Near- versus mid-infrared diffuse reflectance spectroscopy for soil analysis emphasizing carbon and laboratory versus on-site analysis: Where are we and what needs to be done?

Abstract

Over several decades, near infrared (NIR) diffuse reflectance spectroscopy (NIRS) has been shown to be extremely versatile for the rapid analysis of many agricultural materials including forages, foods and grains. More recently, mid-infrared (mid-IR) diffuse reflectance spectroscopy (DRIFTS) and NIRS have come under intense scrutiny for their potential to provide a rapid method for the analysis of soil C, minerals and other soil parameters of interest. Research has demonstrated that for the determination of soil C, DRIFTS is often more accurate and produces more robust calibrations than NIRS when analyzing ground, dry soils under laboratory conditions. However, mid-IR spectra are known to be more affected by moisture and sample preparation than NIR. In reality, DRIFTS is not even considered feasible on samples containing high levels of moisture due to the strong water absorptions in the mid-IR, although the presence of water is also known to often degrade even NIR spectra and subsequent calibrations. While both techniques offer the potential for the analysis of soils on-site, and even in situ, many questions remain to be answered including: 1. What are the advantages and disadvantages of on-site as opposed to laboratory analysis? 2. What are the effects of moisture and particle size on accuracy if samples are to be analyzed on-site? 3. Which spectral range (mid-IR or NIR) is the most effective for in laboratory and/or on-site analysis? 4. Which analytes can be accurately analyzed by NIR and/or mid-IR spectroscopy? 5. What are the effects of different soil types and compositions on the entire process of calibration development? In addition, while DRIFTS has been shown to be advantageous in the laboratory, if samples need to be ground and dried, and instruments purged to obtain useable data, it may not be practical for on-site use. This review will try to answer some of these questions and show where the science stands and what needs to be done before NIRS or DRIFTS can be fully exploited for routine soil analysis.