental data provided by PXRF, reflecting differences in the chemical composition of the soils studied. Summarily, PXRF can be used for in-situ determination of soil pH in arctic environments without the need for sample modification and thawing. Furthermore, use of PXRF for determination of soil pH may provide higher sample throughput than traditional eletrometric-based methods, while generating elemental data useful for the prediction of multiple soil parameters.

Influence of Ice on Soil Elemental Characterization via Portable X-Ray Fluorescence Spectrometry

Field portable X-ray fluorescence （PXRF） spectrometry has become an increasingly popular technique for in-situ elemental characterization of soils. The technique is fast, portable, and accurate, requiring minimal sample preparation and no consumables. However, soil moisture 〉 20% has been known to cause fluorescence denudation and error in elemental reporting and few studies have evaluated the presence of soil moisture in solid form as ice. Gelisols （USDA Soil Taxonomy）, permafrost-affected soils, cover a large amount of the land surface in the northern and southern hemispheres. Thus, the applicability of PXRF in those areas requires further investigation. PXRF was used to scan the elemental composition （Ba, Ca, Cr, Fe, K, Mn, Pb, Rb, Sr, Ti, Zn, and Zr） of 13 pedons in central and northern Alaska, USA. Four types of scans were completed： 1） in-situ frozen soil, 2） re-frozen soil in the laboratory, 3） melted soil/water mixture in the laboratory, and 4） moisture-corrected soil. All were then compared to oven dry soil scans. Results showed that the majority of PXRF readings from in-situ, re-frozen, and melted samples were significantly underestimated, compared to the readings on oven dry samples, owing to the interference expected by moisture. However, when the moisture contents were divided into 〉 40% and 〈 40〈 groups, the PXRF readings under different scanning conditions performed better in the group with 〈 40% moisture contents. Most elements of the scans on the melted samples with 〈 40% moisture contents acceptably compared to those of the dry samples, with R2 values ranging from 0.446 （Mn） to 0.930 （St）. However, underestimation of the melted samples was still quite apparent. Moisture-corrected sample PXRF readings provided the best correlation to those of the dry, ground samples as indicated by higher R2 values, lower root mean square errors （RMSEs）, and slopes closer to 1 in linear regression equations. However, the in-situ （frozen） sample scans did not differ appreciably from the melted sample scans in their correlations to dry sample scans in terms of R2 values （0.81 vs. 0.88）, RMSEs （1.06 vs. 0.85）, and slopes （0.88 vs. 0.92）. Notably, all of those relationships improved for the group with moisture contents 〈 40%.