Soil carbon determination using rapid, inexpensive, non-destructive spectroscopic techniques.

摘要

New methods are required to rapidly and accurately measure soil C at field- and landscape-scales to improve local, regional, and global soil C stock and flux estimates. This research evaluated visible-near infrared diffuse reflectance spectroscopy (VisNIR) and laser-induced breakdown spectroscopy (LIBS) for non-destructive in situ soil carbon determination. 'On-the-go' VisNIR has been proposed as a rapid and inexpensive tool for intensively mapping soil clay and organic carbon concentration. In a direct comparison, lab-based spectral data consistently provided greater prediction accuracy than on-the-go spectral data for Montana cropland soils. The current configuration of on-the-go VisNIR systems allows for rapid field scanning, however on-the-go soil processing could improve predictions. LIBS is an emerging elemental analysis technology with the potential to provide rapid, accurate, and precise analysis of soil constituents. We evaluated LIBS for measuring soil profile C in field-moist, intact soil cores. Results indicate that LIBS can be calibrated to accurately estimate and differentiate between soil total and inorganic C concentrations utilizing stoichiometric relationships between C and elements related to total and inorganic C in the soil matrix. The fundamental principles on which VisNIR and LIBS are based differ in regards to molecular and elemental spectroscopy, respectively, therefore integrating VisNIR and LIBS should theoretically improve soil C predictions compared to individual sensors. Integrating VisNIR and LIBS did not consistently improve soil profile C predictions over individual sensors. In general, SOC was not well characterized using VisNIR, LIBS, or combined VisNIR-LIBS for soils in this study, presumably due to challenges associated with scanning surfaces of intact soil cores, variable SOC chemistry, and lack of SOC variation. Considering the challenging conditions under which VisNIR and LIBS were tested, model calibrations and sensor integration performed admirably. Further testing of combined VisNIR-LIBS under more controlled soil conditions with samples containing greater SOC diversity is necessary to determine the technical potential of the method. Currently, results suggest that in situ VisNIR and LIBS may be best employed as field-stratification tools for targeted conventional soil C measurements. Ultimately, we envision a penetrometer-mounted, integrated VisNIR-LIBS sensor array for rapid soil elemental and molecular characterization at field- and landscape-scales.