Soil characterization using the NIR+ spectrometer
In this case study, we share spectra of various soil samples as measured with the NIRQuest+ spectrometer (900-2500 nm). NIRQuest+ is distinguished by its high sensitivity and great thermal stability. In this experiment, we identified the moisture content of organic and top soil samples using the NIRQuest+2.5 spectrometer (900-2500 nm) as part of a diffuse reflectance system. Determining moisture content in soil has great significance in agriculture, where crop health is largely dependent on maintaining optimal moisture levels, which vary by factors such as the soil and crop type.
Results
Figure 1 shows samples comprising top soil, compost and sand initially were measured with no sample preparation.
Note the two strong moisture peaks. Absorption bands for water typically are seen at 970 nm, 1200 nm, 1450 nm and 1950 nm.
Although the data were collected using diffuse reflectance techniques, we plotted the spectra as absorbance (Log (1/Reflectance). This is because observing the trend in intensity/absorbance is more intuitive than observing the plots as %Reflectance.
Next, we put the samples into a drying oven at 100 °C for two days, weighing and measuring the samples again to determine the moisture content for each. These samples were labeled as Dry Soil. Once the samples were dried, the moisture peaks were not nearly as strong. Still, in our humid Florida environment, the dried samples absorbed at least some water, which is evident inthe spectra.
Figure 3 shows a data plot that has been baseline corrected. As you’ll see, the trend in the reflectance at the two water peaks follows the trend in moisture we determined for our soil samples.
By applying machine learning tools and algorithmic modelling, additional insights could be gleaned from the moisture content spectral data and other characteristics of soil samples.
