224-6 Mid Infrared Reflectance Spectroscopy As an Alternative to Wet Chemistry Methods for Soil Health Assessment for Mollisols of the United States.

See more from this Division: ASA Section: Land Management and Conservation
See more from this Session: Soil Health for Improving Soil Water Dynamics and Agroecosytem Resilience

Tuesday, November 8, 2016: 11:05 AM
Phoenix Convention Center North, Room 221 A

Sonam Sherpa, Cornell University, Ithaca, NY and David W. Wolfe, 117 Plant Sci. Bldg., Tower Road, Cornell University, Ithaca, NY
Abstract:
In recent years there has been a growing interest by farmers to adopt soil management strategies that improve soil health. Assessment methods for measuring soil health incorporate a range of chemical, physical, and biological analytical techniques. The comprehensive nature of soil health assessment can result in costs which far exceed traditional soil chemical analysis. This additional cost can limit a farmer’s ability to adequately sample the various management units of the farm. Mid infrared reflectance spectroscopy (MIR) has been proposed as a safe, rapid, nondestructive, and inexpensive method for quantifying soil properties. In this study we evaluated the potential of using MIR to predict 12 soil properties often utilized for assessing soil health: total nitrogen (TN), total carbon (TC), inorganic carbon (IC), bulk density (BD), soil texture (sand, silt, clay), cation exchange capacity (CEC), pH, water retention 1/3rdbar (WR1/3), water retention 2bar (WR2), and water retention 15bar (WR15) for mollisol soils of the U.S. We selected 5571 samples representing 26 states and 243 counties, allowing for a thorough representation of Mollisols of the country. Soil spectra and reference data were obtained from the National Soil Survey Center in Lincoln NE, USA. Soil properties were predicted from spectra by partial least squares regression. Models for predicting IC, TC, CEC, clay, and sand performed best, r2 = .99, .96, .91, .91, and .90, respectively. While BD, WR1/3, WR15, and WR2 displayed the poorest model performance r2 = .49, .60, .70, and .70, respectively. This study demonstrates the potential of using MIR for rapidly predicting multiple soil properties for samples representing a large geographic area at a precision suitable for most agricultural management applications. Inexpensive rapid characterization of soil properties will allow for better soil management, consequently improving soil quality and reducing nutrient loss and environmental pollution associated with agricultural and other land management practices.

See more from this Division: ASA Section: Land Management and Conservation
See more from this Session: Soil Health for Improving Soil Water Dynamics and Agroecosytem Resilience

<< Previous Abstract | Next Abstract