Show simple item record

dc.contributor.advisorScharenbrock, Bryant C.
dc.contributor.authorBrice, Logan
dc.date.accessioned2023-06-28T19:37:49Z
dc.date.available2023-06-28T19:37:49Z
dc.date.issued2023-05
dc.identifier.urihttp://digital.library.wisc.edu/1793/84396
dc.description.abstractLabile carbon is a portion of soil organic matter (SOM) associated with biological activity and is commonly used as an indicator of soil health. Permanganate oxidizable carbon (POXC) has been proposed as an indicator of labile carbon that reflects a soil's response to management. Permanganate oxidizable carbon provides infield assessment of soil dynamics but relies upon uncertain assumptions related to the chemical composition of oxidized materials and methodological limitations restrict analysis to mineral soils. Rapid assessment and greater understanding of the oxidation reactions of labile fractions of SOM are needed to create a more robust soil health indicator. Recent applications of spectroscopy have demonstrated accurate measurements of SOM fractions with higher sample throughput than POXC. Advantages of spectroscopy include non-destructive analysis and prediction of multiple soil properties from a single spectral scan. This study focuses on predicting POXC with mid-infrared (MIR) diffuse reflectance spectroscopy and the identification of associated functional chemical groups. Spectral data from 710 soil samples were collected by an Agilent 4300 FTIR instrument. Predictive models were built using partial least squares (PLS) regression and performance was evaluated with R2 and root mean square error (RMSE). Six models were built with an All-Soils model and five subsets based on SOM %. Utilizing the entire MIR spectral region, the All-Soils model had R2 = 0.76/0.75 (Cal/Val) with RMSEC = 100.9 and RMSECV = 109.8 mg/kg POXC. The prediction of POXC in this model relied upon positive correlations with the following spectral peaks: 1,565 (amide II), 1,660 (amide I), 1,722 (carboxylic acid), 2,851 and 2,924 (aliphatic symmetric and anti-symmetric). The spectral peaks correlated with POXC prediction varied in other tested models as did SOM %. In general, the SOM subset models gave inaccurate predictions, but revealed changes in the spectral peaks correlated with POXC prediction at certain SOM concentrations. The All-Soils model showed POXC prediction with a handheld MIR instrument can be accurate. There is a need for rapid and accurate quantification of soil health indicators, and we produced predictive models from an instrument applicable for in-field use.en_US
dc.language.isoen_USen_US
dc.publisherCollege of Natural Resources, University of Wisconsin-Stevens Pointen_US
dc.subjectMid-Infrared Diffuse Reflectance Spectroscopyen_US
dc.subjectPermanganate Oxidizable Carbonen_US
dc.titlePREDICTING PERMANGANATE OXIDIZABLE CARBON WITH MID-INFRARED DIFFUSE REFLECTANCE SPECTROSCOPY ACROSS WISCONSIN SOILSen_US
dc.typeThesisen_US


Files in this item

Thumbnail

This item appears in the following Collection(s)

Show simple item record